Beyond butadiene II: Thermal isomerization of the [2 + 2] photodimer of an all-trans-tetraene, (R)-4,4aβ,5,6,10,10a-hexahydro-10aβ-methyl-2(3H)-methyleanthracene, to a 16-membered [8 + 8] cycle

Article abstract of DOI:10.1021/ja011135n

Enthalpies of stabilization of polyenyl radicals of increasing order previously obtained by thermal geometrical isomerization are applied to the ethylene-cyclobutane paradigm. Progressively lower enthalpies of activation for thermal cyclodimerization and its reverse, cycloreversion, are predicted and realized. Photochemical dimerization at -75 °C of the optically pure tetraene of the title (1) at the semicyclic double bond produces in the main only one (4-axx) of the three allowed cyclobutanes (4), to which the tentative configuration anti-exo,exo is assigned. Equilibration among the three cyclobutanes (4), a slower rearrangement to a thermodynamically considerably more stable, [8 + 8] cyclohexadecahexaene (16), and a surprisingly slow fragmentation to 1 are studied kinetically between -42.3 and -8.2 °C. Cycloreversion of the dimer 16 to monomer 1 occurs in the range 60.4-86.6 °C (ΔH? = 31.7 kcal mol^-1, ΔS? = +10.8 cal mol^-1 K^-1). The ratio of the rates of stereomutation and cycloreversion is significantly: Larger in these 1,2-dihexatrienylcyclobutanes than in two less strongly stabilized, previously published examples. The possible extension of Doubleday's calculational finding of entropic control of products from cyclobutane is considered.

Full text of DOI:10.1021/ja011135n

J. Am. Chem. Soc. 2001, 123, 9153-9161
9153
Beyond Butadiene II: Thermal Isomerization of the [2 + 2]
Photodimer of an all-trans-Tetraene,
(R)-4,4aâ,5,6,10,10a-Hexahydro-10aâ-methyl-2(3H)-methyleneanthracene,
to a 16-Membered [8 + 8] Cycle
W. von E. Doering,* Jianing He, and Liming Shao
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138-2902
ReceiVed May 7, 2001
Abstract: Enthalpies of stabilization of polyenyl radicals of increasing order previously obtained by thermal
geometrical isomerization are applied to the ethylene-cyclobutane paradigm. Progressively lower enthalpies
of activation for thermal cyclodimerization and its reverse, cycloreversion, are predicted and realized.
Photochemical dimerization at -75 °C of the optically pure tetraene of the title (1) at the semicyclic double
bond produces in the main only one (4-axx) of the three allowed cyclobutanes (4), to which the tentative
configuration anti-exo,exo is assigned. Equilibration among the three cyclobutanes (4), a slower rearrangement
to a thermodynamically considerably more stable, [8 + 8] cyclohexadecahexaene (16), and a surprisingly
slow fragmentation to 1 are studied kinetically between -42.3 and -8.2 °C. Cycloreversion of the dimer 16
to monomer 1 occurs in the range 60.4-86.6 °C (∆Hq ) 31.7 kcal mol-1, ∆Sq ) +10.8 cal mol-1 K-1). The
ratio of the rates of stereomutation and cycloreversion is significantly larger in these 1,2-dihexatrienylcy-
clobutanes than in two less strongly stabilized, previously published examples. The possible extension of
Doubleday’s calculational finding of entropic control of products from cyclobutane is considered.
Introduction
enthalpy of activation for cycloreversion of these cyclobutanes
has been predicted to be 33.8 kcal mol-1.4 This value agrees
well with experimental values of 34.0 and 35.7 kcal mol-1.9 In
the sterically more complex system 3-methylenecyclohexene/
anti- and syn-dispiro[5.0.5.2]tetradeca-1,8-diene (2; Scheme 1),
experimental enthalpies of activation are 31.7 kcal mol-1 for
cycloreversion and 30.0 kcal mol-1 for stereomutation.10 The
decrement of 2-4 kcal mol-1 is reasonably ascribed to a steric
acceleration.
Enthalpies of stabilization of polyenyl radicals obtained in
one system1-3 and incorporated into a simple paradigm lead to
the prediction of some remarkably low enthalpies of activation
for apposite, thermally induced transformations.4 The paradigm
for the present set of perturbations by polyenes (Scheme 1) is
the two-step, thermal interconversion of cyclobutane and two
molecules of ethylene5 (cyclodimerization to cyclobutane in the
forward direction,6 and cycloreversion or fragmentation to two
molecules of ethylene in the reverse).7 Of the factors ∆∆fH and
T∆∆fS which control the position of equilibrium, enthalpy
Perturbation by two double bonds (butadiene) leads to a
polyene of the third order, hexatriene, and dimers cis- and trans-
1,2-dibutadienylcyclobutane. The prediction of a further lower-
ing of the enthalpy of activation for cycloreversion to 27.0 kcal
mol-1 compares to an experimental value for 3 of 24.3 kcal
mol-1. Here too, the actual value is again lower than the
strongly favors cyclobutane by ∼18 kcal mol-1 8
, whereas
entropy, unfavorable for dimerizations in general, strongly favors
ethylene at temperatures required for convenient rates of
reaction.
predicted one by ∼3 kcal mol-1
.
Perturbation of the paradigm by one double bond generates
butadiene and dimers cis- and trans-1,2-divinylcyclobutane. The
Results: The 16-Membered Ring
* To whom correspondence should be addressed. E-mail: doering@
chemistry.harvard.edu.
In the present work, perturbation of the paradigm is extended
to a polyene of the fourth order (Scheme 1). An enthalpy of
activation of ∆Hq ) 22.4 kcal mol-1 is estimated for fragmenta-
tion of the hypothetical, unsubstituted 1,2-dihexatrienylcyclobu-
tane to a pair of octatetraene molecules.11 In the rigid example
examined here (4 of Scheme 1), the steric factor may be
expected to lower the value to ∼19 kcal mol-1 in analogy to
the two examples noted above.
