An investigation of the solid-state photochemistry of α-mesitylacetophenone derivatives: Asymmetric induction studies and crystal structure-reactivity relationships

Article abstract of DOI:10.1016/S0040-4020(00)00496-8

The photochemical conversion of a series of α-mesitylacetophenone derivatives into 2-indanols via δ-hydrogen abstraction has been investigated in the solid state. A correlation between solid-state reactivity and crystal structure has been established for this type of reaction. For the seven compounds whose crystal structures were determined, the average value of d (C=O···H distance) and L (C=O···CH₃ distance) were 2.77±0.04 A? and 3.42±0.06 A?, and the value of ω (δ-H out of plane angle), Δ(C=O···H angle) and θ (C-H···O angle) were 59±2°, 80±7°and 123±3°, respectively. These parameters depended mainly on the magnitude of the deviation of the carbonyl group from the fully bisecting position over the mesityl ring, which ranged from 9-14°in the case of in the solid state reactive ketones and approached 0°for the unreactive compounds. 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 (13 in total) in the crystalline state gave enantiomeric excesses of up to 90%. The crystal structures of three of the salts were determined and on this basis, the reasons for the selectivity in the crystalline state are discussed. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4020(00)00496-8

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 6739±6751
An Investigation of the Solid-State Photochemistry of
a-Mesitylacetophenone Derivatives: Asymmetric Induction
Studies and Crystal Structure±Reactivity Relationships
*
Eugene Cheung, Katja Rademacher, John R. Scheffer and James Trotter
Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, B.C., Canada V6T 1Z1
Received 7 February 2000; revised 7 March 2000; accepted 9 March 2000
AbstractÐThe photochemical conversion of a series of a-mesitylacetophenone derivatives into 2-indanols via d-hydrogen abstraction has
been investigated in the solid state. A correlation between solid-state reactivity and crystal structure has been established for this type of
reaction. For the seven compounds whose crystal structures were determined, the average value of d (CvO´´´H distance) and L
Ê
Ê
(CvO´´´CH3 distance) were 2.77^0.04 A and 3.42^0.06 A, and the value of v (d-H out of plane angle), D (CvO´´´H angle) and u
(C±H´´´O angle) were 59^28, 80^78 and 123^38, respectively. These parameters depended mainly on the magnitude of the deviation of the
carbonyl group from the fully bisecting position over the mesityl ring, which ranged from 9±148 in the case of in the solid state reactive ketones
and approached 08 for the unreactive compounds. 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 (13 in
total) in the crystalline state gave enantiomeric excesses of up to 90%. The crystal structures of three of the salts were determined and on this
basis, the reasons for the selectivity in the crystalline state are discussed. q 2000 Elsevier Science Ltd. All rights reserved.
Introduction
environment within the crystal takes on the role of an asym-
metric inductor. Spontaneous crystallization of achiral
compounds in chiral space groups, while well documented,2
is rare and unpredictable and therefore unsuitable as a
general method to perform asymmetric induction in the
solid state. It occurred to us that one approach to overcome
this problem lay in the use of ionic chiral auxiliaries.3 By
introducing a carboxylic acid group into the photoreactant
and forming a salt with an optically pure amine, we obtain a
compound that is required to crystallize in a chiral space
group.
Investigations of the photochemical behavior of organic
compounds in a solid environment have been of interest to
theoretical as well as synthetic chemists for some time now.
Our interests have been centered around the pure crystalline
state, since this medium makes it possible to determine the
exact structure under study by using the very powerful tech-
nique of X-ray crystallography. We and others have carried
out extensive research that has de®ned the role of the crystal
lattice in controlling organic photoreactivity.1 Through
correlation of the reactivity pattern for a series of closely
related compounds with their X-ray crystal structures,
detailed mechanistic information on the preferred inter-
atomic distances and angular requirements for a particular
type of reaction can be obtained. This work has also shown
that, compared to the results in ¯uid media, profound
changes in reaction regio- and stereoselectivity can be
achieved by irradiation of molecular crystals. Organic
photochemistry offers a variety of reactions that transform
achiral starting materials into chiral photoproducts, and to
perform these reactions enantioselectively is one of the
challenges facing organic photochemists these days. Our
approach is based on the solid state irradiation of materials
that crystallize in chiral space groups, where the chiral
In this paper, we describe the application of the solid-state
ionic chiral auxiliary approach to asymmetric induction.
The reaction chosen for investigation was the photocycli-
zation of a-mesitylacetophenone 1a leading to 2-phenyl-2-
hydroxy-4,6-dimethylindan (2a), which Wagner reported in
1991 to be highly ef®cient upon irradiation in solution as
well as in the solid state.4 The mechanism is similar to the
Norrish/Yang type II sequence, and involves triplet-
mediated d-hydrogen abstraction via a seven membered
transition state that generates a 1,5-biradical, which then
undergoes ring closure5 (Scheme 1).
A new stereogenic center is created from an achiral
substrate during this reaction. Because only racemic
mixtures were obtained by photolyzing pure a-mesityl-
acetophenone, this prompted us to investigate asymmetric
induction using the ionic chiral auxiliary method.6
Keywords: photochemistry; cyclization; asymmetric induction.
* Corresponding author. Tel.: 1604-822-3496; fax: 1604-822-2847;
e-mail: scheffer@chem.ubc.ca
0040±4020/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00496-8

6740
E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
Scheme 1. Photocyclization of a-mesitylacetophenone.
Although much is known about the geometric requirements
for g-hydrogen abstraction,5 very little comparable data
have been published on the d-H-abstraction process.7 We
therefore decided to perform a crystal structure-solid state
reactivity correlation study prior to the investigation of the
chiral salts, in order to have a better understanding of the
reaction under consideration.
results were obtained in the solid state, and Table 1
summarizes the conversions obtained as a function of
irradiation temperature. The techniques used for solid-
state irradiations are described in the Experimental Section.
Low temperature photolyses were conducted in order to
minimize the loss of topochemical control in the crystal
lattice through melting, as observed in some cases.
The 2-indanols 2b±k were the only photoproducts detect-
able by GC, with one exception. Upon photolysis of keto-
ester 1j to low conversion, we observed two photoproducts
of similar retention time and identical molecular mass (GC/
MS), with 2-indanol 2j still being the major product. At
higher overall conversion the ratio of indanol to unknown
byproduct increases in favor of the former, until it becomes
the only detectable product. Because of the very low yields
we were unable to isolate the byproduct.8
Results and Discussion
Crystal structure±solid-state reactivity
In developing crystal structure±solid-state reactivity
relationships, it is desirable to have a series of closely
related compounds to study, since this permits general
trends to be detected. Accordingly, a-mesitylacetophenones
1b±k (Scheme 2) were synthesized by standard procedures
as outlined in the experimental section. All nine ketones
were crystalline, and we were able to obtain the crystal
structures of seven of them (1b±d, 1f±g, 1j±k). None of
these ketones crystallized in a chiral space group, thus
precluding absolute asymmetric synthesis studies.
The data collected in Table 1 are only a rough estimate of
quantum yields in the solid state, and no rate constants could
be obtained. Given the dif®culties involved in making
quantitative measurements in solids, the observation of
ef®cient photocyclization merely indicates that hydrogen
abstraction can occur.9 Nevertheless, indanol formation
could not occur unless these ketones exist in geometries
that allow formation via d-hydrogen abstraction of
Irradiation of ketones 1b±k in solution led ef®ciently and
exclusively to the corresponding 2-indanols 2b±k. Similar
Scheme 2.
Table 1. Temperature-dependent conversions in the solid-state photolysis of a-mesitylacetophenones 1b±k
T (8C)a
t (h)
1b (%)
1c (%)
1d (%)
1e (%)
1f (%)
1g (%)
1h (%)
1jb (%)
1k (%)
Rt
220
2
2
0
0
96
45
31
5
58
10
76c
21
100
70
28
6
30 (73)
3 (100)
97c
0
a rtˆambient temperature in photolysis box (30±408C).
b % Yield of indanol in brackets (see text).
c Sample showed signs of melting upon photolysis.

E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
6741
Figure 1.
Figure 2.
biradicals in conformations that can then cyclize without too
much molecular motion.
being in the same plane as one of the m-methyl groups. The
least favorable value for k is 908, in which case the carbonyl
group is totally bisecting the mesityl plane.
There are four geometric parameters associated with
the hydrogen abstraction process as established for the
Norrish/Yang type II reaction sequence (Fig. 1).10 The
symbol d is given to the CO´´´H distance, and it seems
reasonable to suggest that the optimum value should be
close to the sum of the van der Waals radii for oxygen
Hydrogen abstraction in the solid state is likely to occur
with a minimum of motion in the carbon framework, and
we assume that ground state geometric data derived from
X-ray crystallography can be used to analyze the behavior
of the excited molecule in the crystal lattice as well.10 The
values of d, L, v, D, u, k and D derived from crystal
structure analyses of ketones 1b±k are summarized in
Table 2.
Ê
and hydrogen, which is 2.72 A. The symbol v refers to
the angle by which the hydrogen atom lies outside the
mean plane of the carbonyl group, and since the n-orbitals
of oxygen, which are generally agreed to be involved in the
abstraction process, lie in this plane, its ideal value is
expected to be 08. The angle D de®nes the CO´´´H angle,
and its optimum value lies between 90±1208, depending on
the hybridization of the non-bonding electrons on oxygen.
The symbol u refers to the CH´´´O angle, which according
to theory, should have an optimum value of 1808.12 Finally,
the symbol D represents the interatomic distance between
The conformations of all these molecules are essentially
the same. The central carbonyl moiety is approximately
coplanar with the p- or m-substituted phenyl group and is
pointed towards the mesityl ring. The most signi®cant
difference lies in the torsion angle k. The two cyano ketones
1c and 1g show the shortest abstraction distances d and L,
which is consistent with the observed torsion angle k in both
compounds deviating by 148 from its bisecting angle of 908.
The angles v, D and u lie within established values.10
Ketones 1c and 1g are expected to show a high reactivity
in the solid state, as was found (Table 1). The asymmetric
unit of keto acid 1d contains four independent molecules.
Only two of these have the carbonyl oxygen tilted more
towards one of the two o-methyl groups (kˆ77 and 798,
Table 2) and the observed abstraction parameters predict
hydrogen abstraction. The other two molecules have the
carbonyl group almost bisecting the mesityl ring plane
and the closest hydrogen (d)/methyl carbon (L) is too far
from the carbonyl oxygen for abstraction to occur. This fact
should reduce the reactivity of p-keto acid 1d. In m-¯uoro
ketone 1f and m-keto ester 1j the closest methyl groups are
disordered and only 77 and 67% of the hydrogens lie close
enough for abstraction to occur (Scheme 2: conformer A
versus C). Even if we expect the methyl groups to rotate
during photolysis, the greater values for L predict a lower
reactivity in comparison to the cyano ketones 1c and 1g, and
that was also found (Table 1). Finally, the crystal structure
of the m-methyl ketone 1k suggests why it should be almost
unreactive in the solid state. With the torsion angle k
between the carbonyl plane and the mesityl ring being
Ê
the two carbon atoms and is expected to be around 3.4 A, the
sum of the van der Waals radii of the two carbon atoms. In
the case of a-mesitylacetophenones, one can expect the
methyl groups to rotate around the single bond, even in
the crystalline state, though they still spend most of the
time at or near the energy minimum (conformation A or C
in Scheme 3). Thus, the above discussed abstraction
parameters are not necessarily ®xed as in the case of more
rigid compounds.
For this reason, we de®ne a new parameter with reference to
Fig. 2 to take the rotational freedom of the methyl groups in
a-mesitylacetophenone derivatives into account. The
distance L between the carbonyl oxygen atom and the
methyl carbon atom does not change upon rotation of
the methyl group and therefore seems to be more valuable
for predicting hydrogen abstraction. The ideal value for L is
Ê
Ê
3.70 A, given by the sum of didealˆ2.72 A and the CH-bond
Ê
length (0.98 A) at the angle uidealˆ1808 (see Fig. 1). Another
parameter, k, has been de®ned to show the magnitude of the
deviation of the carbonyl group from the fully bisecting
position, which is best given by the value of the dihedral
angle between the carbonyl plane and the plane de®ned by
the mesityl ring carbon atoms. The ideal value for k is 08,
which situates the carbonyl group closest to a d-hydrogen by
Scheme 3.

