Photochemistry of 9-methylbicyclo[3.3.1]nonyl aryl ketones - A novel 1,5-disproportionation of 1,4-hydroxy biradicals and asymmetric induction using the solid-state ionic chiral auxiliary method

Article abstract of DOI:10.1139/v05-147

A novel 1,5-disproportionation reaction has been discovered for 1,4-hydroxy biradicals derived from the photolysis of 9-methylbicyclo[3.3.1]nonyl phenyl ketones (1), which undergo mainly Yang cyclization both in solution and the solid state. By applying t

Full text of DOI:10.1139/v05-147

1460
Photochemistry of 9-methylbicyclo[3.3.1]nonyl aryl
ketones — A novel 1,5-disproportionation of 1,4-
hydroxy biradicals and asymmetric induction using
the solid-state ionic chiral auxiliary method1
Shuang Chen, Brian O. Patrick, and John R. Scheffer
Abstract: A novel 1,5-disproportionation reaction has been discovered for 1,4-hydroxy biradicals derived from the
photolysis of 9-methylbicyclo[3.3.1]nonyl phenyl ketones (1), which undergo mainly Yang cyclization both in solution
and the solid state. By applying the solid-state ionic chiral auxiliary method of asymmetric synthesis to the Yang
cyclization, enantiomeric excesses as high as 95% were achieved at high reaction conversions. The origin of the reac-
tion selectivity is discussed with the help of X-ray crystallography. In addition, the solid-state photoreaction of ketone
1b was found to occur in a single crystal-to-single crystal fashion.
Key words: photochemistry, 1,4-hydroxy biradical, disproportionation, asymmetric induction, ionic chiral auxiliary, sin-
gle crystal-to-single crystal reaction.
Résumé : On a découvert une nouvelle réaction de dismutation pour les 1,4-dihydroxy biradicaux obtenus par photo-
lyse des 9-méthylbicyclo[3.3.1]nonyl (phényl) cétones (1) qui subissent principalement une cyclisation de Yang tant en
solution qu’à l’état solide. En appliquant à la cyclisation de Yang la méthode de l’auxiliaire chiral ionique en phase so-
lide, il a été possible d’obtenir des excès énantiomériques pouvant atteindre 95 % avec des taux de conversion élevés.
On discute de l’origine de la sélectivité de la réaction à l’aide de la cristallographie des rayons X. De plus, on a trouvé
que, à l’état solide, la réaction de la cétone 1b s’effectue d’une façon allant d’un cristal unique à un cristal unique.
Mots clés : photochimie, 1,4-dihydroxy biradical, dismutation, induction asymétrique, auxiliaire chiral ionique, réaction
d’un cristal unique à un cristal unique.
[Traduit par la Rédaction] Chen et al. 1472
Introduction
(Scheme 1) (2, 3). In these two previous systems, hydrogen ab-
straction occurred from either a rigid six-membered ring or a
rigid five-membered ring, while in the bicyclo[3.3.1]nonyl
phenyl ketone system hydrogen abstraction occurs from a
relatively flexible six-membered ring, allowing for the ef-
fects of small changes in ring geometry on the hydrogen ab-
straction process and the reaction outcome to be examined.
Moreover, 9-methylbicyclo[3.3.1]nonyl phenyl ketones (1)
are achiral molecules that give mainly chiral Yang
photocyclization products upon photolysis. As hydrogen ab-
straction of either of the two enantiotopic γ-hydrogen atoms
could occur, this allows the enantioselectivity of the reaction
to be studied using the solid-state ionic chiral auxiliary
method (4).
Triplet 1,4-hydroxy biradicals formed in the γ-hydrogen
abstraction step of the photochemical Norrish–Yang type II
reaction have three possible fates: reverse hydrogen transfer
to regenerate the starting ketone in its ground state, cleavage
of the central carbon–carbon bond to form an alkene and an
enol, and Yang cyclization to form a cyclobutanol derivative
(Scheme 1) (1). In this paper, we report a fourth reaction
pathway, a novel 1,5-disproportionation, observed for the
1,4-hydroxy biradicals derived from hydrogen abstraction of
9-methylbicyclo[3.3.1]nonyl phenyl ketones (1). The study
of the photochemistry of 9-methylbicyclo[3.3.1]nonyl phenyl
ketones (1) complements our earlier studies on the solid-
state Norrish–Yang type II photochemistry of adamantyl
phenyl ketones and norbornyl phenyl ketones (2 and 3)
Results and discussion
Received 21 January 2005. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 10 November
2005.
9-Methylbicyclo[3.3.1]nonyl phenyl ketones (1a and 1b)
were prepared by standard synthetic methods (Scheme 2)
from the known compound 9-methylbicyclo[3.3.1]nonyl
methanol (4) (5). Hydrolysis of ester 1b afforded the corre-
sponding keto acid 1c, which was then reacted with various
optically pure amines to form chiral salts 1d–1j through
acid–base chemistry. All these compounds were fully char-
S. Chen, B.O. Patrick, and J.R. Scheffer.2 Department of
Chemistry, University of British Columbia, 2036 Main Mall,
Vancouver, BC V6T 1Z1, Canada.
1This article is part of a Special Issue dedicated to organic
reaction mechanisms.
1
acterized by H and 13C NMR, FT-IR, HR-MS, UV, and ele-
2Corresponding author: (e-mail: scheffer@chem.ubc.ca).
mental analysis (see details in the Experimental section).
Can. J. Chem. 83: 1460–1472 (2005)
doi: 10.1139/V05-147
© 2005 NRC Canada

Chen et al.
1461
Scheme 1.
OH
OH
+
H
3
R
R
hν
O
OH
4
3
CL
CY
1
RHT
R
R
2
RHT = Reverse Hydrogen Transfer
CL = Cleavage
CY = Cyclization
R
R
R
O
O
O
CH3
CH3
CH3
1
3
2
R = H, CO2Me, CO2H, CO2 NH3R*
Scheme 2.
HO
O
PCC
CH3
CH3
r
B
g
M
Ph
O
H
OH
6 (89%)
PCC
1a (95%)
CH3
CH3
i
-
CO2Me
PCC
P
r
CO2Me
M
g
C
l
I
C
HO
4
6 H
5 (99%)
4 C
O
O
O
M
e
CH3
CH3
H
NH3
7 (90%)
H
O
1d R = CO2
H3C C
1h R = CO2
1b (100%)
NH2
N
H2
1. LiOH
2. HCl
H
NH3
H
C
NH3
CH3
1e R = CO2
1f R = CO2
H3C C
1i R = CO2
1j R = CO2
CO2H
R
H3N
HO
NH3
H
O
O
H
CH2OH
C
NH2R*
HO
C
CH3
H3N
C
H
CH3
1g R = CO2
CH2OH
C2H5
1d-1j (71%–90%)
1c (99%)
© 2005 NRC Canada

