Temperature effect on [2+2] intramolecular photocycloadditions

Article abstract of DOI:10.1021/ja9531079

The temperature dependence of intramolecular [2+2] photocycloadditions has been studied for two series of cyclohexenones, α- and β-linked, respectively, to a variously substituted olefin moiety via a trimethylene chain. With one exception, the disappearance quantum yield of the starting material decreases sharply in a narrow temperature range to approximately one-half of its initial value. No such sharp temperature effect was observed on the product ratio, which either remains constant or varies monotonically over the entire temperature range. This indicates that the temperature effects on the disappearance rate and on the product distribution refer to two distinct processes. A mechanism rationalizing this behavior is proposed.

Full text of DOI:10.1021/ja9531079

8278
J. Am. Chem. Soc. 1996, 118, 8278-8284
Temperature Effect on [2 + 2] Intramolecular
Photocycloadditions†
Dan Becker*,‡ and Yael Cohen-Arazi
Contribution from the Department of Chemistry, Technion-Israel Institute of Technology,
Haifa 32000, Israel
ReceiVed September 11, 1995X
Abstract: The temperature dependence of intramolecular [2 + 2] photocycloadditions has been studied for two
series of cyclohexenones, R- and â-linked, respectively, to a variously substituted olefin moiety Via a trimethylene
chain. With one exception, the disappearance quantum yield of the starting material decreases sharply in a narrow
temperature range to approximately one-half of its initial value. No such sharp temperature effect was observed on
the product ratio, which either remains constant or varies monotonically over the entire temperature range. This
indicates that the temperature effects on the disappearance rate and on the product distribution refer to two distinct
processes. A mechanism rationalizing this behavior is proposed.
Introduction
Scheme 1
The temperature effect on [2 + 2] intermolecular photocy-
cloadditions has been examined in detail by de Mayo.1,2 Less
attention has been paid to the effect of temperature on the
corresponding intramolecular photocycloadditions; we are aware
of only one such communication, published by Jones3 in 1976.
Loutfy and de Mayo2 observed a temperature effect on both
the product ratio and the product quantum yield in certain [2 +
2] intermolecular photocycloadditions. For example, when the
temperature is increased from -102 °C to 27 °C, the product
quantum yield for the addition of cyclopentenone to cyclohexene
drops by a factor of up to 1.6, and the cis/trans product ratio
decreases. In the addition of cyclopentenone to cyclopentene,
the product quantum yield decreases by a factor of up to 2.6
when the temperature is increased from -71 °C to 27 °C. In
several other [2 + 2] intermolecular photocycloadditions the
product quantum yield increases or remains unchanged with
increasing temperature, depending on the structures of the enone
and olefin. The authors concluded that the temperature effect
on the product quantum yield is not related to the greater
viscosity at low temperatures, but to the activation energy
difference between biradical closure and fission (k2/k3 in Scheme
1). In contrast, the temperature effect on the product ratio was
said to arise from the temperature dependence of the relative
rates of closure from different conformations of the biradical,
i.e., cis-cyclization has the lower activation energy and therefore
presumably predominates at lower temperatures.
quantum yield, which depends on the rate difference between
biradical closure and fission, is temperature independent.
In this paper we describe a systematic study of the temper-
ature effect on the disappearance rate of the starting material
and on the product selectivity in [2 + 2] intramolecular
photocycloaddition of cyclohexenones, R- or â-linked to a
variously substituted olefin moiety Via a trimethylene chain.
Results
All irradiations were carried out in n-hexane at constant
concentration (0.5-0.7 mM) under nitrogen at several temper-
atures ((0.5 °C, internal monitoring) using a uranium glass filter
(λ > 330 nm). The irradiations were continuous, and the
disappearance of the starting material was followed by GC
analysis of aliquots withdrawn during irradiation at various time
intervals. The disappearance rate of the starting material (V) at
each temperature was determined4 in triplicate from the slope
of the linear plot of ln(C) versus time, where C is the starting
material’s concentration. The disappearance rate (V) is propor-
In contradiction to the findings of Loutfy and de Mayo,
Jones,3 who studied the temperature effect on the [2 + 2]
intramolecular photocycloaddition of cyclopentenone to an
olefin in a rigid system, found that the rate constant for the
triplet enone reaction (k1 in Scheme 1) increases as the
temperature is increased from 4 °C to 51 °C, whereas the product
tional to the corresponding quantum yield (φ), where φ ) V/Iabs
,
and Iabs is the intensity of the light absorbed.
It should be emphasized, that the immersion lamp intensity
was constant in all irradiations. This was checked by irradiating
samples of 1 (Scheme 2) at 20 °C at identical concentration
before each irradiation and verifying that the rate of its
disappearance was unchanged. The lamp was sufficiently well
insulated, so we could assume that its intensity was not affected
by changing the temperature of the cooling liquid. Furthermore,
the absorbance spectra of all the substances studied in this work
were similar due to their common enone chromophor and the
remote changes in structure, so that we could assume that the
intensity of the light absorbed (Iabs) in all irradiations was
approximately constant. Thus, the ratio of the quantum yields
† Prefatory comment: The study described in this paper is part of a
comprehensive experimental investigation of photochemical [2 + 2]
cycloadditions carried out in this laboratory, that was discontinued at the
death of the senior author. The present results are sufficiently complete,
however, to allow novel conclusions to be drawn about the mechanism of
intramolecular [2 + 2] photocycloadditions, which are amenable to
computational confirmation.
‡ Deceased, 7/4/1994.
X Abstract published in AdVance ACS Abstracts, July 15, 1996.