6742
E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
Table 2. Crystallographically derived CvO´´´Hd-abstraction geometries for ketones 1b±k
Ketonea
v (8)
D (8)
u (8)
k (8)
Occ.b
Ê
d (A)
Ê
L (A)
Ê
D (A)
1b
1c
2.93
2.71
3.31
2.79
2.77
3.15
2.80
3.10
3.56
3.33
3.84
3.38
3.46
3.77
3.44
59
61
42
59
56
60
57
48
77
86
86
88
65
61
83
66
122
121
117
119
118
122
125
102
85
76
90
76
79
88
81
3.45
3.32
3.78
3.37
3.53
3.31
3.43
1dc
1fd
0.77
0.23
1g
1jd
2.73
2.83
3.22
3.03
3.38
3.52
64
59
60
61
59^2
0
34±64
83
74
85
69
80^7
90±120
76±92
124
128
99
123
123^3
180
113±117
76
81
3.30
3.40
0.67
0.33
1k
Averagee
Ideal10
3.66
3.42^0.06
3.70
87
78^2
0
3.39
2.77^0.04
2.72
2.54±2.90
3.39^0.06
#3.4
2.95±3.10
±
±
±
g-H-Exp.10
±
±
a Only the parameters for the closest methyl group and its closest hydrogen are shown.
b Occupancy factor to signify population of the disordered component.
c Four independent molecules in the asymmetric unit.
d Rotational disorder in the o-methyl groups.
e Average of best values for reacting ketones 1c±j.
close to 908, the abstraction distances d and L are com-
parably long. In fact, ketone 1k shows the longest abstrac-
tion distances in this series of compounds. The observed
high yields for 1k at room temperature are most likely due
to melting with a concomitant loss of topochemical control
resulting from the breakdown of the ordered crystal lattice.
Upon photolysis at 2208C, where melting is suppressed, no
reaction in ketone 1k was observed. The lack of reactivity of
the p-¯uoro ketone 1b in the solid state cannot be explained
suf®ciently by looking at the crystal structure. From the
observed abstraction parameters it is predicted to react
slowly, but a total lack of reactivity in the solid state is
unexpected. Analysis of the crystal packing pattern of
ketone 1b in comparison with the other ketones 1c±k
does not show any difference in the free space around a
molecule within the crystal. This excludes the possibility
of steric interactions between two molecules in the crystal
being responsible by preventing even the minimum motion
required for reaction to occur. On the other hand, looking at
the electronic features there is no reason to expect a p-¯uoro
substituent to have a large effect on the energy levels and
therefore alter the photochemical behavior of a phenyl
ketone.11 The question why ketone 1b does not react in
the solid state remains unsolved.
hydrogen shift to form an alkene has to overcome a more
strained 7-membered transition state.
Asymmetric induction
Salts S1±10 and S11±13 formed between carboxylic acids
1d and 1h and a number of commercially available, enan-
tiomerically pure amines (Table 3) were prepared and crys-
tallized. X-Ray quality crystals were obtained for three of
the saltsÐthe R-(1)-a,4-dimethylbenzylamine salt (S3),
the R-(1)-N,a-dimethylbenzylamine salt (S4) and the
(1S,2R)-(1)-norephedrine salt (S6).
Solid state irradiations were conducted for various periods
of time and at different temperatures on 5±10 mg crystalline
samples of salts S1±13 sandwiched between two Pyrex
plates.13 The extent of conversion was estimated by gas
chromatography following workup with diazomethane,
and the enantiomeric excesses in which photoproduct 2e
or 2j was formed in each case was determined using chiral
HPLC. The results of the photochemical studies in the solid
state are compiled in Table 3.
In several cases high enantiomeric excesses with over 90%
ee could be obtained. As expected, irradiation of salts S1
and S2 as well as S5 and S6 formed between carboxylic acid
1d and the two optical antipodes of the same amine give rise
to enantiomeric 2-indanols of similar ee. A well-behaved
system like this allows access to both enantiomers by simple
exchange of the ionic chiral auxiliary. An increase in
conversion leads to decreasing ee values, due to the loss
of topochemical control resulting from the breakdown of
the ordered crystal lattice as starting material converts into
photoproduct (although no melting was observable). As
shown for some salts in Table 3, low temperature photolysis
can be used to compensate this effect. By using different
solvents for recrystallization we were able to obtain
dimorphic crystals of salt S13, needles and ¯akes. Upon
photolysis the needles reacted faster and led to better
enantiomeric excess at low conversion. Comparison of
By comparing the hydrogen abstraction geometries given in
Table 2 with those from previous studies on the Norrish/
Yang type II reaction in the solid state,10 we ®nd an excel-
lent agreement. The abstraction distance d, as well as the
angles v, D and u, come close to the previously reported
values, and they show the same tendency as the two new
parameters L and k, introduced to take the rotational free-
dom of the methyl groups in a-mesitylacetophenones into
account. It seems logical that cyclization occurs even
though the distances between the two radical centers in
Ê
1,5 biradicals (Dˆ3.30±3.53 A) are longer than in the corre-
10
Ê
sponding 1,4-biradicals (Dˆ2.95±3.10 A). Unlike 1,4-
biradicals, in which cleavage competes strongly with cycli-
zation to a strained cyclobutane, 1,5 biradicals cyclize to
relatively strain-free rings, and the possible competing 1,6-

E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
Table 3. Solid-state photolysis results for chiral salts of keto acid 1d and 1h
6743
Entry
Keto-acid
Ionic chiral auxiliary
Time (min)
Temp. (8C)a
Convers. (%)b
ee of 2e/2j (%)c
HPLC Peakd
1
S1
1d
(S)-(2)-a-Methylbenzylamine
30
08
120
30
06
120
180
rt
rt
220
rt
rt
220
rt
69
16
61
53
18
74
89
69
94
83
66
92
75
83
S2
1d
1d
(R)-(1)-a-Methylbenzylamine
2
1
S3e
(R)-(1)-a,4-Dimethyl-
benzylamine
07
280
rt
rt
32
14
92
0
S4e
S5
1d
1d
(R)-(1)-N,a-Dimethyl-
benzylamine
(1R,2S)-(2)-Norephedrine
60
06
140
130
06
62
60
08
120
240
240
60
120
16
rt
rt
219
rt
rt
219
rt
rt
219
rt
rt
rt
56
12
21
80
12
30
62
12
38
13
81
40
63
12
10
2
82
97
95
80
90
98
4
16
12
39
13
19
43
69
77
0
1
2
2
S6e
S7
1d
1d
(1S,2R)-(1)-Norephedrine
(1S,2S)-(1)-2-Amino-1-phenyl-
1,3-propanediol
S8
S9
1d
1d
(1R,2S)-(1)-Aminoindanol
cis-(2)-Myrtanylamine
1
2
S10
1h
(S)-(2)-a-Methylbenzylamine
rt
rt
219
rt
1
310
210
S11
S12
1h
1h
(R)-(1)-N,a-Dimethyl-
benzylamine
(1R,2S)-(2)-Norephedrine
±
2
90
30
90
60
rt
rt
218
rt
69
36
17
86
37
48
55
18
S13
1h
(1R,2S)-(1)-Aminoindanol,
2
needles
Needles
Flakes
Flakes
05
45
335
rt
rt
rt
32
26
73
79
47
37
a rtˆambient temperature in photolysis box (30±408C).
b Percentage of total GC integral due to compound 1e or 1j. Remaining integral due .98% within experimental error to photoproduct 2e or 2j.
c Enantiomeric excesses measured by chiral HPLC.
d Peak 1 indicates major enantiomer is ®rst peak eluted from chiral HPLC column.
e Crystal structure obtained.
salts S1±10 prepared from para-carboxylic acid 1d, and
salts S11±13 formed between the same optically pure
amines and meta-carboxylic acid 1h reveals that enantio-
selectivity depends not only on the chiral ionic auxiliary but
on the location of the carboxylic acid group in the reactant
as well. Small structural changes in the acid as well as in the
chiral base have a considerable in¯uence on reactivity and
selectivity in the d-hydrogen abstraction-cyclization
sequence of a-mesitylacetophenones in the solid state. By
optimizing the chiral auxiliary used we were able to achieve
over 80% ee at 80±90% conversion, which indicates
synthetic value.
There is a plausible explanation as to how the observed solid
state enantioselectivity is governed by structural factors. In
mesitylacetophenone there are two `enantiotopic' methyl
groups available for d-hydrogen abstraction by the excited
ketone to form the 1,5-biradical. If we label the methyl
Scheme 4.