1462
Can. J. Chem. Vol. 83, 2005
Fig. 1. ORTEP representation of (a) ketone 1a and (b) ketone 1b.
Table 1. Photolysis of ketones 1a and 1b in solution and the
crystalline state.
(a)
Products (%)b
Time Conversion
Ketone Medium (h)
(%)a
65
9a 10a 9b 10b
1a
1b
CH3CNc
62
2
1
2
8
24
1
89
100
100
98
85
92
92
3
3
5
Crystalc
CH3CNc
Crystalc
91
90
4
5
24
100
aPercentage of total GC integral owing to the disappearance of the cor-
responding starting material.
bPercentage of total GC integral owing to the corresponding product.
cThe combined percentage of the reported compounds is less than unity
because of some unidentified GC peaks.
(b)
results of solid-state photolysis of ketones 1a and 1b were
strikingly similar to their solution counterparts, with 1a and
1b affording photoproducts 9a and 9b (90%–92%) and 10a
and 10b (5%). All the photoproducts were fully character-
1
1
ized by 1D H and 13C (BB and APT) NMR, 2D H–1H
COSY, H–13C HMQC, HMBC and NOESY, FT-IR, HR-
1
MS, as well as by elemental analysis. The structures of
photoproducts 10a, 9b, and 10b were also confirmed by X-
ray crystallography (Fig. 2). The details of the separation
conditions along with the analytical data for each of the
products are given in the Experimental section.
As described above, 9-methylated bicyclo[3.3.1]nonyl
phenyl ketones 1a and 1b underwent Yang photocyclization
and yielded endo-aryl cyclobutanols 9a and 9b as major
photoproducts, both in solution and the solid state. In this
case, it is relevant to ask why neither exo-aryl cyclobutanols
nor photocleavage products were obtained. To answer this
question, consider the idealized reaction pathway shown in
Scheme 4. X-ray crystallography shows that in the crystal-
line state ketones 1a and 1b adopt chair–chair conformations
in which the benzoyl group is in the axial position and the
mean plane of the carbonyl group is roughly orthogonal to
the plane bisecting the bicyclo[3.3.1]nonane ring between γ-
hydrogen atoms Hx and Hy. In this case, Hx is in closer prox-
imity to the carbonyl oxygen than Hy (for Hx: d = 2.50 Å in
ketone 1a, and 2.52 Å in ketone 1b; for Hy: d = 3.49 Å in
ketone 1a, and 3.69 Å in ketone 1b). The other γ-hydrogen
abstraction parameters (ω, ∆, and θ) in Table 2 also suggest
that Hx is in a more favorable position for hydrogen abstrac-
tion than Hy (7).
Upon photoexcitation and hydrogen abstraction, 1,4-
hydroxy biradical intermediate 11 will be generated. Since
hydrogen abstraction in the solid state is considered to occur
with minimum movement of the heavy atoms, geometric
data derived from the ground-state ketone can be used to an-
alyze the behavior of the corresponding 1,4-hydroxy
biradical intermediate (8). For this purpose, we define two
torsional angles with reference to Fig. 3a, which is a depic-
tion of the 1,4-hydroxy biradicals derived from ketones 1a
and 1b. Angle ϕ1 is defined as the dihedral angle between
the C2—C3 bond and the p orbital on C1, and the angle ϕ4
is likewise defined as the analogous angle between the C2—
C3 bond and the p orbital on C4. The p orbitals at C1 and
Single crystal X-ray crystallographic determinations were
also performed for compounds 1a and 1b (Fig. 1).
The photochemistry of aryl ketones 1a and 1b was ex-
plored in both solution and the crystalline state. The reaction
products were readily obtained by preparative scale solution
photolysis and the solid-state reactions were performed on
an analytical scale to determine the mode of reactivity in
that medium. The details of the experiments as well as the
apparatus used are given in the Experimental section. Ta-
ble 1 and Scheme 3 summarize the photochemical results in
both media. As shown in Scheme 3 and Table 1, like the
analogous compounds 2 and 3 (2, 3), solution photolysis of
ketones 1a and 1b primarily yielded Yang photocyclization
products 9a and 9b.
Interestingly, photolysis of ketones 1a and 1b also af-
forded small amounts (<5%) of the unusual 1,5-dispropor-
tionation products 10a and 10b (Scheme 3). These minor
products are presumably formed by a novel disproportiona-
tion of the intermediate 1,4-hydroxy biradical 8 in which a
hydrogen atom on C5 is transferred to the radical center on
C1. As far as we are aware, this type of disproportionation
reaction is unknown for 1,4-hydroxy biradicals, although the
corresponding process has been observed for a 1,5-hydroxy
biradical (6). The stereochemistry of the major photopro-
ducts (9a and 9b) was later determined to be that in which
the aryl group is in the endo position, although generation of
exo-aryl products is theoretically possible. It is interesting to
note that in the photochemistry of 9-methyl substituted ke-
tones 1a and 1b, no Norrish type I or type II photocleavage
products could be identified, although traces (<1%) of sev-
eral unknown photoproducts could be detected by GC. The
© 2005 NRC Canada

Chen et al.
1463
Scheme 3.
R
HO
CH3
n
o
i
t
9
a
z
i
l
R
c
R
y
c
g
n
a
s
Y
O
1
HO
hν (Pyrex)
2
CH3
solution
CH3
5
3
R
4
1
,
5
d
i
p
r
o
p
1
8
o
r
t
i
o
H
n
a
t
i
o
n
HO
CH3
a R = H
b R = CO2Me
10
Scheme 4.
Ar
Ar
1
O
O
Hy
Hx
Hy
Hx
2
Me
3
Me
4
1a, 1b
hν
Ar OH
1
1
Ar
OH
HO
Ar
1
2
Hy
Hy
Hy
2
2
3
3
3
Me
Me
Me
4
4
4
12
13
11
endo-Aryl cyclobutanol 9a, 9b
exo-Aryl cyclobutanol
Cleavage product
© 2005 NRC Canada

1464
Can. J. Chem. Vol. 83, 2005
Fig. 2. ORTEP representation of photoproducts (a) 10a, (b) 9b,
and (c) 10b.
Table 2. Hydrogen atom abstraction parameters for compounds
1a and 1b.
(a)
Ketone
H
d (Å)a
ω (°)b
∆ (°)c
86
θ (°)d
1a
Hx
Hy
Hx
Hy
2.50
57
91
63
94
0
118
105
119
105
180
3.49
41
1b
2.52
78
3.69
35
Ideal value
2.7 0.2
90–120
aC=O···Hγ distance.
bDeviation of Hγ from the mean plane of the carbonyl group.
cC=O···Hγ angle.
dC-Hγ···O angle.
Fig. 3. (a) Idealized 1,4-hydroxybiradical intermediate. (b) De-
piction of the angle β.
(b)
Ar
1
OH
2
1
Ar
2
OH
3
3
β
4
4
(c)
5
5
(b)
(a)
Table 3. Geometric data for ketones 1a and 1b.
Ketone
ϕ1 (°)
cosϕ 1
ϕ 4 (°)
cosϕ 4
β (°)
D (Å)
1a
1b
65
59
0.42
0.52
35
36
0.82
0.81
35
28
3.01
2.90
With cosϕ4 ≈ 80% and cosϕ1 ≈ 50% (only 50% of maxi-
mum overlap between the p orbital on C1 and the C2—C3
bond), the geometry of biradical intermediate 11 is poor for
cleavage, as cleavage requires good overlap of both p
orbitals with the cleaving C2—C3 bond. But this biradical
has a good geometry for cyclization, with β ≈ 30° (indicating
there is reasonably good overlap of the two p orbitals at C1
and C4) and D = 2.90 Å in ketone 1b, and 3.01 Å in ketone
1a. As a result, cyclization of biradical 11 with retention of
configuration at C1 and C4 leads to the predominant endo-
aryl photoproducts 9a and 9b. On the other hand, for cleav-
age to occur, an approximate 90° rotation of the aryl and
hydroxyl groups about the C1—C2 bond of biradical 11 is
required to align the p orbital on C1 with the C2—C3 σ
bond (see Scheme 4, structure 12). An additional 90° rota-
tion of the aryl and hydroxyl groups about the C1—C2 bond
leads to biradical 13, closure of which would give exo-aryl
cyclobutanol products. Both processes require rotations
about the C1—C2 bond and are unfavorable because of
Ar/α-CH3 eclipsing, a phenomenon known as the “α-alkyl
group effect” (12). Therefore, endo-aryl cyclobutanols 9a
C4 are assumed to be orthogonal to the O-C1-C2 and C3-
C4-C5 planes, respectively. It is generally believed that
cleavage of 1,4-hydroxy biradicals is favored by good over-
lap between the p orbitals on C1 and C4 with the C2—C3
bond (9). The overlap between these orbitals is proportional
to cosϕ1 and cosϕ4 (10), and the best geometry for cleavage
can be expected when ϕ1 and ϕ4 = 0° and cosϕ1 = cosϕ4 = 1.
The values of cosϕ1 and cosϕ4 for the two ketones investi-
gated in the present study are given in Table 3. Also shown
in Table 3 are the values of the angle β, defined as the dihed-
ral angle between the p orbital on C1 and the C2–C4 vector
(Fig. 3b). The most favorable geometry for cyclization exists
when β = 0°, i.e., when the p orbital on C1 is pointing di-
rectly at the p orbital on C4. Table 3 also gives the values of
the distance (D), defined as the distance between C1 and C4.
For Yang photocyclization to be successful, the radical-
containing carbon atoms C1 and C4 must be close to one an-
other, i.e., D < 3.4 Å, which is the sum of the van der Waals
radii for two carbon atoms (11).
© 2005 NRC Canada