(1) de Mayo, P. Acc. Chem. Res. 1971, 4, 41.
(2) Loutfy, R. O.; de Mayo, P. J. Am. Chem. Soc. 1977, 99, 3559.
(3) Jones, G.; Ramachandran, B. R. J. Photochem. 1976, 5, 341.
(4) The experimental error in determining the rate is (10%.
S0002-7863(95)03107-6 CCC: $12.00 © 1996 American Chemical Society

[2 + 2] Intramolecular Photocycloadditions
J. Am. Chem. Soc., Vol. 118, No. 35, 1996 8279
Figure 1. Dependence of V(1) (9) and V(2) (0) on temperature.
Figure 3. Dependence of V(10) (b) and V(11) (O) on temperature.
Scheme 3
a Reference 22.
decreased viscosity of the solution as the temperature is
increased.2 However, the dependence of the viscosity on the
temperature is smoothly exponential, so it would be unreason-
able to ascribe the sharp temperature effect on the rates to it.
Moreover, irradiation of 1 and 2 at 20 °C in pentane, hexane,
decane, dodecane, and tetradecane showed that the temperature
effect is not related to viscosity changes. For example, V(1) in
decane at 20 °C (η ) 0.92 cp)7 is smaller by a factor of 1.2
than V(1) in hexane at 20 °C (η ) 0.33 cp). The viscosity of
hexane at -55 °C is approximately 0.84 cp,8 so that V(1) in
hexane is about twice as large as in decane under comparable
conditions of viscosity. Hence, de Mayo’s conclusion2 that
decreasing product quantum yield in intermolecular photocy-
cloadditions with increasing temperature is independent of
viscosity changes, holds for the disappearance quantum yield
of the starting material in the corresponding intramolecular
photocycloadditions as well.
We also considered the possibility that the temperature effect
on the disappearance rates between -15 °C and 5 °C might be
related to the decreased absorbance as the temperature is
increased. However, the absorbance changes in this narrow
temperature range are negligible.
The temperature effect on the disappearance rates of the
R-substituted cyclohexenones (Scheme 3) was more variable
but, except for 16, the trends were qualitatively similar to that
of the â-substituted cyclohexenones.
Figure 2. Dependence of V(1) (9), V(2) (0), V(3) (b), V(4) ([) and
V(5) (O) on temperature.
Scheme 2
of two reactions (φ(i)/φ(j)) could be equated with the ratio of
the corresponding disappearance rates (V(i)/V(j)).
As can be seen in Figure 1, the dependence of the disap-
pearance rates of 1 and 2 (V(1) and V(2), respectively) (Scheme
2) on temperature is identical; both of them decreases sharply
in a narrow temperature range above -15 °C to approximately
one-half of their initial values. The quantum yield for disap-
pearance of 1 (φ(1)) in hexane at 20 °C had been determined5
actinometrically to be 0.5 ( 0.05, so the 1.8-fold larger value
of V(1) at -55 °C brings φ(1) up to approximately 0.9 at that
temperature.
The temperature range between -15 and 5 °C, where the
rates decrease sharply, is the interesting one. We therefore
irradiated all of the â-substituted cyclohexenones 1, 2, 3,6 4,
and 5 (Scheme 2) in this region, extended to 20 °C. The same
temperature effect was found on the disappearance rates of all
the â-substituted cyclohexenones, as can be seen in Figure 2.
We considered the possibility that the temperature effect on
the disappearance rates in this region might be related to the
Separate irradiation of 10-15 (Scheme 3) were carried out
under the usual conditions. Here too, as can be seen in Figure
3, the disappearance rates of 10 and 11 drop to approximately
one-half of their initial values as the temperature is increased
above -35 °C. Here the maximum is shifted to a lower
temperature, so the subsequent drop of the rate with increasing
(7) The viscosity values were taken from the CRC Handbook of
Chemistry and Physics, 72nd ed.; Lide, D. R., Ed. CRC Press, Inc.: Boston
1991-1992, pp 6-156.
(8) This viscosity value was calculated according to equation (4) in Allan,
J. M.; Teja, A. S. Can. J. Chem. Eng. 1991, 69, 986.
(5) Nagler, M. D. Sc. Thesis 1978, Haifa, Israel.
(6) Synthetic restrictions did not permit preparation of 3 in greater than
83% purity, accompanied by 17% of 2 and this mixture was irradiated.

8280 J. Am. Chem. Soc., Vol. 118, No. 35, 1996
Figure 4. Dependence of V(12) (4) and V(13) (b) on temperature.
Figure 5. Dependence of V(16) ([) on temperature.
Becker and Cohen-Arazi
starting material.10 De Mayo2 has concluded that biradical
closure to products (k2) governs the temperature dependence
of the quantum yield for product formation in intermolecular
photocycloadditions. This is not the case with respect to the
disappearance rate of starting materials, because k2 is irrelevant
to it, assuming that the 1,4-biradical does not revert to the triplet.
Moreover, as will be discussed below, regardless of whether or
not biradical formation is reversible, if k2 were the temperature-
dependent rate coefficient we would expect the product ratio
to be sharply temperature dependent in the region between -15
and 5 °C, as is the reaction efficiency. However, this is not so:
the product ratio varies monotonically with temperature, as can
be seen in Figures 6 and 8. Thus, it is obvious that k2 does not
represent the rate determining step.