6744
E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
Figure 3.
group proximal to the re face of the carbonyl group as a and
the methyl group proximal to the si face as b as shown in
Scheme 4, abstraction of a hydrogen atom from group a(b)
should lead to the S(R)-enantiomer, assuming that, with
molecular motions severely restricted in the crystalline
state, rotation at the biradical stage around the single
bonds between the carbonyl group and the mesityl ring to
form the opposite enantiomer does not occur. The reactant
may adopt a conformation in the crystal lattice that situates
the carbonyl oxygen closer to one o-methyl group than the
other. In fact, the previously discussed crystal structures of
ketones 1b±k showed that the carbonyl group is indeed not
always perfectly perpendicular to the mesityl ring plane as
indicated by k±908 (Table 2).
preferentially from methyl group a, and this predicts that
the (S)-enantiomer of photoproduct 2e should be formed in
excess. Formation of the (R)-enantiomer would require a
1808 rotation of the acyl group, which is topochemically
unlikely in the crystalline state. These ®ndings clearly
explain the high enantioselectivity observed for salt S3.
By comparison with the geometric parameters for ketones
1b±k, salt S3 shows the shortest distances and best angular
values observed in this series. This no doubt helps to
account for the relatively short photolysis times needed
(Table 3).
Support for the picture presented above comes from the
crystal structure of salt S6. In this case, the absolute confor-
mation of the reacting molecule shows the carbonyl group
pointing towards Hb rather than Ha, as suggested by the
dihedral angle k being 1028, and hydrogen Hb is clearly
favored to be abstracted. Therefore photolysis of salt S6
leads to the optical antipode (peak 2) of indanol 2e, and
the assumption that was made earlier concerning the
restricted motion of the biradical in the solid state appears
justi®ed.
Fig. 3 shows the molecular conformations of a-mesityl-4-
carboxyacetophenone in the chiral crystal lattices of salts S3
and S6. Since the absolute con®gurations of the ammonium
cations are known, the absolute conformations of the react-
ing ketones can be assigned unequivocally. The abstraction
parameters are compiled in Table 4, together with those for
salt S4.
The X-ray crystal structure of salt S3 clearly shows a hydro-
gen atom on methyl group a being in a favorable position for
d-hydrogen abstraction (Table 4). The distance d between
Finally, the crystal structure of salt S4 explains why photo-
lysis of this material in the solid state gives only racemic
product 2e. Crystals of the R-(1)-N,a-dimethylbenzyl-
amine salt (S4) contain equal amounts of two independent
conformational isomers of the mesitylacetophenone moiety
in the asymmetric unit. The two conformers have a near
mirror image relationship and each slightly favors abstrac-
tion of the opposite d-hydrogen. Thus abstraction of Ha is
favored for one molecule and abstraction of Hb for the other,
Ê
the carbonyl oxygen and Ha is over 1 A shorter than the
corresponding distance to Hb, and the angles are also at
better settings for abstraction. The twisted conformation is
emphasized by the 158 deviation of the dihedral angle k
from its bisecting angle of 908. According to the observed
crystal structure data, hydrogen abstraction should occur
Table 4. Crystallographically derived CvO´´´Hd-abstraction geometries
Ê
d (A)
Ê
L (A)
Ê
D (A)
Salt
S3
d-Hydrogen
v (8)
D (8)
u (8)
k-90 (8)
215
21
Ha
Hb
Ha
Hb
2.60
3.66
3.07
3.14
3.11
3.05
3.29
3.74
50
45
58
37
34
51
90
53
60
80
90
52
128
114
126
126
124
126
3.46
3.69
3.32
3.79
3.79
3.26
S4a
0
Ha
Hb
3.71
12
0
S6
Ha
Hb
d-H
3.68
2.69
2.85^0.21
3.38
53
56
54^3
68
84
72^16
108
127
127^1
112
|7|^6
3.72
3.45
3.37^0.08
Aver.b
3.53^0.20
a Two independent molecules in the asymetric unit.
b Average values with standard deviations taken from the six reacting ketones 1c±j.

E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
6745
and the overall result is formation of racemic product 2e.
Similar observations have been made by us previously,
although in those cases some asymmetric induction was
still observed (ee,30% at low conversion) due to the
reaction taking place in the chiral environment of the crystal
lattice.14 A totally racemic outcome of the photolysis of salt
S4 was unexpected and may be due to a non-stereospeci®c
reaction course after all. It should also be noted that, with
the carbonyl group almost bisecting the mesityl ring (kˆ89/
928), the overall observed geometric parameters are poor for
hydrogen abstraction to occur, and this explains the lower
reactivity of salt S4 compared to salts S3 and S6.
5554). Column chromatography was performed by using
230±400 mesh silica gel (Merck 9384 or Silicycle) slurry
packed with the eluting solvent. For ¯ash chromatography
compressed air was used. High-pressure liquid chroma-
tography was performed on a Waters 600E system
controller connected to a tunable UV detector (Waters
486). For the determination of enantiomeric excesses, a
Chiralcelw OD column (250£4.4 mm2) from Chiral
Technologies, Inc. was used.
Preparation of starting materials
Mesitylacetic acid was prepared as described in Organic
Synthesis,15 and its acid chloride derivative was prepared
according to the procedure of Lutz and Hinkley.16 As far
as we are aware, the remainder of the compounds described
below are new.
Conclusion
In summary this paper demonstrates once more the extent to
which reactivity and selectivity in the solid state are
controlled by the conformation adopted by each component
in the crystal lattice. As a result, the ionic chiral auxiliary
method allows us to govern enantioselectivity in solid state
photochemical reactions in a straightforward way.
a-Mesityl-4-¯uoroacetophenone (1b). Following the
procedure described by Leibovitch et al.,3b 10 g (50 mmol)
of mesitylacetic acid chloride in 200 mL of anhydrous THF
was added to 31 mL of a 2 M diethyl ether solution of
4-¯uorobenzenemagnesium bromide (61 mmol) at 2788C.
After slow warming to room temperature, the red-brown
mixture was carefully hydrolized with water and acidi®ed
with 50% aqueous HCl. Standard workup (extraction with
diethyl ether, washing, drying, and removal of the solvents)
and recrystallization from ethanol afforded 9.2 g (36 mmol,
72%) of the desired ketone 1b as light yellow needles: mp
131±1338C; 1H NMR (400 MHz, CDCl3) d 8.11±8.07 (2H,
m, aryl-H), 7.19±7.13 (2H, m, aryl-H), 6.89 (2H, s, mesityl-
H), 4.29 (2H, s, CH2), 2.27 (3H, s, p-CH3), 2.17 (6H, s,
o-CH3); 13C NMR (75 MHz, CDCl3) d 20.24 (2), 20.90
(2), 39.18 (1), 115.55 and 115.84 (2JC±Fˆ22 Hz, negative),
128.10 (2), 128.89 (2), 130.59 and 130.71 (3JC±Fˆ6 Hz,
negative), 133.49 and 133.54 (4JC±Fˆ3 Hz, positive),
Experimental
General information
Commercial spectral grade solvents were used for photo-
chemical experiments unless otherwise stated. For synthetic
use tetrahydrofuran, diethyl ether, hexane and benzene were
dried over sodium/benzophenone complex; dimethyl sulfate
and dichloro methane were dried over calcium hydride.
Most starting materials were purchased from Sigma-Aldrich
Canada, Ltd., and some were bought from Lancaster
Synthesis, Inc.
136.39 (1), 136.72 (1), 164.03 and 167.40 (1JC±F
ˆ
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 ®lms between two NaCl 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 (Bruker) and 300 MHz (Varian) instruments.
13C spectra were run under broadband 13C{1H} decoupling
and assignments, where given, are supported by the attached
proton test. For APT, (1) signi®es C or CH2 and (2)
signi®es 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 spectro-
meter with xenon bombardment 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 analysis
was performed on a Hewlett-Packard 5890 instrument ®tted
with a ¯ame ionization detector and a Hewlett-Packard 3392
integrator. The following 15 m£0.25 mm capillary columns
from J & W Scienti®c, Inc. were used: DB-5, HP-5.
Analytical thin-layer chromatography was carried out on
commercial precoated silica gel plates (Merck, type
225 Hz, positive), 195.53(1); IR (KBr) 2916, 1680
(CvO), 1600, 1506, 1486, 1413, 1327, 1229 (C±F), 1201,
1159, 994, 852, 841, 815, 604 cm21; MS (EI) m/z 256
(M1), 133 (base) 123, 95; HRMS 256.12601 (calcd for
C17H17FO, 256.12634). Anal. Calcd for C17H17FO: C,
79.66; H, 6.69. Found: C, 79.55; H, 6.61.
a-Mesityl-4-cyanoacetophenone (1c). In a ¯ame-dried
100 mL ¯ask equipped with condenser and stirrer, 2.69 g
(10.5 mmol) of ketone 1b and 1.08 g (21.6 mmol) of sodium
cyanide were dissolved in 70 mL of anhydrous dimethyl
sulfoxide. The mixture was stirred at 1208C for 18 h until
no starting material remained, as indicated by TLC (20%
ethyl acetate±petroleum ether 35±60). The dark red
solution was poured into 100 mL of water and extracted
with ether (3£200 mL). Standard workup afforded a yellow
solid, which was puri®ed by ¯ash chromatography (SiO2,
15% ethyl acetate±petroleum ether 35±60). Recrystalli-
zation from methanol afforded 2.13 g (8.1 mmol, 77%) of
the desired ketone 1c as light yellow prisms: mp 129±
1318C; 1H NMR (400 MHz, CDCl3) d 8.13 (2H, d,
Jˆ8 Hz, aryl-H), 7.79 (2H, d, Jˆ8 Hz, aryl-H), 6.90 (2H,
s, mesityl-H), 4.32 (2H, s, CH2), 2.28 (3H, s, p-CH3), 2.17
(6H, s, o-CH3); 13C NMR (75 MHz, CDCl3) d 20.2 (2),
20.9 (2), 39.6 (1), 116.3 (1), 117.9 (1), 128.9 (2),
128.4 (2), 128.5 (2) 132.5 (1), 136.6 (1), 136.7 (1),