Chen et al.
1465
Scheme 5.
R
R
1
1
3
O
HO
hν (Pyrex)
2
2
H
5
H
CH3
CH3
3
4
4
5
1
8
a R = H
b R = CO2Me
Ring flip
R
R
H
1
HO
H
HO
1,5 disproportionation
2
5
CH3
CH3
3
4
14
10
and 9b are the major products formed in the photoreaction
of ketones 1a and 1b.
tricyclic phenyl ketones, but also shows that small changes in
the flexibility of the bicyclic ring system can lead to a new re-
action pattern, in this case a novel 1,5-disproportionation re-
action.
Formation of minor photoproducts 10a and 10b observed
in the photolysis of ketones 1a and 1b can be seen to be the
result of a 1,5-disproportionation of biradical intermediate
14, which in turn is formed by a ring flip of initially formed
biradical 8 (Scheme 5).3 X-ray crystallographic analysis
showed that the distance between C1 and the equatorial hy-
drogen atom on C5 in ketones 1a and 1b is 4.08 and 4.02 Å,
respectively. It is unlikely that 1,5-disproportionation could
take place directly from the chair–chair biradical 8 without a
significant geometry change of the bicyclic ring. On the
other hand, a ring flip of biradical 8 at the C5 end leads to
the boat–chair biradical 14 in which the flagpole hydrogen
atom on C5 is brought close to the radical center on C1 (13),
and a 1,5-disproportionation process should easily occur.
Because this ring-flipping process is energetically unfavor-
able, in an established equilibrium, the population of bira-
dical 14 is less than that of biradical 8. Moreover, because
1,5-disproportionation requires breaking a C—H bond, it is
likely to be a slower process than Yang cyclization. There-
fore, only small amounts of disproportionation products 10a
and 10b were observed in the photolysis of ketones 1a and
1b.
To apply the solid-state ionic chiral auxiliary method to
asymmetric induction in the photochemistry of 9-methyl-
bicyclo[3.3.1]nonyl phenyl ketones, a number of crystalline
salts (1d–1j, Scheme 2) were prepared by reaction of keto
acid 1c with commercially available optically pure amines.
The amines were selected according to two criteria: (i) the
pKa of the reacting acid should be lower than that of the
protonated amine by about 3 pKa units (14, 15) because this
ensures complete proton transfer from the acid to the amine
instead of formation of a neutral hydrogen bonded complex;
(ii) the amine selected should not possess chromophores that
interfere with the aryl ketone absorption at 320–330 nm. The
formation of an ammonium carboxylate linkage between the
acid and amine was confirmed by the melting point and FT-
IR spectrum of each salt. The melting points of all the salts
are different from both those of keto acid 1c and the corre-
sponding amines. In the IR spectrum of keto acid 1c, the
C=O stretches of the ketone and carboxylic acid exist as
broadened bands at 1695 and 1680 cm–1, respectively. In the
spectra of the salts, these absorptions are replaced by a sin-
gle sharper band at ~1675 cm–1 corresponding to the C=O
stretch of the ketone. At the same time, two new intense
bands (1300–1600 cm–1) owing to the symmetric and asym-
metric stretches of the carboxylate anion are observed (16).
The 1:1 acid–base stoichiometry of all the salts was con-
In comparison with our previous studies on the photo-
chemistry of adamantyl phenyl and norbornyl phenyl ke-
tones (2 and 3) in which Yang cyclization was found to be
the sole reaction pathway (no 1,5-disproportionation product
was obtained owing to the rigidity of adamantyl and
norbornyl ring skeletons) (2, 3), the present study not only
corroborates this main reaction pattern for bicyclic and
1
firmed by H NMR spectroscopy, ESI (or LSI-MS) mass
spectrometry, and elemental analysis.
3 An alternative mechanism for the formation of minor photoproducts 10a and 10b involving initial δ-hydrogen atom abstraction followed by
disproportionation is deemed unlikely on two counts: the general and well-established kinetic preference for γ over δ-hydrogen abstraction
(1a), and in the specific case of ketones 1a and 1b, impossibly long C=O···Hδ distances (4.40 and 4.38 Å, respectively).
© 2005 NRC Canada

1466
Can. J. Chem. Vol. 83, 2005
Table 4. Solid-state photolysis of optically pure salts 1d–1j.
To determine whether the ionic chiral auxiliary has any ef-
fect on the product distribution or the steric course of the
photoreaction in solution, a number of the optically active
salts were photolyzed in a mixture of acetonitrile and water.
Following irradiation, the reaction mixture was treated with
excess ethereal diazomethane solution to form the corre-
sponding methyl esters 9b and 10b. All the solution-state re-
actions essentially afforded the same product ratio of
cyclobutanol 9b and alkene 10b as that observed in the solu-
tion photolysis of keto ester 1b. Photoproduct 9b was subse-
quently determined to be racemic. The enantiomeric excess
of photoproduct 10b was not measured owing to its small
quantity in the reaction mixture. These results show that, in
solution, the anisotropy of the chiral ammonium ion has no
influence on the reaction selectivity.
Each of the salts listed in Scheme 2 was irradiated in the
crystalline state. No color change or melting of the crystals
was observed during the reaction, and following irradiation,
the crystals were dissolved in ethyl acetate or methanol and
treated with excess ethereal diazomethane to form the corre-
sponding methyl esters 9b and 10b. Following diazomethane
treatment, the organic layer containing the esterified photo-
products and starting material was washed with water and
filtered through a short-path silica gel column to remove the
chiral auxiliary. The mixture was then analyzed by GC for
yield and conversion. Owing to the overlap of the product
and starting ketone peaks in the chiral HPLC analysis, achiral
preparative HPLC was used to isolate photoproduct 9b (Wa-
ters RadialpakTM µPorasilTM silica column, 10% EtOAc in
hexanes, 5 mL/min). Following purification, the enantio-
meric excess of photoproduct 9b was measured using chiral
HPLC (Chiralcel® OC column, 2% isopropanol in hexanes,
0.5 mL/min). The results of the solid-state photolysis of
chiral salts 1d–1j are summarized in Table 4. Irradiation of
the salts in the crystalline state leads to asymmetric induc-
tion with product ees ranging from 3% to 95%. The highest
ee was observed for salt 1i formed between acid 1c and (R)-
(–)-1-cyclohexylethylamine, where an enantiomeric excess
of 95% was achieved at 38% reaction conversion.
Salt
Temp. (°C) Time (min) Conv. (%)a ee (%)b
[α]c
1d
r.t.d
–20
–20
–20
–70
r.t.d
–20
–20
–20
–70
r.t.d
–20
–20
–20
r.t.d
–20
–20
r.t.d
–20
–20
–70
r.t.d
–20
–20
–20
r.t.d
–20
–20
–70
120
15
30
240
180
120
15
>99
31
60
>99
44
>99
36
72
89
93
83
55
73
95
98
21
81
94
26
87
56
>99
38
64
97
>99
35
84
68
70
91
86
81
86
73
90
86
86
80
49
86
60
54
25
38
29
60
94
77
93
36
95
84
70
3
+
+
+
+
+
–
–
–
–
–
+
+
+
+
–
–
–
+
+
+
+
–
–
–
–
+
+
+
+
1e
1f
30
60
120
120
60
180
300
120
30
240
120
30
180
180
120
10
1g
1h
1i
1j
30
120
120
60
240
180
17
7
6
aPercentage of total GC integral owing to the disappearance of the cor-
responding starting material.
bThe enantiomeric excess for photoproduct 9b was determined using a
Chiralcel® OC HPLC column.
cSign of rotation of 9b at the sodium D-line.
dRoom temperature.
their X-ray crystal structures; thus, we were not able to cor-
relate the observed enantioselectivity with the solid-state
structure. Nevertheless, the high ees observed in the solid
state can be explained by a conformational effect, a situation
that has been seen in other solid-state enantioselective
Norrish–Yang type II reactions (17). In the solid state, the
role of the ionic chiral auxiliary is to preorganize the
photoreactant in a homochiral conformation in which the
carbonyl group of the substrate is positioned to abstract only
one of the two diastereotopic γ hydrogens in the initial step
of the photoreaction. With topochemical control persisting in
the solid state, formation of a single enantiomer of the prod-
uct is achieved following removal of the chiral auxiliary,
thereby transforming conformational chirality into configu-
rational chirality.
Among the salts studied, salt 1j showed a particularly low
enantioselectivity (3% ee). The low enantioselectivity ob-
served is likely due to the phenomenon of conformational
enantiomerism (18) in which equal amounts of both
enantiomeric conformers of the photolabile carboxylate exist
within the crystal lattice. Each conformer abstracts one of
the two diastereotopic γ hydrogens and forms opposite enan-
tiomers of the same photoproduct after workup. In addition,
Photolysis of the (S)-(–)-1-phenylethylamine and (R)-(+)-
1-phenylethylamine salts (1d and 1e) afforded nearly equal
optical yields of the enantiomers of cyclobutanol 9b (after
diazomethane workup). This allows direct access to both en-
antiomers by simple exchange of the ionic chiral auxiliary.
Table 4 also shows that the product ee declines with increas-
ing reaction 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. As
the reaction proceeds, the rigidity and regularity of the crys-
tal lattice is gradually lost. As a result, the topochemical
control is reduced and enantioselectivity declines. Low-
temperature photolysis can be used to partially compensate
this effect. As shown in Table 4, for reactions conducted at
reduced temperature (–20 °C), enantioselectivities are better
than those obtained at room temperature. Specifically, at
–20 °C, quantitative conversion to cyclobutanol 9b from salt
1d occurred in 81% ee in comparison to 70% ee at room
temperature. Further reducing the reaction temperature to
–70 °C provided limited improvement in enantioselectivity,
while much longer irradiation times were required.
To date, salts 1d–1j have resisted attempts to determine
© 2005 NRC Canada