The initial bonding step (k1) is an activated process. If k1
were rate determining for the disappearance of starting material,
the normal monotonic increase of rate with temperature should
have been observed. Such an increase is approximated only in
the case of the R-substituted cyclohexenone 16, which has a
tert-butyl group on the olefin (Figure 5), and is manifested
residually at the lowest temperatures in the case of the other
R-substituted cyclohexenones. We can assume that the tem-
perature-dependent process in these cases is the primary
bonding, which is presumably subject to steric effects. In the
case of the â-substituted cyclohexenones, where a constant
reaction rate is observed at low temperatures for both the
unsubstituted compound 1 and the tert-butyl substituted com-
pound 2 (Figure 1), the steric effect on the primary bonding is
evidently insufficient to affect the temperature dependence of
the disappearance rate.
Regardless of the different behavior at low temperatures, a
similar temperature effect on the reaction efficiency in the region
between -15 and 5 °C was observed in all cases (except 16).
We can therefore conclude that this temperature effect is roughly
independent of the reactant’s structure. Consequently, in this
region, the temperature-dependent process must precede the
primary bonding step (k1 in Scheme 1).
temperature is more gradual. The quantum yield for the
disappearance of 10 (φ(10)) in hexane at 20 °C had been
determined9 actinometrically to be 0.5 ( 0.05, so that φ(10)
exhibits a maximum at -35 °C that is very close to unity.
In the case of the methyl analogues 12-15 (Scheme 3), we
observed a greater similarity to the temperature dependence of
the disappearance rates of the â-substituted cyclohexenones, as
can be seen in Figure 4 for compounds 12 and 13. Their
disappearance rates drop to approximately one-half of their
initial values as the temperature is increased above -15 °C.
However, a slight increase of the rates was observed between
-55 and -15 °C, in contrast to the constant reaction rate, which
was observed in the case of the â-substituted cyclohexenones
(Figure 1). Compounds 14 and 15 behaved similarly to 12 and
13.
This slight increase of the disappearance rate at low temper-
atures in the case of the R-substituted cyclohexenones is much
more pronounced when a bulky tert-butyl group is substituted
on the olefin, as can be seen in Figure 5 for compound 16
(Scheme 3). V(16) increases with increasing temperature, except
in the limited region between -5 and 0 °C, where a local
decrease is observed, presumably due to a vestigial temperature
effect, similar to that observed in the other cases. V(16) is at
least one order of magnitude smaller than V(11) (≈V(14) ≈
V(15)), which is slightly smaller than V(10) (≈V(12) ≈ V(13))
at all temperatures (Figure 3); thus, the quantum yield for the
disappearance of 16 is smaller than 0.05 at all temperatures.
Hence, if our model is restricted to Scheme 1, the reduction
of the reaction efficiency in the region between -15 and 5 °C
must be due either to a decrease in the rate of intersystem
crossing (kisc) or to an increase in the rate of the thermal decay
of the triplet enone to its ground state (k′d),11 since radiative
decay is unlikely to be markedly temperature dependent. The
decrease in the disappearance rate in our case is very sharp,
and it is unreasonable to assume that thermal decay of the active
triplet suddenly becomes efficient at a particular temperature
near -15 °C. Furthermore, the decrease in the rate stops at 5
°C, when it has reached one-half of its initial value, and then
remains constant or begins to increase slightly. It is unlikely
that the efficiency of thermal decay would become constant at
an arbitrary temperature.
In intermolecular photocycloadditions, the quantum yield of
intersystem crossing from the singlet to the triplet enone (kisc
in Scheme 1) is equal to unity,12 and it is reasonable to assume
that it has the same value in intramolecular additions as well.
The triplet yield can therefore not be improved by lowering
(10) Becker, D.; Nagler, M.; Sahali, Y.; Haddad, N. J. Org. Chem. 1991,
56, 4537.
(11) (a) Schuster, D. I.; Dunn, D. A.; Heibel, G. E.; Brown, P. B.; Rao,
J. M.; Woning, J.; Bonneau, R. J. Am. Chem. Soc. 1991, 113, 6245. (b)
Schuster, D. I.; Brown, R. H.; Resnick, B. M. J. Am. Chem. Soc. 1978,
100, 4504. (c) Wagner, P. J.; Bucheck, D. J. J. Am. Chem. Soc. 1969, 91,
5090.
Discussion
In our systems, k3 (see Scheme 1) is insignificant because
(12) (a) de Mayo, P.; Pete, J. P.; Tchir, M. F. Can. J. Chem. 1968, 46,
2535. (b) de Mayo, P.; Nicholson, A. A.; Tchir, M. F. Can. J. Chem. 1970,
48, 225.
there is no more than 10% reversion of the 1,4-biradical to the
(9) Denekamp, C. M. Sc. Thesis 1992, Haifa, Israel.

[2 + 2] Intramolecular Photocycloadditions
J. Am. Chem. Soc., Vol. 118, No. 35, 1996 8281
Scheme 4
the temperature, so it cannot be responsible for the observed
increase in the reaction efficiency with lowering the temperature
as well.
Obviously, the model in Scheme 1 cannot explain our
experimental results. Consequently, we must assume that there
is an additional temperature-dependent process that occurs
before the primary bonding step. This may be a partial transition
(kt in Scheme 4) of the active triplet (Ta) to an inactive triplet
(Tna), which is more stable than Ta, so that the reverse transition
from Tna to Ta is inefficient. The inactive triplet cannot undergo
primary bonding, but can only decay (k′′d) to its ground state.
If this is the case, then the sharp decrease at -15 °C of the
disappearance rate to approximately one-half of its initial value
might be due to this transition, where -15 °C is a sufficiently
high temperature for it to occur. At temperatures above 5 °C,
the two competing processessprimary bonding and the transi-
tion from the active triplet to the inactive oneshave similar
rates, so that they compensate one another and the reaction
efficiency remains approximately constant or varies slowly.