6746
E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
139.8 (1), 196.0 (1); IR (KBr) 2921, 2232 (CN), 1686
(CvO), 1608, 1485, 1404, 1327, 1292, 1214, 995, 858,
838, 823, 700, 584 cm21; MS (EI) m/z 263 (M1), 133
(base) 115, 91; HRMS 263.13056 (calcd for C18H17NO,
263.13101). Anal. Calcd for C18H17NO: C, 82.09; H, 6.51;
N, 5.32. Found: C, 82.10; H, 6.36; N, 5.21.
workup afforded after chromatography (silica gel, 2% ethyl
acetate±petroleum ether 35±60) and recrystallization from
methanol 444 mg (1.7 mmol, 41%) of the desired ketone 1f
1
as colorless cubes: mp 110±1138C; H NMR (400 MHz,
CDCl3) d 7.85 (1H, dt, Jˆ8, 1 Hz, aryl-H), 7.74±7.71
(1H, m, aryl-H), 7.50±7.45 (1H, m, aryl-H), 7.31±7.26
(1H, m, aryl-H), 6.89 (2H, s, mesityl-H), 4.29 (2H, s,
CH2), 2.27 (3H, s, p-CH3), 2.17 (6H, s, o-CH3); 13C NMR
(75 MHz, CDCl3) d 20.26 (2), 20.94 (2), 39.45 (1),
114.70 and 114.99 (2JC±Fˆ22 Hz, negative), 120.00 and
120.28 (2JC±Fˆ22 Hz, negative), 123.74 and 123.78
(4JC±Fˆ3 Hz, negative), 128.85 (2), 130.28 and 130.39
(3JC±Fˆ8 Hz, negative), 136.51 (1), 136.73 (1), 139.09
and 139.17 (3JC±Fˆ6 Hz, positive), 161.22 and 164.51
(1JC±Fˆ248 Hz, positive), 195.92(1); IR (KBr) 2915,
1682 (CvO), 1589, 1486, 1442, 1413, 1331, 1251 (C±F),
1169, 1152, 875, 787, 681 cm21; MS (EI) m/z 256 (M1),
133 (base), 123; HRMS 256.12615 (calcd for C17H17FO,
256.12634). Anal. Calcd for C17H17FO: C, 79.66; H, 6.69.
Found: C, 79.58; H, 6.77.
a-Mesityl-4-carboxyacetophenone (1d). To a 500 mL
¯ask equipped with a condenser and stirrer was added
2.7 g (10.3 mmol) of ketone 1c, 75 mL of a 25% aqueous
sodium hydroxide solution and 250 mL of methanol. The
mixture was re¯uxed for 4 h, cooled to room temperature
and strongly acidi®ed with 50% aqueous HCl. The aqueous
layer was extracted with diethyl ether (1.5 L), and standard
workup afforded after recrystallization from benzene 2.85 g
(8.0 mmol, 78%) of the desired keto acid 1d as colorless ®ne
1
needles: mp 185±1888C; H NMR (400 MHz, CDCl3) d
11.1 (1H, s (br), COOH), 8.59 (2H, d, Jˆ8 Hz, aryl-H),
8.49 (2H, d, Jˆ8 Hz, aryl-H), 7.24 (2H, s, mesityl-H),
4.71 (2H, s, CH2), 2.63 (3H, s, p-CH3), 2.55 (6H, s,
o-CH3); 13C NMR (100 MHz, CDCl3) d 20.3 (2), 20.9
(2), 39.8 (1), 128.1 (2), 128.8 (2), 128.9 (2), 130.6
(2) 133.0 (1), 136.6 (1), 136.8 (1), 141.1 (1), 171.2
(1), 196.9 (1); IR (KBr) 2918, 2601, 2454, 1689 (CvO,
COOH), 1425, 1291, 1214, 995 cm21; MS (EI) m/z 282
(M1), 149, 133 (base); HRMS 282.12609 (calcd for
C18H18O3, 282,12558). Anal. Calcd for C18H18O3: C,
76.57; H, 6.43. Found: C, 76.37; H, 6.47.
a-Mesityl-3-cyanoacetophenone (1g). 2-Mesitylethanol
(5.8 g, 35.4 mmol) was converted into the corresponding
mesitylacetaldehyde by standard Swern oxidation (4.5 g,
78% after ¯ash chromatography on silica gel, eluent 10%
ethyl acetate±petroleum ether 35±60).17 A solution of 1.0 g
(6.4 mmol) of mesitylacetaldehyde in 5 mL of THF was
added slowly at 2908C (diethyl ether±liquid nitrogen) to
a THF±hexane solution of 3-cyanophenyl lithium (freshly
prepared from 3-bromobenzonitrile (1.2 g) and 4.3 mL of a
1.6 M in THF solution of n-butyl lithium following the
procedures described by Parham et al.18). After slow
warming to room temperature the mixture was carefully
hydrolized with water and acidi®ed with 50% aqueous
HCl. Standard workup afforded 1.6 g (6.1 mmol, 95%) of
the secondary alcohol as a yellow oil. Without further
puri®cation, this material was converted into ketone 1g by
standard pyridinium chlorochromate oxidation (PCC on
Celite in dichloro methane).19 Flash chromatography (10±
20% ethyl acetate±petroleum ether 35±60) yielded 1.0 g
(3.8 mmol, 58%) of ketone 1g as a light yellow powder.
Recrystallization from ethanol gave colorless prisms: mp
a-Mesityl-4-carbomethoxyacetophenone (1e). A solution
of 284 mg (1.0 mmol) of keto acid 1d and 100 mg of
p-toluensulfonic acid in 15 mL of methanol was re¯uxed
for 4.5 h until no starting material remained as indicated
by TLC (20% ethyl acetate±petroleum ether 35±60). The
desired keto ester 1e crystallized in ®ne needles from the
reaction mixture. Recrystallization from methanol afforded
191 mg (0.7 mmol, 64%) of the desired keto-ester as color-
less ®ne needles: mp 127±1288C; 1H NMR (400 MHz,
CDCl3) d 8.16±8.08 (4H, m, aryl-H), 6.89 (2H, s, mesityl-
H), 4.33 (2H, s, CH2), 3.95 (3H, s, COOCH3), 2.27 (3H, s,
p-CH3), 2.17 (6H, s, o-CH3); 13C NMR (100 MHz, CDCl3) d
20.3 (2), 20.9 (2), 39.7 (1), 52.4 (2), 127.9 (2), 128.9
(2), 129.9 (2), 133.9 (1), 136.5 (1), 136.7 (1), 140.4 (1),
166.2 (1), 196.8 (1); IR (KBr) 2953, 1718 (COOCH3),
1685 (CvO), 1486, 1440, 1407, 1328, 1288 (CH3), 1213,
1198, 1114, 995, 856, 766, 696 cm21; MS (EI) m/z 296
(M1), 163, 133 (base); HRMS 296.14156 (calcd for
C19H20O3, 296.14124). Anal. Calcd for C19H20O3: C,
77.00; H, 6.80. Found: C, 76.84; H, 6.75.
1
134±1368C; H NMR (400 MHz, CDCl3) d 8.33 (1H, t,
Jˆ1 Hz, aryl-H), 8.27 (1H, dt, Jˆ8, 1 Hz, aryl-H), 7.86
(1H, dt, Jˆ8, 1 Hz, aryl-H), 7.63 (1H, t, Jˆ8 Hz, aryl-H),
6.90 (2H, s, mesityl-H), 4.31 (2H, s, CH2), 2.26 (3H, s,
p-CH3), 2.16 (6H, s, o-CH3); 13C NMR (75 MHz, CDCl3)
d 20.2 (2C, negative), 20.9 (2), 39.4 (1), 113.2 (1), 117.9
(1), 128.3 (1), 128.9 (2), 129.7 (2) 131.7 (2), 132.0 (2),
136.0 (2), 136.6 (1), 136.7 (1), 137.7 (1), 195.2 (1); IR
(KBr) 2915, 2230 (CN), 1699 (CvO), 1482, 1451, 1337,
1239, 1172, 1006, 856, 803, 739, 682 cm21; MS (EI) m/z
263 (M1), 133 (base) 115, 91; HRMS 263.13069 (calcd for
C18H17NO, 263.13101). Anal. Calcd for C18H17NO: C,
82.09; H, 6.51; N, 5.32. Found: C, 81.96; H, 6.61; N, 5.36.
a-Mesityl-3-¯uoroacetophenone (1f). To 260 mg of
magnesium turnings, activated with 10 mL of 1,2-dibromo-
ethane or iodine, in 5 mL of anhydrous ether was added
1.04 g (5.9 mmol) of 1-bromo-3-¯uoro benzene at a rate
to keep the ether at re¯ux. After the addition was complete,
the mixture was re¯uxed for another 20 min. An additional
15 mL of anhydrous ether was added and the solution was
cooled to 2788C (acetone/dry ice), whereupon 825 mg
(4.2 mmol) of mesitylacetic acid chloride dissolved in
5 mL of ether was transferred into the cooled ¯ask. After
slow warming to room temperature the mixture was care-
fully hydrolized with 15 mL 1N hydrochloric acid. The
aqueous layer was extracted with diethyl ether and standard
a-Mesityl-3-carboxyacetophenone (1h). To a 500 mL
¯ask equipped with a condenser and stirrer was added
958 mg (4.6 mmol) of ketone 1g, 25 mL of a 25% aqueous
sodium hydroxide solution and 85 mL of methanol. The
mixture was re¯uxed for 4 days, then cooled to room
temperature and strongly acidi®ed with aqueous 50% HCl.
The aqueous layer was extracted with diethyl ether (0.5 L)