(1) Doering, W. von E.; Kitagawa, T. J. Am. Chem. Soc. 1991, 113,
4288-4297.
(2) Doering, W. von E.; Sarma, K. J. Am. Chem. Soc. 1992, 114, 4,
6038-6043.
(3) The “order”, n, defines polyenes and polyenyl radicals containing n
double bonds.
(4) Doering, W. von E.; Belfield, K. D.; He, J.-n. J. Am. Chem. Soc.
1993, 115, 5414-5421.
(5) Doering, W. von E. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 5279-
5283.
(6) Quick, L. M.; Knecht, D. A.; Back, M. H. Int. J. Chem. Kinet. 1972,
4, 61-68.
(7) (a) Genaux, C. T.; Kern, F.; Walters, W. D. J. Am. Chem. Soc. 1953,
75, 6169-6199. (b) Butler, J. N.; Ogawa, R. B. J. Am. Chem. Soc. 1963,
85, 3346-3348. (c) Vreeland, R. W.; Swinehart, D. F. J. Am. Chem. Soc.
1963, 85, 3349-3353. (d) Beadle, P. C.; Golden, D. M.; King, K. D.;
Benson, S. W. J. Am. Chem. Soc. 1972, 94, 2943-2947.
(8) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of
Organic Compounds, 2nd ed.; Chapman and Hall: London, 1986.
(9) Hammond, G. S.; DeBoer, C. D. J. Am. Chem. Soc. 1964, 86, 899-
902.
(10) Doering, W. v. E.; Ekmanis, J. L.; Belfield, K. D.; Kla¨rner, F.-G.;
Krawzcyk, B. J. Am. Chem. Soc. 2001, 123, 5532-5541.
(11) The procedure is depicted explicitly in Scheme 2 of ref 4.
10.1021/ja011135n CCC: $20.00 © 2001 American Chemical Society
Published on Web 08/25/2001

9154 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
VonDoering et al.
Scheme 1
Figure 1. NMR chemical shifts (δ scale, 500 MHz) assigned to
hydrogen atoms (ppm, italics) and to carbon atoms for 1′-13C and 16′-
13C, along with diagrams of [4 + 4] and [6 + 6] head-to head dimers
of 1′, and the [8 + 8] head-to-tail dimer 16′.
tetraene 1′ does not yield a cyclobutane, but compound 16′ as
a yellow precipitate. Its ultraviolet absorption spectrum (λmax
278 nm, log ꢀ 4.40) is consistent with that of a conjugated
triene,14 while its mass spectrum (m/e 396) points to a dimer.
In its 13C NMR spectrum, 15, not 30, identifiable absorptions
indicate a structure of 2-fold symmetry, while two pairs of three
peaks (146.3, 139.7, and 138.3 ppm) can be identified as
quaternary olefinic carbon atoms by application of the DEPT
method. A dimer of the [8 + 8] type, but not of the [6 + 6], [4
1
+ 4], or [2 + 2] type, satisfies these observations. In its H
NMR spectrum, the most important feature is a hydrogen atom
at 2.65 ppm identified as allylic H-7 by homonuclear decoupling
of olefinic H-8 at 5.32 ppm, and by the effect of homonuclear
decoupling of H-7 on two aliphatic H atoms at 1.45 and 1.99
ppm (see Figure 1).
Tetraene 1′ (demethyl-1; throughout, the prime will denote
compounds without the methyl groups) was the initial subject
of our study. It had already been prepared as a relatively
unstable, colorless, crystalline compound of mp 110 °C from
an available ketone, 7′ 1 (demethyl-7), by application of a
conventional Wittig reaction.12 Although its structure followed
quite reasonably from the method of preparation, additional
support came from its mass, UV, and 13C NMR (normal
abundance) spectra, in which three quaternary olefinic carbon
atoms were recognized by the DEPT technique (Figure 1).
More rigorous confirmation of structure is obtained from a
sample of 1′ enriched with 13C (99.4%) in the exo-methylene
Although these observations are consistent with a 16-
membered cyclic structure, they do not suffice to distinguish
between head-to-head and head-to-tail alternatives. Crystal-
lographic analysis might have resolved the issue, but we have
been unable to obtain a suitable single crystal. Examination of
16′-13C, prepared from 1′-13C labeled in the exocyclic methylene
group, has, however, allowed the distinction to be made. All
the peaks in the spectrum of unlabeled 16′ can be seen again in
the spectrum of 16′-13C, except for the peak at 36.88 ppm, which
is overwhelmed by the much stronger peak at 36.91 ppm. The
resonance of the quaternary carbon atom at 138.3 ppm is split
into a doublet (J ) 40 Hz) by one-bond, carbon-carbon
coupling. There are some small peaks around these peaks which
might be due to weak two-bond, carbon-carbon coupling.
Observation of only a single coupling between the 13C-labeled
methylene and an adjacent quaternary carbon is strongly
supportive of a head-to-head structure for the dimer. A head-
to-tail structure would have been expected to show coupling
1
position. This tetraene, 1′-13C, shows in its H NMR spec-
trum a typical 13C-1H coupling constant of 156 Hz,13 and a
strong resonance at 110.5 ppm for the exo-methylene group in
its 13C NMR spectrum, which is otherwise identical to that of
the unlabeled sample. Two quaternary carbon atoms, at 144.6
ppm (C-2) and 141.3 ppm (C-9a), are coupled to the exo-
methylene-13C carbon atom by 72 and 7.9 Hz, respectively,
presumably as a consequence of one-bond and three-bond,
carbon-carbon couplings, respectively, two-bond coupling
usually being small.
1
with two adjacent carbon atoms. In the H NMR spectrum of
16′-13C, the expected large (120 Hz) coupling between 13C and
its two protons (2.02 and 2.04 ppm) remains unchanged on
homonuclear decoupling by irradiation of the proton at C-7 (2.65
ppm). This observation is consistent with the methylene-13C
On irradiation at 0 °C in toluene-d8 with a medium-pressure,
mercury lamp and benzophenone as sensitizer, unlabeled
(12) We are indebted to Dr. Keshab Sarma for the procedure for this
preparation.