Chen et al.
1467
Fig. 4. ORTEP representation of: (a) ketone 1b showing γ hydro-
gens Hx and Hy; (b) mixed crystal composed of 75% of ketone
1b (gray) and 25% of photoproduct 9b (black); (c) crystal pack-
ing diagram of the mixed crystal.
Since attempts to grow crystals of salt 1j suitable for X-ray
analysis were not successful, the exact reason for the low
enantioselectivity of salt 1j remains to be determined.
During our photochemical studies, we noticed that crys-
tals of ketone 1b did not change in appearance upon irradia-
tion. This suggested the occurrence of a single-crystal-to-
single-crystal reaction (19), which allowed the structures of
both reactant and product to be obtained at various stages of
reaction. Figure 4a shows the crystal structure of ketone 1b
prior to photolysis, and Fig. 4b depicts the superposition of
ketone 1b and its photoproduct in the mixed crystal follow-
ing 25% conversion to the corresponding cyclobutanol 9b.
The extent of conversion was determined through refinement
of the atom site occupancy factors in the X-ray crystallo-
graphic study. Figure 4a predicts preferential abstraction of
Hx over Hy, and in agreement with this, the molecular struc-
ture of the photoproduct shows that the carbon atom to
which Hx was attached is in fact part of the newly formed
four-membered ring. The crystal structures also reveal that
closure of the biradical does occur with “retention of config-
uration” at the carbonyl carbon. The reactant and product
have similar geometries in which the aryl group portion of
the molecule remains essentially unchanged during photo-
lysis, while an upward movement of the γ carbon occurs to
allow for cyclobutane ring formation. The corresponding al-
teration of the bicyclic ring skeleton is significant but a close
atom-for-atom correspondence between the photoreactant
and photoproduct is maintained. Also observed was a change
in the orientation of the carbon–oxygen bond during the
change in hybridization at the carbonyl carbon from sp2 to
sp3. Accompanying this was a change in the length of the
C—O bond from 1.22 to 1.33 Å. While these motions are
significantly greater than those accompanying other uni-
molecular single crystal-to-single crystal photoreactions,4
the overall packing arrangement of the crystal (Fig. 4c) re-
mains unchanged, and as a result, the process still occurs in
a topochemical fashion.
(a)
(b)
(c)
(c)
Summary
The photochemistry of several 9-methylbicyclo[3.3.1]-
nonyl phenyl ketones was studied in solution and the solid
state. In both media, Yang photocyclization was found to be
the major process for α-methylated bicyclo[3.3.1]nonyl
phenyl ketones (1a and 1b), with endo-aryl cyclobutanols
(9a and 9b) being the major products. An unusual 1,4-
hydroxy biradical disproportionation reaction was also ob-
served during the photolysis of ketones 1a and 1b, although
it was a minor process (<5%). By utilizing the ionic chiral
auxiliary method, high enantioselectivities (up to 95%) were
achieved for a number of the optically active salts on
photolysis in the solid state. In addition, structure–reactivity
correlation studies were performed based on the crystal
crystal disorder could also be a cause for the low enantio-
selectivity observed in the photolysis of salt 1j. In a disor-
dered crystal, two or more conformations of the reactive ion
are distributed randomly throughout the crystal (static disor-
der) or are thermally interconverting (dynamic disorder).
structure data of ketones 1a and 1b to rationalize their pho
-
tochemical behavior. The solid-state photoreaction of ketone
1b was found to occur in a topochemically controlled, single
crystal-to-single crystal fashion.
4 For examples of unimolecular single-crystal-to-single-crystal photochemical reactions, see refs. 2 and 3.
© 2005 NRC Canada

1468
Can. J. Chem. Vol. 83, 2005
(EtOAc – petroleum ether). IR (KBr, cm–1): 3395, 2943,
2912, 2868, 1491, 1461, 1020, 892, 761, 732, 703, 567. H
Experimental section
1
Commercial spectral grade solvents were used for photo-
chemical experiments unless otherwise stated. For synthetic
use tetrahydrofuran and diethyl ether were dried over the so-
dium ketyl of benzophenone, dichloromethane was dried
over calcium hydride, and acetonitrile was distilled from
P2O5. All reactions were performed under a nitrogen atmo-
sphere unless otherwise noted. Melting points were deter-
mined on a Fisher-Johns hot-stage apparatus and were
uncorrected. Analytical TLC was performed on 0.20 mm sil-
ica gel 60-F plates and visualized under UV light and (or) by
staining with I2 or phosphomolybdic acid. Preparative chro-
matography was performed using either the flash column
method with silica gel (particle size 230–400 mesh) or radial
elution chromatography on a Harrison Research Chroma-
otron (model 7924T) with self-coated silica gel plates (1 or
2 mm thickness with EM Science silica gel 60 PF254 con-
taining gypsum 7749-3). IR spectra were recorded on a
PerkinElmer 1710 Fourier transform spectrometer. UV spec-
tra were recorded on a PerkinElmer Lambda-4B UV–vis
spectrometer at 25 °C. High-resolution mass spectra were
obtained from a Kratos MS 50 instrument using electron im-
pact (EI) ionization at 70 eV, a Kratos MS 80 spectrometer
using desorption chemical ionization (CI) with the ionizing
gas noted, a Kratos Concept IIHQ hybrid spectrometer using
liquid secondary ionization (LSI-MS), or a Micromass LCT
spectrometer using electrospray ionization (ESI). Elemental
analyses were performed by the University of British Co-
lumbia (UBC) microanalytical service on a Carlo Erba CHN
NMR (CDCl3, 400 MHz) δ: 0.85 (s, 3H), 1.27 (br, 1H), 1.39
(m, 1H), 1.50–1.62 (m, 4H), 1.66 (s, 1H), 1.68–1.99 (m,
5H), 2.00–2.20 (m, 2H), 2.37 (m, 1H), 5.40 (s, 1H), 7.23 (m,
1H), 7.29 (m, 2H), 7.37 (m, 2H). 13C NMR (CDCl3,
100 MHz) δ: 16.54 (-ve), 20.58, 20.87, 27.20, 27.60, 27.74,
27.85, 32.24 (-ve), 32.92 (-ve), 39.65, 74.13 (-ve), 126.93
(-ve), 127.47 (-ve), 127.55 (-ve), 142.45. HR-MS (CI) calcd.
for C17H24O: 244.1827; found: 244.1834. Anal. calcd. for
C17H24O: C 83.55, H 9.90; found: C 83.95, H 10.02.
9-Benzoyl-9-methylbicyclo[3.3.1]nonanes (1a)
A mixture of PCC (1.15 g, 5.33 mmol) and Celite® 545
(2.30 g) was ground with a mortar and pestle until homoge-
neous. This solid was added to a solution of alcohol 6
(651 mg, 2.66 mmol) in anhydrous dichloromethane
(30 mL) and the reaction mixture was stirred for 10 h at
room temperature. The reaction mixture was filtered through
a column of Florisil® and the remaining solids well triturated
with anhydr. Et2O. Removal of solvent in vacuo and silica
gel chromatography (3% Et2O in petroleum ether) gave
ketone 1a (612 mg, 95%) as a white solid, mp 67 to 68 °C
(EtOAc – petroleum ether). UV–vis (2.39 × 10–4 mol/L,
MeOH, nm ((mol/L)–1 cm–1)): 285 (1178), 330 (720). IR
(KBr, cm–1): 2960, 2913, 2865, 1672, 1444, 1266, 1232,
1216, 1124, 963, 932, 721, 696, 663. 1H NMR (CDCl3,
300 MHz) δ: 1.30–1.93 (m, 13H), 2.02–2.17 (m, 2H), 2.34
(br, 2H), 7.33–7.46 (m, 3H), 7.70–7.75 (m, 2H). 13C NMR
(CDCl3, 75 MHz) δ: 20.20, 20.98, 23.63 (-ve), 26.82, 29.02,
33.90 (-ve), 50.79, 127.84 (-ve), 127.97 (-ve), 130.83 (-ve),
139.34, 209.84. HR-MS (EI) calcd. for C17H22O: 242.1671;
found: 242.1670. Anal. calcd. for C17H22O: C 84.25, H 9.15;
found: C 84.45, H 9.13. This structure was confirmed by X-
ray crystallographic analysis.
1
Model 1106 analyzer. H NMR spectra were obtained at ei-
ther 300 or 400 MHz on Bruker AV-300 or Bruker AV-400
instruments. 13C NMR spectra were recorded at either 75 or
100 MHz. Some spectra are supported by data from the at-
tached proton test (APT). Where these are given, (-ve) de-
notes a negative APT peak corresponding to a methine (CH)
or methyl (CH3) carbon centre, while no assignment signi-
fies a quaternary or methylene (CH2) carbon center.
(9-Methyl)bicyclo[3.3.1]non-9-yl (p-carbomethoxy)phenyl
methanol (7)
A mixture of PCC (3.71 g, 17.2 mmol) and Celite® 545
(5.00 g) was ground with a mortar and pestle until homoge-
neous. This solid was added to a solution of alcohol 4 (5)
(1.81 g, 10.8 mmol) in anhydrous dichloromethane (120 mL)
and the mixture was stirred for 2 h at room temperature. The
reaction mixture was filtered through a column of Florisil®
and the remaining solids well triturated with anhydr. Et2O.
Removal of solvent in vacuo gave aldehyde 5 (1.79 g, 99%)
as colorless oil, which was used without further purification.
To a cold (–40 °C) solution of methyl 4-iodobenzoate
(2.85 g, 10.9 mmol) in THF (60 mL) was added isopropyl-
magnesium chloride (5.7 mL of a 2.0 mol/L solution in THF,
11.4 mmol) over 10 min. The reaction was stirred in the cold
(–40 °C) for 1 h. A solution of aldehyde 5 (1.73 g,
10.4 mmol) previously obtained in THF (20 mL) was added.
The reaction was stirred at –40 °C for 3 h and quenched
with 5% aq. NH4Cl (100 mL). THF was removed in vacuo
and the resultant aqueous phase was extracted with Et2O
(3 × 70 mL). The combined ethereal extracts were washed
with brine (2 × 100 mL), followed by drying (MgSO4) and
removal of the solvent in vacuo. Silica gel chromatography
(14% EtOAc in petroleum ether) afforded alcohol 7 (2.82 g,
90%) as a white powder, mp 135–137 °C (EtOAc – petro-
9-Methylbicyclo[3.3.1]nonyl carboxaldehyde (5) and 9-
methylbicyclo[3.3.1]nonyl phenyl methanol (6)
A mixture of PCC (1.40 g, 6.48 mmol) and Celite® 545
(2.79 g) was ground with a mortar and pestle until homoge-
neous. This solid was added to a solution of alcohol 4 (5)
(545 mg, 3.24 mmol) in anhydrous dichloromethane
(40 mL) and the mixture stirred for 2 h at room temperature.
The reaction mixture was filtered through a column of
Florisil® and the remaining solids triturated well with
anhydr. Et2O. Removal of solvent in vacuo gave aldehyde 5
(538 mg, 100%) as colorless oil, which was found to be un-
stable and was used without further purification. To a cold
(0 °C) solution of phenyl magnesium bromide in THF
(25 mL) was added aldehyde 5 (538 mg, 3.24 mmol in 5 mL
of THF). The reaction was stirred in the cold (0 °C) for 2 h
and quenched with 5% aq. NH4Cl solution (15 mL). THF
was removed and the resultant aqueous phase was extracted
with Et2O (3 × 20 mL). The combined ethereal extracts were
washed with H2O (20 mL) and brine (20 mL), followed by
drying (MgSO4) and removal of the solvent in vacuo. Silica
gel chromatography (5% EtOAc in petroleum ether) afforded
alcohol 6 (706 mg, 89%) as a white solid, mp 118–118.5 °C
© 2005 NRC Canada

