Mediation of conformationally controlled photodecarboxylations of chiral and cyclic aryl esters by substrate structure, temperature, pressure, and medium constraints

Article abstract of DOI:10.1021/ja049688w

An aryl alkanoate, 2,4,6-trimethylphenyl (S)-(+)-2-methylbutyrate, whose ester group has a chiral center alpha to the carbonyl carbon and in which photo-Fries rearrangements are blocked by methyl substituents, undergoes facile photodecarboxylation under a variety of conditions and with complete retention of configuration. In fact, the decarboxylation process has many of the attributes of a symmetry-allowed suprafacial [1,3]sigmatropic rearrangement. The process requires concerted extrusion of carbon dioxide in a spiro-lactonic transition state, which has been investigated using high level DFT and CIS calculations: thermally less stable s-cis conformers in the ground and excited singlet states play an important role in determining the competitive efficiency of the process. Conformational control has also been imposed by substrate structure, solvent interactions, temperature, and applying external pressure, as well as using constraining media such as cyclodextrins and polyethylene films. The results are correlated with steady-state and dynamic fluorescence measurements at various temperatures in order to investigate further degrees to which ground and excited singlet state conformations affect the different photoreactivity channels available to the aryl esters.

Full text of DOI:10.1021/ja049688w

Published on Web 07/01/2004
Mediation of Conformationally Controlled
Photodecarboxylations of Chiral and Cyclic Aryl Esters by
Substrate Structure, Temperature, Pressure, and Medium
Constraints
Tadashi Mori,*,† Richard G. Weiss,‡ and Yoshihisa Inoue*,†,§
Contribution from the Department of Molecular Chemistry, Osaka UniVersity,
2-1 Yamada-oka, Suita 565-0871, Japan; Department of Chemistry, Georgetown UniVersity,
Washington, D.C., 20057-1227; and Entropy Control Project, ICORP, JST, 4-6-3 Kamishinden,
Toyonaka 560-0085, Japan
Received January 17, 2004; E-mail: tmori@chem.eng.osaka-u.ac.jp
Abstract: An aryl alkanoate, 2,4,6-trimethylphenyl (S)-(+)-2-methylbutyrate, whose ester group has a chiral
center alpha to the carbonyl carbon and in which photo-Fries rearrangements are blocked by methyl
substituents, undergoes facile photodecarboxylation under a variety of conditions and with complete retention
of configuration. In fact, the decarboxylation process has many of the attributes of a symmetry-allowed
suprafacial [1,3]sigmatropic rearrangement. The process requires concerted extrusion of carbon dioxide in
a spiro-lactonic transition state, which has been investigated using high level DFT and CIS calculations:
thermally less stable s-cis conformers in the ground and excited singlet states play an important role in
determining the competitive efficiency of the process. Conformational control has also been imposed by
substrate structure, solvent interactions, temperature, and applying external pressure, as well as using
constraining media such as cyclodextrins and polyethylene films. The results are correlated with steady-
state and dynamic fluorescence measurements at various temperatures in order to investigate further
degrees to which ground and excited singlet state conformations affect the different photoreactivity channels
available to the aryl esters.
Introduction
phenyl esters, the acyl radical then adds at the ortho or para
position of phenoxy (based on their high spin densities) or
Photo-Fries rearrangements (Scheme 1, paths A and A′) are
the most common and best investigated photoreactions of aryl
alkanoates.1-9 The reaction is generally devoid of large amounts
of side products (other than those from separation of the radical
pair and subsequent H-atom abstractions), and the quantum
yields are usually 0.1-0.3. CIDNP experiments indicate that
aryloxy-carbonyl bond cleavage proceeds from excited singlet
states, leading to acyl-aryloxy radical pairs.10-12 In simple
reforms a bond with the phenoxy oxygen (regenerating the initial
ester in its ground state). The so-formed ortho or para adducts
are keto intermediates that tautomerize to the isolated acylphenol
photoproducts: a flash photolysis study of phenyl acetate has
detected the presence of two transient species (presumably the
keto intermediates) whose different rates of decay can be
correlated with the rates of appearance of the o- and p-acetyl-
phenols.13,14
Another potential photoreaction of aryl esters, decarboxylation
(Scheme 1, path C), has been investigated to a much smaller
extent, largely because it usually represents a minor reaction
pathway during the photolyses. However, when the structures
of aromatic esters or their environments are “engineered”
correctly, photodecarboxylation can be an important reaction
component.15 For example, irradiations of 2-tert-butylphenyl and
2,4,6-trimethylphenyl pivalates and 3,5-di-tert-butylphenyl ben-
zoate have been reported to give 1%, 5%, and 7-14% yields,
respectively, of the corresponding decarboxylation products.16,17
† Osaka University.
‡ Georgetown University.
§ ICORP.
(1) Miranda, M. A.; Galindo, F. In CRC Handbook of Organic Photochemistry
and Photobiology, 2nd ed.; Horspool, W. M., Lenci, F., Eds.; CRC Press
LLC: Boca Raton, FL, 2004; Chapter 42.
(2) Miranda, M. A.; Galindo, F. In Molecular and Supramolecular Photo-
chemistry; Ramamurthy, V., Schanze, K. S., Eds.; Marcel Dekker: New
York, 2003; Vol. 9, pp 43-131.
(3) Guisnet, M.; Perot, G. Fine Chem. Hetero. Catal. 2001, 211-216.
(4) Rusu, E.; Comanita, E.; Onciu, M. Roum. Chem. Q. ReV. 2000, 7, 241-
250.
(5) Pincock, J. A. In CRC Handbook of Organic Photochemistry and
Photobiology; Horspool, W. M., Song, P.-S., Eds.; CRC Press: Boca Raton,
FL, 1995; pp 393-407.
(6) Martin, R. Org. Prep. Proced. Int. 1992, 24, 369-435.
(7) Pfau, M.; Julliard, M. Bull. Soc. Chim. Fr. 1977, 785-802.
(8) Coyle, J. D. Chem. ReV. 1978, 78, 97-123.
(13) Arai, T.; Tobita, S.; Shizuka, H. J. Am. Chem. Soc. 1995, 117, 3968-
3975.
(14) Kalmus, C. E.; Hercules, D. M. J. Am. Chem. Soc. 1974, 96, 449-456.
(15) Bradshaw, J. S.; Loveridge, E. L.; White, L. J. Org. Chem. 1968, 33, 4127-
4132.
(16) Finnegan, R. A.; Knutson, D. Chem. Commun. 1966, 172-173.
(17) Finnegan, R. A.; Knutson, D. Chem. Ind. (London) 1965, 1837-1838.
(9) Bellus, D. AdV. Photochem. 1971, 8, 109-159.
(10) Gritsan, N. P.; Tsentalovich, Y. P.; Yurkovskaya, A. V.; Sagdeev, R. Z. J.
Phys. Chem. 1996, 100, 4448-4458.
(11) Adam, W. J. Chem. Soc., Chem. Commun. 1974, 289-290.
(12) Adam, W.; Arce de Sanabia, J. J. Org. Chem. 1973, 38, 2571-2572.
9
10.1021/ja049688w CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 8961-8975
8961

A R T I C L E S
Mori et al.
Scheme 1. Photo-Fries Rearrangements and Photodecarboxylation of Aryl Esters
Scheme 2. Photodecarboxylation of Aryl Esters (Additional
Photoproducts from Photo-Fries Reactions Are Not Shown)
Prolonged irradiation of p-tolyl ferrocenoate also yielded
ferrocenoic acid and p-tolylferrocene in 2% and 5% yields,
respectively.18 Photoreaction of the coumarin derivative, di-
acetylhexahydromarmesin, in ethanol afforded 4% of decar-
boxylation product together with photo-Fries rearrangement and
cage-escape products (30% and 13% yields, respectively).19 In
all of these cases, the yield of photodecarboxylation is low, and
concomitant formation of photo-Fries and cage-escape products
were observed. Recently, we have prepared 2,4,6-trimethyl-
phenyl (mesityl) esters, which generally provide good yields
of decarboxylation products upon irradiation.20,21 For instance,
photolysis of mesityl cyclohexanecarboxylate in acetonitrile
affords the decarboxylation product in good yield in the absence
of acid, while, in the presence of a catalytic amount of acid
and alcohol, the same substrate undergoes efficient transesteri-
fication.20
Here, we investigate the photodecarboxylation of chiral aryl
ester mesityl (S)-(-)-2-methylbutanoate (1), 2,4,6-trimethyl-
phenyl trans-4-methylcyclohexanecarboxylate (4), and a struc-
turally constrained cyclic ester, 3,4,5,6-tetrahydro-2H-1-ben-
zoxocin-2-one (6), under a variety of conditions to distinguish
whether a concerted or a stepwise mechanism is operative and
to determine how the photodecarboxylation pathway can be
optimized by varying several environmental reaction conditions
(e.g., temperature, solvent, pressure, and confinement in cyclo-
dextrins and polyethylene films). We did not try to analyze for
the other fragments from the cage escape products; we focused
on the aromatic part of the fragments throughout the studies.
The decarboxylation process has many of the attributes of a
symmetry-allowed suprafacial [1,3]sigmatropic rearrangement.22-24
The natures and the divisions among the various photoreactions
available to 1 (and to 4 and 6) must depend on the ground and
excited states conformational equilibria as well as the rates at
which various conformers interconvert; in that regard, the s-cis
conformer must be the immediate precursor of the decarboxy-
lation if it is concerted and proceeds with retention of config-
uration of the chiral center (Scheme 1). It is the only
conformation capable of attaining directly the spiro-lactonic
transition state that allows expulsion of carbon dioxide while
retaining the stereochemical integrity of the chiral carbon center.
The properties of the excited singlet state precursors to the
reactions have been interrogated as well by steady-state
fluorescence measurements at various temperatures and dynam-
ics of the fluorescence, and their equilibria (especially between
the key s-cis conformer and other conformers such as the
s-trans) have been investigated computationally. Complete
stereospecificity and reaction specificity, leading exclusively to
the photodecarboxylation product, have been achieved during
irradiation of 1 in unstretched high-density polyethylene films.
Results and Discussion
Irradiation at 254 nm of chiral ester 1 in acetonitrile at -25
°C under an argon atmosphere leads to efficient decarboxylation
to yield alkylmesitylene 2 in good yields. A smaller amount of
2,4,6-trimethylphenol (mesitol, 3) was also obtained, in which
the cage-escape process is also operative (Scheme 2). Time
profiles for the distribution and appearance of the products
(Figure 1) were monitored by GC. Prolonged irradiation (>40
min) did not increase the yield of photoproducts despite the
continued loss of starting ester 1. Secondary reactions, such as
photoisomerization of the alkylbenzene21 or dimerization of the
phenol, must become competitive with formation of additional
2 and 3 from remaining 1 under these conditions. Irradiations
conducted for 1 h under a variety of experimental conditions
(i.e., high conversions of 1) resulted in poor material balances
(30-60%) but did provide large amounts of photodecarboxyl-
ation product for stereochemical analyses; material balances
were almost quantitative after 10-min irradiation periods and
(18) Finnegan, R. A.; Mattice, J. J. Tetrahedron 1965, 21, 1015-1026.