(13) Breitmeyer, E.; Voelter, W. Carbon-13 NMR Spectroscopy; Verlag
Chemie: Weinheim, 1978; Chapter 3, Section 2.4.
(14) Hofman, H. J. Tetrahedron Lett. 1964, 2329-2331.

Isomerization of the [2 + 2] Photodimer of a Tetraene
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9155
Scheme 2
exclusive, product is a thermally unstable compound, 4a, which
rearranges slowly to 4b and, more slowly, to 4c even while its
NMR spectrum is being observed at -40 °C. Over timesthe
more rapidly, the higher the temperaturesall three stereoisomers
of 4 are replaced by the macrocycle 16 and, to a much smaller
extent, monomer 1.
Lacking appropriate experimental facilities, we have not been
able to separate, alone isolate, the three unstable components
of 4. Their presence (and relative concentrations) are inferred
1
by H NMR spectroscopy, while their structures as the three
possible cyclobutanes resulting from dimerization at the exo-
cyclic double bond are assigned by analogy with the photo-
chemistry of the two lower relatives of 1 of orders 2 and 3 (see
Scheme 1).4,10 An exciplex constructed from two parallel
monomers overlapping so as to minimize steric repulsion
between methyl groups might be expected to favor an anti-
exo,exo configuration, and lead predominantly to cyclobutane
4a (4-axx in Figure 2).
In the plot of concentration against time in Figure 3 (-29.4
°C), stereomutation of 4a to 4b is seen to be initially the fastest
reaction among the three cyclobutanes. As this interconversion
approaches equilibrium, 4b passes through a maximum as it
contributes increasingly to the slower generation of 4c and 16
(and to a much lesser extent 1). Thereafter, as pictured in a
similar diagram at -8.2 °C, the three cyclobutanes 4 appear to
reach a steady state in which all three contribute to the
rearrangement to 16. In this pseudoequilibrium, the concentra-
tions of 4a and 4c are essentially equal, while that of 4b is
favored by a factor of ∼2, incidentally the same as the statistical
factor. Note that the two methyl groups in all configurations of
4 are far apart (see the 3D representation of 4c (4-ann) in Figure
2) and that this qualitative impression is confirmed by the
essentially identical steric energies calculated for the three
stereoisomers by molecular mechanics (Figure 4).
The prospect of extracting the 12 specific rate constants
implied in Figure 2, and of obtaining their temperature depend-
ences with acceptable accuracy given only one of the compo-
nents of 4 as starting material, though possible in theory, cannot
be realized in practice. Nonetheless, the kinetics of the system
has been studied, encouraged in part by a question recently
brought into focus by the work of Doubleday (vide infra).
The kinetics of the system 4, 16, and 1 was examined
extensively over a range of temperature from -42.3 to -8.2
°C in 17 independent runs. Several not fully resolved diffi-
culties were encountered, including the poorly defined time
taken for a sample of 1, freshly irradiated at -75 °C, to reach
thermal equilibrium in the NMR spectrometer used for analy-
sis, inaccuracies in the measurement of temperature in the
NMR tube, and fluctuations in reaction temperature over time.
The set of data relating relative concentrations and time
(Table SI-1 of the Supporting Information) is given in part to
satisfy the intrepid reader who may wish to explore alternate
approaches to handling the data for this seriously underdeter-
mined system.15
groups not being attached to C-7, and is fully consistent with
the head-to-head structure (see Figure 1 for numbering and
collection of resonances).
In the 13C NMR spectrum of unpurified 16′-13C, four peaks,
at 36.82, 36.58, 36.14, and 35.90 ppm, of lower intensity
correspond to two methylene-13C carbon atoms, appearing as
two doublets at 36.70 and 36.02 ppm, each having a coupling
constant of 30 Hz. These peaks are probably associated with
contamination by minor dimers of unsymmetrical structure, that
is, having 30 spectroscopically distinguishable types of carbon
atoms. Although cycloreversion of 16′ to 1′ occurs at temper-
atures above 60 °C, preliminary kinetic studies revealed
substantial deviation from first-order behavior. Contamination
by one or more minor isomers differing in the rates of
fragmentation by a factor of ∼2 seemed to be the cause.
In the hope of circumventing this obstacle to generating
reliable activation parameters, attention was directed to optically
pure 1, the 10a-methyl analogue of 1′. By this device, which
had been used successfully in earlier work, the number of
possible, stereoisomeric products of photodimerization is re-
duced from the six for racemic material to three.4
Preparation of optically active 1 was accomplished by two
closely related routes (Scheme 2). Both started from the Revial-
Pfau compound (R)-5, readily available in optically pure form.4
One way around [(R)-5, -9, and -10] gives 1 in relatively low
yield, while the other [(R)-6, -7, and -8] gives a good yield.
Irradiation at 0 °C of the monomeric tetraene 1 generates a single
compound, 16, the physical properties of which are so similar
to those of 16′ that a 16-membered structure can be confidently
assumed.
Thermal decomposition of 16 to 1 follows a now strictly first-
order rate law. The kinetics has been studied without complica-
tion in the range 60-87 °C. Specific rate constants and
activation parameters are given in Table 1. The positive entropy
of activation, ∆Sq ) 11.7 eu, accords with a loosening of the
system as breaking a single bond generates a diradical inter-
mediate in the cycloreversion.
Treatment of the sum of the rearrangements of 4a, 4b, and
4c (∑4) to 16 as a first-order transformation leads to the specific
rate constants in Table 2, while a temperature dependence of
acceptable precision is given in Table 3. The simplification of
combining all stereoisomers of 4 into a single entity is easier
to accept at the higher temperatures than at the lower ones. At
the higher temperatures, a steady state among ∑4 appears to
The Four-Membered Ring Compounds
(15) Attempts to use a more realistic kinetic model including up to seven
differential equations were unproductive. The program did not lead to
convergence, and seemed to require negative rate constants in its efforts to
minimize the sum of differences.