The low quantum yield (<0.05) in the case of 16 (Figure 5)
indicates that at all temperatures, thermal decay of the active
triplet (Ta) is much more facile than primary bonding, which
becomes more favorable as the temperature is increased. At
-5 °C we observe a small decrease in the rate, presumably due
to a vestigial temperature effect, similar to that described above,
which results from a possible partial transition from the active
triplet to the inactive one.
Semiempirical calculations at a relatively high level of
configuration interaction (MOPAC 93: AM1 (CI ) 6))13
supports the mechanism described in Scheme 4. Two stable
triplets formed from 3 (Z) were identified. The first (T1), in
which the unpaired spin is concentrated on the cyclohexenone
moiety, is separated by a potential barrier of 5 ( 1 kcal/mol
from a second triplet (T2), in which the unpaired spin is
concentrated on the olefin moiety in the side chain, that is
twisted out of planarity. T2 is more stable than T1 by 11 ( 1
kcal/mol, sosif we suppose that T1 is the active triplet (Ta) and
T2 is the inactive one (Tna)sthen the reverse endothermic
transition T2 f T1 is indeed inefficient, as proposed in Scheme
4. The more efficient thermal decay of T2 to the ground state
than that of T1 can be explained by their different geometries
and location of the spin. Thermal decay of a triplet in an open-
chain olefin is more efficient than that of one located in a ring,
due to the greater flexibility of the open-chain olefin, that
facilitates vibronically induced intersystem crossing from the
triplet T2 to the ground state; consequently, T2 has a shorter
lifetime.14 Thus, the reason for the inactivity of T2 is its more
rapid thermal decay than reaction with the enone moiety.
Preliminary attempts to measure the lifetime of either triplet
formed from 1 by laser experiments were unsuccessful. Only
Figure 6. Temperature dependence of the photoadduct ratio 6/7 (endo/
exo) on irradiation of 2 (E-isomer) (b) and 3 (Z-isomer) (4).
one signal was observed; the lifetime of the species responsible
for it is shorter than 1 ns15 at room temperature and is therefore
not measurable within the time resolution of the equipment
presently available to us (≈1 ns nitrogen laser). This is not
surprising in view of Gleiter’s finding that the lifetime of a
similarly substituted cyclohexenone is 90 ps.16 Moreover, it
appears that the quantum yield of the phosphorescence in the
case of 1 is very low, making lifetime determinations very
difficult.
â-Substituted Cyclohexenones. In the case of 2 and 3 the
ratio of photoadducts 6/7 (endo/exo) is similar from irradiation
of both isomers at all temperatures (Figure 6),17 and an identical
temperature dependence of the product ratio was obtained for
both of them; the endo/exo ratio increases monotonically with
temperature from 0.55 at -80 °C to 1.4 at 100 °C; at
approximately 16 °C the ratio is 1.0.
It is reasonable to assume that 2 and 3 undergo primary
bonding at the â-carbon, providing the 1,4-biradicals I and II,
respectively (Scheme 5),18 which are in rapid equilibrium and
can close to the endo and exo products, respectively. Since
the product ratio is temperature-dependent, the temperature
dependence of the rates of closure to the two stereoisomers endo
and exo must be different.
The excess of the endo product in the irradiation of 2 (the
E-isomer) above 16 °C might have been explained, by primary
bonding at the R-carbon of 2 (E) at the higher temperatures
(>16 °C), in addition to predominant primary bonding at the
â-carbon, so that an excess of endo product results. Hence, in
the case of 3 (Z) we would expect an excess of exo product
above 16 °C. Since this is not the case, as can be seen in Figure
6, the possibility of primary bonding at the R-carbon can be
excluded.
Following the modified Eyring approach19 of eq 1, a linear
dependence of the ln(P) value on the reciprocal of the temper-
ature was found in all of the cases. P is defined by eq 2, where
k and k′ are the overall rate constants for the formation of the
diastereomeric products endo and exo respectively.
(15) Very short lifetime in comparison to the lifetime of the triplet formed
from 2-cyclohexenone measured by Schuster as 29 ns in cyclohexane: ref
11a.
(16) Gleiter, R.; Fischer, E. Chem. Ber. 1992, 125, 1899.
(17) The experimental error in determining the product ratio is (5%.
(18) The biradical conformations in Scheme 5 are analogous to those in
Griesbeck’s work, where triplet-singlet biradical inversion geometries were
taken in account: (a) Griesbeck, A. G.; Mauder, H.; Stadtmu¨ller, S. Acc.
Chem. Res. 1994, 27, 70. (b) Griesbeck, A. G.; Stadtmu¨ller, S. J. Am. Chem.
Soc. 1991, 113, 6923. (c) Griesbeck, A. G.; Stadtm¸ller, S.; Busse, H.;
Bringmann, G.; Buddrus, J. Chem. Ber. 1992, 125, 933.
(13) Full details will be given in a subsequent paper (Halevi, E. A.;
Cohen-Arazi, Y.). A referee requested a simulation of the kinetic data in
order to verify the “second triplet hypothesis”. This will be carried out as
soon as all of the potential barriers of the processes along the mechanism
have been computed.
(14) This is consistent with Bonneau’s conclusion that triplet lifetimes
of open-chain alkenones are shorter than those of cyclohexenones: Bonneau,
R. J. Am. Chem. Soc. 1980, 102, 3816.
(19) Eyring, H. J. Chem. Phys. 1935, 3, 107.