(19) Ishii, H.; Sekiguchi, F.; Ishikawa, T. Tetrahedron 1981, 37, 285-290.
(20) Mori, T.; Wada, T.; Inoue, Y. Org. Lett. 2000, 2, 3401-3404.
(21) Mori, T.; Takamoto, M.; Wada, T.; Inoue, Y. Photochem. Photobiol. Sci.
2003, 2, 1187-1199.
(22) Cookson, R. C.; Kemp, J. E. G. J. Chem. Soc., D 1971, 385-386.
(23) Cookson, R. C.; Hudec, J.; Sharma, M. M. J. Chem. Soc., D 1971, 108.
(24) Cookson, R. C.; Hudec, J.; Sharma, M. M. J. Chem. Soc., D 1971, 107-
108.
9
8962 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
Scheme 3. Potential Routes for Loss of Chirality in the Decarboxylation Product Assuming a Radical Mechanism
scission. The dominant fate of such radical pair intermediate(s)
is to recombine and reform the starting esters.20,21,33,34 The
involvement of a large component of reaction proceeding via
such radical species prevents us from performing studies using
techniques such as laser flash photolysis and transient EPR
spectroscopy to ascertain the details of the (minor) photode-
carboxylation mechanism.
As a substitute mechanistic tool, we have followed the
stereochemical course of the photodecarboxylation of a proto-
typical chiral probe, mesityl (S)-(-)-2-methylbutanoate (1), as
a substitute mechanistic tool.35 If radical intermediates are
involved in the decarboxylation, the resultant photoproduct
should suffer loss of some chiral integrity. Such an approach
was employed qualitatively by Finnegan and Knutson in 1967.36
They isolated a 15-20% yield of the decarboxylation product
after irradiating optically active 3,5-di-tert-butylphenyl (S)-
(-)-2-methylbutanoate in dioxane. The optical rotatory disper-
sion (ORD) spectrum of the product was a positive curve,
suggesting the (S)-(-) configuration (i.e., net retention) based
on the ORD spectrum of the phenyl analogue. Although the
authors suspected that the degree of optical purity of the
decarboxylation product was high, no quantification was pos-
sible with the information available. Such an approach using a
chiral probe to determine the nature of the involvement by the
radical species has been recently adopted for other systems.37-39
In the present system, if photodecarboxylation involves a
radical pair (i.e., cleavage of 1, loss of CO2 from one of the
fragments, and then recombination), the chiral purity of 2 should
be less than that of 1 because some of the radicals are expected
to rotate or move from one face to the other (without rotating)
before combining (Scheme 3). There is ample precedent for loss
of at least partial stereointegrity in analogous thermal and
photochemical reactions.40-45 The enantiomeric excess (ee) of
Figure 1. Product distribution from (S)-1 in acetonitrile under argon at
-25 °C versus irradiation time. Black: (S)-1. Blue: 2. Red: 3. Open
circle: material balance (1 + 2 + 3).
remained good (>85%) after 20-40 min of photolysis. There-
fore, irradiations where mechanistic information was sought
from the product distributions were usually stopped after 10-
20% conversions of ester, corresponding to 5-10 min of
irradiation at room temperature and longer periods at subambient
temperatures.
Retention of Chiral Configuration during Photodecar-
boxylation of 1. The mechanism of decarboxylation of benzylic-
type molecules has been well documented, and the reaction has
exploited to synthesize difficult to make species, such as
cyclophanes.25-28 The mechanism involves homolysis to form
benzyl radical pairs, most of which react within the solvent cage
(to yield the decarboxylation product(s)) faster than escaping
from it.29-32 In contrast, the mechanism of decarboxylation of
aryl esters has been less investigated even though the photo-
decarboxylation has been observed, especially when the path-
ways leading to photo-Fries products are fully or partially
blocked by substituents of the aryl portions of the esters. Even
in those cases, the major process occurring in the excited singlet
states remains the initial step in the photo-Fries reactions, â-bond
(33) Mori, T.; Takamoto, M.; Saito, H.; Furo, T.; Wada, T.; Inoue, Y. Chem.
Lett. 2004, 33, 254-255.
(34) Mori, T.; Takamoto, M.; Saito, H.; Furo, T.; Wada, T.; Inoue, Y. Chem.
Lett. 2004, 33, 256-257.
(25) Pincock, J. A. Acc. Chem. Res. 1997, 30, 43-49.
(35) Mori, T.; Saito, H.; Inoue, Y. Chem. Commun. 2003, 2302-2303.
(36) Finnegan, R. A.; Knutson, D. J. Am. Chem. Soc. 1967, 89, 1970-1972.
(37) Bhanthumnavin, W.; Bentrude, W. G. J. Org. Chem. 2001, 66, 980-990.
(38) Bhanthumnavin, W.; Arif, A.; Bentrude, W. G. J. Org. Chem. 1998, 63,
7753-7758.
(26) Hibert, M.; Solladie, G. J. Org. Chem. 1980, 45, 4496-4498.
(27) Truesdale, E. A. Tetrahedron Lett. 1978, 3777-3780.
(28) Kaplan, M. L.; Truesdale, E. A. Tetrahedron Lett. 1976, 3665-3666.
(29) Givens, R. S.; Matuszewski, B.; Levi, N.; Leung, D. J. Am. Chem. Soc.
1977, 99, 1896-1903.
(39) Gao, F.; Boyles, D.; Sullivan, R.; Compton, R. N.; Pagni, R. M. J. Org.
Chem. 2002, 67, 9361-9367.
(30) Givens, R. S.; Matuszewski, B. J. Am. Chem. Soc. 1975, 97, 5617-5619.
(31) Givens, R. S.; Matuszewski, B.; Neywick, C. V. J. Am. Chem. Soc. 1974,
96, 5547-5552.
(40) Lee, K.-W.; Horowitz, N.; Ware, J.; Singer, L. A. J. Am. Chem. Soc. 1977,
99, 2622-2627.
(32) Givens, R. S.; Oettle, W. F. J. Am. Chem. Soc. 1971, 93, 3301-3302.
(41) Johnson, R. A.; Seltzer, S. J. Am. Chem. Soc. 1973, 95, 938-939.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8963

A R T I C L E S
Mori et al.
Figure 2. Chiral GC separation of 2 on a SUPELCO beta-DEX 325 column (30 m × 0.25 mm; thickness 0.25 µm, column temp ) 100 °C). For details,
see Experimental Section in the Supporting Information. (a) Racemic 2. (b) Photoproduct 2 obtained upon photolysis of (S)-1. (c) Typical chromatogram of
photoproduct mixture. (d) Coinjection of racemic 2 and photolysis mixture. Note that the earlier retention time peak from 2 is the larger.
Table 1. Solvent and Temperature Effects on Irradiation of (S)-1a
irrad
time
(min)
yields (%)c
[1]
(mM)
T
(°C)
conversion
(%)
2/3
ratio
ee of
2 (%)d
e
solvent
solvent propertiesb
2
3
MB
acetonitrile
0.34 1.344 35.94 3.95 45.6 14.1 2.0 -25 60
73
59
61
67
35
53
14
32
9
41
36
4.2
1.9
7.4
>99.5 48
>99.9 39
>99.5 32
2.0
25 60
25 60
25 60
75 60
93
2.0
2.0
2.0
3.1
76f
76g
89
0.85 >99.5 61
1.5
5.92 >99
5.21 >99
>99.5 28
0
0
5.0 19.7 ( 0.7
5.0 19.3 ( 0.5
78.8 ( 2.6
58.9 ( 4.1
13.3 ( 1.0
11.3 ( 1.2
92
70
acetonitrile
(purified)h
methylcyclohexane 0.73 1.423 2.02
0.34 1.344 35.94 3.95 45.6 14.1 3.1
0
3.1 -95 24 h not detected not detected not detected ndj
3.1 -80 24 h 0.1 ( 0.1
not detected not detected ndj
-45 30 8.9 ( 0.4 78.6 ( 11.6 19.5 ( 3.0
ndj
ndj
ndj
ndj
98
98
99
97
93
92
87
98
93
47
72
71
83
72
79
45
4.03 >99
5.32 >99
6.82 >99
6.05 >99
4.58 >99
2.81 >99
2.23 >99
3.26 >99
4.08 >99
3.94 >99
4.06 >99
4.70 >99
2.05 >99
1.19 >99
7.12 >99
1.44 >99
-25 10.0 7.3 ( 0.2
-10 10.0 14.4 ( 0.4
82.5 ( 5.0
86.6 ( 2.7
82.9 ( 2.1
76.5 ( 3.1
68.0 ( 2.1
60.3 ( 2.6
82.9 ( 2.1
74.7 ( 5.6
74.7 ( 4.7
37.8 ( 3.8
58.0 ( 5.3
58.3 ( 4.1
15.5 ( 1.4
12.7 ( 0.7
13.7 ( 0.6
16.7 ( 0.5
24.2 ( 1.6
27.0 ( 3.2
13.7 ( 0.6
22.9 ( 1.5
18.3 ( 1.8
9.6 ( 0.4
0
+25
+60
+95
0
5.0 15.1 ( 0.4
5.0 15.6 ( 0.4
5.0 13.5 ( 0.3
5.0 8.5 ( 0.2
5.0 15.8 ( 1.3
5.0 14.2 ( 0.8
5.2 26.4 ( 0.3
5.0 17.0 ( 0.5
5.0 14.8 ( 0.1
5.0 12.6 ( 0.6
5.0 16.9 ( 0.6
5.0 9.0 ( 0.3
10.0 7.2 ( 0.8
diethyl ether
0.24 1.352 4.42 1.11 34.5 19.2 3.3
tetrahydrofuran
hexane (dried)i
methanol (dried)i
diglyme
propionitrile
dichloromethane
1,4-dioxane
0.58 1.406 7.47 1.69 37.4
0.31 1.375 1.89 31
0.59 1.328 32.66 1.69 55.4
0.46 1.408 7.3 1.92 38.6
20
0
30
3.3
3.1
3.1
3.1
0
0
0
0
0
0
0
14.3 ( 0.2
12.4 ( 0.2
0.41 1.366 28.86 4.04 43.6 16.1 3.4
0.45 1.424 8.93 1.62 40.7 3.3
1
56.0 ( 13.8 27.3 ( 0.1
39.3 ( 2.1
69.1 ( 2.6
1.44 1.422 2.27 0.46
0.66 1.252 1.57
36
14.3 3.1 +10
33.0 ( 1.1
9.7 ( 1.0
perfluorohexane
3.1
0
a All irradiations were carried out with a 30 W low-pressure mercury lamp (Eikosha PIL-type) in Pyrex tubes (diameter ) 1 cm) under an argon atmosphere
(λ ) 254 nm), unless otherwise stated. b Solvent properties in reference shown: viscosity (mPa s), reflactive index, permittivity, dipole moment, Reichardt’s
polarity ET(30), and donor number. From Abboud, J.-L. M.; Notario, R. Pure Appl. Chem. 1999, 71, 645-718. c Absolute yields of 2 and 3 based on
consumed 1, from an average of at least three separate experiments, except for 60-min irradiations. d %ee of product 2, determined by chiral GC. Error )
f
(0.2%. e Material balance (%). Saturated with oxygen (ca. 8 mM). g In the presence of TEMPO (2.0 mM). h Acetonitrile (RdH, Chromasolv) was purified.