Irradiation of the monomeric, optically pure tetraene 1 at -75
°C in toluene-d8 reveals a richer story. The major, perhaps

9156 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
VonDoering et al.
Table 1. Specific Rate Constants and Activation Parameters for the Cycloreversion of the [8 + 8] Dimer 16 to Two Molecules of 1 in
Benzene-d6
a,b
a,b
a,b
T, °C
k1
T, °C
k1
T, °C
k1
60.4 ( 0.9
60.4 ( 0.9
0.290 ( 0.012
0.296 ( 0.017
68.8 ( 0.3
68.8 ( 0.3
0.993 ( 0.022
0.994 ( 0.040
86.6 ( 0.1
86.6 ( 0.1
10.26 ( 0.29
10.66 ( 0.58
Arrhenius Parametersb: Ea ) 32.51 ( 0.44 kcal mol-1, log A ) 15.77 ( 0.28
Eyring Parametersb,c: ∆H ) 31.8 ( 0.4 kcal mol-1, ∆S ) +11.7 ( 1.3 cal mol-1 K-1
a k in units of 10-5 s-1
.
b Standard deviations are at the 95% confidence level. c Calculated at 346.7 K.
Figure 2. Kinetic scheme for the six interconversions relating 4a, 4b,
and 4c, their three ring enlargements to 16, and their three fragmenta-
tions each to two molecules of 1. A 3-D representation of 4c (4-ann),
the stereoisomer expected to show the greatest steric repulsion between
methyl groups, is shown at the bottom.
have been reached, but not at the lower temperatures (see Figure
3). The accuracy of the results is doubtless lower than the
precision.
Data relating to the interconversion of 4a and 4b are handled
as a system of reversible first-order reactions. Reasonable rate
constants are obtained if only the early points are used where
a linear approximation can be visually entertained (see the plot
of data at -29.4 °C in Figure 3). This model is an oversimpli-
fication owing to several factors: both 4a and 4b are rearranging
simultaneously to 16, 4c, and 1 in that order, but not necessarily
at equal rates, the model is applicable only over a limited range
of temperature, and a value for the unknown equilibrium
constant must be assumed. The statistical factor for 4b leads us
arbitrarily to take K ) 2.
The generation of 4c is sensibly slower, and is likewise
handled by using only early points where approximation of a
first-order interconversion of (4a + 4b) and 4c can be enter-
tained. A value of K ) 1/3 is assumed in the calculations.
Fragmentation to 1 is so much slower that its relative
concentrations are small and subject to large analytical errors.
The resulting low degree of precision is reflected in high
Figure 3. Fractions of 4a (diamonds), 4b (squares), 4c (triangles), 16
(times signs), and 1 (circles) (ordinate) against reaction time in seconds
(abscissa) at two temperatures.
uncertainties in the rate constants and derived activation
parameters. Recall that 1 can also be formed from 16 at higher
temperatures (Table 1). This mode has a calculated specific rate
constant at -8.2 °C of ∼1 × 10-11 s-1, and therefore makes a
negligible contribution to the formation of 1 at the lower
temperatures.
All worthy rate constants are collected in Table 2, while
derived activation parameters are collected in Table 3, and
depicted graphically in Figure 5.

Isomerization of the [2 + 2] Photodimer of a Tetraene
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9157
explained that values in column 4 are the difference between
values in columns 2 and 3, and reflect the superiority of MM2
EVBH for the calculation of elements of strain that elude the
gev method.
MM2 calculations of steric energy (Figure 4) have already
shown their usefulness in supporting the intuitive conclusion
that the methyl groups in 4 are too far apart to give rise to
significant thermochemical differences among the three stereo-
isomers. In contrast, thermochemical calculations indicate very
large unfavorable interactions of the methyl groups in the two
unseen configurational isomers of 16, and afford a rationalization
for the appearance of 16 in only one configuration.
The relatively much higher steric energies of the 8- and 12-
membered analogues of 4 and 16 not only render their
appearance highly unlikely (under thermodynamic control) but
also render improbable their involvement as nonisolable inter-
mediates in the formation of 16 from 4 in a succession of two
Cope rearrangements. If the enthalpy of activation for the Cope
Figure 4. Steric energies calculated by MM2 for four ring sizes of
the symmetrical, head-to-head, C30 dimers of tetraene 1 [(R,R) config-
uration, kcal mol-1].
rearrangement of cis-divinylcyclobutane, 23 kcal mol-1 9
, is
Table 2. Specific Rate Constants (s-1) from Kinetic Studies of the
System 1, 4, and 16 in Toluene-d8 (Figure 2)
taken as a lower estimate, addition of calculated incremental
strain energies of more than 25 kcal mol-1 in the 8- and 12-
ring compounds (Figure 4) places them out of reach of an
intramolecular rearrangement rapid at 0 °C. The situation stands
in contrast to the ring expansions so elegantly developed by
Wender and co-workers.20
4a
(4a + 4b) (4a + 4b + 4c) (4a + 4b + 4c)
T, °C 1/T, K
T 4ba,b
T 4ca,c
f 16
f 1
-42.3 0.004332 4.333E-5 6.949E-6
-42.3 0.004332 4.675E-5 8.919E-6
-35.3 0.004204 1.895E-4 2.735E-5
-35.3 0.004204 1.749E-4 2.630E-5
-35.3 0.004204 2.047E-4 2.487E-5
-29.4 0.004103 6.059E-4 1.194E-4
-27.2 0.004066 5.931E-4 9.228E-5
-27.2 0.004066 5.701E-4 1.242E-4
-25.1 0.004031 1.009E-3 1.747E-4
-24.9 0.004028 1.074E-3 1.854E-4
2.122E-6
1.220E-6
7.092E-6
6.990E-6
6.462E-6
1.602E-5
1.877E-5
2.119E-5
2.267E-5
3.472E-5
3.955E-5
4.294E-5
7.210E-5
1.023E-4
6.648E-5
1.396E-4
2.097E-4
A heat of formation of the hypothetical, bisheptatrienyl
intermediary diradical (Figure 4) can be estimated from the
(MM2EVBH) heat of formation of 1 by a straightforward
transference of the stabilization energy ascribed to the hep-
tatrienyl radical to the dimerization of 1. The difference between
the estimated heat of formation of 16 (∆fH° ) +35.8 kcal
mol-1) and that of the diradical (∆fH° ) +67.8 kcal mol-1
)
-22.1 0.003983
-22.1 0.003983
-16.3 0.003893
-11.7 0.003825
-11.7 0.003825
-11.7 0.003825
-8.2 0.003774
2.657E-4
3.587E-4
5.420E-4
5.485E-4
3.724E-7
6.068E-7
8.037E-7
3.331E-6
1.438E-6
2.357E-6
3.682E-6
then corresponds to an estimate of the enthalpy of activation
for fragmentation of 32.0 kcal mol-1. Agreement with the
experimental value of +31.7 kcal mol-1 is good. The unsatis-
factory agreement obtained by employing the values for the heat
of formation of 16 estimated by the gev method is owed to the
inability of this method to cope with the strain energy in the
16-membered ring.