E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
6747
and standard workup afforded 1.56 g (4.4 mmol, 95%) of
desired keto acid 1h as a white powder. Recrystallization
from benzene gave colorless ®ne needles: mp 197±1998C;
1H NMR (400 MHz, CDCl3) d 8.81 (1H, t, Jˆ1 Hz, aryl-H),
8.34 (H, dt, Jˆ8, 1 Hz, aryl-H), 8.30 (1H, dt, Jˆ8, 1 Hz,
aryl-H), 7.63 (1H, t, Jˆ8 Hz, aryl-H), 6.90 (2H, s, mesityl-
H), 4.39 (2H, s, CH2), 2.19 (3H, s, p-CH3), 2.17 (6H, s,
o-CH3); 13C NMR (100 MHz, CDCl3) d 20.2 (2), 20.9
(2), 39.9 (1), 128.8 (2), 128.9 (2), 129.2 (2), 130.0
(2) 133.0 (2), 136.5 (1), 136.8 (1), 137.6 (1), 170.8
(1), 196.3 (1); IR (KBr) 2918, 2654, 2590, 1685/1604
(CvO, COOH), 1446, 1413, 1312, 1216, 752 cm21; MS
(EI) m/z 282 (M1), 149, 133 (base); HRMS 282.12521
(calcd for C18H18O3, 282.12558). Anal. Calcd for
C18H18O3: C, 76.57; H, 6.43. Found: C, 76.61; H, 6.55.
Pyrex immersion well. For preparative-scale liquid-phase
photolyses, Pyrex tubes were used and the solution was
purged with nitrogen for at least 30 min prior to photolysis.
Workup consisted of solvent removal in vacuum followed
by silica gel radial chromatography. For solid-state photo-
lyses, 5±10 mg samples were 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 microscope plates were Scotch-
taped together at the top and bottom ends, place in a poly-
ethylene bag, and thoroughly degassed with nitrogen. The
bag was then sealed under a positive pressure of nitrogen
with a heat-sealing device and placed 5 cm from the immer-
sion well. Low temperature solid-state photolyses were
performed by placing the double sealed bags as close as
possible to the immersion well in a cryostated methanol
bath (Neslab Cryocool CC-100 II). Temperatures were
maintained within 28C of the designated values. Workup
of the solid-state photolysis samples consisted of washing
the solid off the plate with ether into a suitable container and
determining the conversion by gas chromatography.
a-Mesityl-3-carbomethoxyacetophenone (1j). A sample
of 400 mg of keto acid 1h (1.4 mmol) was treated with an
approximately 0.5 M solution of diazomethane in diethyl
ether. Radial chromatography afforded 379 mg of keto
ester 1j (1.3 mmol, 90%) as a white powder. Recrystalli-
zation from diisopropyl ether gave colorless cubes: mp
123±1258C; 1H NMR (400 MHz, CDCl3) d 8.72±8.71
(1H, m, aryl-H), 8.27±822 (2H, m, aryl-H), 7.58 (1H, t,
Jˆ8 Hz, aryl-H), 6.89 (2H, s, mesityl-H), 4.36 (2H, s,
CH2), 3.96 (3H, s, COOCH3), 2.27 (3H, s, p-CH3), 2.18
(6H, s, o-CH3); 13C NMR (100 MHz, CDCl3) d 20.3 (2),
20.9 (2), 39.5 (1), 52.4 (2), 128.8 (2), 128.9 (2), 129.1
(2), 130.7 (1), 132.2 (2), 133.7 (2), 136.4 (1), 136.7 (1),
166.3 (1), 196.3 (1); IR (KBr) 2909, 1724 (COOCH3),
1693 (CvO), 1488, 1441, 1327, 1306, 1276 (CH3), 1210,
1190, 1117, 1075, 1000, 958, 853, 753, 683 cm21; MS (EI)
m/z 296 (M1), 163, 133 (base); HRMS 296.14093 (calcd
for C19H20O3, 296.14124). Anal. Calcd for C19H20O3: C,
77.00; H, 6.80. Found: C, 77.32; H, 6.68.
2-(40-Fluorophenyl)-2-hydroxy-4,6-dimethylindane (2b).
A solution of 254 mg of ketone 1b in 40 mL of acetonitrile
was irradiated through Pyrex for 7 h. GC indicated a 100%
conversion, and radial chromatography (silica gel, 20%
ethyl acetate±petroleum ether 35±60) of the reaction
mixture afforded 163 mg of indanol 2b (64%) as a colorless
oil.1H NMR (400 MHz, CDCl3) d 7.61±7.50 (2H, m, aryl-
H), 7.10±7.02 (2H, m, aryl-H), 6.90 (1H, s, indanyl-H), 6.87
(1H, s, indanyl-H), 3.31 (2H, AB quart., Jˆ16 Hz, CH2),
3.24 (2H, AB quart., Jˆ16 Hz, CH2), 2.35 (3H, s, CH3), 2.25
(3H, s, CH3); 13C NMR (100 MHz, CDCl3) d 18.98 (2),
21.19 (2), 47.4 (1), 49.48 (1), 82.68 (1), 114.86 and
115.07 (2JC±Fˆ21 Hz, negative), 122.97 (2), 126.84 and
126.92 (3JC±Fˆ8 Hz, negative), 128.75 (2), 134.07 (1),
136.51 (1), 136.87 (1), 140.59 (1), 141.68 (1), 160.65
and 163.09 (1JC±Fˆ245 Hz, positive); IR (neat) 3401
(OH), 2916, 1604, 1510, 1481, 1230 (C±F), 1161, 1046,
837, 815, 574 cm21; MS (EI) m/z 256 (M1), 133 (base)
123, 95; HRMS 256.12647 (calcd for C17H17FO,
256.12634). Anal. Calcd for C17H17FO: C, 79.66; H, 6.69.
Found: C, 79.70; H, 6.67.
a-Mesityl-3-methylacetophenone (1k). To 931 mg
(4.7 mmol) of mesitylacetic acid chloride in 20 mL of anhy-
drous THF was added 5.2mL of a 1 M solution of 3-tolyl-
magnesium chloride (61mmol) in THF at 2788C. After
slow warming to room temperature, the red±brown mixture
was carefully hydrolyzed with water and acidi®ed with
aqueous 50% HCl. Standard workup and puri®cation by
¯ash chromatography (silica gel, 2% ethyl acetate±petro-
leum ether 35±60) afforded 426 mg (1.7 mmol, 36%) of the
desired ketone 1k. Recrystallization from ethanol afforded
colorless cubes: mp 75±788C; 1H NMR (400 MHz, CDCl3)
d 7.91±7.87 (2H, m, aryl-H), 7.43±7.37 (2H, m, aryl-H),
6.90 (2H, s, mesityl-H), 4.36 (2H, s, CH2), 2.45 (3H, s, aryl-
CH3), 2.27 (3H, s, p-mesityl-CH3), 2.17 (6H, s, o-mesityl-
CH3); 13C NMR (75 MHz, CDCl3) d 20.2 (2), 20.9
(2),21.4 (2) 39.3 (1), 125.2 (2), 128.5 (2), 128.7 (2),
129.3 (1), 133.8 (2), 136.2 (2), 136.8 (2), 137.1 (1),
138.4 (1), 197.2 (1); IR (KBr) 2918, 1683 (CvO), 1601,
1486, 1423, 1324, 1246, 1158, 861, 790, 692 cm21; MS (EI)
m/z 252 (M1), 133, 119 (base). 91; HRMS 252.15211
(calcd for C18H20O, 252.15141). Anal. Calcd for
C19H20O3: C, 85.67; H, 7.99. Found: C, 85.79; H, 8.02.
2-(40-Cyanophenyl)-2-hydroxy-4,6-dimethylindane (2c).
A solution of 200 mg of ketone 1c in 40 mL of acetonitrile
was irradiated through Pyrex for 4 h. GC indicated a 100%
conversion, and after evaporation of the solvent, recrystal-
lization of the resulting yellow oil from hexane afforded
120 mg of indanol 2c (60%) as a colorless solid: mp 131±
1
1328C, H NMR (400 MHz, CDCl3) d 7.68±7.61 (4H, m,
aryl-H), 6.91 (1H, s, indanyl-H), 6.88 (1H, s, indanyl-H),
3.33 (2H, AB quart., Jˆ16 Hz, CH2), 3.26 (2H, AB quart.,
Jˆ16 Hz, CH2), 2.31 (3H, s, CH3), 2.22 (3H, s, CH3); 13C
NMR (100 MHz, CDCl3) d 19.0 (2), 21.2 (2), 48.4 (1),
49.6 (1), 82.7 (1), 110.8 (1), 118.8 (1), 123.0 (2), 125.9
(2), 129.0 (2), 132.1 (2), 134.2 (1), 135.9 (1), 137.2 (1),
140.0 (1), 151.4 (1); IR (KBr) 3408 (OH), 2947, 2901,
2234 (CN), 1610, 1505, 1478, 1406, 1377, 1219, 1051,
960, 852, 830, 591, 565 cm21; MS (EI) m/z 263 (M1),
133 (base); HRMS 263.13135 (calcd for C18H17NO,
263.13101). Anal. Calcd for C18H17NO: C, 82.09; H, 6.51;
N, 5.32. Found: C, 81.88; H, 6.61; N, 5.38.
Photochemical procedures
All photolyses were performed using a 450 W Hanovia
medium-pressure mercury lamp placed in a water-cooled,