See Experimental Section in the Supporting Information for details. i Hexane or methanol (RdH, Chromasolv) was distilled from CaH2 under Ar. j Not
determined.
the photodecarboxylation product 2 in our experiments was
determined by chiral GC (Figure 2). A near baseline separation
of the racemic product 2 was achieved and a <0.2% standard
deviation of the error in the enantiomeric ratios could be
maintained for the racemates. The detection limit of each
enantiomer was ∼1% at the lower conversions of 1 and was
0.5% at higher conversions (e.g., 1-h irradiations). In some cases,
the product ee was also determined on a chiral HPLC (see
Supporting Information for detailed conditions of analyses of
ee’s). None of the other isomers were detected, and the ee of 2
remained >99% when 1 was irradiated for 1 h in acetonitrile
at temperatures between 75 and -25 °C. The addition of a
radical trapping agent, TEMPO or molecular oxygen, to the
acetonitrile solutions had no effect on the ee (Table 1). In one
experiment at 25 °C, a known amount of racemic 2 was added
and the mixture was analyzed by GC to ascertain the purity of
the product 2. The calculated ee of the photochemically
produced 2 was >99.9%. The ee of 2 remained >99% regardless
of the conversion of 1 and when any of the other experimental
(42) Tsolis, A.; Mylonakis, S. G.; Nieh, M. T.; Seltzer, S. J. Am. Chem. Soc.
1972, 94, 829-833.
(43) Greene, F. D.; Berwick, M. A.; Stowell, J. C. J. Am. Chem. Soc. 1970, 92,
867-874.
(44) Engstrom, J. P.; Greene, F. D. J. Org. Chem. 1972, 37, 968-972.
(45) Koenig, T.; Owens, J. M. J. Am. Chem. Soc. 1973, 95, 8484-8486.
9
8964 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
point. As reported previously,21,56 the photoproduct distributions
from aryl esters can depend on the method that some solvents
are purified. Indeed, the photoproduct distributions from 1, 4,
and 6, but not the enantio- or diastereo-purity of the decar-
boxylation products, did depend (reproducibly within the same
solVent batch) on the source of some solvents or how they were
purified. However, the results could be reproduced whenever a
solvent from the same batch was employed for the photolysis.
Except as indicated, the data reported here were obtained from
one batch for each solvent. We suspect that the source of this
dependence is trace impurities within the solvents that react
rapidly with the radical intermediates, but the results indicate
that “purification” increases the concentrations of these species!
The highest yields of 2 were obtained in methylcyclohexane
and in acetonitrile, the 2/3 ratio being 6.1 and 5.9, respectively,
at 0 °C. Purification of acetonitrile (see the Experimental Section
in the Supporting Information for details) increased the relative
yield of the cage-escape product slightly; the ratio of 2/3 was
5.2, but surprisingly, the material balance was significantly lower
than that when the irradiation was conducted in nonpurified
acetonitrile. Better yields of photo-Fries rearrangement products
from 2-naphthyl esters have been reported in spectrograde
hexane than in purified hexane.56 Reaction of 1 in (nonpurified,
not spectrograde) propionitrile afforded a lower 2/3 ratio with
lower material balance than that in acetonitrile. Reaction in
distilled hexane also afforded a lower relative yield of 2 and a
much lower material balance (47%) than in undistilled hexane.
The highest 2/3 ratio from reaction in an isotropic solution,
7.1, was found in 1,4-dioxane (at 10 °C). In dichloromethane,
the relative yields of 2 and 3 were almost equal and the yield
of 3 was actually higher than that of 2 in perfluorohexane, where
the material balance was poor. Although the ionic liquid, 1-ethyl-
3-methyl-1H-imidazolium tetrafluoroborate, has been used as
a “green” photochemical solvent for several transformations,57-59
no conversion of 1 was detected after prolonged irradiation,
probably as a result of absorption of the radiation by the
medium. No direct correlation between the product distributions
and solvents parameters such as viscosity and polarity was
obvious to us.
Figure 3. Circular dichroism spectrum of 2 from photolysis of (S)-1 (1
mM) in hexane at 25 °C.
parameters (such as temperature, solvent, or pressure) was
changed.
Although the data presented thus far demonstrate the stereo-
specificity of the photodecarboxylation reaction, they do not
distinguish between complete retention and the less likely
possibility, complete inversion. The specific rotation [R]D in
hexane of the photodecarboxylation product 2 (+11.9°) is nearly
the same as that of (S)-(+)-1 (+11.7° in chloroform), and
although they depend somewhat on solvent,46 they remain
similar within any one solvent. This result suggests that the
decarboxylation process occurs with retention of configuration.
This assertion has been confirmed by circular dichroic (CD)
spectroscopy of the product 2 in hexane (Figure 3). A weak
1
positive Cotton effect in the Lb band region and fairly strong
positive and negative Cotton effects at shorter wavelengths were
found: molar circular dichroism (∆ꢀ) values were +0.063 (264
nm), +1.78 (224 nm), and -1.16 (210 nm) M-1 cm-1, and these
extrema correspond to the maxima in the UV absorption
spectrum (log ꢀ ) 3.35 (268 nm), 4.68 (222 nm, sh), and 5.39
(203 nm)). According to the benzene chirality and benzene
sector rules,47-52 a positive Cotton effect in the 1Lb band region
must be from the (S)-isomer. The CD spectrum was also very
similar to other chiral (S)-alkenylbenzenes.53 Our conclusion
that the photodecarboxylation proceeds with net retention of
configuration is further demonstrated by the diastereospecific
photodecarboxylation of 4 (vide infra). These results exclude
involvement of radical pairs during the photodecarboxylation
of aryl esters, at least those like the ones investigated here.
Solvent Effects on the Competition between Photodecar-
boxylation and Photo-Fries Processes. The distribution of
photo-Fries products and the ratio of decarboxylation to Fries
product yields are known to be solvent dependent.54,55 Table 1
summarizes the results from irradiations of 1 for 5 min at 0 °C
(typically 10∼20% conversions) in a variety of solvents; reaction
was performed at 10 °C in 1,4-dioxane because of its melting
Like the photo-Fries rearrangement, the photodecarboxylation
reaction is believed to occur from excited singlet states.31,60
Attempted triplet sensitization of 1 (3 mM) in acetone (ET )
332 kJ/mol) as solvent and sensitizer, using >280 nm where
acetone absorbs strongly and 1 does not, was unsuccessful. No
conversion was detected even after 6 h of irradiation. After
irradiation for 24 h, small amounts of unknown products were
observed, but no 2 or 3 could be detected. The estimated triplet
energy of the ester 1 is ∼330 kJ/mol.61 Other triplet sensitizers
whose energies are probably somewhat lower than the triplet
of 1, such as acetophenone (ET ) 310 kJ/mol) and benzophe-
none (ET ) 287 kJ/mol), were also unsuccessful. Similarly,
reactions of aryl esters 1 and 4 were not hindered by addition
of naphthalene (ET ) 253 kJ/mol, ca. 10 mM), a known
quencher of triplet states of aromatic esters.62 Although these
(46) Baxter, J. G.; Robeson, C. D.; Taylor, J. D.; Lehman, R. W. J. Am. Chem.
Soc. 1943, 65, 918-924.
(47) Smith, H. E. In Circular Dichroism, Principles and Applications, 2nd ed.;
Berova, N., Nakanishi, K., Woody, R. W., Eds.; John Wiley & Sons: New
York, 2000; Chapter 14, pp 397-429.
(48) Rumbero, A.; Borreguero, I.; Sinisterra, J. V.; Alcantara, A. R. Tetrahedron
1999, 55, 14947-14960.
(56) Cui, C.; Wang, X.; Weiss, R. G. J. Org. Chem. 1996, 61, 1962-1974.
(57) Gordon, C. M. NATO Sci. Ser. II 2003, 92, 365-383.
(58) Gordon, C. M.; McLean, A. J.; Muldoon, M. J.; Dunkin, I. R. ACS Symp.
Ser. 2002, 818, 428-443.
(49) Smith, H. E.; Neergaard, J. R. J. Am. Chem. Soc. 1997, 119, 116-124.
(50) Smith, H. E.; Neergaard, J. R. J. Am. Chem. Soc. 1996, 118, 7694-7701.
(51) Michals, D.; Smith, H. E. Chirality 1993, 5, 20-23.
(52) Smith, H. E. Chem. ReV. 1998, 98, 1709-1740.
(59) Hondrogiannis, G.; Lee, C. W.; Pagni, R. M.; Mamantov, G. J. Am. Chem.
Soc. 1993, 115, 9828-9829.
(53) Lardicci, L.; Salvadori, P.; Caporusso, A. M.; Menicagli, R.; Belgodere,
E. Gazz. Chim. Ital. 1972, 102, 64-84.
(60) Budac, D.; Wan, P. J. Photochem. Photobiol., A 1992, 67, 135-166.
(61) Kanda, Y.; Shimada, R.; Kakenoshita, Y. Spectrochim. Acta 1963, 19,
1249-1260.
(54) Finnegan, R. A.; Knutson, D. Tetrahedron Lett. 1968, 3429-3432.
(55) Hageman, H. J. Tetrahedron 1969, 25, 6015-6024.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8965

A R T I C L E S
Mori et al.
Table 2. Temperature Dependence on the Photochemistry of cis/trans Mixtures of 4 in Methylcyclohexanea
Z/E ratio
yields (%)c
T
(°C)
irrad
(min)
conversionb
of recovered
Z/E ratio
of 5
material
balance (%)
(%)
4
5
3
5/3 ratio
-80
-45
-25
-10
+25
+60
+95
24 h
30
10
10
5
1.4 ( 0.1
5.2 ( 0.1
14.6 ( 0.6
13.3 ( 0.5
14.8 ( 0.6
20.5 ( 1.4
10.7 ( 0.3
63.7:36.3
63.2:36.8
63.2:36.8
63.4:36.6
63.7:36.3
63.8:36.2
63.5:36.5
not detected
92.4 ( 10.9
78.9 ( 2.1
77.9 ( 4.3
75.1 ( 11.5
61.9 ( 7.8
57.6 ( 5.1
not detected
16.0 ( 9.2
12.2 ( 0.7
13.5 ( 0.9
21.9 ( 0.9
27.7 ( 1.5
33.7 ( 0.9
n.d.d
5.78
6.47
5.77
3.43
2.23
1.71
n.d.d
98
104
91
66:34
63:37
64:36
64:36
63:37
64:36
91
97
5
90
5
91
a Initial concentration of 4 was 3.6 mM (Z/E ) 64.2:35.8). For irradiation conditions, see footnote a in Table 1. b All reactions were repeated at least 3
times, and the results were averaged. c Absolute yields based on consumed 1. d Not determined.
argon, decarboxylation product 5 (23% yield; 32% based on
consumption of 4) and 2,4,6-trimethylphenol (13%) were
isolated after purification on silica gel column chromatography.