a Rate constants calculated for the reversible, first-order model:
ln[(KAt + Bt)(/KAo + Bo)] ) (k-1 + k1)/t. b Assumed K ) 2. c Assumed
K ) 1/3.
In essence, this work has added a third, successful test of the
thesis that stabilization energies of polyenyl radicals evaluated
in one reaction can be transferred to a substantially different
system. Extension to a polyene of order 4 has led to a large
acceleration qualitatively in accord with prediction. The range
of temperature covering convenient rates of stereomutation
has decreased from 350 to 400 °C for the paradigm, through
70-110 °C for 1,2-divinyl, to -20 to 0 °C for 1,2-dibutadienyl,
and now to -40 to -20 °C for the 1,2-ditrienyl system.
Quantitative confirmation of the thesis has been less precise
than would have been wished for two reasons. The first applies
to the entire series. Predictions have been based on unsubstituted
models, whereas in practice more highly substituted, rigidly all-
trans examples have been studied to avoid numerous side
reactions. The resulting measured enthalpies of activation have
been 3-4 kcal mol-1 lower than predicted. These discrepancies
are reasonably ascribed to a small steric acceleration, a
proposition that finds support in the difference of 9.4 kcal mol-1
between the two heats of formation given for 4a in column 4
of Table 4. Roughly half (twice the discrepancy of 2.6 kcal
mol-1 for 1 in column 4) comes from an intrinsic strain in the
hydroanthracene system that is conserved on ring opening. The
Discussion: Thermochemical Considerations
Thermochemistry in systems such as this one provide
important arguments for and against hypothetical structures, and
facilitate predictions of the effect of perturbations on reactions
of paradigms. Heats of formation have been estimated by both
the method of group equivalent values (gev)16 and the molecular
mechanical program of Allinger17 as elaborated by W. R. Roth
(MM2EVBH). The former seeks to reproduce the available
thermochemical experimental data18 as generalized partial heats
of formation for simple groups, subject to several empirical
corrections of necessarily limited scope. The latter has been
optimized for hydrocarbons by parametrization to a body of
data enlarged by the inclusion of otherwise excluded heats of
hydrogenation, and by addition of the extended valence bond
approach of Malrieu and Maynau.19 Between the contents of
Table 4 and Scheme SI-1, the resulting thermochemical quanti-
ties and their derivation seem transparent. It needs only to be
(16) Franklin, J. L. Ind. Eng. Chem. 1949, 41, 1070-1076. Benson, S.
W. Thermochemical Kinetics, 2nd ed.; Wiley: New York, 1976.
(17) Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 8127-8134.
(18) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of
Organic Compounds, 2nd ed.; Chapman and Hall: London, 1986.
(19) Sa¨ıd, M.; Maynau, D.; Malrieu, J.-P.; Garcia Bach, M.-A. J. Am.
Chem. Soc. 1984, 106, 571-579, 580-587.
(20) Wender, P. A.; Sieburth, S. McN.; Petraitis, J.; Singh, S. K.
Tetrahedron 1981, 37, 3967-3975. Wender, P. A.; Ternansky, R. J.;
Sieburth, S. McN. Tetrahedron Lett. 1981, 37, 4319-4322.

9158 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
VonDoering et al.
∆Sq c,d
Table 3. Arrhenius Parameters from the Kinetics of the System 4, 16, and 1
reaction
4a f 4b
(4a + 4b) f 4c
(4a + 4b + 4c) f 16
(4a + 4b + 4c) f 1
16 f 1
Eaa,b
log A
∆Hq c
20.0 ( 1.9
18.7 ( 2.3
15.9 ( 1.5
19.8 ( 8.6
32.5 ( 0.4
14.60 ( 1.7
14.14 ( 2.1
9.39 ( 1.3
10.9 ( 7.2
15.8 ( 0.3
19.5 ( 1.9
18.2 ( 2.3
15.4 ( 1.5
19.3 ( 8.6
31.8 ( 0.4
+6.6 ( 8.0 (-1.1)
-2.4 ( 9.4 (-4.9)
-17.2 ( 5.9 (-8.5)
-10.3 ( 33.2 (-17.4)
+11.7 ( 1.3
a In kcal mol-1
.
b Standard deviations at the 95% confidence level. c Calculated at -25.6 °C. d Values in parentheses are based on the weighted
mean, ∆Hq ) 17.56 kcal mol-1
.
Entropic Control of Products?
In a series of exceptionally penetrating, theoretical explora-
tions of the role of diradical intermediates in the thermal
transformations of cyclopropane and cyclobutane, Doubleday
has reached a striking conclusion that choices among competing
pathways out of a “common diradical intermediate” (“caldera”)21
are under the control of entropy alone.22 In cyclobutane, more
apposite to the present work, these paths or “exit channels”
involve internal rotations about the C1-C2 and C2-C3 bonds
followed by either reclosure or fragmentation to ethylenes.23
His work brings a question into sharp focus: Are cyclobutanes
in general subject to control by entropy alone, or is this attribute
confined to the paradigmatic unsubstituted cyclobutane?