Chen et al.
1469
leum ether). IR (KBr, cm–1): 3516 (br), 2950, 2913, 2868,
1724, 1610, 1490, 1460, 1436, 1283, 1192, 1111, 1034, 1018,
909, 869, 812, 770, 736, 715, 573. 1H NMR (CDCl3,
300 MHz) δ: 0.79 (s, 3H), 1.90–2.40 (m, 15H), 3.88 (s, 3H),
5.43 (s, 1H), 7.42 (d, J = 8.4 Hz, 2H), 7.93 (d, J = 8.4 Hz,
2H). 13C NMR (CDCl3, 75 MHz) δ: 16.56 (-ve), 20.48,
20.78, 27.15, 27.62 (br), 27.76, 32.19 (-ve), 32.90 (-ve),
39.91, 52.00 (-ve), 73.82 (-ve), 127.49 (-ve), 128.68 (-ve),
147.91 (br), 167.07. HR-MS (CI) calcd. for C19H30NO3:
320.2226 (M + NH4)+; found: 320.2222. Anal. calcd. for
C19H26O3: C 75.46, H 8.67; found: C 75.48, H 8.73.
286.1569; found: 286.1567. Anal. calcd. for C18H22O3: C
75.50, H 7.74; found: C 75.23, H 7.67.
(S)-(–)-1-Phenylethylamine salt (1d)
Keto acid 1c (86 mg, 0.30 mmol) and (S)-(–)-1-phenyl-
ethylamine (40 µL, 38 mg, 0.31 mmol) were dissolved in a
hot mixture of acetonitrile and methanol. Upon cooling to
room temperature, filtration gave salt 1d as colorless needles
(98 mg, 80%), mp 211–213 °C (MeOH–CH3CN). IR
(KBr, cm–1): 3447, 2915, 2862, 2767, 1667, 1615, 1518,
1455, 1395, 1232, 1219, 1124, 1092, 965, 932, 844, 825,
1
764, 745, 695, 539. H NMR (CD3OD, 300 MHz) δ: 1.30–
2.25 (m, 18H), 2.33 (br, 2H), 4.42 (q, J = 6.8 Hz, 1H), 7.41
(m, 5H), 7.73 (d, J = 8.4 Hz, 2H), 7.95 (d, J = 8.4 Hz, 2H).
13C NMR (CD3OD, 75 MHz) δ: 20.96 (-ve), 21.26, 22.01,
24.16 (-ve), 27.80, 30.11, 35.25 (-ve), 52.05, 52.26 (-ve),
127.60 (-ve), 128.44 (-ve), 129.89 (-ve), 130.05 (-ve),
130.26 (-ve), 140.11, 141.37, 141.99, 173.98, 212.09. HR-
MS (LSI-MS) calcd. for C26H34NO3: 408.2539 (M + H+);
found: 408.2538. Anal. calcd. for C26H33NO3: C 76.62, H
8.16, N 3.44; found: C 76.34, H 8.15, N 3.43.
9-(p-Carbomethoxybenzoyl)-9-methylbicyclo[3.3.1]nonane
(1b)
A mixture of PCC (3.09 g, 14.3 mmol) and Celite® 545
(5.0 g) was ground with a mortar and pestle until homoge-
neous. This solid was added to a solution of alcohol 7
(2.71 g, 8.96 mmol) in anhydrous dichloromethane
(100 mL) and the resulting mixture was stirred for 12 h at
room temperature. The reaction mixture was filtered through
a column of Florisil® and the remaining solids triturated well
with anhydr. Et2O. Removal of solvent in vacuo gave ketone
1b (2.69 g, 100%) as a white solid, mp 137 to 138 °C
(R)-(+)-1-Phenylethylamine salt (1e)
1
(EtOAc–hexanes). H NMR (CDCl3, 300 MHz) δ: 1.48 (s,
Keto acid 1c (57 mg, 0.20 mmol) and (R)-(+)-1-
phenylethylamine (27 µL, 25 mg, 0.21 mmol) were dis-
solved in a hot mixture of acetonitrile and methanol. Upon
cooling to room temperature, filtration gave salt 1e as color-
less needles (64 mg, 79%), mp 211–213 °C (MeOH–
CH3CN). IR (KBr, cm–1): 3446, 2915, 2864, 2766, 1668,
1615, 1519, 1456, 1396, 1232, 1219, 1124, 1092, 966, 933,
3H), 1.32–1.94 (m, 10H), 2.08 (m, 2H), 2.29 (br, 2H), 3.92
(s, 3H), 7.74 (d, J = 8.5 Hz, 2H), 8.03 (d, J = 8.5 Hz, 2H).
13C NMR (CDCl3, 75 MHz) δ: 20.10, 20.89, 23.64 (-ve),
26.71, 29.03, 33.74 (-ve), 51.04, 52.32 (-ve), 127.69 (-ve),
129.27 (-ve), 131.89, 143.20, 166.36, 209.71. IR (KBr, cm–1):
2950, 2906, 2857, 1722, 1666, 1497, 1435, 1401, 1279,
1235, 1190, 1108, 1017, 960, 934, 862, 738, 703. UV–vis
(1.46 × 10–4 mol/L, MeOH, nm ((mol/L)–1 cm–1)): 285
(1459), 330 (194). HR-MS (EI) calcd. for C19H24O3:
300.1725; found: 300.1729. Anal. calcd. for C19H24O3: C
75.97, H 8.05; found: C 75.80, H 8.00. This structure was
confirmed by X-ray crystallographic analysis.
1
844, 825, 764, 744, 695, 539. H NMR (CD3OD, 300 MHz)
δ: 1.30–2.25 (m, 18H), 2.33 (br, 2H), 4.42 (q, J = 6.8 Hz,
1H), 7.41 (m, 5H), 7.73 (d, J = 8.4 Hz, 2H), 7.95 (d, J =
8.4 Hz, 2H). 13C NMR (CD3OD, 75 MHz) δ: 20.96 (-ve),
21.26, 22.01, 24.16 (-ve), 27.80, 30.11, 35.25 (-ve), 52.05,
52.26 (-ve), 127.60 (-ve), 128.44 (-ve), 129.89 (-ve), 130.05
(-ve), 130.26 (-ve), 140.20, 141.37, 141.99, 173.98, 212.09.
HR-MS (ESI) calcd. for C26H34NO3: 408.2539 (M + H+);
found: 408.2544. Anal. calcd. for C26H33NO3: C 76.62, H
8.16, N 3.44; found: C 76.35, H 8.46, N 3.47.
9-(p-Carboxybenzoyl)-9-methylbicyclo[3.3.1]nonane (1c)
To a solution of ester 1b (2.45 g, 8.16 mmol) in THF
(50 mL) was added a solution of lithium hydroxide
monohydrate (3.42 g, 81.6 mmol, in 25 mL of water). The
reaction was stirred at room temperature for 20 h. The reac-
tion solution was diluted with Et2O (200 mL) and extracted
with water (3 × 100 mL). The combined aqueous extracts
were acidified with concd. HCl and extracted with Et2O (3 ×
100 mL). The combined organic extracts were washed with
water (100 mL) and brine (100 mL), then dried over MgSO4.
Removal of solvent in vacuo yielded acid 1c (2.31 g, 99%)
as a white powder, mp 224–226 °C (EtOH–H2O). IR
(KBr, cm–1): 3080–2720 (br), 2950, 2906, 2863, 1695, 1680,
1568, 1505, 1431, 1401, 1316, 1295, 1214, 1126, 968, 933,
(1S,2R)-(–)-cis-1-Amino-2-indanol salt (1f)
Salt 1f was prepared by dissolving 57 mg (0.20 mmol) of
keto acid 1c and 31 mg (0.21 mmol) of (1S,2R)-(–)-cis-1-
amino-2-indanol in a hot mixture of acetonitrile and metha-
nol. Upon cooling to room temperature, filtration gave salt
1f as an off-white powder (71 mg, 81%), mp 190 °C (dec)
(MeOH–CH3CN). IR (KBr, cm–1): 3235, 2913, 2862, 2625,
1670, 1619, 1580, 1535, 1452, 1397, 1267, 1216, 1093, 966,
1
932, 844, 823, 803, 741, 567. H NMR (CD3OD, 300 MHz)
δ: 1.28–2.00 (m, 13H), 2.18 (m, 2H), 2.33 (br, 2H), 3.01 (dd,
J = 5.9, 16.3 Hz, 1H), 3.22 (dd, J = 6.4, 16.3 Hz, 1H), 4.54
(d, J = 5.9 Hz, 1H), 4.70 (dd, J = 5.9, 11.3 Hz, 1H), 7.31 (m,
3H), 7.46 (d, J = 7.3 Hz, 2H), 7.73 (d, J = 8.4 Hz, 2H), 7.96
(d, J = 8.4 Hz, 2H). 13C NMR (CD3OD, 75 MHz) δ: 21.24,
22.00, 24.15 (-ve), 27.80, 30.10, 35.24 (-ve), 40.09, 52.05,
58.71 (-ve), 72.08 (-ve), 126.14 (-ve), 126.64 (-ve), 128.40
(-ve), 129.44 (-ve), 129.89 (-ve), 130.77 (-ve), 138.47,
141.28, 141.99, 142.77, 174.11, 212.11. HR-MS (ESI) calcd.
for C27H34NO4: 436.2488 (M + H+); found: 436.2486. Anal.
1
860, 807, 761, 731, 698, 543. H NMR (DMSO, 300 MHz)
δ: 1.20–2.16 (m, 15H), 2.22 (br, 2H), 7.82 (d, J = 8.2 Hz,
2H), 7.98 (d, J = 8.2 Hz, 2H). (No COOH signal was ob-
served owing to proton exchange with a trace amount of wa-
ter in the solvent). 13C NMR (DMSO, 75 MHz) δ: 19.66,
20.41, 23.05 (-ve), 26.16, 28.47, 33.10 (-ve), 50.22, 127.71
(-ve), 129.12 (-ve), 132.78, 142.24, 166.66, 209.02. UV–vis
(1.68 × 10–4 mol/L, MeOH, nm ((mol/L)–1 cm–1)): 285
(2994), 335 (333). HR-MS (EI) calcd. for C18H22O3:
© 2005 NRC Canada