The product 5 was determined to be only the trans-isomer by
1H NMR spectroscopy: the coupling constant of the triple-triplet
at δ 2.90, J ) 12.4 and 3.7 Hz, corresponding to the proton
adjacent to the aromatic moiety is indicative of axial-axial and
axial-equatorial proton couplings (see Supporting Information
for details). This assignment is consistent with our previous
conclusions that photodecarboxylation is concerted and proceeds
with retention of configuration.35
The influence of temperature on irradiations of 4 was also
examined using a Z/E mixture of esters in methylcyclohexane
(Table 2). The trends are very similar to those mentioned above
for the photolysis of 1. At low temperature, -80 °C, the reaction
proceeded very slowly, and the best yield of photodecarboxy-
lation product was obtained at -25 °C. The relative yield of 5
decreased at either higher or lower temperatures; the “critical
temperature” of 4 is slightly lower than that for 1. Again, a
plot of the relative yields of 5 versus the inverse of temperature
could be fitted to two linear segments (Figure 4); similar
mechanistic changes must be occurring in 1 and 4.
Interestingly, the Z/E ratios of recovered 4 were always
slightly lower than the initial value, 1.79, because the reaction
of the Z-isomer, which is sterically more constrained,63-65 was
slightly more efficient than that of the E-isomer. However, the
differences were so small that Z/E ratios of the decarboxylation
photoproduct were essentially the same as the initial ratio of
the ester at all temperatures examined.
Photolyses of 3,4,5,6-Tetrahydro-2H-1-benzoxocin-2-one
(6). Photodecarboxylation of some lactones has been reported
but only for a limited set of compounds.66,67 The more common
photochemical course for benzolactones with five-, six- (such
as 2,3-dihydrocoumarin), or seven-member rings is transesteri-
fication in the presence of an excess of alcohol.68 Here, we have
examined the photochemistry of a benzolactone with an eight-
membered ring, 3,4,5,6-tetrahydro-2H-1-benzoxocin-2-one 6,
expecting that its favored conformations might be more
amenable to concerted decarboxylation.
Figure 4. Arrhenius-type plots of the temperature dependence of relative
yields of photodecarboxylation products from irradiations of (a) 1, (b) 4,
and (c) 6 in methylcyclohexane solutions and (d) of 1 in PE46(u) film.
Best linear fits are drawn as well. Data below and above -10 °C were
used for individual fits for 1 and 4 in methylcyclohexane. The datum at 95
°C was omitted in fitting data from 6.
experiments do not rule out a triplet photodecarboxylation
mechanism, they and the complete retention of configuration
of the decarboxylation products strongly implicate the singlet
manifolds; only an adiabatic transformation of triplet 1 to triplet
2 could accommodate the stereochemistry of the reaction.
Influence of Temperature on the Competition between
Photodecarboxylation and Photo-Fries Reactions of 1. The
dependence of temperature on the ratio of 2/3 was examined in
(one batch of spectrograde) liquid methylcyclohexane between
-95 and +95 °C. The temperature dependence of the fluores-
cence spectra of 1 was also studied in the same solvent between
-120 and +100 °C (vide infra). Photolyses below -80 °C were
extremely slow, and only trace conversions of 1 were observed
at -80 °C after 24 h irradiation. By contrast, 15% conversion
was achieved after 5-min irradiation at 0 °C. Reaction became
faster as the temperature was increased, but it slowed slightly
above 25 °C. Similarly, the highest 2/3 ratio, 6.8, was obtained
at 0 °C, and a plot of the relative yield of 2 versus reciprocal
temperature could be fitted to two linear segments (Figure 4).
A change in the relative efficiencies of the two photoprocesses
responsible for 2 and 3 is clearly occurring near 0 °C. Regardless
of the reason for this interesting temperature effect, the cage-
escape product 3 cannot be avoided simply by changing
temperature (or solvent).
Irradiation of 6 gave the corresponding decarboxylation
product, tetrahydronaphthalene 7, in good yields despite the
(63) Van de Graaf, B.; Baas, J. M. A.; Wepster, B. M. Recl. TraV. Chim. Pays-
Bas 1978, 97, 268-273.
(64) Booth, H.; Gidley, G. C. Tetrahedron Lett. 1964, 1449-1455.
(65) Beckett, C. W.; Pitzer, K. S.; Spitzer, R. J. Am. Chem. Soc. 1947, 69, 2488-
2495.
(66) Wan, P.; Budac, D. In CRC Handbook of Organic Photochemistry and
Photobiology; Horspool, W. M., Song, P.-S., Eds.; CRC Press: Boca Raton,
FL, 1995; pp 384-392.
(67) Suginome, H. Chem. Acid DeriV. 1992, 2, 1107-1198.
(68) Gutsche, C. D.; Oude-Alink, B. A. M. J. Am. Chem. Soc. 1968, 90, 5855-
5861.
Photodecarboxylation of 2,4,6-Trimethylphenyl trans-4-
Methylcyclohexanecarboxylate (4). The photodecarboxylation
of trans-4 has been examined on a preparative scale in purified
acetonitrile at 25 °C (Scheme 2). After 3-h irradiation under
(62) Plank, D. A. Tetrahedron Lett. 1969, 10, 4365-4368.
9
8966 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
Table 3. Temperature Dependence of the Photochemistry of 6 in
was evaluated at λex ) 260 nm by comparison with the value
for mesitylene in hexane (ΦFL ) 0.088).69
Methylcyclohexanea
material
The spectra of aryl esters 1, 4, and 6 were also compared in
acetonitrile. Their absorption spectra are quite similar in shape,
but some differences in vibrational structures are apparent. Molar
extinction coefficients of 6 were almost twice as large as those
of 1 and 4. The fluorescence wavelength maxima of 1 and 4
were virtually the same, but that of lactone 6 was blue-shifted
somewhat, suggesting that its relaxed excited-state and ground-
state structures are more similar than those of the conforma-
tionally more labile 1 and 4. As mentioned above, the most
favored ground-state conformation of 6 is close to the structure
necessary for excited-state decarboxylation. Because the fluo-
rescence spectra were structureless, the singlet energies of 1, 4,
and 6, 423, 423, and 431 kJ mol-1, respectively, were estimated
from the midpoint of the maxima of the lowest energy
absorption and fluorescence band.
T
(°C)
irradiation
time (min)
conversionb
(%)
yield of
7 (%)
balance
(%)
-80
-45
-25
-10
+25
+60
+95
24 h
10.0
10.0
5.0
5.0
5.0
41.0 ( 3.0
27.9 ( 0.9
41.6 ( 2.1
33.2 ( 1.6
29.1 ( 0.9
27.0 ( 1.7
17.2 ( 0.8
19.5 ( 10.8
30.7 ( 0.8
42.5 ( 1.4
45.5 ( 5.0
51.3 ( 2.3
55.8 ( 4.6
72.4 ( 7.3
20
31
43
46
51
56
72
5.0
a Initial concentration of 6 was 3.7 mM. For irradiation conditions, see
footnote a in Table 1. b All reactions were repeated at least 3 times, and
the results were averaged.
Fluorescence lifetimes were obtained by the time-correlated
single-photon counting technique (see Figure S-1 in Supporting
Information for details). The decay constants of 1 in acetonitrile
and methylcyclohexane were indistinguishable, but ΦFL was
much higher in methylcyclohexane. The quantum yield for
consumption of 1 in acetonitrile at 254 nm excitation was
determined to be ∼0.02. The fluorescence lifetime and quantum
yield of 1 and 4 are similar and larger than those of 6, as
expected from the richer n,π* nature and/or the lactone’s higher
reactivity.
Temperature Dependence of the Fluorescence of 1. In
hopes of understanding better the interesting dependence of the
photoproduct distribution from 1 on temperature, we investigated
its fluorescence in (liquid) methylcyclohexane between -120
and +100 °C as well (Table 5 and Figure 6). As expected, the
overall emission intensity decreased with increasing temperature
due to increases in the rates of nonradiative processes.70-72
A
linear correlation between fluorescence intensity and reciprocal
temperature was observed in a temperature range of -120 to 0
°C. Above 0 °C, a somewhat stronger emission intensity than
the expected (extrapolated) values were observed and another
linear correlation is obtained (Figure 6).
Figure 5. UV-vis absorption (top) and excitation and fluorescence
(bottom) spectra of (a) 1, (b) 4, and (c) 6 (1 mM) in acetonitrile at 25 °C.
Excitation and emission wavelengths were 266 and 290 nm, respectively.
Fluorescence maxima shifted from 288 nm at -120 °C to
295 nm at +100 °C (∆λ ) 7 nm). These wavelength shifts
suggest that the emission at higher temperatures occurs from a
more relaxed excited singlet state and/or to a less relaxed ground
state than at lower temperatures. One attractive hypothesis is
that the s-cis conformation, which is a local minimum in the
ground state but not the global energy minimum, fluoresces at
relatively longer wavelengths than the s-trans and contributes
to a greater extent at higher temperatures. The calculated
distribution of ground-state conformers in the “gas phase”
suggests that the population of s-cis conformer is essentially
negligible (∼0.01%) at room temperature. The energy differ-
ences of the s-trans and s-cis conformers are sufficiently large
to preclude a significant population being present in the ground
states. Each ground-state conformer would find it difficult to
convert to another conformer within the short lifetime of the
availability of the ortho and para positions of the benzo ring
for photo-Fries type rearrangements (Table 3). The other
(unindentified) products are probably derived from ketene
intermediates. Qualitatively, conversion of 6 was more efficient
than that of 1 or 4: after 5-min irradiation in methylcyclohexane
at 25 °C, 33% of 6 was converted to products; under the same
experimental conditions, 16% of 1 and 15% of 4 were converted.
In contrast to the sluggish photoreactivity of 1 and 4 at low
temperatures, decarboxylation of lactone 6 occurred even at
-80 °C, albeit slowly, and 41% conversion was achieved after
24 h.
Absorption and Fluorescence Spectra of 1, 4, and 6. The
absorption and fluorescence spectra of 1 were measured in
methylcyclohexane and compared with those obtained in the
more polar solvent acetonitrile (Table 4 and Figure 5). The UV-
vis spectra of 1 were almost the same in both solvent, but the
1Lb band was slightly blue-shifted in acetonitrile, suggesting
partial n,π* nature of the transition. Absorption maxima at ca.
275 nm were vibrationally structured to some extent. The
corresponding fluorescence spectra were measured in the same
solvents, and the quantum yield of fluorescence ΦFL ) 0.027
(69) Froehlich, P. M.; Morrison, H. A. J. Phys. Chem. 1972, 76, 3566-3570.
(70) Tanaka, N.; Yamazaki, H.; Sakuragi, H.; Tokumaru, K. Bull. Chem. Soc.
Jpn. 1994, 67, 1434-1440.
(71) de Lange, M. C. C.; Thorn Leeson, D.; van Kuijk, K. A. B.; Huizer, A. H.;
Varma, C. A. G. O. Chem. Phys. 1993, 177, 243-256.