6748
E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
2-(40-Carbomethoxyphenyl)-2-hydroxy-4,6-dimethylindane
(2e). A solution of 200 mg of ketone 1e in 40 mL of aceto-
nitrile was irradiated through Pyrex for 4 h. GC indicated a
100% conversion, and radial chromatography (silica gel,
10% ethyl acetate±petroleum ether 35±60) of the reaction
mixture afforded 120 mg of indanol 2e (60%) as a colorless
oil. 1H NMR (400 MHz, CDCl3) d 8.01 (2H, dt, Jˆ9, 2 Hz,
aryl-H), 7.63 (2H, dt, Jˆ9, 2 Hz, aryl-H), 6.91 (1H, s,
indanyl-H), 6.87 (1H, s, indanyl-H), 3.90 (3H, s, OCH3),
3.35 (2H, AB quart., Jˆ16 Hz, CH2), 3.31 (2H, AB quart.,
Jˆ16 Hz, CH2), 2.38 (1H, s (br), OH), 2.31 (3H, s, CH3),
2.22 (3H, s, CH3); 13C NMR (100 MHz, CDCl3) d 19.0 (2),
21.2 (2), 48.1 (1), 49.8 (1), 52.0 (2), 82.9 (1), 122.9 (2),
125.1 (2), 128.8 (2), 129.6 (2), 134.1 (1), 136.3 (1),
136.9 (1), 140.4 (1), 151.2 (1), 166.9 (1); IR (neat)
3441 (OH), 2949, 1723 (COOMe), 1610, 1436, 1407,
1280, 1192, 1110, 1020, 857, 776, 710 cm21; MS (EI) m/z
296 (M1), 278, 263, 163 (base), 133; HRMS 296.14158
(calcd for C19H20O3, 296.14124). Anal. Calcd for
C19H20O3: C, 77.00; H, 6.80. Found: C, 77.01; H, 6.77.
indanyl-H), 3.33 (2H, AB quart., Jˆ16 Hz, CH2), 3.25
(2H, AB quart., Jˆ16 Hz, CH2), 2.31 (3H, s, CH3), 2.22
(3H, s, CH3); 13C NMR (75 MHz, CDCl3) d 19.0 (2),
21.2 (2), 48.1 (1), 49.7 (1), 82.4 (1), 112.2 (1), 118.9
(1), 122.6 (2), 128.9 (2), 129.0 (2), 129.7 (2), 130.7 (2),
134.1 (1), 135.9 (1), 137.2 (1), 140.0 (1), 147.6 (1); IR
(KBr) 3450 (OH), 2916, 2227 (CN), 1477, 1438, 1376,
1342, 1320, 1284, 1247, 1152, 1048, 904, 847, 802, 693,
606, 543, 523 cm21; MS (EI) m/z 263 (M1), 245, 133
(base); HRMS 263.13097 (calcd for C18H17NO,
263.13101). Anal. Calcd for C18H17NO: C, 82.09; H, 6.51;
N, 5.32. Found: C, 81.95; H, 6.40; N, 5.22.
2-(30-Carbomethoxyphenyl)-2-hydroxy-4,6-dimethyl-
indane (2j). A solution of 114 mg of ketone 1j in 35 mL of
acetonitrile was irradiated through Pyrex for 2.5 h. GC indi-
cated a 100% conversion, and radial chromatography (silica
gel, 5% ethyl acetate±petroleum ether 35±60) of the
reaction mixture afforded 85 mg of indanol 2j (75%) as a
colorless solid: mp 151±1538C; 1H NMR (400 MHz,
CDCl3) d 8.26 (1H, t, Jˆ2 Hz, aryl-H), 7.95 (1H, dt, Jˆ8,
1 Hz, aryl-H), 7.80±7.78 (1H, m, aryl-H), 7.43 (1H, t,
Jˆ8 Hz, aryl-H), 6.93 (1H, s, indanyl-H), 6.87 (1H, s,
indanyl-H), 3.90 (3H, s, OCH3), 3.37 (2H, AB quart.,
Jˆ16 Hz, CH2), 3.29 (2H, AB quart., Jˆ16 Hz, CH2),
2.31 (3H, s, CH3), 2.26 (1H, s (br), OH), 2.23 (3H, s,
CH3); 13C NMR (75 MHz, CDCl3) d 19.0 (2), 21.2 (2),
48.0 (1), 49.6 (1), 52.1 (2), 82.8 (1), 123.0 (2), 126.4
(2), 128.3 (2), 128.4 (2), 128.8 (2), 129.8 (2), 130.1 (1),
134.1 (1), 136.4 (1), 136.9 (1), 140.4 (1), 146.3 (1),
167.1 (1); IR (KBr) 3461 (OH), 1697 (COOMe), 1447,
1347, 1294, 1232, 1209, 848, 763, 695 cm21; MS (EI) m/z
296 (M1), 278, 263, 163 (base), 133; HRMS 296.14064
(calcd for C19H20O3, 296.14124). Anal. Calcd for
C19H20O3: C, 77.00; H, 6.80. Found: C, 76.89; H, 6.66.
2-(40-Carboxyphenyl)-2-hydroxy-4,6-dimethylindane (2c).
A solution of 30 mg of ketone 1d in 5 mL of acetonitrile was
irradiated through Pyrex for 1 h. GC indicated, after treat-
ment with diazomethane (0.5 M in diethyl ether), a 100%
conversion, and radial chromatography (silica gel, 10%
ethyl acetate±petroleum ether 35±60) of the reaction
mixture afforded ester indanol 2e.
2-(30-Fluorophenyl)-2-hydroxy-4,6-dimethylindane (2f).
A solution of 110 mg of ketone 1f in 40 mL of acetonitrile
was irradiated through Pyrex for 1 h. GC indicated a 100%
conversion, and radial chromatography (silica gel, 5% ethyl
acetate±petroleum ether 35±60) of the reaction mixture
1
afforded 105 mg of indanol 2f (96%) as a colorless oil. H
NMR (400 MHz, CDCl3) d 7.33±7.29 (3H, m, aryl-H),
6.99±6.94 (1H, m, aryl-H), 6.93 (1H, s, indanyl-H), 6.88
(1H, s, indanyl-H), 3.34 (2H, AB quart., Jˆ16 Hz, CH2),
3.26 (2H, AB quart., Jˆ16 Hz, CH2), 2.33 (3H, s, CH3), 2.24
(3H, s, CH3); 13C NMR (75 MHz, CDCl3) d 18.98 (2),
21.19 (2), 47.95 (1), 49.59 (1), 82.66 (1), 112.32 and
112.61 (2JC±Fˆ22 Hz, negative), 113.67 and 113.92
(2JC±Fˆ21 Hz, negative), 120.65 and 120.69 (4JC±Fˆ
3 Hz, negative), 122.93 (2), 128.76, 129.66 and 129.77
(3JC±Fˆ8 Hz, negative), 134.06 (1), 136.35 (1), 136.91
(1), 140.41 (1), 148.69 and 148.76 (3JC±Fˆ7 Hz, nega-
tive), 161.16 and 164.42 (1JC±Fˆ244 Hz, positive); IR
(neat) 3370 (OH), 2916, 1615, 1588, 1484, 1442, 1377,
1273, 1214, 1046, 873, 843, 786, 696 cm21; MS (EI) m/z
256 (M1), 238, 223, 133 (base) 123; HRMS 256.12610
(calcd for C17H17FO, 256.12634). Anal. Calcd for
C17H17FO: C, 79.66; H, 6.69. Found: C, 79.58; H, 6.77.
2-(30-Carboxyphenyl)-2-hydroxy-4,6-dimethylindane (2h).
A solution of 30 mg of ketone 1h in 5 mL of acetonitrile was
irradiated through Pyrex for 1h. GC indicated, after treat-
ment with diazomethane (0.5 M in diethyl ether), a 100%
conversion, and radial chromatography (silica gel, 5% ethyl
acetate±petroleum ether 35±60) of the reaction mixture
afforded ester indanol 2j.
2-(30-Methylphenyl)-2-hydroxy-4,6-dimethylindane (2k).
A solution of 136 mg of ketone 1k in 35 mL of acetonitrile
was irradiated through Pyrex for 1 h 45 min. GC indicated a
98% conversion, and radial chromatography (silica gel, 5%
ethyl acetate±petroleum ether 35±60) of the reaction
mixture afforded 129 mg of indanol 2j (95%) as a colorless
oil: 1H NMR (400 MHz, CDCl3) d 7.44 (1H, s, aryl-H), 7.38
(1H, d, Jˆ8 Hz, aryl-H), 7.28 (1H, t, Jˆ8 Hz, aryl-H), 7.12
(1H, d, Jˆ8 Hz, aryl-H), 6.95 (1H, s, indanyl-H), 6.90 (1H,
s, indanyl-H), 3.37 (2H, AB quart., Jˆ16 Hz, CH2), 3.29
(2H, AB quart., Jˆ16 Hz, CH2), 2.40 (3H, s, aryl-CH3),
2.33 (3H, s, indanyl-CH3), 2.26 (3H, s, indanyl-CH3),
2.05(1H, s (br), OH); 13C NMR (75 MHz, CDCl3) d 19.0
(2), 21.2 (2), 21.6 (2), 47.7 (1), 49.4 (1), 83.0 (1), 122.1
(2), 122.9 (2), 125.9 (2), 127.8 (2), 128.2 (2), 128.6 (2),
134.0 (1), 136.7 (1), 136.7 (1), 137.8 (1), 140.8 (1),
145.8 (1); IR (KBr) 3392 (OH), 2916, 1608, 1482, 1377,
1251, 1212, 1046, 901, 849, 829, 786, 706 cm21; MS (EI)
m/z 296 (M1), 278, 263, 163 (base), 133; HRMS 252.15157
2-(30-Cyanophenyl)-2-hydroxy-4,6-dimethylindane (2g).
A solution of 121 mg of ketone 1g in 40 mL of acetonitrile
was irradiated through Pyrex for 21 h. GC indicated a 95%
conversion, and radial chromatography (silica gel, 10%
ethyl acetate±petroleum ether 35±60) of the reaction
mixture afforded 68 mg of indanol 1g (57%) as a light
1
yellow solid: mp 107±1108C, H NMR (400 MHz, CDCl3)
d 7.89 (1H, t, Jˆ1 Hz, aryl-H), 7.78 (1H, dt, Jˆ8, 1 Hz,
aryl-H), 7.55 (1H, dt, Jˆ8, 1 Hz, aryl-H), 7.45 (1H, t,
Jˆ8 Hz, aryl-H), 6.91 (1H, s, indanyl-H), 6.88 (1H, s,