8282 J. Am. Chem. Soc., Vol. 118, No. 35, 1996
Becker and Cohen-Arazi
∆∆Hq
R
1
T
∆∆Sq
R
ln(P) ) -
×
+
(1)
(2)
k(endo)
k′(exo)
P )
Figure 7 shows the temperature dependence of diastereose-
lectivity on irradiation of 2 in the form of an Eyring diagram.20
The values of the activation parameter differences ∆∆Hq and
∆∆Sq (Tables 1 and 2) were derived from the slope and the
intercept on the y-axis, respectively.
According to the values referring to 2 and 3 in Table 1, the
activation enthalpy (∆Hq) for exo closure is smaller than that
for endo closure. This is consistent with the excess of the exo
product at low temperatures and its decrease with temperature
(Figure 6). At higher temperatures (>16 °C) the endo product
is in excess, because the activation entropy (∆Sq) for endo
closure is more positive than that for exo closure (Table 1).
Presumably, this is so because the rotation of the tert-butyl
group, and/or of its methyl groups, is more restricted during
exo closure, as can be seen in Scheme 5.
Contrary to 2 and 3, in the case of 4 and 5 the ratio of
photoadducts 8/9 (endo/exo) is identical for both isomers (≈1.1)
at all temperatures. The methyl group is probably insufficiently
bulky to cause a significant difference in either the activation
enthalpy or entropy of closure to either the endo or the exo
product (see Table 1), so the endo/exo ratio is near unity and
essentially temperature independent.
r-Substituted Cyclohexenones. A monotonic temperature
effect on the product ratio was also observed in several
irradiations of R-substituted cyclohexenones. The ratio of
photoadducts 20/21 (endo/exo) on irradiation of 12 (the E-
isomer) was dependent on the temperature, decreasing mono-
tonically from 9.3 at -55 °C to 6.5 at 20 °C;10,21 approximately
a 1.4-fold decrease (Figure 8). On irradiation of 13 (the
Z-izomer) the ratio 20/21 was also temperature dependent,
decreasing monotonically from 2.7 at -55 °C to 1.8 at 20
°C;10,21 a 1.5-fold decrease, similar to that obtained for the
E-isomer.
Figure 7. Temperature dependence of diastereoselectivity on irradiation
of 2 (b) in the form of an Eyring diagram.
Table 1. Activation Parameter Differences for the â-Substituted
Cyclohexenones: ∆∆Hq ) ∆Hq(endo) - ∆Hq(exo); ∆∆Sq )
∆Sq(endo) - ∆Sq(exo)
∆∆Hq (kJ/mol)
∆∆Sq (J/mol × K)
2
3
4
5
2.9
10.2
2.9
0.0 ( 0.1
0.0 ( 0.1
10.4
0.0 ( 0.1
0.0 ( 0.1
Table 2. Activation Parameter Differences for the R-Substituted
Cyclohexenones: ∆∆Hq ) ∆Hq(endo) - ∆Hq(exo); ∆∆Sq )
∆Sq(endo) - ∆Sq(exo)
∆∆Hq (kJ/mol)
∆∆Sq (J/mol × K)
11
12
13
14
15
16
0.0 ( 0.1
-2.3
0.8
8.1
-2.3
-3.2
-2.5
1.4
0.6
1.2
0.0 ( 0.1
0.0 ( 0.1
On irradiation of 14 (the E-isomer), the ratio of photoadducts
22/23 (endo/exo) was dependent on the temperature, decreasing
monotonically from 7.0 at -55 °C to 4.3 at 20 °C;22 a 1.6-fold
decrease, similar to that obtained for 12 and 13. On irradiation
of 15 (the Z-izomer) the ratio 22/23, which had been deter-
mined22 to be 1.8 at 20 °C, remained essentially constant
throughout the temperature range, and on irradiation of 16 the
ratio 24/25 (endo/exo), which had been determined9 to be 3.3
at 20 °C, also remained constant at all temperatures. On
irradiation of 11, the ratio of photoadducts 18/19 (anti/syn),
which had been determined22 to be 2.3 at 20 °C, remained
constant throughout the temperature range as well.
Thus, in the case of the R-substituted cyclohexenones a large
excess of endo product is observed, and this excess decreases
with temperature in the case of 12, 13, and 14. The R-substi-
tuted cyclohexenones probably undergo primary bonding at the
R-carbon of the enone (Scheme 6) to yield two biradical
rotamers, which can close either to the endo or to the exo product
(in analogy to biradicals I and II in Scheme 5).
Figure 8. Temperature dependence of the photoadduct ratio 20/21
(endo/exo) on irradiation of 12 (E-isomer) ([).
14. This is consistent with the excess of the endo product and
its decrease with temperature. In the case of 11, 15, and 16,
where no temperature effect was observed (the activation
enthalpy is equal for both products), the excess product (anti
or endo, respectively) must be ascribed to the entropy difference
(Table 2).
According to Table 2, the activation enthalpy for endo closure
is smaller than that for exo closure in the case of 12, 13, and
(20) (a) Pelzer, R.; Scharf, H.-D.; Buschmann, H.; Runsink, J. Chem.
Ber. 1989, 122, 1187. (b) Buschmann, H.; Scharf, H.-D.; Hoffmann, N.;
Plath, M. W.; Runsink, J. J. Am. Chem. Soc. 1989, 111, 5367.
(21) The ratio of photoadducts published in ref 10 were found to be
incorrect, and were redetermined in this work according to NMR, GC, and
GC-MS data.
On irradiation of 13 and 15 (the Z-isomers), the endo/exo
product ratio is smaller than that obtained from the E-isomers
12 and 14. A possible explanation might be that the Z-isomers
undergo primary bonding at the â-carbon of the enone, yielding
(22) Becker, D.; Haddad, N. Tetrahedron 1993, 49, 947.