(72) Werner, T. C.; Lyon, D. B. J. Photochem. 1982, 18, 355-364.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8967

A R T I C L E S
Mori et al.
Table 4. Photophysical Data of Esters 1, 4, and 6a
absorption
fluorescence
additive
(amount)
λmax
(nm)
λmax
(nm)
λmax
(nm)
quantum
yield, ΦFL
τs
(ns)
E b
s
ester
solvent
log ꢀ
log ꢀ
(kJ mol-1
)
(S)-1
methyl-
275.2
2.48
266.6
2.52
291.2
0.027
1.6
423
cyclohexane
acetonitrile
water
274.6
274.0
2.42
266.0
265.8
2.48
291.0
288.2
0.017
0.005
1.7
0.2
423
426
â-CD
(10 mM)
γ-CD
water
273.8
265.6
293.4
0.006
0.2
422
(10 mM)
trans-4
6
acetonitrile
acetonitrile
274.6
272.4
2.50
2.78
266.0
265.6
2.55
2.80
291.0
283.6
0.013
0.0014
1.8
0.3
423
431
a At 25 °C for 1 mM ester under air. b Estimated from midpoint of absorption and fluorescence maxima.
Table 5. Temperature Dependence of Fluorescence Spectra of
(S)-1a
T
(°C)
λ
max(emission),
λ
max(excitation 1),
λ
max(excitation 2),
I(λ1/λ2)c
b
nm
Irel
nm
nm
-120
-100
-80
-60
-40
-20
0
+20
+40
+60
+80
+100
287.8
288.2
289.0
289.2
290.0
290.4
291.6
292.0
292.8
293.4
294.0
294.6
5.63
4.71
3.86
3.03
2.37
1.80
1.34
265.0
265.2
265.4
265.8
266.0
266.2
266.4
266.8
267.0
267.4
267.6
267.6
273.8
274.0
274.2
274.4
274.6
274.6
274.8
274.8
274.4
274.4
274.2
274.0
1.046
1.042
1.039
1.038
1.038
1.041
1.045
1.065
1.088
1.095
1.098
1.109
≡ 1
0.772
0.587
0.444
0.358
a In methylcyclohexane for 1 mM solutions under air. b Relative
fluorescence intensities obtained from integrated areas of spectra. c Ratio
of intensities at 266 and 274 nm in excitation spectra.
singlet excited states; on this basis, we conclude that the
observed fluorescence spectrum is mainly from the s-trans
conformer. As the temperature is increased, the population of
the less stable s-cis conformer in the excited state increases (vide
infra). We believe an overall bathochromic shift in fluorescence
maxima at higher temperature is derived from the partial
contribution from excited s-cis conformers.
The positions of peaks in the excitation spectrum of 1 also
shifted slightly to the red with increasing temperature. More
pronounced changes were apparent in the ratios of the peak
intensities at ca. 266 and 274 nm (Table 5); they increased from
1.04 at -100 °C to 1.11 at +100 °C. Because no analogous
changes in the absorption spectrum could be detected, we
conclude that more than one ground-state conformer is being
excited and that equilibration among them is incomplete during
the excited-state lifetime.
The temperature dependencies of the fluorescence quantum
yields (ΦFL), excitation spectra, and product distributions suggest
that a different mix of excited species is present at lower and
higher temperatures. We choose to interpret these species as
predominantly the s-cis and s-trans conformers, with the content
of the more energetic s-cis increasing at higher temperatures.
Photolysis of 1 in Solutions of Cyclodextrins. Although their
range of cavity sizes and shapes is rather limited (in contrast to
the cavities available in the family of zeolites73-77), cyclodextrins
are very useful prototypes for supramolecular complexation and
Figure 6. Temperature dependence of excitation and fluorescence spectral
intensities of 1 (1 mM) in methylcyclohexane (a). Excitation and emission
wavelengths were 266 and 290 nm, respectively. Plot of the integrated
fluorescence intensities versus reciprocal temperature (b). The intersection
of the two linear segments is at ca. 10 °C.
photochemical studies because they are readily available, easily
used, transparent into the UV region, and chiral.78-84 A few
photodecarboxylations of aryl esters have been reported in other
supramolecular hosts, including polyethylene films,85,86 SDS
micelles,87 and cation-exchanged faujasites.88
In view of the demonstrated sensitivity of the photoproduct
distribution from 1 to its environment, we have irradiated
(75) Turro, N. J. Acc. Chem. Res. 2000, 33, 637-646.
(76) Scaiano, J. C.; Garcia, H. Acc. Chem. Res. 1999, 32, 783-793.
(77) Ramamurthy, V.; Garcia-Garibay, M. A. Comput. Supramol. Chem. 1996,
7, 693-719.
(78) Ueno, A.; Ikeda, H. Mol. Supramol. Photochem. 2001, 8, 461-503.
(79) Bortolus, P.; Monti, S. AdV. Photochem. 1996, 21, 1-133.
(80) Bortolus, P.; Grabner, G.; Koehler, G.; Monti, S. Coord. Chem. ReV. 1993,
125, 261-268.
(81) Ramamurthy, V.; Eaton, D. F. Acc. Chem. Res. 1988, 21, 300-306.
(82) Sivaguru, J.; Natarajan, A.; Kaanumalle, L. S.; Shailaja, J.; Uppili, S.; Joy,
A.; Ramamurthy, V. Acc. Chem. Res. 2003, 36, 509-521.
(83) Turro, N. J. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 4805-4809.
(84) Inoue, Y. Chem. ReV. 1992, 92, 741-770.
(73) Tung, C.-H.; Wu, L.-Z.; Zhang, L.-P.; Chen, B. Acc. Chem. Res. 2003, 36,
39-47.
(74) Weiss, R. G.; Ramamurthy, V.; Hammond, G. S. Acc. Chem. Res. 1993,
26, 530-536.
9
8968 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
Table 6. Irradiations of Racemic 1 in Aqueous and Aqueous Methanol Solutions of â- and γ-Cyclodextrinsa
combined
yield of
products
(%)
irradiation
time (min)
conversion
(%)
2/3
ratio
ee of 2b
(%)
c
host
solvent
MB
â-cyclodextrin
H2O
20
40
60
60
20
20
43
62
80
97
77
70
6
9
11
9
0
4
1.8
2.3
2.3
2.0 (R)
0.8 (R)
0.4 (S)
14.1 (R)
ndd
63
47
31
12
23
34
H2O/MeOH (1:1)
H2O
H2O/MeOH (1:1)
2.5
γ-cyclodextrin
ndd
<0.05e
ndd
a Initial concentrations of racemic substrate and â- or γ-cyclodextrin were 0.5 and 2.5 mM, respectively. Reactions at 25 °C under argon. For details, see
footnote a in Table 1 and Experimental Section in the Supporting Information. b Percentage of ee determined by chiral GC. Error ) (1.0%. c Material
balance (%). d Not determined. e Photodecarboxylation product 2 was not detected by GC.
Figure 7. Induced CD spectra of racemic 1 (1 mM in H2O) in the presence
Figure 8. Normalized fluorescence spectra of 1 (1 mM) at 25 °C. Excitation
of (a) 5 mM and (b) 10 mM â-cyclodextrin and in the presence of (c) 5
and emission wavelengths were 266 and 290 nm, respectively. (a) In
mM and (d) 10 mM γ-cyclodextrin.
methylcyclohexane. (b) In 10 mM aqueous â-cyclodextrin. (c) In 10 mM
aqueous γ-cyclodextrin. (d) In amorphous PE0 film (ca. 50 mmol/kg‚film).
Note that the emission and excitation spectra in PE0 are somewhat distorted
racemic 1 in cyclodextrin cavities. No chiral induction should
be observed in 2 after partial conversion of 1 if the photode-
carboxylation process is concerted unless one of the enantiomers
of 1 is more photoreactiVe in cyclodextrin complexes or is
complexed to a different extent than the other.
by self-absorption phenomena.
aqueous solutions of â- and γ-cyclodextrins; very small changes
were detected in the excitation spectra. In the presence of excess
â-cyclodextrin, a 3-nm hypsochromic shift was observed in the
emission spectrum, suggesting that the ester portion of 1 is in
a relatively polar environment (i.e., outside or at the surface of
the cavity). The emission maximum was shifted ca. 2 nm to
the red upon complexation by γ-cyclodextrin, indicating that
the chromophore is inside the cavity. In both types of cyclo-
dextrins, the fluorescence decay constants of 1 in water (1 mM)
in the presence of cyclodextrins (10 mM), ca. 0.2 ns, were
significantly smaller than the 1.6-1.7-ns values observed in
methylcyclohexane.
Results from photolyses of racemic 1 in aqueous â- and
γ-cyclodextrin solutions at 25 °C are summarized in Table 6.
Low (6-11%) combined yields of 2 and 3 and in much lower
2/3 ratios than those in organic solvents were obtained in
â-cyclodextrin solutions of water or water/methanol, and the
detectable unidentified byproducts were also observed in smaller
amounts. Apparently, 1 undergoes photoreactions more ef-
ficiently with cyclodextrin than it does intramolecularly. Similar
low yields have been reported for irradiations in aqueous
â-cyclodextrin solutions of other aryl esters for which decar-
boxylation is a very minor reaction pathway.92-95 The prefer-
ences for â-bond scission and formation of 3 were expected
Initially, the induced CD spectra of 1 were examined
spectroscopically in aqueous solutions of â- and γ-cyclodextrins
(Figure 7). As the cyclodextrin concentration was increased from
1 to 10 mM, a weak, but appreciable, negative Cotton effect
was induced in the 230-280 nm region, where the aryl ester
absorbs the light. This indicates that the aryl ester 1 forms
complexes with these cyclodextrins and the chromophoric
moiety of 1 is included in the cyclodextrin cavity at least to
some extent. Unfortunately, due to the weakness of the signals
and the low solubility of 1 in water, it was not possible to
evaluate the association constant K or to assess the exact
alignment of the ester and cyclodextrins by applying the
empirical Harata89 and Kodaka rules.90,91
The inclusion of 1 in cyclodextrin cavities has been inves-
tigated also by absorption and fluorescence spectroscopies
(Table 4 and Figure 8). The low solubility of 1 precluded our
measurement of its fluorescence decay or its absorption and
fluorescence spectra in water alone; comparisons of spectral
shifts here are based on values in methylcyclohexane. Absorp-
tion spectra of 1 are hypsochromically shifted by ∼1 nm in
(85) Gu, W.; Abdallah, D. J.; Weiss, R. G. J. Photochem. Photobiol., A 2001,
139, 79-87.
(86) Gu, W.; Weiss, R. G. Tetrahedron 2000, 56, 6913-6925.
(87) Singh, A. K.; Sonar, S. M. Synth. Commun. 1985, 15, 1113-1121.
(88) Lalitha, A.; Pitchumani, K.; Srinivasan, C. Tetrahedron 2001, 57, 4455-
4459.
(89) Harata, K.; Uedaira, H. Bull. Chem. Soc. Jpn. 1975, 48, 375-378.
(90) Kodaka, M. J. Am. Chem. Soc. 1993, 115, 3702-3705.