E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
6749
(calcd for C18H20O, 252.15141). Anal. Calcd for C19H20O3:
C, 85.67; H, 7.99. Found: C, 85.45; H, 8.02.
washed with ether to give 129 mg (0.30 mmol, 96%) of salt
S5 (colorless powder, methanol): mp 168±1708C (isopropa-
nol), MS (FAB, 1LSIMS) m/z 434 (M111), 283, 260, 152,
134, 118 (matrix: thioglycerol); Anal. Calcd for C27H31NO4:
C, 74.79; H, 7.21; N, 3.23. Found: C, 74.77; H, 7.14; N,
3.20.
Preparation of salts
Salts were prepared by dissolving equimolar amounts of
keto acid and amine in the solvent indicated. The initial
precipitates were collected and recrystallized from the
indicated solvents. Satisfactory NMR and IR spectra were
recorded for all salts. The NMR spectra showed the change
in the carboxylic acid carbonyl signal expected for a carb-
oxylate anion. In the IR, salt formation resulted in a
characteristic change in the carboxylic acid OH stretching
band at 3500±2400 cm21, which was replaced by multiple
combination bands for ammonium in the 3200±2220 cm21
region. In addition, the carboxylic acid carbonyl stretching
vibration at 1690 cm21 was replaced by two bands charac-
teristic for carboxylate: a strong absorption in the 1650±
The (1S,2R)-(1)-norephedrine salt S6 was prepared ana-
logously (95% yield) and also gave satisfactory MS and
elemental analysis results. X-Ray quality crystals were
grown by dissolving the crude salt in a mixture of metha-
nol/ethanol 3:1 and slow evaporation of the solvents.
(1S,2S)-(1)-2-Amino-1-phenyl-1,3-propandiol salt S7 of
keto acid 2d. A solution of 67 mg (0.24 mmol) of keto acid
2d in 5 mL of diethyl ether was added to a solution of 49 mg
of (1S,2S)-(1)-2-amino-1-phenyl-1,3-propandiol (Aldrich)
in 5 mL of diethyl ether. The resulting precipitate was
®ltered off and washed with ether to give 91 mg
(0.20 mmol, 87%) of salt S7: mp 176±1788C (needles,
benzene/ethyl acetate 1:1), MS (FAB, 1LSIMS) m/z 450
(M111), 283, 265, 168, 150, 133, 120 (matrix: thio-
glycerol); Anal. Calcd for C27H31NO5: C, 72.12; H, 6.95;
N, 3.12. Found: C, 71.82; H, 6.94; N, 3.07.
1550 cm21 region, and a weaker one near 1400 cm21
.
(S)-(2)- and (R)-(1)-a-Methylbenzylamine salt S1 and
S2 of keto acid 2d. To a solution of 50 mg (0.18 mmol) of
keto acid 2d in 5 mL of diethyl ether was added 50 mL of
optically pure (S)-(2)-a-methylbenzylamine (Aldrich). The
resulting precipitate was ®ltered off and washed with ether
to give 59 mg (0.15 mmol, 83%) of salt S1 (colorless
powder, ethyl acetate): mp 142±1458C (CH3CN), MS
(FAB, 1LSIMS) m/z 404 (M111), 283, 230, 149, 133,
122, 105 (matrix: thioglycerol); Anal. Calcd for
C26H29NO3: C, 77.38; H, 7.25; N, 3.47. Found: C, 77.60;
H, 7.27; N, 3.32.
(1R,2S)-(1)-cis-1-Amino-2-indanol salt S8 of keto acid
2d. To a solution of 88 mg (0.31 mmol) of keto acid 2d in
5 mL of diethyl ether was added a solution of 51 mg of
(1R,2S)-(1)-cis-1-amino-2-indanol (Aldrich) in methanol/
ethyl acetate. The resulting precipitate was ®ltered off and
washed with ether to give 101 mg (0.23 mmol, 74%) of salt
S8: mp 170±1738C (needles, isopropanol), MS (FAB,
1LSIMS) m/z 432 (M111), 283, 271, 150, 133, 105
(matrix: thioglycerol); Anal. Calcd for C27H29NO4: C,
75.14; H, 6.78; N, 3.25. Found: C, 75.07; H, 6.64; N, 3.21.
The (R)-(1)-a-methylbenzylamine salt S2 was prepared
analogously (90% yield) and also gave satisfactory MS
and elemental analysis results.
(S)-(2)-Myrtanylamine salt S9 of keto acid 2d. To a solu-
tion of 91 mg (0.32 mmol) of keto acid 2d in 5 mL of
diethyl ether was added 90 mL (S)-(2)-myrtanylamine
(Aldrich). The resulting precipitate, which was formed
after 30 min, was ®ltered off and washed with ether to
give 98 mg (0.23 mmol, 71%) of salt S9: mp 154±1578C
(powder, ethanol), MS (FAB, 1LSIMS) m/z 436 (M111),
343, 295, 283, 265, 154, 133, 95, 81 (matrix: thioglycerol);
Anal. Calcd for C28H37NO3: C, 77.19; H, 8.57; N, 3.22.
Found: C, 77.49; H, 8.52; N, 3.22.
(R)-(1)-a,4-Dimethylbenzylamine salt S3 of keto acid
2d. To a solution of 92 mg (0.33 mmol) of keto acid 2d in
5 mL of diethyl ether was added 60 mL of optically pure
(R)-(1)-a,4-dimethylbenzylamine (Aldrich). The resulting
precipitate was ®ltered off and washed with ether to give
121 mg (0.15 mmol, 88%) of salt S3: mp 170±1728C
(plates, benzene/isopropanol 2:1), MS (FAB, 1LSIMS)
m/z 418 (M111), 283, 271, 136, 133, 119 (matrix: thio-
glycerol); Anal. Calcd for C27H31NO3: C, 77.67; H, 7.48;
N, 3.35. Found: C, 77.72; H, 7.32; N, 3.35.
(S)-(2)-a-Methylbenzylamine salt S10 of keto acid 2h.
To a solution of 62 mg (0.22 mmol) of keto acid 2h in 5 mL
of diethyl ether/ethyl acetate (2:1) was added 70 mL of
optically pure (S)-(2)-a-methylbenzylamine (Aldrich). A
white precipitate formed after reducing the volume of
solvent by 50%. The solid was ®ltered off and washed
with ether to give 47 mg (0.12 mmol, 55%) of salt S10:
mp 136±1388C (needles, ethyl acetate); MS (FAB,
1LSIMS) m/z 404 (M111), 305, 283, 265, 133, 122, 105
(matrix: thioglycerol); Anal. Calcd for C26H29NO3: C,
77.38; H, 7.25; N, 3.47. Found: C, 77.01; H, 7.45; N, 3.41.
(R)-(1)-N,a-Dimethylbenzylamine salt S4 of keto acid
2d. To a solution of 84 mg (0.30 mmol) of keto acid 2d in
5 mL of diethyl ether was added 60 mL of optically pure
(R)-(1)-N,a-dimethylbenzylamine (Aldrich). The resulting
precipitate was ®ltered off and washed with ether to give
115 mg (0.28 mmol, 93%) of salt S4: mp 144±1468C
(cubes, ethyl acetate), MS (FAB, 1LSIMS) m/z 418
(M111), 307, 283, 265, 136, 133, 105 (matrix: thio-
glycerol); Anal. Calcd for C27H31NO3: C, 77.67; H, 7.48;
N, 3.35. Found: C, 77.41; H, 7.67; N, 3.37.
(1R,2S)-(2)- and (1S,2R)-(1)-Norephedrine salt S5 and
S6 of keto acid 2d. A solution of 87 mg (0.31 mmol) of keto
acid 2d in 5 mL of diethyl ether was added to a solution of
53 mg of (1R,2S)-(2)-norephedrine (Aldrich) in 5 mL of
diethyl ether. The resulting precipitate was ®ltered off and
(R)-(1)-N,a-Dimethylbenzylamine salt S11 of keto acid
2h. To a solution of 51 mg (0.18 mmol) of keto acid 2h in
5 mL of diethyl ether was added 50 mL of optically pure
(R)-(1)-N,a-dimethylbenzylamine (Aldrich). The resulting