1470
Can. J. Chem. Vol. 83, 2005
calcd. for C27H33NO4: C 74.45, H 7.64, N 3.22; found: C
74.38, H 7.68, N 3.23.
(1R,2R)-(–)-2-Amino-1-phenyl-1,3-propanediol salt (1j)
Salt 1j was prepared by dissolving keto acid 1c (57 mg,
0.20 mmol) and (1R,2R)-(–)-2-amino-1-phenyl-1,3-pro-
panediol (35 mg, 0.21 mmol) in a hot mixture of acetonitrile
and methanol. Upon cooling to room temperature, filtration
gave salt 1j as a white powder (76 mg, 84%), mp 142 to
143 °C (MeOH–CH3CN). IR (KBr, cm–1): 3367, 2914, 2864,
1668, 1582, 1542, 1493, 1453, 1386, 1229, 1124, 1039, 968,
960, 932, 868, 840, 803, 746, 700, 572, 540. 1H NMR
(CD3OD, 300 MHz) δ: 1.25–1.73 (m, 8H), 1.51 (s, 3H), 1.89
(m, 2H), 2.17 (m, 2H), 2.33 (br, 2H), 3.26 (m, 1H), 3.47 (m,
2H), 4.73 (d, J = 8.6 Hz, 1H), 7.37 (m, 5H), 7.72 (d, J =
8.6 Hz, 2H), 7.96 (d, J = 8.6 Hz, 2H). 13C NMR (CD3OD,
75 MHz) δ: 21.23, 21.98, 24.15 (-ve), 27.81, 30.10, 35.26
(-ve), 52.06, 60.23, 60.29 (-ve), 72.54 (-ve), 127.86 (-ve),
128.42 (-ve), 129.51 (-ve), 129.74 (-ve), 129.89 (-ve),
141.38, 141.97, 142.23, 174.13, 212.15. HR-MS (ESI) calcd.
for C27H36NO5: 454.2593 (M + H+); found: 454.2599. Anal.
calcd. for C27H35NO5: C 71.50, H 7.78, N 3.09; found: C
71.34, H 8.05, N 3.41.
(R)-(–)-2-Amino-1-butanol salt (1g)
Salt 1g was prepared by dissolving 57 mg (0.20 mmol) of
keto acid 1c and 19 mg (0.21 mmol) of (R)-(–)-2-amino-1-
butanol in hot methanol. Upon cooling to room temperature,
filtration gave salt 1g as an off-white powder (53 mg, 71%),
mp 159–162 °C (MeOH). IR (KBr, cm–1): 3367, 2960, 2910,
2871, 1681, 1667, 1595, 1524, 1463, 1387, 1263, 1232,
1
1217, 1124, 1083, 964, 931, 869, 803, 747, 570. H NMR
(CD3OD, 300 MHz) δ: 1.01 (t, J = 7.6 Hz, 3H), 1.25–2.00
(m, 15H), 2.18 (m, 2H), 2.33 (br, 2H), 3.08 (m, 1H), 3.54
(dd, J = 6.7, 11.7 Hz, 1H), 3.75 (dd, J = 3.7, 11.7 Hz, 1H),
7.73 (d, J = 8.5 Hz, 2H), 7.96 (d, J = 8.5 Hz, 2H). 13C NMR
(CD3OD, 75 MHz) δ: 10.16 (-ve), 21.25, 22.00, 23.67, 24.16
(-ve), 27.81, 30.11, 35.25 (-ve), 52.05, 55.98 (-ve), 61.99,
128.44 (-ve), 129.88 (-ve), 141.43, 141.96, 174.17, 212.11.
HR-MS (ESI) calcd. for C22H34NO4: 376.2488 (M + H+);
found 376.2488. Anal. calcd. for C22H33NO4: C 70.37, H
8.86, N 3.73; found: C 70.28, H 8.90, N 3.76.
Photochemistry
Both solution- and solid-state photolyses were carried out
by using a 450 W Hanovia medium-pressure mercury lamp
in a water-cooled immersion well. Light from the lamp was
filtered through a Pyrex (λ ≥ 290 nm) immersion well. For
preparative scale solution-state photolyses, samples were
dissolved in HPLC grade or spectral grade solvents and were
purged with nitrogen for at least 15 min prior to irradiation.
The reactions were performed either in sealed reaction ves-
sels or under a positive pressure of nitrogen. Yields and con-
versions were calculated based on the mass of the isolated
products. For analytical solid-state photolyses, the procedure
consisted of crushing ca. 5 mg of the salt to be photolyzed
between two Pyrex microscope slides, taping the plates to-
gether, sealing the resulting sandwiches in polyethylene bags
under a positive nitrogen atmosphere, and irradiating the en-
sembles with the output from a water-cooled 450 W
medium-pressure mercury lamp. Photolyses were generally
carried out at room temperature. For some runs that required
low temperature during photolysis, a Cryocool CC-100 II
immersion cooling system (Neslab Instrument Inc.) was
used with ethanol as the coolant. The salts were irradiated
for varying lengths of time to determine the dependence of
the ee values on the extent of conversion. To achieve high
conversions it was usually necessary to rotate the sample
180° midway through the irradiation, exposing the rear side
of the slide to the light and allowing as much of the sample
as possible to react. For neutral molecules, the sample was
directly analyzed by GC. The reaction mixtures containing
chiral organic salts were first derivatized to their correspond-
ing methyl esters by treatment with excess ethereal diazome-
thane solution and then washed with water and subjected to
short-path silica gel chromatography to remove the chiral
auxiliary. The mixtures were then analyzed by GC for con-
version as well as product composition. Yields and conver-
sions of the solid-state reactions were determined based on
the average integration of at least two GC analyses. The
overall precision of the reported GC results is estimated to
be 1%.
L-Prolinamide salt (1h)
Salt 1h was prepared by dissolving keto acid 1c (57 mg,
0.20 mmol) and L-prolinamide (24 mg, 0.21 mmol) in a hot
mixture of acetonitrile and methanol. Upon cooling to room
temperature, filtration gave salt 1h (72 mg, 90%) as a white
powder, mp 210–211.5 °C (MeOH–CH3CN). IR (KBr, cm–1):
3365, 3232, 2935, 2909, 2871, 1715, 1680, 1631, 1591,
1544, 1383, 1217, 1015, 963, 931, 827, 804, 744, 668, 641,
1
567. H NMR (CD3OD, 300 MHz) δ: 1.20–2.00 (m, 16H),
2.17 (m, 2H), 2.33 (br, 2H), 2.39, (m, 1H), 3.31 (m, 2H),
4.19 (dd, J = 6.4, 8.4 Hz, 1H), 7.73 (d, J = 8.4 Hz, 2H), 7.97
(d, J = 8.4 Hz, 2H). 13C NMR (CD3OD, 75 MHz) δ: 21.23,
21.99, 24.15 (-ve), 25.38, 27.79, 30.10, 31.25, 35.24 (-ve),
47.34, 52.07, 60.99 (-ve), 128.48 (-ve), 129.93 (-ve), 140.68,
142.21, 173.01, 173.71, 212.07. HR-MS (ESI) calcd. for
C23H33N2O4: 401.2440 (M + H+); found: 401.2451. Anal.
calcd. for C23H32N2O4: C 68.97, H 8.05, N, 6.99; found: C
69.10, H 8.09, N 6.98.
(R)-(–)-1-Cyclohexylethylamine salt (1i)
Salt 1i was prepared by dissolving keto acid 1c (57 mg,
0.20 mmol) and (R)-(–)-1-cyclohexylethylamine (31 µL,
27 mg, 0.21 mmol) in hot methanol. Upon cooling to room
temperature, filtration gave salt 1i as a white powder (71 mg,
86%), mp 192–195 °C (MeOH). IR (KBr, cm–1): 3443,
2929, 2862, 2603, 2551, 2361, 2342, 2203, 1666, 1635,
1578, 1537, 1450, 1382, 1263, 1232, 1216, 1126, 964, 932,
1
868, 821, 804, 745, 568. H NMR (CD3OD, 300 MHz) δ:
0.98–2.00 (m, 27H), 2.18 (m, 2H), 2.33 (br, 2H), 3.06 (m,
1H), 7.73 (d, J = 8.4 Hz, 2H), 7.96 (d, J = 8.4 Hz, 2H). 13C
NMR (CD3OD, 75 MHz) δ: 16.07 (-ve), 21.24, 22.00, 24.15
(-ve), 26.92, 27.00, 27.08, 27.81, 28.82, 30.01, 30.11, 35.26
(-ve), 42.75 (-ve), 52.05, 53.39 (-ve), 128.42 (-ve), 129.85
(-ve), 141.61, 141.90, 174.20, 212.09. HR-MS (ESI) calcd.
for C26H40NO3: 414.3008 (M + H+); found: 414.3009. Anal.
calcd. for C26H39NO3: C 75.50, H 9.50, N 3.39; found: C
75.31, H 9.55, N 3.46.
© 2005 NRC Canada