(91) Kodaka, M. J. Phys. Chem. 1991, 95, 2110-2112.
(92) Veglia, A. V.; de Rossi, R. H. J. Org. Chem. 1993, 58, 4941-4944.
(93) Veglia, A. V.; Sanchez, A. M.; De Rossi, R. H. J. Org. Chem. 1990, 55,
4083-4086.
(94) Syamala, M. S.; Rao, B. N.; Ramamurthy, V. Tetrahedron 1988, 44, 7234-
7242.
(95) Ohara, M.; Watanabe, K. Angew. Chem. 1975, 87, 880-881.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8969

A R T I C L E S
Mori et al.
Table 7. Irradiations of (S)-1 in Polyethylene Filmsa
irrad
time
(min)
mass
balance
(%)
yields (%)d
crystc
(%)
VPE
[1],
thickness
(µm)
T
(°C)
conversion
(%)
ee of
2 (%)
c
b
3
film
(Å )
mmol (kg film)-1
2
3
2/3 ratio
PE0(u)
PE46(u)
0
46
177
139
2.0
2.7
1.2 × 103
23
2
23
60
90
40
60
20
40
60
40
40
60
60
60
60
28 ( 6
23 ( 1
13 ( 1
48 ( 10
26 ( 3
11 ( 2
31 ( 5
14 ( 2
28 ( 5
17 ( 3
31 ( 5
19 ( 2
20 ( 3
59 ( 7
32 ( 4
18 ( 3
45 ( 20
18 ( 7
42 ( 23
19 ( 7
3 ( 1
13 ( 2
10 ( 1
7 ( 1
4.5
8.7
3.1
2.5
0.96
1.4
1.9
0.62
0.59
82
>98
>98
>98
>98
>98
>98
>98
>98
>98
>98
>95
>98
>98
92
89
93
64
83
97
78
87
74
90
88
88
90
82
5 ( 2
65
23
19 ( 7
31 ( 17
10 ( 4
5 ( 2
PE46(s)
47
121
2.7
34
PE50(u)
PE50(s)
PE68(u)
PE68(s)
PE74(u)
PE74(s)
50
58
68
84
74
70
144
141
124
113
129
120
5.0
5.0
19
19
12
25
16
21
12
17
13
23
23
23
23
23
23
2 ( 1
4 ( 1
1 ( 1
44 ( 5
51 ( 8
32 ( 4
38 ( 13
11 ( 2
1 ( 1
4.8
82
4.0
12
10 ( 3
a Irradiated with a low-pressure Hg lamp (ethanol filter, λ ) 254 nm); samples under vacuum unless stated otherwise. Each datum is the average of at
least 9 separate analyses of 3 films. c Crystallinities and calculated mean hole free volumes from positron annihilation studies. Gu, W.; Hill, A. J.; Wang,
X.; Cui, C.; Weiss, R. G. Macromolecules 2000, 33, 7801-7811. d Absolute yields based on consumed 1.
based on the mode of complexation mentioned above (enhanced
population to the extended conformation). When the aryl portion
is not inside the cyclodextrin torus, the s-cis conformation
required for photodecarboxylation is difficult to attain; the more
extended s-trans conformation that favors 3 should be even more
preferred than in organic solvents.
properties, the percent of crystallinity, is indicated in each as a
suffix to their PE acronyms, and “u” or “s” in parentheses refers
to unstretched or cold-stretched films (see Experimental Section
in Supporting Information for details); other physical parameters
of these films have been reported elsewhere.99
There are reasons to believe that PE films may be more
amenable to directing the photochemistry of 1 than the other
media examined above. Some conformational control of decar-
boxylation upon irradiation of 1-naphthyl esters in PE films
and at various temperatures has been reported.100 In addition,
the fates of the prochiral radical pair intermediates from
decarbonylation of the initial acyl/1-naphthoxy radical pairs,
generated during irradiation of 1-naphthyl (R)-2-phenylpro-
panoate, have been used as a diagnostic of the motional freedom
within cavities of PE films.101 These radical pair recombinations
occur with significant stereoselectivity, but they are not ste-
reospecific.
Results from the irradiations are summarized in Table 7. The
material balances were quite high at <30% conversion (where
most of the irradiations were analyzed to avoid secondary
reactions) in all cases. The only photoproducts detected were 2
and 3, and the enantiomeric excess of 2 was >98%, as in other
media (above the detection limit; no other enantiomer was
detected). In completely amorphous PE0, which lacks interfacial
sites,99 the 2/3 ratio was quite similar to that in methylcyclo-
hexane solution at a comparable temperature. The large differ-
ences between the 2/3 ratios from PE46(u) and PE50(u), films
of nearly the same crystallinity, indicate that the similarity
between the ratios in methylcyclohexane and PE0 is fortuitous.
There is no clear trend between PE crystallinity or any other
single physical property of the films we can identify and the
relative yields of 2 (Table 7). The natures of the reaction cavities
of each film, as well as the effect that they have on the
conformations of 1, must be determined by a combination of
the properties. However, the two unstretched high density PE
films, PE68(u) and PE74(u), afforded 2 (ee was above detection
limit: >98%) almost exclusively (>98%); only trace amounts
of 3 were detected. The reaction cavities of the two high-density
PE films either force all of the molecules of 1 to adopt s-cis
A small enantiomeric enrichment in the (R)-isomer of 2 from
irradiation of racemic 1 in aqueous â-cyclodextrin was observed
at low conversions; the enantioselectivity at high conversions
was negligible. In aqueous methanol solutions of â-cyclodextrin,
14% ee of 2, enriched in the (R)-enantiomer, was found at 97%
conversion. Given the preponderance of evidence for the
concertedness and retention of configuration of the photode-
carboxylation, these results strongly suggest that one of the
enantiomers of 1 reacts more rapidly than the other and that
the reaction pathways of complexed and uncomplexed 1 differ.
In an aqueous or aqueous methanol solution of γ-cyclodextrin,
1 was consumed upon irradiation, but no 2 was detected. No
other low-mass products were observed (GC-MS analysis) in
the aqueous medium, and only a 4% yield of 3 was found in
the aqueous methanolic mixture. It appears that 1 reacts
preferentially with the γ-cyclodextrin to which it is bound; the
CD spectroscopic results above indicate that the aryl portion of
1, at least, is within the torus, and spectroscopic analyses of
the cyclodextrins after irradiation indicate that aromatic groups
are covalently attached.
Photolysis of 1 in Polyethylene Films. In all of the
experiments above, varying solvent and temperature, cleavage
of 1 leading to 3 could not be eliminated. Thus, we investigated
a different strategy, involving polyethylene (PE) films as the
reaction medium, to improve the relative yields of 2 and to
suppress the appearance of byproducts such as those found in
cyclodextrin solutions. The PE films apply “passive” pressure
to guest molecules96-98 that can influence ground-state confor-
mations and rates of diffusion of pairs of species within one
reaction cavity. Both are capable of directing the reactions of
molecules in their excited states. The films employed here span
a wide range of morphologies. One of their most important
(96) Gu, W.; Warrier, M.; Schoon, B.; Ramamurthy, V.; Weiss, R. G. Langmuir
2000, 16, 6977-6981.
(99) Wang, C.; Xu, J.; Weiss, R. G. J. Phys. Chem. B 2003, 107, 7015-7025.
(100) Gu, W.; Abdallah, D. J.; Weiss, R. G. J. Photochem. Photobiol., A 2001,
139, 79-87.
(97) Gu, W.; Warrier, M.; Ramamurthy, V.; Weiss, R. G. J. Am. Chem. Soc.
1999, 121, 9467-9468.
(98) Gu, W.; Weiss, R. G. J. Photochem. Photobiol., C 2001, 2, 117-137.
(101) Xu, J.; Weiss, R. G. Org. Lett. 2003, 5, 3077-3080.
9
8970 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
Table 8. Effect of Pressure on the Photochemistry of 1a
conformations or prohibit the radical pairs from the alternative
reaction pathway leading eventually to 3. We would discard
the latter possibility to force the reformation of 1, because the
overall rate of the reaction in films was almost the same to (or
slightly faster than) those observed in isotropic solutions,
although the absolute rates were not obtained because of
technical problems.
yieldsb (%)
c
pressure
(MPa)
conversion
(%)
ee of 2
(%)
d
2
3
2/3 ratio
MB
0.1
23
43.5
35.8
8.8f
8.7
3.0
<0.1
<0.1
5.0
12
>50
>50
>99
>99
>99
>99
89
200
350
400
40
76
n.d.e
26
n.de
82
29.2
a Irradiation with an ultrahigh-pressure Hg lamp in a quartz cuvette
installed in a high-pressure vessel for 1.8 mM of 1 in MeCN for 20 h under
argon at 25 °C. See Supporting Information for details. b Absolute yields
based on consumed 1. c %ee determined by chiral GC. Error ) (0.5%.
d Material balance (%). e Not determined. f Based on amount of substrate
used.
Stretching films affects leads to lower 2/3 ratios in all of the
films employed. Film stretching of partially crystalline PE films
redistributes microcrystallites more evenly within the amorphous
domains, decreases mean hole free volumes, increases the area
of the interfacial region and, thus, the fraction of guest molecules
in interfacial cavities.99 Stretching films also affect the shapes
of the reaction cavities in each film, making them presumably
more extended than in unstretched films. Distortion of the cavity
shapes may be responsible for the increased relative yields of
3 in the stretched films because it should favor the more
extended s-trans conformers. Despite these perturbations, the
formation of 2 remains stereospecific; a radical mechanism must
not be involved, even in the confining reaction cavities of the
various unstretched and stretched PE films.
We were unable to obtain reliable fluorescence spectra of 1
in all of the PE films except PE0 at a high doping concentration,
∼50 mmol (kg film)-1 (Figure 8). The shapes of the emission
and the excitation spectra of 1 are almost identical to those in
methylcyclohexane, although there appears to be some self-
absorption in the PE films that distorts the blue edge of the
emission and the red edge of the excitation. However, the
fluorescence decay histograms in PE0 are biexponential; the
best fits have decay constants of 2.9 ns (37%) and 4.8 ns (63%).
One explanation of these decays is that 1 is distributed in two
different site types in PE0 films, perhaps near and far from
regions of high chain branching. Regardless, the much larger
decay constants in PE0 than in solution suggest that the esters
are conformationally dissimilar in the films and in solution.
The influence of temperature between 2 and 65 °C on the
photodecarboxylation of 1 in PE46(u) films has been compared
with results from irradiations in methylcyclohexane (Figure 4).
An Arrhenius-type plot of the relative yield of 2 (or the 2/3
ratio) is linear in PE46(u) but is clearly nonlinear in methyl-
cyclohexane. The linearity of the plot in the polyethylene film
is consistent with very minor conformational changes of 1 during
its short excited singlet lifetime and, therefore, maintenance of
the ground-state s-cis/s-trans equilibrium ratio. Substantial
constraints to motions of 1 are imposed by the reaction cavities
of the PE46(u).