6750
E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
precipitate was ®ltered off and washed with ether to give
57 mg (0.14 mmol, 77%) of salt S11: mp 141±1448C
(cubes, ethyl acetate); MS (FAB, 1LSIMS) m/z 418
(M111), 307, 283, 265, 136, 105 (matrix: thioglycerol);
Anal. Calcd for C27H31NO3: C, 77.67; H, 7.48; N, 3.35.
Found: C, 77.29; H, 7.50; N, 3.35.
Crystal data for 1c: C18H17NO, Mˆ263.34, transparent,
yellow pyramid 1.00£1.00£0.50 (methanol), monoclinic,
Ê
Ê
Ê
P21/a; aˆ10.586(1) A, bˆ8.377(2) A, cˆ17.539(2) A,
Ê 3
bˆ107.325(8)8, Vˆ1484.9 9(4) A , Zˆ4, rcalcdˆ1.18 g/
cm3, 3437 re¯ections collected, Rˆ4.5%.
Crystal data for 1d: C18H18O3, Mˆ282.34, transparent,
(1R,2S)-(2)-(1)-Norephedrine salt S12 of keto acid 2h.
A solution of 61 mg (0.22 mmol) of keto acid 2h in 5 mL of
diethyl ether was added to a solution of 62 mg of (1R,2S)-
(2)-norephedrine (Aldrich) in 5 mL of diethyl ether. The
resulting precipitate was ®ltered off and washed with ether
to give 75 mg (0.17 mmol, 77%) of salt S12: mp 85±908C
(powder, isopropanol), MS (FAB, 1LSIMS) m/z 434
(M111), 283, 265, 214, 152, 134, 133, 105 (matrix: thio-
glycerol); Anal. Calcd for C27H31NO4: C, 74.79; H, 7.21; N,
3.23. Found: C, 75.00; H, 7.06; N, 3.16.
yellow prism 0.30£0.25£0.20 (methanol); monoclinic,
Ê
Ê
Ê
P21/c, aˆ12.590(1) A, bˆ9.201(1) A, cˆ52.835(4) A,
bˆ94.451(8)8, Vˆ6102(1) A , Zˆ16, rcalcdˆ1.23 g/cm3,
Ê 3
13816 re¯ections collected, Rˆ4.6%, four independent
molecules in the unit cell.
Crystal data for 1f: C17H17FO, Mˆ256.32, transparent,
colorless cube 0.25£0.25£0.25 (methanol), orthorhombic,
Ê
Ê
Ê
Pbca, aˆ99.985(3) A, bˆ33.586(7) A, cˆ8.457(4) A, Vˆ
2836(1) A , Zˆ8, rcalcdˆ1.20 g/cm3, 2951 re¯ections
Ê 3
collected, Rˆ4.9%.
(1R,2S)-(1)-cis-1-Amino-2-indanol salt S13 of keto acid
2h. To a solution of 60 mg (0.21 mmol) of keto acid 2h in
5 mL of diethyl ether was added a solution of 57 mg of
(1R,2S)-(1)-cis-1-amino-2-indanol (Aldrich) in methanol/
diethyl ether. The solvent was removed by rotary evapo-
ration and the resulting oil recrystallized from isopropanol/
petroleum ether 35±60 to give ®ne needles (49 mg,
0.10 mmol, 48%). Recrystallisation of the resulting oil
from acetonitrile afforded a ¯aky solid: mp 100±1038C,
MS (FAB, 1LSIMS) m/z 432 (M111), 283, 271, 150,
133, 105 (matrix: thioglycerol); Anal. Calcd for
C27H29NO4: C, 75.14; H, 6.78; N, 3.25. Found: C, 74.84;
H, 6.64; N, 3.39.
Crystal data for 1g: C18H17NO, Mˆ263.34, transparent,
colorless prism 0.40£0.25£0.15 (ethanol), monoclinic,
Ê
Ê
Ê
P21/a, aˆ7.962(1) A, bˆ15.840(3) A, cˆ12.119(3) A, bˆ
107.40(1)8, Vˆ1458.5(5) A , Zˆ4, rcalcdˆ1.20 g/cm3, 3294
Ê 3
re¯ections collected, Rˆ4.9%.
Crystal data for 1j: C19H20NO3, Mˆ296.37, transparent,
colorless prism 0.20£0.10£0.10 (isopropyl ether/ethyl ace-
ꢀ
Ê
Ê
tate), triclinic, P1, aˆ9.099(2) A, bˆ11.302(2) A, cˆ
Ê
8.406(2) A; aˆ100.84(2)8 bˆ109.16(1)8, gˆ78.88(2)8,
Vˆ794.0(3) A , Zˆ2, rcalcdˆ1.24 g/cm3, 3460 re¯ections
Ê 3
collected, Rˆ5.3%.
Crystal data for 1h: C18H20O, Mˆ252.36, transparent,
Photolysis of salts S1±13
colorless cube 0.25£0.25£0.25 (ethanol), monoclinic, C2/
Ê
Ê
Ê
c, aˆ10.554(4) A, bˆ15.238(4) A, cˆ18.885(3) A; bˆ
Crystals of salts S1±S13 (5±10 mg) were photolyzed
between Pyrex microscope slides for varying lengths of
time and at different temperatures (see Table 3) as described
under Photochemical Procedures. The solid mixtures were
removed from the slides with methanol, and after removal of
the solvent, treated with an excess ethereal diazomethane to
convert the acids into the corresponding methyl esters. The
extent of conversion was analyzed by GC. Only starting
materials and 2-indanols were observed. The enantiomeric
excesses were determined on chiral HPLC (eluent hexane/
isopropanol, Ciralcelw OD column).
95.88(2)8; Vˆ3021(1) A , Zˆ8, rcalcdˆ1.11 g/cm3, 3395
Ê 3
re¯ections collected, Rˆ5.3%.
Crystal data for S3: C27H31NO3, Mˆ417.55, transparent,
colorless plate 0.65£0.30£0.10 (benzene/isopropanol),
Ê
Ê
monoclinic, P21, aˆ7.182(1) A, bˆ6.048(1) A, cˆ
Ê
Ê 3
27.779(2) A, bˆ97.094(9)8, Vˆ1197.4(2) A , Zˆ2,
rcalcdˆ1.16 g/cm3, 2911 re¯ections collected, Rˆ4.4%.
Crystal data for S4: C27H31NO3, Mˆ417.55, transparent,
colorless prism 0.50£0.25£0.25 (ethyl acetate), triclinic,
Ê
Ê
Ê
P1, aˆ10.519(3) A, bˆ13.927(5) A, cˆ9.501(3) A, aˆ
Ê 3
99.57(3)8, bˆ114.97(2)8, gˆ101.68(3)8; Vˆ1184.9(8) A ,
X-Ray crystallographic analyses
Zˆ2, rcalcdˆ1.17 g/cm3, 5103 re¯ections collected, Rˆ
6.2%; two independent molecules in the unit cell.
The crystals were grown by slow evaporation of the listed
solvents. The crystallographic data were obtained with a
Rigaku AFC-6 four cycle diffractometer (Cu-Ka-radiation,
graphite-monochromator) at 293 K. The structures were
solved using direct methods and re®ned on F using teXsan.20
The atomic coordinates were deposited with the Cambridge
Crystallographic Data Centre, University Chemical Labora-
tory, 12 Union Road, Cambridge CB2 IEZ, UK.
Crystal data for S6: C27H31NO3, Mˆ433.55, transparent,
colorless prism 0.35£0.10£0.05 (methanol/ethanol 3:1),
Ê
Ê
monoclinic, C2; aˆ17.732(5) A, bˆ5.612(2) A, cˆ
Ê
Ê
25.720(4) A, bˆ109.18(2)8, Vˆ2417(1)A, Zˆ4, rcalcd
1.19 g/cm3, 2841 re¯ections collected, Rˆ4.5%.
ˆ
Crystal data for 1b: C17H17FO, Mˆ256.32, transparent,
colorless plate 0.50£0.25£0.10 (ethanol); monoclinic;
Acknowledgements
Ê
Ê
Ê
P21/a; aˆ10.151(3) A, bˆ8.687(4) A, cˆ16.802(4) A,
Acknowledgement is made to the donors of the Petroleum
Research Fund, administered by the American Chemical
Society, for partial support of this research. Financial
Ê 3
bˆ104.99(2)8, Vˆ1431.1(7) A , Zˆ4, rcalcdˆ1.19 g/cm3,
3275 re¯ections collected, Rˆ4.3%.

E. Cheung et al. / Tetrahedron 56 (2000) 6739±6751
6751
conversion and total absence at high conversion. Although Wagner
never observed enol formation upon photolysis of a-mesitylaceto-
phenone, it is possible that the byproduct in the photolysis of keto
ester 1j could be an enol as well.
support by the Natural Sciences and Engineering Research
Council of Canada is also grateful acknowledged. We thank
the DAAD (German Academic Exchange Service) for a
postdoctoral fellowship for K. R.
9. The possibility of reverse hydrogen transfer, although shown by
Wagner to be small in a-mesitylacetophenone (1a),4 is another
unknown factor affecting the quantum yields for ketones 1b±k.
10. Intensive research has been done on the Norrish/Yang type II
reaction regarding geometric requirements: Ihmels, H.; Scheffer,
J. R. Tetrahedron 1999, 55, 885 and references therein.
11. (a) Wagner, P. J.; Truman, R. J.; Scaiano, J. C. J. Am. Chem.
Soc. 1985, 107, 7093±7097. (b) Wagner, P. J.; Kemppainen, A. E.;
Schott, H. N. J. Am. Chem. Soc. 1973, 95, 5604±5614.
12. Dorigo, A. E.; Houk, K. N. J. Am. Chem. Soc. 1987, 109,
2195±2197.
References
1. For an excellent overview of the ®eld see: Photochemistry in
Organized and Constrained Media, V. Ramamurthy, Ed.; VCH:
New York, 1991.
2. Collet, A.; Wilen, S. H. 4. Enantiomers, Racemates and
Resolution; Wiley: New York, 1981, pp 14±18.
3. For previous examples of the utilization of the chiral auxiliary
approach to asymmetric synthesis in solid-state organic photo-
chemistry see: (a) Cheung, E.; Netherton, M. R.; Scheffer, J. R.
J. Am. Chem. Soc. 1999, 121, 2919±2920. (b) Leibovitch, M.;
Olovsson, G.; Scheffer, J. R.; Trotter, J. J. Am. Chem. Soc. 1998,
120, 12755±12769. (c) Gamlin, J. N.; Jones, R.; Leibovitch, M.;
Patrick, B.; Scheffer, J. R.; Trotter, J. Acc. Chem. Res. 1996, 29,
203 and references therein.
13. As has been observed in other systems, irradiation of these
optically active salts in solution leads only to racemic photo-
products. See Ref. 3c.
14. (a) Jones, R.; Scheffer, J. R.; Trotter, J.; Yang, J. Tetrahedron
Lett. 1992, 33, 5481. (b) Koshima, H.; Maeda, A.; Masuda, N.;
Matsuura, T.; Hirotsu, K.; Okada, K.; Mizutani, H.; Ito, Y.; Fu,
T. Y.; Scheffer, J. R.; Trotter, J. Tetrahedron: Asymmetry 1994, 5,
1415. (c) Cheung, E.; Kang, T.; Raymond, J. R.; Scheffer, J. R.;
Trotter, J. Tetrahedron Lett. 1999, 40 (50), 8729±8732.
15. Fuson, R. C.; Rabjohn, N. Org. Synth. Col. 1995, III, 557±560.
16. Lutz, R. E.; Hinkley, D. F. J. Am. Chem. Soc. 1950, 72, 4091±
4092.
4. Wagner, P. J.; Meador, M. A.; Zhou, B.; Park, B.-S. J. Am.
Chem. Soc. 1991, 113, 9630±9639.
5. For a general review of hydrogen abstraction reactions see
Wagner, P. J., Park, B.-S. In Organic Photochemistry, Padwa,
A., Ed.; Marcel Dekker: New York, Vol. 11, 1991; p 227.
6. Portions of this work have appeared in preliminary communi-
cation form: Cheung, E.; Rademacher, K.; Scheffer, J. R.; Trotter,
J. Tetrahedron Lett. 1999, (50), 8733±8736.
17. (a) Parish, J. R., von Doering, W. E. J. Am. Chem. Soc. 1967,
89, 5505. (b) Hamada, Y.; Shibata, M.; Sugiura, T.; Kato, S.;
Shioiri, T. J. Org. Chem. 1987, 52, 1252±1255.
7. (a) Wagner, P. J.; Zhou, B.; Hasegawa, T.; Ward, D. L. J. Am.
Chem. Soc. 1991, 113, 9640±9654. (b) Irngartinger, H.; Fettel, P.
W.; Siemund, V. Eur. J. Org. Chem. 1998, 2079.
18. Parham, W. E.; Jones, L. D. J. Org. Chem. 1976, 41, 1187±
1191.
8. Wagner observed upon photolysis of a-(2,4,6-triisopropyl-
phenyl)acetophenone the formation of an enol, 1-phenyl-2-
(2,4,6-triisopropylphenyl)ethanol, from a 1,4-hydrogen shift in
the biradical stage in solution as well as in the solid state.4 These
stilbene type enols were found to be photolabile and convert
back into ketone, which would explain their appearance at low
19. Corey, E. J.; Suggs, J. W. Tetrahedron Lett. 1975, 2647.
20. Molecular Structure Corporation (1992±1997). teXsan. Single
Crystal Structure Analysis Software. Version 1.7, MSC, 3200
Research Forest Drive, The Woodlands, TX 77381, USA.