Chen et al.
1471
130.32 (-ve), 149.34, 166.64. HR-MS (EI) calcd. for
C19H24O3: 300.1725; found: 300.1726. Anal. calcd. for
C19H24O3: C 75.97, H 8.05; found: C 75.96, H 8.28. This
structure was confirmed by X-ray crystallographic analysis.
Preparative photolysis of ketone 1a
A solution of ketone 1a (485 mg, 2.00 mmol) in
acetonitrile (40 mL) was purged with nitrogen for 30 min
and irradiated for 7 h. Removal of the solvent in vacuo fol-
lowed by silica gel chromatography (6% EtOAc in petro-
leum ether) afforded starting material 1a (17 mg, 4%),
cyclobutanol 9a (442 mg, 91%), and alkene 10a (19 mg,
4%).
Data for alkene 10b
Melting point: 148–150 °C (CHCl3–hexanes). IR
(KBr, cm–1): 3524, 3017, 2950, 2921, 2862, 1723, 1706,
1
1610, 1436, 1281, 1111, 1016, 860, 772, 724, 705, 612. H
NMR (C6D6, 400 MHz) δ: 0.85 (s, 3H), 1.03–1.12 (m, 2H),
1.20–1.30 (m, 3H), 1.48–1.80 (m, 4H), 2.35 (br, 1H), 2.53
(br, 1H), 3.53 (s, 3H), 4.30 (s, 1H), 5.62 (m, 1H), 5.83 (m,
1H), 7.29 (d, J = 8.3 Hz, 2H), 8.13 (d, J = 8.3 Hz, 2H). 13C
NMR (C6D6, 75 MHz) δ: 15.71 (-ve), 16.52, 24.57, 29.36,
32.11, 32.90 (-ve), 36.16 (-ve), 40.09, 51.57 (-ve), 75.22
(-ve), 128.05 (-ve), 128.55 (-ve), 129.04 (-ve), 129.19 (-ve),
129.35, 148.50, 166.82. HR-MS (CI) calcd. for C19H28NO3:
Data for cyclobutanol 9a
Colorless oil. IR (KBr, cm–1): 3441, 3027, 2923, 2874,
1
1479, 1452, 996, 882, 775, 702. H NMR (C6D6, 400 MHz)
δ: 0.89 (m, 1H), 1.14 (m, 1H), 1.23–1.27 (m, 1H), 1.28 (s,
3H), 1.33–1.39 (m, 2H), 1.52–1.77 (m, 5H), 1.82–1.97 (m,
2H), 2.60 (m, 1H), 2.71 (m, 1H), 7.07–7.18 (m, 5H). 13C
NMR (CDCl3, 75 MHz) δ: 17.77 (-ve), 18.29, 20.23, 22.55,
24.53, 27.56, 36.11 (-ve), 37.62 (-ve), 46.26, 46.37 (-ve),
82.64, 125.29 (-ve), 127.17 (-ve), 128.65 (-ve), 144.04. HR-
MS (EI) calcd. for C17H22O: 242.1671; found: 242.1672.
Anal. calcd. for C17H22O: C 84.25, H 9.15; found: C 84.00,
H 9.13.
+
318.2069 (M + NH4 ); found: 318.2069. Anal. calcd. for
C19H24O3: C 75.97, H 8.05; found: C 75.89, H 8.36. This
structure was confirmed by X-ray crystallographic analysis.
X-ray crystal structure analyses
Crystal data for compounds 1a, 1b, 9b, 10a, and 10b were
collected using a Rigaku/AFC7 four-circle diffractometer
equipped with an ADSC Quantum CCD detector. The X-ray
source was graphite-monochromated Mo Kα radiation (λ =
0.710 73 Å). Structures were solved by means of direct
methods (SHELXS-97 (20) for 1a, SIR92 (21) for 1b, 9b,
10a, and 10b), and all refinements were carried out by full-
matrix least-squres procedures (SHELXL-97) (22) using
anisotropic displacement parameters for all non-hydrogen at-
oms. Hydrogen atoms were introduced in calculated posi-
tions and only one overall isotropic displacement parameter
was refined. The CCDC reference numbers are CCDC
255968–255972 for compounds 1a, 1b, 9b, 10a, and 10b,
respectively, and 260038 for the mixed crystal resulting from
partial solid-state photolysis of ketone 1b.5
Data for alkene 10a
Melting point: 55–56.5 °C (EtOAc – petroleum ether). IR
(KBr, cm–1): 3431, 3057, 3015, 2940, 2918, 2864, 1492,
1
1472, 1453, 1015, 899, 763, 712, 608. H NMR (CD2Cl2,
300 MHz) δ: 0.90 (s, 3H), 1.18–1.43 (m, 4H), 1.52–2.04 (m,
5H), 2.48 (m, 1H), 2.70 (m, 1H), 4.97 (br, 1H), 5.73 (m,
1H), 6.00 (m, 1H), 7.20–7.40 (m, 5H). 13C NMR (CD2Cl2,
75 MHz) δ: 15.77 (-ve), 16.57, 24.75, 29.52, 32.32, 33.22
(-ve), 35.48 (-ve), 40.04, 75.92 (-ve), 127.23 (-ve), 127.77
(-ve), 128.12 (-ve), 128.97 (-ve), 129.16 (-ve), 143.07. HR-
MS (EI) calcd. for C17H22O: 242.1671 (M+); found:
242.1672. Anal. calcd. for C17H22O: C 84.25, H 9.15; found:
C 84.05, H 9.15. This structure was confirmed by X-ray
crystallographic analysis.
Preparative photolysis of ketone 1b
Crystal structure determination of ketone 1a
A solution of ketone 1b (501 mg, 1.67 mmol) in
acetonitrile (50 mL) was purged with nitrogen for 30 min
and irradiated for 6 h. Removal of the solvent in vacuo fol-
lowed by silica gel chromatography (11% EtOAc in petro-
leum ether) afforded cyclobutanol 9b (446 mg, 89%) and
alkene 10b (27 mg, 5%).
Single crystals of 1a were recrystallized from EtOAc –
petroleum ether, mounted in inert oil, and transferred to the
cold nitrogen stream of the diffractometer. Crystal data:
C17H22O, MW = 242.35, triclinic, a = 6.938 Å, b = 10.104 Å,
c = 10.197 Å, α = 101.473°, β = 98.041°, γ = 104.520°, V =
664.48 Å3, T = 173 K, space group P, Z = 2, µ(Mo Kα) =
0.073 mm–1, reflections measured = 4092, unique = 2552,
R1 = 0.0402, wR2 = 0.1053, GoF = 1.046.
Data for cyclobutanol 9b
Melting point: 149 to 150 °C (EtOAc – petroleum ether).
IR (KBr, cm–1): 3482, 2958, 2927, 2879, 1723, 1609, 1439,
1279, 1106, 1007, 861, 781, 712. 1H NMR (C6D6, 400 MHz)
δ: 0.85 (m, 1H), 0.98 (m, 1H), 1.20 (s, 3H), 1.21–1.35 (m,
3H), 1.45–1.65 (m, 5H), 1.83 (m, 2H), 2.52 (m, 1H), 2.66
(m, 1H), 3.53 (s, 3H), 7.09 (d, J = 8.2 Hz, 2H), 8.11 (d, J =
8.2 Hz, 2H). 13C NMR (C6D6, 100 MHz) δ: 17.70 (-ve),
18.61, 20.46, 22.87, 24.61, 27.73, 36.19 (-ve), 37.62 (-ve),
46.42, 46.79 (-ve), 51.61 (-ve), 81.87, 125.73 (-ve), 129.45,
Crystal structure determination of ketone 1b and its
single crystal-to-single crystal reaction
Single crystals of 1b were recrystallized from EtOAc–
hexanes, mounted in inert oil and transferred to the cold ni-
trogen stream of the diffractometer. Crystal data:
C19H24O3, MW = 300.38, monoclinic, a = 6.5508(3) Å, b =
11.0233(6) Å, c = 21.4693(13) Å, β = 92.627(2)°, V =
5 Supplementary data for this article are available on the Web site or may be purchased from the Depository of Unpublished Data, Document
Delivery, CISTI, National Research Council Canada, Ottawa, ON K1A 0S2, Canada. DUD 4027. For more information on obtaining mate-
rial refer to http://cisti-icist.nrc-cnrc.gc.ca/irm/unpub_e.shtml. CCDC 255968–255972 and 260038 contain the crystallographic data for this
manuscript. These data can be obtained, free of charge, via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge Crystallo-
graphic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax +44 1223 336033; or deposit@ccdc.cam.ac.uk).
© 2005 NRC Canada