A theoretical answer to the question could come from
calculations on 1,2-dimethylcyclobutane. For an experimentally
based answer, careful studies of the temperature dependence in
prudently selected systems of not obviously concerted thermal
rearrangements of two, or preferably more, disparate exit
channels may be helpful. In principle, the present example
appears to be ideal. It has three dissimilar exit channels:
stereomutation, a novel ring enlargement, and fragmentation.
In practice, obtaining Arrhenius parameters for the 12 specific
rate constants at a high enough accuracy to elucidate the balance
between control by entropy and enthalpy is problematical, if
not fantastical.
Figure 5. ln k (ordinate) versus 1/T (K) (abscissa) for the reactions of
4a to 4b (diamonds), (4a + 4b) to 4c (squares), (4a + 4b + 4c) to 16
(triangles), and (4a + 4b + 4c) to 1 (times signs) (see Table 3).
Table 4. Thermochemistry of 1, 4a, 16, and the Diradical in
Common
At this point, we can do little more than explore the
consequences of accepting the generality of Doubleday’s
conclusion that all three processes would have had in common
a single enthalpy of activation for entrance into the caldera,
and that the weighted mean of 17.6 kcal mol-1 would be its
value! The corresponding set of controlling entropies of
activation then becomes calculable (given in Table 3, last
column, in parentheses). Those relating to the interconversions
among 4 are small, and consistent with those found for the two
lower homologues. That for rearrangement to the conforma-
tionally more rigid 16 issnot unreasonablysmore negative.
That for fragmentation to 1, which might intuitively have been
expected to be positive, is provocatively negative. The unex-
pectedly slow rate of cleavage of 4 to 1 appears to fall in line
with a positive correlation between increasing stabilization in
the diradical intermediates and ratios of stereomutation to
fragmentation, kst/kfr. In the sequence 1,2-dideuteriocyclobu-
tane,24 2, 2(Ph), 3, and 4 (Scheme 1), these ratios, although of
varying accuracy and precision, are, respectively, ∼0.3 (380
a,b
a,c
a,d
compound
∆fH(Roth)
∆fH(Benson)
∆∆fH(strain)
1
4a
16
diradical
27.6
50.7
35.8
67.8f
25.0
41.3e
15.1
62.6f
2.6
9.4
20.7
5.2
a In kcal mol-1 b Heats of formation calculated by the molecular
.
mechanical program MM2EVBH. c Heats of formation estimated by
the group equivalent values of Benson. d The difference between values
in columns 2 and 3 is identified with the energy of strain unable to be
included in the Benson value. e This value includes cyclobutane strain
of 26.2 kcal mol-1 f Estimated by the addition to twice the heat of
.
formation of 1 of 43.5 kcal mol-1 (2C2H4 f -(CH2)4-), 2K (conjugative
interaction 3.75 kcal mol-1), and 2(SE3) (-19.2 kcal mol-1, the
stabilization energy of a heptatrienyl radical).
remainder of 4.2 kcal mol-1 arises from the tetrasubstituted
nature of the cyclobutane ring, and is available sterically to
accelerate the opening of the ring to the diradical.
The second reason applies to the current example in which
experimental difficulties arising from the inconveniently low
temperatures required for kinetic studies have been incompletely
resolved. The four enthalpies of activation have fallen in the
range of 16-20 kcal mol-1 in satisfactory agreement with the
predicted value of 22.4 kcal mol-1, when decreased as noted
(21) A sobriquet to encompass “common diradical intermediate”,
“continuous diradical”, and “twixtyl”.
(22) (a) Doubleday, C., Jr.; Camp, R. N.; King, H. F.; Page, M.; McIver,
J. W., Jr.; Mullally, D. J. Am. Chem. Soc. 1984, 106, 447-448. (b)
Doubleday, C., Jr.; Page, M.; McIver, J. W., Jr. J. Mol. Struct. (THEOCHEM)
1988, 163, 331-341. (c) Doubleday, C., Jr. J. Am. Chem. Soc. 1993, 115,
11968-11983. (d) Doubleday, C., Jr. Chem. Phys. Lett. 1995, 233, 509-
513. (e) Doubleday, C., Jr. J. Phys. Chem. 1996, 100, 15083-15086.
(23) Unobserved pathways such as rearrangement to butene, or frag-
mentation to butadiene and dihydrogen, which we guess are under enthalpic
control, are not included in the survey.
above by the steric correction of 3-4 kcal mol-1 4,10
The
.
weighted mean of these activation enthalpies, 17.6 kcal mol-1
agrees surprisingly well with the difference, 17.1 kcal mol-1
,
,
between the estimated heat of formation of the diradical (+67.8
kcal mol-1) and the heat of formation of 4a (50.7 kcal mol-1
)
(MM2EVBH values in Table 4, column 2).
(24) Chikos, J. S. J. Org. Chem. 1979, 44, 780-784.

Isomerization of the [2 + 2] Photodimer of a Tetraene
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9159
1′ is unstable at room temperature, gradually being converted to a
yellow polymer, which can, however, be removed by flash chroma-
tography with hexane on silica gel, or basic alumina.
°C), ∼0.5 (100 °C), ∼6 (43.6 °C), ∼25 (43.6 °C), and ∼4 ×
103 (-25.1 °C).
In terms of entropy alone as the controlling factor, we might
then speculate that the number of simultaneous adjustments in
bond distances required to transit from polyenic cyclobutane
to polyenyl radical increases with order. A convincing answer
to the question of where along the continuum from complete
control by enthalpy to complete control by entropy the
transformations in the present system lie will require greater
accuracy than we have been able to provide.
The facile thermal rearrangement of 4 to 16 appears to be a
novel ring enlargement, but is in some respects quite similar to
the dimerizations developed by Glatzhofer and Longone in the
paracyclophane series.25 An important difference lies in the
method of generation. In contrast to the purely aliphatic polyenes
of our series, their polyenes dimerize thermally without being
obliged to pass over a cyclobutane stage, presumably owing to
the extra stabilization afforded to the diradical stage by the
development of aromatic character.