Effects of Pressure on the Photochemistry of 1. Pressure
effects on (asymmetric) photochemical reactions have been
discussed in some depth recently.105-108 Pressure-induced
changes in activation parameters109 may be of use to mediate
photoproduct selectivities of 1 in acetonitrile, and experiments
at pressures up to 400 MPa have been conducted (using
procedures described previously105,107) to determine the mag-
nitude of such effects (Table 9). At 350-400 MPa, the 2/3 ratios
were >50 (i.e., no 3 was detectable) and the ee of 2 remained
>99%.
Because the relative molecular volume calculated at the PM3
level for the s-cis conformer of 1, the putative precursor of 2,
is much smaller than that of the s-trans conformer, the putative
precursor of 3 (Table 8), the equilibrium between the two favors
the s-cis as pressure is increased and free volume is decreased.
In addition, high pressure would retard the separation of radical
pairs from photo-Fries type processes (should they be formed
from s-trans conformers), thereby attenuating the yield of 3 and
accelerating the reformation of 1. Although more experiments
should be conducted in the future to quantify and analyze more
deeply the pressure-volume relationship, the results in hand
again indicate that the photoselectivity of 1 can be mediated
by altering the s-cis/s-trans ground-state equilibrium. Whereas
polyethylene films rely on “passive” pressure to effect (what
may be) the equilibrium changes, these experiments demonstrate
that “active” pressure can do the same.
Conformational Study of Aryl Ester 1. (i) By DFT
Calculation of Conformational Energies. In order for photo-
decarboxylation to be concerted and proceed with retention of
configuration, reaction must occur from the s-cis conformation
of the excited singlet state of 1. At least in “gas phase”
calculations, the s-trans conformation is much more stable than
the s-cis in both the ground and excited singlet states. Given
that the quantum yields for decarboxylation are much larger
than the fraction of 1 calculated to be s-cis in the ground state,
a large fraction of the excited singlet s-trans species must change
to the s-cis before reacting, and the s-cis must then react more
rapidly than its return to the s-trans in order that photodecar-
boxylation may be a major pathway. These conditions must be
These constraints, coupled with cavities that either cause the
s-cis/s-trans ratio to be very large or force initially formed
radical pairs from incipient photo-Fries lyses of s-trans con-
formers to reform 1, explain why reaction can be channeled to
yield almost exclusively 2 in the unstretched high-density PE
films. Such selectivity is synthetically useful because it is a two-
step, high-yield synthesis of enantiopure alkylmesitylenes.
Currently available methods for syntheses of similar molecules,
but in lower optical purity, involve sophisticated transition metal
catalyzed hydrogenations.102-104
(105) Kaneda, M.; Nakamura, A.; Asaoka, S.; Ikeda, H.; Mori, T.; Wada, T.;
Inoue, Y. Org. Biomol. Chem. 2003, 1, 4435-4440.
(106) Inoue, Y.; Wada, T.; Asaoka, S.; Sato, H.; Pete, J.-P. Chem. Commun.
2000, 251-259.
(107) Inoue, Y.; Matsushima, E.; Wada, T. J. Am. Chem. Soc. 1998, 120, 10687-
10696.
(102) Hou, D.-R.; Reibenspies, J.; Colacot, T. J.; Burgess, K. Chem.sEur. J.
2001, 7, 5391-5400.
(108) Heritage, T. W. Aspects of arene-ethene photocycloadditions and thermal
asymmetric reactions under high pressure; University Reading: Reading,
U.K., 1991.
(103) Forman, G. S.; Ohkuma, T.; Hems, W. P.; Noyori, R. Tetrahedron Lett.
2000, 41, 9471-9475.
(109) Van Eldik, R.; Asano, T.; Le Noble, W. J. Chem. ReV. 1989, 89, 549-
(104) Gross, Z.; Ini, S. Org. Lett. 1999, 1, 2077-2080.
688.
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8971

A R T I C L E S
Mori et al.
Table 9. DFT Calculated Properties of the Four Conformers of 1 at Local Minimaa
optimized
SCF energy
(hartree)
zero-point
correction
(hartree/particle)
Gibbs free energy
with thermal correction
(hartree)
∆G
(25 °C)
(kJ mol-1
b
population
(%)
µ
V
3
conformation
)
(debye)
(Å )
for 1
s-trans-anti
s-trans-syn
s-cis-anti
s-cis-syn
-696.228 774 9
-696.227 067 6
-696.218 470 5
-696.209 203 9
0.264 057
0.264 463
0.264 25
-695.969 893 4
-695.967 788
-695.959 399 8
-695.949 775 4
≡0
90.3
9.7
0.01
∼0
1.52
1.77
4.02
4.12
269
333
236
330
+5.52
+27.5
+52.7
0.264 615
for 6
s-trans
s-cis
-577.026 457 1
-577.020 743 9
0.173 215
0.173 428
-576.856 637 1
-576.850 715 1
≡0
99.8
0.2
2.33
4.13
221
215
+15.5
a At B3LYP/6-311G+(2d,p)//B3LYP/6-31G(d) level. Zero-point energies were scaled by 0.9804 and corrected. See also Figures 9 and 10 for the calculated
conformations. 1 Hartree ) 2625.5 kJ mol-1 b van der Waals volumes calculated by the PM3 method for the DFT optimized conformations.
.
met because a large part of the photoreactions of 1 does go
through the photodecarboxylation route under almost all of the
experimental conditions we have explored.
Structural modifications that stabilize s-cis conformers at the
expense of s-trans ones of aryl amides and aryl esters are
known.110-112 For instance, although acetanilide exists almost
completely in the s-trans form, introduction of alkyl groups in
the ortho positions of the phenyl ring leads to detectable amounts
of the s-cis form.113,114 Recently, s-cis/s-trans conformers of
the triplet state of alkyl phenylglyoxylates were observed by
means of time-resolved FTIR spectroscopy.115 Both conforma-
tions are known to exhibit similar (but appreciably different)
reactivities in intermolecular hydrogen abstractions, but only
the s-cis undergoes Norrish Type II photoelimination reactions.
Although the short-lived excited singlet state of 1 cannot be
detected by this technique, we believe that similar fast equilibra-
tion of conformers is operative. Therefore, DFT and CIS
calculations116 of the conformers of 1 in the excited singlet state
have been conducted.
Initially, ground-state geometries were optimized at the
B3LYP/6-31G(d) level for the s-cis and s-trans conformations
of 1. They afforded four conformational energy minima (Table
9 and Figure 9). Among them, the two s-cis conformers are
potential immediate precursors of the spiro-lactonic transition
state required for concerted decarboxylation. All of the con-
formers have significantly larger molecular volumes than the
mean free void volumes of the PE films (113-177 Å3),101,117
and thus, their reaction cavities are expected to impose some
Figure 9. Possible conformations in the ground and singlet excited states
of 1. Optimized structures of s-cis and s-trans conformations were calculated
at the B3LYP/6-31G(d) level for the ground states at the CIS(Nstates )
8)/6-31+G(d) level for the excited state.
constraints on shape changes of 1. The optimized geometries
at B3LYP/6-31G(d) level were further subjected higher-level
(B3LYP/6-311+(2d,p)) single-point energy calculations that
were corrected for their zero-point energies by applying an
empirical factor of 0.98. The calculated (gas phase) distributions
of the conformers using these energies predict only ca. 0.01%
equilibrium population of s-cis at 25 °C. The quantum yield
for decarboxylation of 1 in methylcyclohexane at 0 °C, ∼0.2%,
is much larger than the calculated ground-state population,
suggesting (as noted above) that some s-trans/s-cis equilibration
in the excited singlet state is occurring. Although the calculated
(gas phase) and methylcyclohexane equilibria probably differ
to some extent, the quantum yield value is the lower limit of
the actual s-cis content in the excited state because not all of
the excited singlets that attain the appropriate conformation may
undergo decarboxylation.
(110) Sakamoto, K.; Oki, M. Bull. Chem. Soc. Jpn. 1974, 47, 2623-2624.
(111) Oki, M.; Nakanishi, H. Bull. Chem. Soc. Jpn. 1971, 44, 3148-3151.
(112) Nakanishi, H.; Fujita, H.; Yamamoto, O. Bull. Chem. Soc. Jpn. 1978, 51,
214-218.
(113) Kessler, H.; Rieker, A. Justus Liebigs Ann. Chem. 1967, 708, 57-68.
(114) Kessler, H.; Rieker, A. Z. Naturforsch., B: Chem. Sci. 1967, 22, 456-
457.
(115) Merzlikine, A. G.; Voskresensky, S. V.; Danilov, E. O.; Rodgers, M. A.
J.; Neckers, D. C. J. Am. Chem. Soc. 2002, 124, 14532-14533.
(116) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa,
J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.;
Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo,
J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi,
R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G.
A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.;
Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.;
Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham,
M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian
03, revision A.1; Gaussian, Inc.: Pittsburgh, PA, 2003.
Energies of the singlet excited-state conformers of 1 were
also calculated at the CIS/6-31+G(d) level with Nstate ) 8
(117) Mori, T.; Inoue, Y.; Weiss, R. G. Org. Lett. 2003, 5, 4661-4664.
9
8972 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
Figure 10. Calculated HOMO and LUMO (corresponding to the 60th and 61st molecular orbitals) of energy-minimized conformers of ground-state 1
calculated at the B3LYP/6-31G(d) level. Numbers under figures are calculated energies of the levels in eV.
Figure 11. Energy diagrams in the ground- and excited-state conformers of (a) aryl ester 1 and (b) lactone 6. A relative energetic difference calculated by
DFT method is shown in kJ mol-1. Singlet excitation energies shown are experimental and calculated (in parentheses) values.
(Figure 9). Although the s-trans conformer is energetically
favored again in the excited singlet state, each of the ground-
state s-cis and s-trans conformations had a local minimum in
the excited singlet state at a comparable geometry. In the s-cis
conformation, the distance between the ipso carbon on the
aromatic ring and the carbon atom R to the carbonyl of the ester
(i.e., the two atoms that make a σ-bond in 2) moves from 2.929
Å in the ground state to a shorter distance, 2.913 Å, in the
excited singlet state. The shorter distance facilitates concerted
extrusion of CO2. Appreciable interaction between the σ(CO-
C)-orbital and the π-orbital of the aromatic ring prior to
electronic excitation of the s-cis-anti conformation aslo supports
the cheletropic concerted decarboxylation mechanism.
Molecular orbital calculations have been also performed on
the four conformers of 1 at the B3LYP/6-31G(d) level (Figure
10). HOMO-LUMO energy gaps were calculated to be 6.26,
6.22, 6.18, and 6.12 eV for the s-trans-anti, s-trans-syn, s-cis-
anti, and s-cis-syn conformations, respectively. Energy differ-
ences between the four conformers in their excited singlet states
have been estimated from these HOMO-LUMO energy gaps
and calculated gas-phase ground-state energies (Figure 11). In
the excited state, the Boltzmann distribution at room temperature
indicates 0.2% of the s-cis conformer assuming that complete
equilibrium is attained. The value is small but much larger than
the calculated population in the ground state. Therefore, the
excited-state potential energy surface is predicted to facilitate
formation of the s-cis conformer.
populations of s-cis and s-trans conformers in the ground state.