[2 + 2] Intramolecular Photocycloadditions
J. Am. Chem. Soc., Vol. 118, No. 35, 1996 8283
Scheme 5
Scheme 8
Scheme 6
resolution MS were measured on a Varian MAT-711 instrument, and
GC analyses were carried out on an HP 5890 instrument using the
capillary column DB210, 30 m, 0.25 mm with helium as carrier gas.
A Perkin-Elmer 298 instrument was used for IR, and a Hewlett Packard-
diode array spectrophotometer-8452 ≈ was used for UV.
Irradiations. Irradiations were carried out in hexane as a solvent
under a nitrogen atmosphere. The commercial hexane was purified
by shaking with concentrated sulfuric acid and then with 10% sodium
carbonate, irradiation through quartz for 4 h using a 450-W Hanovia
lamp, repeated shaking with concentrated sulfuric acid, and distillation.
The photoreactions were carried out in concentrations of 0.5-0.7 mM
using an 80W Hanau mercury vapor lamp (Q-81) with an uranium glass
filter (λ > 330 nm). The photoreactions were followed by GC analysis.
A HAAKE F3 circulator was used for cooling the irradiating solution
using methanol as the cooling liquid. The temperature monitoring
((0.5 °C) was internal using a thermocouple (type K).
Scheme 7
(E)-6,6-Dimethyl-4-hepten-1-ol (26). Kocien’nski’s23 procedure
was used for the preparation of 26. n-Butyllithium (43.5 mL, 1.7 M
in hexane, 74 mmol) was added dropwise to a suspension of CuCN
(1.8 g, 12.8 mmol) in dry ether (45 mL) at -40 °C under nitrogen.
The resultant mixture was slowly warmed to -20 °C and 3,4-dihydro-
[2H]pyrane (2.25 mL, 24.6 mmol) in dry ether (30 mL) was added.
The reaction mixture was stirred for 30 min at -20 °C and 2 h at
room temperature. Water (2 mL) was added, and the resultant mixture
was poured into a 10% solution of ammonia in a saturated aqueous
solution of ammonium chloride (60 mL). The product was extracted
with ether (3 × 25 mL) and dried over MgSO4, and the solvent was
removed under reduced pressure. The resultant oil was purified by
distillation (40 °C, 0.1 mmHg) to give 2.3 g of pure 26 in 65% yield:
1H NMR (200 MHz, CDCl3) δ 0.96 (s, 9 H), 1.61 (qu, 2 H), 2.05 (q,
2 H), 3.62 (t, 2 H), 5.3 (dt, 1 H), 5.42 (d, 1 H).
(E)-7-Bromo-2,2-dimethyl-3-heptene (27). Methanesulfonyl chlo-
ride (3.5 mL, 45 mmol) was added dropwise to a solution of 26 (3.96
g, 27.9 mmol) and triethylamine (9.7 mL, 70 mmol) in THF (28 mL)
at 0 °C under nitrogen. The reaction mixture was stirred for 1 h at 0
°C and 30 min at room temperature and then quenched with water (14
mL). The solvent was removed under reduced pressure, and the product
was extracted with CH2Cl2 (3 × 15 mL), washed with 5% HCl (10
mL) and NaHCO3 saturated aqueous solution (5 mL), and dried over
MgSO4. The solvent was removed to give 4.9 g of crude mesylate,
which was used in the next step without further purification. LiBr
(anhydrous, 3.6 g, 41.4 mmol) was added to a solution of this mesylate
(4.9 g) in THF (12 mL), and the reaction mixture was stirred overnight
under nitrogen. Water (14 mL) was added, and the solvent was
removed under reduced pressure. The product was extracted with
CH2Cl2 (3 × 15 mL). The organic phases were combined, washed
with water (10 mL), and dried over MgSO4. The crude bromide was
chromatographed over silica gel (eluent hexane) to give 4.0 g of pure
27 in 70% yield: 1H NMR (200 MHz, CDCl3) δ 0.96 (s, 9 H), 1.88
(qu, 2 H), 2.09 (q, 2 H), 3.41 (t, 2 H), 5.26 (dt, 1 H), 5.52 (d, 1 H).
Ethanal-3-bromopropylethylacetal (28). p-TSA‚H2O (12 mg) and
3-bromopropanol (8 g, 85 mmol) were added to a flask containing ethyl
only the exo product, in addition to primary bonding at the
R-carbon (Scheme 7). If so the E-isomers might also undergo
primary bonding at the â-carbon, increasing the excess of the
endo product.
Obviously, further investigation is needed in order to clarify
the mechanism of this reaction. The computations that are now
in progress13 may indicate the direction that it should take.
Preparation of Materials
The synthesis of 2 and 3 by the reaction of the corresponding
Grignard reagent with the enol ether 32 followed by hydrolysis,
is shown in Scheme 8, as well as the synthesis of the E- and
Z-bromo-olefins (27 and 31, respectively) substituted with a
bulky tert-butyl group.
Syntheses of compounds 1, 4, 5, and 10-15 and experimental
data for their photoproducts are available in refs 10 and 22.
Compound 16 was synthesized according to ref 10 using the
E-bromoolefin 27.
Experimental Section
General Data. Nuclear magnetic resonance spectra were obtained
on a Bruker AM-400 400-MHz instrument equipped with an Aspect
3000 computer and on a Bruker AM-200 200-MHz instrument. High-
(23) Kocien’nski, P.; Wadman, S. J. Am. Chem. Soc. 1989, 111, 2363.