Whether the small, extra degree of stabilization in the next
homologue in our series, a pentaene, would be enough to
facilitate thermal dimerization is unpredictable. That a 20-
membered ring should be formed by way of the cyclobutane
seems highly likely. The calculated difference in steric energy
between the 20-membered and related 16-membered ring is -38
kcal mol-1, more than enough to favor the former under
thermochemical control, even without the extra fillip of two
more units of conjugative interaction.
1′-13C. (b) A 5 mL sample of a 2.5 M solution of n-butyllithium
(n-BuLi) (Aldrich) in hexane was added slowly over a 5 min period to
a cooled (ice-water bath) suspension of methyl-13C-triphenylphos-
phonium iodide (13C, 99.4%, 2.00 g, 5.00 mmol) in 15 mL of
tetrahydrofuran (THF). To the resulting solution was then slowly added
ketone 8′ (0.51 g, 2.5 mmol) in 3 mL of THF. After being stirred for
3 h as the temperature rose to ambient, the reaction mixture was treated
with 0.5 mL of water, evaporated to dryness, and triturated with 5 mL
of hexane. The hexane solution was passed through a basic alumina
column (2.5 cm i.d. × 20 cm, hexane as eluent), and concentrated to
yield 0.31 g (62%) of 4,4a,5,6,10,10a-hexahydro-2(3H)-methylene-13C-
1
anthracene (1′-13C) as colorless crystals: mp 110 °C; H NMR 6.09
(d, 1H, J ) 7.7 Hz, H-8), 5.96 (s, 1H, H-9), 5.84 (m, 2H, H-7 and
H-1), 4.79 (d, 1H, J13C,H ) 156 Hz, H-2a), 4.73 (d, 1H, J13C,H ) 157
Hz, H-2a), 2.48-2.02 (m, 6H, H-3,3′,4a,10a,6,6′), 1.92-1.79 (m, 3H,
H-4′,10′,5′), 1.36-1.22 (m, 2H, H-4,5), 1.08 (q, 1H, J ) 12.4 Hz,
H-10); 13C NMR 144.4 (q, d, J ) 71.7 Hz), 141.2 (q, d, J ) 7.8 Hz),
140.5 (q), 130.4 (t), 129.0 (t), 125.7 (t), 125.3 (t), 110.1 (s), 37.9 (s),
35.8 (t), 35.6 (t), 30.7 (s), 30.6 (s), 30.3 (s), 26.0 (s); MS m/z ) 199
(M+, 100%), 184 (- 13CH2, 8%), 143 (- 13CH2/(CH3)2CH, 28%).
Photodimerization of 4,4a,5,6,10,10a-Hexahydro-2(3H)-methyl-
eneanthracene (1′). (a) A solution of 440 mg of freshly chromato-
graphed tetraene 1′ in 6 mL of benzene was sealed in a Pyrex ampule
after three freeze-pump-thaw degassing cycles. The ampule was
irradiated in an ice bath with a medium-pressure mercury lamp for 7
h. Filtration through a 25-50 microfilter funnel yielded 16′ as a pale
yellow precipitate (46.8 mg, 10%). The benzene solution was flash
chromatographed on silica gel (25 cm × 2.5 cm i.d., ether/hexane
(1:3)) to give 94.5 mg of a mixture of unreacted 1′, its dimer 16′, and
180 mg of polymer, removable by washing the column with CHCl3.
Experimental Section
1
General Methods. H NMR and 13C NMR spectra were measured
(b) A solution of 80.0 mg (0.40 mmol) of freshly chromatographed
1′-13C, 2.0 mg of benzophenone, and 2.0 mL of benzene was irradiated
as in (a). The content of the ampule was concentrated to a residue,
which was triturated with 10 mL of ether, leaving a pale yellow
precipitate. Filtration and washing with ether provided 66.2 mg of an
impure 16′-13C (82.7% yield): 1H NMR 5.40 (m, 4H), 5.23 (m, 2H),
2.56 (m, 2H), 1.0-2.1 (m, 24H); 13C NMR 146.3 (q), 139.7 (q), 138.3
(q, d, J ) 40 Hz), 127.7 (t), 122.3 (t), 121.2 (t), 44.6 (t), 37.8 (t),
36.91 (t), 36.88 (s), 35.4 (s), 30.3 (s), 30.0 (s), 28.3 (s), 27.0 (s); UV-
vis 258 (4.35), 278 (4.40); MS m/z ) 398 (M+, 59%), 199 (1/2M+,
100%).
on a Bruker AM-500 (1H, 500 MHz; 13C, 125.8 MHz) instrument in
CDCl3, unless otherwise noted, and are reported in parts per million
from tetramethylsilane (δ) [CDCl3 (7.24 ppm), benzene-d6 (7.14 ppm)].
Differentiation among the types of carbon atoms in the 13C NMR is
based on the DEPT technique [q, quaternary; t, tertiary; s, secondary;
p, primary].13 Spin-lattice relaxation times (T1) were determined by
the inversion-recovery method on vacuum-sealed solutions in toluene-
d8. IR spectra, reported in cm-1, were recorded on a Nicolet Impact
400D FT-IR instrument. Samples, if liquid, were measured as a thin
film on a NaCl plate and, if solid, as a thin layer prepared by evaporation
of a solution in CHCl3 on a NaCl plate. UV-vis spectra were measured
in spectrograde hexane on a Hewlett-Packard 8452A diode array
spectrophotometer or a Varian Cary 1E UV-vis spectrophotometer
[λmax, nm (log ꢀ)]. High-resolution mass spectra were measured on a
JEOL AX 505 spectometer equipped with a data recovery system.