1472
Can. J. Chem. Vol. 83, 2005
1548.70(14) Å3, T = 173 K, space group P21/c, Z = 4,
µ(Mo Kα) = 0.085 mm–1, reflections measured = 13 627,
unique = 3307, R1 = 0.0411, wR2 = 0.1059, GoF = 1.046.
Same crystal photolyzed to 25% conversion: C19H24O3,
MW = 300.38, monoclinic, a = 6.5305(2) Å, b =
11.2200(4) Å, c = 21.4348(7) Å, β = 91.892(1)°, V =
1569.72(1) Å3, T = 173 K, space group P21/c, Z =
4, µ(Mo Kα) = 0.085 mm–1, reflections measured = 24 848,
unique = 2768, R1 = 0.0424, wR2 = 0.1125, GoF = 1.058.
4. For general reviews of the solid-state ionic chiral auxiliary
method of asymmetric synthesis, see: (a) J.N. Gamlin, R.
Jones, M. Leibovitch, B.O. Patrick, J.R. Scheffer, and J. Trot-
ter. Acc. Chem. Res. 29, 203 (1996); (b) J.R. Scheffer. Can. J.
Chem. 79, 349 (2001).
5. D. Braga, S. Chen, H. Filson, L. Maini, M.R. Netherton, B.O.
Patrick, J.R. Scheffer, C. Scott, and W. Xia. J. Am. Chem. Soc.
126, 3511 (2004).
6. P.J. Wagner and G. Laidig. Tetrahedron Lett. 32, 895 (1991).
7. The preference for abstraction of Hx over Hy in the case of ke-
tones 1a and 1b is in accord with previous work from our lab-
oratory on the geometric requirements for Norrish–Yang type
II hydrogen abstraction. See: H. Ihmels and J.R. Scheffer. Tet-
rahedron 55, 885 (1999).
Crystal structure determination of ketone 9b
Single crystals of 9b were recrystallized from EtOAc –
petroleum ether, mounted in inert oil, and transferred to the
cold nitrogen stream of the diffractometer. Crystal data:
C19H24O3, MW = 300.38, triclinic, a = 6.2815(4) Å, b =
7.2477(6) Å, c = 18.7120(17) Å, α = 79.329(9)°, β =
82.478(10)°, γ = 69.225(8)°, V = 780.75(11) Å3, T = 173 K,
space group P, Z = 2, µ(Mo Kα) = 0.085 mm–1, reflections
measured = 3144, unique = 3144, R1 = 0.0367, wR2 =
0.1027, GoF = 1.086.
8. E. Cheung, M.R. Netherton, J.R. Scheffer, and J. Trotter. Org.
Lett. 2, 77 (2000).
9. P.J. Wagner, P.A. Kelso, and A.E. Kempainen. J. Am. Chem.
Soc. 90, 5896 (1968).
10. J.K. Burdett. Molecular shapes. Wiley-Interscience, New York.
1980. p. 6.
11. A. Bondi. J. Phys. Chem. 68, 441 (1964).
12. (a) F.D. Lewis and T.A. Hilliard. J. Am. Chem. Soc. 92, 6672
(1970); (b) F.D. Lewis and R.A. Ruden. Tetrahedron Lett. 715
(1971); (c) P.J. Wagner, P.A. Kelso, A.E. Kemppainen, J.M.
McGrath, H.N. Schott, and R.G. Zepp. J. Am. Chem. Soc. 94,
7506 (1972).
13. The distance between the flagpole hydrogen atom on C5 and
the radical center on C1 in boat–chair biradical 14 is estimated
to be 2.3 Å, as the calculated flagpole hydrogen to hydrogen
distance in the boat conformation of cyclohexane is 2.36 Å.
See: F. Freeman, A. Phornvoranunt, and W.J. Hehre. J. Phys.
Org. Chem. 12, 176 (1999).
Crystal structure determination of ketone 10a
Single crystals of 10a were recrystallized from EtOAc –
petroleum ether, mounted in inert oil, and transferred to the
cold nitrogen stream of the diffractometer. Crystal data:
C17H22O, MW = 242.35, monoclinic, a = 10.4796(10) Å,
b = 19.2773(18) Å, c = 13.6457(13) Å, β = 93.507(3)°, V =
2751.5(5) Å3, T = 173 K, space group P21/c, Z = 8,
µ(Mo Kα) = 0.070 mm–1, reflections measured = 23 886,
unique = 5933, R1 = 0.0618, wR2 = 0.1553, GoF = 1.128.
14. (a) M.C. Etter and D.A. Adsmond. J. Chem. Soc. Chem.
Commun. 589 (1990); (b) S.L. Johnson and K.A. Rumon. J.
Phys. Chem. 69, 74 (1965).
Crystal structure determination of ketone 10b
Single crystals of 10b were recrystallized from CHCl3–
hexanes, mounted in inert oil, and transferred to the cold ni-
trogen stream of the diffractometer. Crystal data:
C19H24O3, MW = 300.38, monoclinic, a = 15.1864(11) Å,
b = 6.8150(5) Å, c = 14.9981(10) Å, β = 90.383(4)°, V =
1552.20(19) Å3, T = 173 K, space group P21/n, Z = 4,
µ(Mo-Kα) = 0.085 mm–1, reflections measured = 3331,
unique = 3331, R1 = 0.0378, wR2 = 0.0949, GoF = 0.928.
15. For exceptions to this rule, see: R.-F. Liao, J.W. Lauher, and
F.W. Fowler. Tetrahedron, 52, 3153 (1996).
16. D. Dolphin and A. Wick. Tabulation of infrared spectral data.
Wiley, New York. 1977. p. 295.
17. For a similar situation in which the enantioselectivity of a
Norrish–Yang photoreaction in the crystalline state is governed
by preferential abstraction of the geometrically favored γ-hy-
drogen atom, see ref. 2 as well as: S. Chen, B.O. Patrick, and
J.R. Scheffer. J. Org. Chem. 69, 2711 (2004).
Acknowledgements
18. E. Cheung, T. Kang, M.R. Netherton, J.R. Scheffer, and J.
Financial support from the Natural Sciences and Engi-
neering Research Council of Canada (NSERC) is gratefully
acknowledged.
Trotter. J. Am. Chem. Soc. 122, 11 753 (2000).
19. For a discussion of photochemical solid-to-solid reactions (in-
cluding single crystal-to-single crystal transformations), see:
A.E. Keating and M.A. Garcia-Garibay In Molecular and
supramolecular photochemistry. Vol. 2. Edited by V.
Ramamurthy and K.S. Schanze. Marcel Dekker, New York.
1998. Chap. 5.
20. (a) G.M. Sheldrick. SHELXS-97: A program for automatic so-
lution of crystal structures [computer program]. University of
Göttingen, Göttingen, Germany. 1997. Release 97-2; (b) G.M.
Sheldrick. Acta Crystallogr. Sect. D, D49, 18 (1993).
21. A. Altomare, G. Cascarano, C. Giacovazzo, and A. Gualardi.
J. Appl. Crystallogr. 26, 343 (1993).
References
1. For general reviews of the Norrish–Yang type II reaction, see:
(a) P.J. Wagner and B.S. Park. In Organic photochemistry. Vol.
11. Edited by A. Padwa. Marcel Dekker, New York. 1991.
Chap. 4; (b) A.G. Griesbeck, S. Buhr, M. Fiege, H.
Schmickler, and J. Lex. J. Org. Chem. 63, 3847 (1998).
Cyclobutanol products in Norrish type II photochemistry were
first reported by: N.C. Yang and D.H. Yang. J. Am. Chem.
Soc. 80, 2913 (1958).
22. G.M. Sheldrick. SHELXL-97: A program for crystal structure
refinement [computer program]. University of Goettingen,
Germany. 1997. Release 97-2.
2. M. Leibovitch, G. Olovsson, J.R. Scheffer, and J. Trotter. J.
Am. Chem. Soc. 120, 12755 (1998).
3. B.O. Patrick, J.R. Scheffer, and C. Scott. Angew. Chem. Int.
Ed. 42, 3775 (2003).
© 2005 NRC Canada