8284 J. Am. Chem. Soc., Vol. 118, No. 35, 1996
Becker and Cohen-Arazi
vinyl ether (12.2 g, 169 mmol) at 0 °C under nitrogen. The reaction
mixture was warmed to 20 °C and stirred for 1 h. An additional amount
of p-TSA‚H2O (9 mg) was added slowly. The reaction mixture was
cooled to 10 °C, and a saturated aqueous solution of K2CO3 (0.9 mL)
was added. After stirring for 5 min the solution was dried with K2CO3
powder and filtered, and the solvent was removed under reduced
pressure. The crude product was purified by distillation (50 °C, 0.2
mmHg) to give 11.8 g of 28 in 84% yield: 1H NMR (200 MHz, CDCl3)
δ 1.2 (t, 3 H), 1.3 (d, 3 H), 2.09 (qu, 2 H), 3.51 (t, 2 H), 3.64 (m, 4 H),
4.67 (q, 1 H).
to magnesium (0.15 g, 6.25 mmol) in dry ether (3 mL) at room
temperature under nitrogen. After stirring the reaction mixture for 0.5
h, 32 (0.55 g, 3.9 mmol) in dry ether (3 mL) was added slowly. After
additional stirring for 1 h, the reaction mixture was quenched with 5%
hydrochloric acid (10 mL) at 0 °C. The product was extracted with
ether (3 × 20 mL), washed with water, and dried over MgSO4. The
solvent was removed under reduced pressure, and the crude oil was
chromatographed over silica gel (eluent hexane:CH2Cl2 ) 1:1) to give
0.61 g of pure 2 in 70% yield: IR (CHCl3) 1660 cm-1; UV (hexane)
1
λmax ) 266 nm (ꢀ ) 12 000); H NMR (400 MHz, CDCl3) δ 0.96 (s,
Ethanal-6,6-dimethyl-4-heptynethylacetal. n-Butyllithium (26 mL,
1.15 M in hexane, 30 mmol) and HMPA (6 mL) were added dropwise
to a solution of tert-butyl acetylene (2 g, 24.4 mmol) in THF (10 mL)
at -40 °C under nitrogen. After stirring for 1 h, 28 (6.3 g, 30 mmol)
was added slowly. The reaction mixture was stirred for 1 h, after which
the cooling bath was removed and the reaction mixture was kept at
room temperature overnight. The reaction flask was cooled (5 °C),
and aqueous 5% HCl (25 mL) was added. The solvent was removed
under reduced pressure, and the product was extracted with ether (3 ×
15 mL), washed with saturated NaHCO3 (10 mL), and dried over
MgSO4. The solvent was removed under reduced pressure to give 5.6
g of the desired product in 90% yield: 1H NMR (200 MHz, CDCl3) δ
1.16 (s, 9 H), 1.2 (t, 3 H), 1.3 (d, 3 H), 1.7 (qu, 2 H), 2.22 (t, 2 H),
3.51 (t, 2 H), 3.6 (q, 2 H), 4.68 (q, 1 H).
6,6-Dimethyl-4-heptyn-1-ol (29). A mixture of ethanal-6,6-di-
methyl-4-heptynethylacetal (4.6 g, 22 mmol) in methanol (50 mL) and
32% HCl (11 mL) was stirred for 24 h at room temperature. Water
(25 mL) was added, and the solvent was removed under reduced
pressure. The product was extracted with CH2Cl2 (3 × 20 mL), washed
with saturated NaHCO3 (10 mL), and dried over MgSO4. The crude
alcohol was chromatographed over silica gel (eluent hexane:CH2Cl2
) 1:1) to give 2.7 g of 29 in 88% yield: 1H NMR (200 MHz, CDCl3)
δ 1.16 (s, 9 H), 1.7 (qu, 2 H), 2.6 (t, 2 H), 3.74 (t, 2 H).
9 H), 1.25 (qu, 2 H), 1.95 (q, 4 H), 2.20 (m, 2 H), 2.35 (m, 4 H), 5.26
(dt, 1 H), 5.41 (d, 1 H), 5.80 (s, 1 H); HRMS calcd for C15H24O
220.3540, found 220.1812. Anal. Calcd for C15H24O: C, 81.76; H,
10.98. Found: C, 82.11; H, 11.26.
3-[(Z)-6,6-Dimethyl-4-heptenyl]-2-cyclohexen-1-one (3). Enone 3
was prepared from 31 in 73% yield as described for 2: 1H NMR (400
MHz, CDCl3) δ 1.03 (s, 9 H), 1.53 (qu, 2 H), 1.96 (q, 2H), 2.21 (m,
4 H), 2.33 (m, 2 H), 5.07 (td, 1 H), 5.33 (d, 1 H), 5.86 (s, 1 H); HRMS
calcd for C15H24O 220.3540, found 220.1838. Anal. Calcd for
C15H24O: C, 81.76; H, 10.98. Found: C, 81.75; H, 11.19.
2-[(E)-6,6-Dimethyl-4-heptenyl]-2-cyclohexen-1-one (16). Enone
16 was prepared from 27 in 60% yield as described for 12 in ref 10:
1H NMR (400 MHz, CDCl3) δ 0.96 (s, 9 H), 1.43 (qu, 2 H), 1.95 (qu,
4 H), 2.14 (t, 2 H), 2.30 (q, 2 H), 2.39 (t, 2 H), 5.26 (dt, 1 H), 5.41 (d,
1 H), 6.67 (t, 1 H); HRMS calcd for C15H24O 220.3540, found 220.1841.
Anal. Calcd for C15H24O: C, 81.76; H, 10.98. Found: C, 82.15; H,
11.20.