The relative contribution from the s-cis (and presumably anti,
as judged from the energy calculations) conformer is increased
at higher temperatures.
Rotation of the methyl groups at the 2- and 6-positions of
the aromatic ring of 2 is predicted to be slightly hampered at
room temperature according to molecular dynamics calculations
(using the MM2 force field). To test these calculations, line
shape analyses of the 1H NMR peaks of 2 and 5 in chloroform-d
were made over a range of temperatures. They reveal that
rotations of the ortho-methyl groups are somewhat restricted
in both compounds: from Eyring-type plots, ∆H‡ ) 53.1 kJ
mol-1 and ∆S‡ ) -25 J mol-1 K-1 (Tc ) 11 °C) for 2 (Figure
S-2 in the Supporting Information) and ∆H‡ ) 57.3 kJ mol-1
and ∆S‡ ) -23 J mol-1 K-1 (Tc ) 37 °C) for 5. These results
help to explain the origin of the slope changes detected in plots
a and b of Figure 4: photodecarboxylation may be facilitated
at lower temperatures due, at least partially, to the restricted
movement of the methyl groups; at elevated temperatures, their
rotation may hamper the approach of the carbon R to the
carbonyl to the ipso carbon of the aromatic ring in the s-cis
conformations.
The s-cis conformation is probably much easier to achieve
in lactone 6 than in either 1 or 4. However, our calculations of
the energies of the various conformations of 6 indicate that its
preferred conformation is not completely appropriate to effect
photodecarboxylation and that small conformational changes are
required to attain the correct atomic orientations (Figure 12).
The temperature dependence on the efficiency of formation of
7 is consistent with the energy barrier to the necessary
conformation being small. Indeed, the highest yield of 7, 72%,
was achieved in methylcyclohexane at 95 °C. This result is
rather amazing because the efficiency of â-bond cleavage, the
The s-cis conformers are also predicted to have much lower
excitation energies and, thus, absorb selectively at longer
wavelengths than the s-trans conformers (Figure 11). In the
context of these calculations, the aforementioned shifts of
excitation maxima in fluorescence measurements upon changing
the temperature can be explained on the basis of changing
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8973

A R T I C L E S
Mori et al.
Figure 12. Possible conformations involved in the photoreactions of 6
optimized at the B3LYP/6-31G(d) level.
initial step in photo-Fries reactions, also increases drastically
with increasing temperature. However, this amazement should
be tempered by the probable increase in the rate of recombina-
tion of the (intramolecular) radical pair to reform 7 at higher
temperatures. The exact cause of the high yields of 7 from
irradiations of 6 at high temperatures deserves further scrutiny.
Two minimum energy conformations of lactone 6 have been
calculated at the B3LYP/6-31G(d) level (Figure 12). Single-
point calculations at the B3LYP/6-311G+(2d,p) level with zero-
point energies were then performed. The energy difference, 15.5
kJ mol-1, corresponds to a 0.2% equilibrium population of the
more energetic s-cis conformer at 25 °C. In the singlet excited
state, the calculated distribution of the s-cis is slightly lower
than that in the ground state, but it is much higher than for 1.
In the s-cis conformation, the distance between the ipso carbon
of the aromatic ring and the carbon R to the carbonyl of 6, 2.885
Å, is significantly shorter than that in the s-cis conformation of
1, 2.929 Å. Both the higher content of the s-cis conformer of 6
and its more favorable distance facilitate photodecarboxylation.
A principal reason the eight-membered lactone 6 underwent
facile photodecarboxylation and the analogous five-, six-, and
seven-membered lactones do not is that the latter cannot form
s-cis conformers easily.
(ii) By VCD Spectroscopy. Recently, information about
conformations of chiral molecules, including esters,118-121 in
solution have been deduced by analyses of vibrational circular
dichroic (VCD) spectra combined with DFT calculations.122-127
This approach has been applied to the conformational analysis
of 1. The VCD spectrum of 1 in dichloromethane-d2 was
compared with simulated spectra from the four conformations
identified above (Figure 13). Relative contributions from each
conformer were estimated as a Boltzmann distribution based
on the gas-phase calculations at the B3LYP/6-311G+(2d,p) level
with zero-point energy corrections (Table 9). As expected, the
main contributor in the simulated VCD spectrum that best
Figure 13. (a) Experimental VCD spectrum of (S)-1 in dichloromethane-
d2 at ambient temperature (solid line) and the B3LYP/6-31G(d) calculated
VCD spectrum of (S)-1 (dashed line). The calculated spectrum is based on
the fractional contribution from ground-state energies of the calculated VCD
spectrum of each conformation as shown in part b. Black: s-trans-anti.
Red: s-trans-syn. Blue: s-cis-anti. Green: s-cis-syn.
Scheme 4. Proposed Conformational Control of
Photodecarboxylation and Cage-Escape Process
approximates the experimental data is from the s-trans-anti form.
However, some of the peaks observed require the inclusion of
contributions from the other conformers. For example, a weak,
positive Cotton effect observed at 1770 cm-1 cannot be
reproduced by the calculated signals of the s-trans-anti form
alone; it could be reproduced only by including contributions
from syn conformers.
Concluding Remarks
A proposed photodecarboxylation mechanism that accounts
for the body of the results presented here is illustrated in Scheme
4. A spiro-lactonic transition state, which presumably forms
from a singlet excited-state s-cis conformer and loses a molecule
of carbon dioxide in a concerted manner, predicts products with
complete retention of stereochemistry. The decarboxylation
process has many of the characteristics of symmetry-allowed
suprafacial [1,3]sigmatropic rearrangements. The results indicate
that no radical intermediates are involved in decarboxylation,
but they are in the competing photo-Fries type cleavage that
occurs from the s-trans conformers of the excited singlet state.
The s-trans conformers are thermodynamically more stable than
the s-cis in the ground and excited states. Despite this,
experimental conditions can be found for which photodecar-
boxylation occurs to the near exclusion of the photo-Fries
(118) Devlin, F. J.; Stephens, P. J.; Oesterle, C.; Wiberg, K. B.; Cheeseman, J.
R.; Frisch, M. J. J. Org. Chem. 2002, 67, 8090-8096.
(119) Gigante, D. M. P.; Long, F.; Bodack, L. A.; Evans, J. M.; Kallmerten, J.;
Nafie, L. A.; Freedman, T. B. J. Phys. Chem. A 1999, 103, 1523-1537.
(120) Bursi, R.; Devlin, F. J.; Stephens, P. J. J. Am. Chem. Soc. 1990, 112,
9430-9432.
(121) Polavarapu, P. L.; Ewig, C. S.; Chandramouly, T. J. Am. Chem. Soc. 1987,
109, 7382-7386.
(122) Urbanova, M.; Setnicka, V.; Volka, K. Chem. Listy 2002, 96, 301-304.
(123) Stephens, P. J.; Devlin, F. J.; Aamouche, A. ACS Symp. Ser. 2002, 810,
18-33.
(124) Nafie, L. A.; Freedman, T. B. In Circular Dichroism, Principles and
Applications, 2nd ed.; Berova, N., Nakanishi, K., Woody, R. W., Eds.;
John Wiley & Sons: New York, 2000; pp 97-131.
(125) Mason, S. Enantiomer 1998, 3, 283-297.
(126) Yang, D.; Rauk, A. ReV. Comput. Chem. 1996, 7, 261-301.
(127) Buckingham, A. D. Faraday Discuss. 1995, 99, 1-12.
9
8974 J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004

Conformationally Controlled Photodecarboxylations
A R T I C L E S
cleavage products. The reason for this may be conformational
control imposed by the media or forced recombination of radical
pairs to yield starting ester. Where both photoproducts are
observed, the relative amounts of photodecarboxylation product
increase with increasing temperature. One important contributing
reason for this observation is the greater importance of the s-cis
conformer, especially in the excited state, at higher temperatures.
In several of the media where both products are obtained, the
rates at which the s-trans undergoes cleavage and then the
radical pair separates must increase with temperature. The fact
that the relative yields of photodecarboxylation products increase
even under these conditions strongly suggests the increased
importance of s-cis conformations.
The photochemistry of chiral ester 1 has been studied in detail.
Under a wide range of experimental conditions, including
changes in temperature, solvent, and pressure and in the presence
of radical trapping reagents, the decarboxylation product is
formed with complete retention of chirality. Only an s-cis
conformation is reasonable as the immediate precursor of the
product-forming transition state which, again, is most easily
assigned to a spiro-lactonic structure. DFT and CIS calculations
reveal that the s-cis conformations are higher in energy than
the more stable s-trans conformers, but the energy differences
are compatible with involvement by the s-cis in the reactions.
The presence of some s-cis conformers is further indicated by
VCD spectroscopic measurements. The importance of confor-
mational equilibria in the ground (and probably excited singlet)
states in most of the media investigated is also indicated by the
efficient photodecarboxylation of the eight-membered ring
lactone 6, especially at elevated temperatures. The constraints
imposed by reaction cavities of polyethylene films, especially
those of higher crystallinity, may well be too large to allow
significant equilibration of conformers within the short excited
singlet state lifetimes of the aryl esters.
much more efficient reaction sequence than other currently
available routes for the syntheses of the same molecules.
Especially, our irradiations of 1 in unstretched high-density PE
films or in acetonitrile under high pressure provide very high
yields of the photodecarboxylation product 2 and almost none
of the competing cleavage product 3; in the PE films, 2 was
isolated in >98% yield (based on consumed 1) and >98% ee.
As such, each of these drastically attenuates “side reactions”,
allowing easy isolation of 2, and, we believe, many other
products similar to it from reactions of appropriate precursor
aryl esters by a simple irradiation of the doped films. The results
presented here demonstrate that seemingly slight changes of the
reaction conditions can lead to large changes in the conforma-
tional contents of the aryl esters and the division between their
competing photoreaction pathways. Conformational control via
confining media as a tool in synthetic organic photochemistry
should be exploited much more in the future than it has been
to date.
Acknowledgment. Financial support of this work by a Grant-
in-Aid for Scientific Research from the Ministry of Education,
Culture, Sports, Science, and Technology of Japan (No.
16750034) and the 21st Century COE for Integrated EcoChem-
istry (to T.M.) and the U.S. National Science Foundation (to
R.G.W.) is gratefully acknowledged. We thank Dr. Hiroshi
Izumi at the National Institute of Advanced Industrial Sciences
and Technology for measurements made with VCD spectros-
copy.
Supporting Information Available: Details of experimental
procedures, figures of exponential fittings of fluorescence decay
profiles to time-correlated single photon counting measurements,
and B3LYP/6-31G(d) or CIS/6-31G(d) optimized geometries
of conformers of aryl ester 1 (Z-matrix) in their ground and
excited states, and optimized ground-state geometries of lactone
6. This material is available free of charge via the Internet at
http://pubs.acs.org.
From a synthetic standpoint, the stereospecific conversion of
a readily available chiral acid to an ester and then transformation
of the ester to an enantiopure arylalkane is a much simpler and
JA049688W
9
J. AM. CHEM. SOC. VOL. 126, NO. 29, 2004 8975