Irradiation of Enone Olefin 2. Enone 2 (2.7 mg, 12.3 µmol) was
dissolved in 20 mL of purified hexane and irradiated for 1 h. The
photoadducts 6 and 7 were formed in an approximately 1:1 ratio at 16
°C and were isolated in 93% yield. The two isomers were separated
on silica gel (eluent hexane:CH2Cl2 ) 8:1). 6: 1H NMR (400 MHz,
CDCl3) δ 0.86 (s, 9 H), 1.30-1.51 (m, 3 H), 1.52-1.66 (m, 4 H),
1.72-1.90 (m, 3 H), 2.01-2.10 (m, 2 H), 2.42 (t, 1 H), 2.52 (bd, 1 H),
2.66 (d, 1 H). 7: 1H NMR (400 MHz, CDCL3) δ 0.87 (s, 9 H), 1.31
(m, 1 H), 1.51-1.65 (m, 3 H), 1.67-1.76 (m, 2 H), 1.77-1.89 (m, 2
H), 2.03-2.26 (m, 3 H), 2.30 (d, 1 H), 2.43 (t, 1 H), 2.53 (dd, 1 H),
2.57 (d, 1 H).
(Z)-6,6-Dimethyl-4-hepten-1-ol (30). 6,6-Dimethyl-4-heptyn-1-ol
29 (0.39 g, 2.8 mmol) was reduced in methanol (150 mL) under
hydrogen pressure (4 atm) by using a Lindlar catalyst (Fluka AG) (39
mg). The reaction was stopped after 36 h, the catalyst was filtered
off, and the solvent was removed under reduced pressure to give 0.36
g (91% yield) of a Z/E-isomer mixture in a 80:20 ratio, respectively,
by GC analysis. Spectral data of the Z isomer 30: 1H NMR (200 MHz,
CDCl3) δ 1.09 (s, 9 H), 1.59 (qu, 2 H), 2.25 (q, 2 H), 3.67 (t, 2 H),
5.12 (dt, 1 H), 5.31 (d, 1 H).
Irradiation of Enone Olefin 3. A mixture of 83% enone 3 and
17% enone 26 (2.8 mg, 12.7 µmol) was irradiated as described for 2.
The photoadducts 6 and 7 were formed in approximately 1:1 ratio at
16 °C and were isolated in 92% yield.
(Z)-7-Bromo-2,2-dimethyl-3-heptene (31). The Z-alcohol 30 was
converted to the corresponding bromide 31 in 68% yield as described
for the E-isomer 27. 1H NMR (200 MHz, CDCl3) δ 1.09 (s, 9 H), 1.9
(qu, 2 H), 2.32 (q, 2 H), 3.38 (t, 2 H), 5.08 (dt, 1 H), 5.36 (d, 1 H).
3-Ethoxy-2-cyclohexen-1-one (32). 3-Ethoxy-2-cyclohexen-1-one
was prepared according to the following procedure.24 A solution of
1,3-cyclohexadione (5.3 g, 47 mmol) and p-TSA‚H2O (0.23 g, 1.2
mmol) in ethanol (25 mL) and benzene (90 mL) was refluxed for 1 h,
and the azeotrope composed of benzene, alcohol, and water was
removed. When the temperature of the vapor reached 78 °C, the
distillation was stopped, and the residual solution was washed with
10% aqueous NaOH saturated with NaCl (4 × 10 mL) and water (3 ×
5 mL). The resulting oil was distilled (67 °C, 0.4 mmHg) to give 4.9
g of pure 32 in 75% yield: 1H NMR (200 MHz, CDCl3) δ 1.32 (t, 3
H), 1.93 (qu, 2 H), 2.33 (m, 4 H), 3.85 (q, 2 H), 5.29 (s, 1 H).
3-[(E)-6,6-Dimethyl-4-heptenyl]-2-cyclohexen-1-one (2). A solu-
tion of 27 (0.89 g, 4.3 mmol) in dry ether (2 mL) was added dropwise
Irradiation of Enone Olefin 16. Enone 16 (2.7 mg, 12.3 µmol)
was irradiated for 5 h as described for 2. The photoadducts 24 and 25
were formed in a 3.3:1 ratio, respectively, at 20 °C and were isolated
in 90% yield. The two isomers were separated on silica gel (eluent
hexane:CH2Cl2 ) 9:1). 24: 1H NMR (400 MHz, CDCl3) δ 0.86 (s, 9
H), 1.50 (dd, 1 H), 1.62 (m, 1 H), 1.66-1.75 (m, 4 H), 1.85-2.03 (m,
4 H), 2.23 (m, 1 H), 2.39 (m, 1 H), 2.45 (m, 1 H), 2.78 (t, 1 H). 25:
1H NMR (400 MHz, CDCL3) δ 0.87 (s, 9 H), 1.50 (dd, 1 H), 1.62 (m,
1 H), 1.66-1.75 (m, 4 H), 1.85-2.03 (m, 4 H), 2.23 (m, 1 H), 2.39
(m, 1 H), 2.45 (m, 1 H), 2.83 (t, 1 H).
Acknowledgment. Y.C.-A. would like to express her deep
gratitude to Professor E. A. Halevi for the semiempirical
calculations and for very useful advice, to Professor M. Rubin
for valuable discussions, and to Professor M. Ottolenghi for
making his nitrogen laser available to her.
(24) Gannon, W. F.; House, H. O. Organic Syntheses; Wiley: New York,
Collect. Vol. 5, p 539.
JA9531079