Specific rotations were measured on a Perkin-Elmer Polarimeter 241;
enantiomeric excess (ee) was determined with a Hewlett-Packard 5890
Series II gas chromatograph employing a Chiraldex G-TA column (20
m × 0.25 mm i.d. × 0.125 mm film, Advanced Separation Technology,
Inc.). Melting points (mp’s) were taken with a Thiele tube, and are not
corrected. A 450 W, medium-pressure HANOVIA mercury lamp, model
679A36, or a GE 275 W sun lamp was used for photochemical
reactions.
(R)-4,4aâ,5,6,10,10a-Hexahydro-10aâ-methyl-2(3H)-anthra-
cenone [(R)-8]. (a) In an initial experiment, the starting material was
(R,S)-4,4a,5,6,7,8-hexahydro-4aâ-methyl-2(3H)-naphthalenone [(R,S)-
5], prepared according to a previously described procedure:4 To a
solution of its lithium enolate, prepared from 1.0 g of (R,S)-5 in 10
mL of THF and lithium diisopropylamide (LDA) (from 4.8 mL of 2.5
M n-BuLi (2 equiv) and 2.52 mL of diisopropylamine), was added
dropwise at -78 °C under a stream of N2 over a 10 min period a
solution of 0.5 mL (1 equiv) of methyl vinyl ketone (MVK) in 12 mL
of anhydrous THF. After 20 min of stirring, the solution was quenched
with 2.0 mL of 10% aqueous HCl, warmed to room temperature, and
extracted with 10 mL of ether. Washed four times with 5 mL each of
H2O and dried over Mg2SO4, the ether solution was concentrated to a
residue, distillation of which at 120 °C and 0.5 mmHg (Kugelrohr)
afforded 0.28 g of unreacted (R,S)-5 and, at 200 °C, 0.60 g of (R,S)-
4,4aâ,5,6,7,8-hexahydro-3-[4′-(2′-ketobutyl)]-4aâ-methyl-2(3H)-naph-
thalenone [(R,S)-9] (36%): 1H NMR (300 MHz) 5.67 (s, 1H), 2.60
(m, 1H), 2.45 (m, 2H), 2.25 (m, 1H), 2.15 (s, 3H), 2.10 (m, 1H), 1.90
(m, 1H), 1.80-1.55 (m, 6H), 1.30 (m, 2H), 1.25 (s, 3H). The use of 2
equiv of MVK and a 65 min reaction time gave more byproducts of
condensation with two and three molecules of MKV.
4,4a,5,6,10,10a-Hexahydro-2(3H)-methyleneanthracene (1′). (a)
This tetraene was prepared from its precursor ketone 8′ (Scheme 1) by
a Wittig reaction2 developed by Keshab Sarma (see (b) below for
details).12 The yield depends on the time of reaction following addition
of the ketone at 0 °C to the ylide solution, and the time of stirring
thereafter at room temperature. Duplicate runs for 1 h at 0 °C and 2 h
at room temperature afford yields of 47% and 53% of theory,
respectively, while runs for 3 and 17 h at room temperature lead to
lower yields of 32% and 27%, respectively. After chromatography on
silica gel, colorless crystals of 1′ were obtained: mp 105-106 °C;
UV-vis 294 (4.24), 306 (4.41), 322 (4.38).
A similar procedure was applied to (R)-5 ([R]20D -229° (ethanol, c
1.1); ee > 99%; 1H NMR 5.72 (s, 1H, vinyl H), 2.50-1.34 (m, 12H),
1.23 (s, 3H, CH3); IR 2927, 2859, 1677, 1627, 1448, 1326, 1225, 1186,
858). A 5 g sample of (R)-5, 1.7 equiv of LDA, and 0.8 equiv of MVK
(70 min reaction time) gave (R)-9 in 31% theoretical yield. For the
(25) Glatzhofer, D. T.; Longone, D. T. Tetrahedron Lett. 1983, 24, 4413-
4416.

9160 J. Am. Chem. Soc., Vol. 123, No. 37, 2001
VonDoering et al.
next step, a procedure of Revial and Pfau was followed.26 A solution
of 2.54 g of (R)-9, 10.0 mL of methanol, and 15 drops of 25% NaOH/
MeOH, heated at 60 °C for 18 h, gave 2.42 g of impure (R)-10 as a
yellow oil, which was recrystallized from 5.0 mL of ether at 4 °C to
give 0.9 g (second crop 0.4 g) of yellow crystals of (R)-(-)-
short column of basic alumina (5 g), and concentrated in vacuo.
Chromatography on basic alumina (pentane) afforded crude tetraene,
which was recrystallized twice from pentane at -40 to -50 °C to give
165 mg (84%) of colorless needles: mp 83.0-83.5 °C; [R]20D -87.8°
1
(hexane, c 0.109); H NMR (toluene-d8) 6.05 (dd, 2H, J ) 9.9, 2.5
4,4aâ,5,6,7,8,10,10a-octahydro-10aâ-methyl-2(3H)-anthracenone [(R)-
Hz), 5.99 (s, 1H), 5.76 (s, 1H), 5.67-5.63 (m, 1H), 4.90 (s, 1H), 4.81
(s, 1H), 2.49-2.38 (m, 2H), 2.46-2.37 (m, 1H), 2.21-2.13 (m, 1H),
1.98-1.92 (m, 2H), 1.63-1.58 (m, 1H), 1.34-1.22 (m, 4H), 1.07 (t,
1H, J ) 12.7 Hz), 1.00 (s, 3H); IR 2926, 2849, 1607, 1579, 1448,
919, 870, 793, 645; UV-vis 323 (4.647), 308 (4.679), 294 (4.486).
In CDCl3, the 1H NMR spectrum reveals a mixture of 1 and its acid-
catalyzed rearrangement product, (R)-(-)-4,4aâ,5,6,10,10a-hexahydro-
2,10aâ-dimethylanthracene: 1H NMR (CDCl3, 300 MHz) 5.95 (m, 1H,
H-3), 5.86 (s, 1H, H-1), 5.25 (s, 1H, H-9), 5.65 (m, 2H, H-7, H-8),
2.45 (m, 1H), 2.20 (m, 3H), 1.80 (s, 3H, CH3-2), 1.67 (dd, 1H), 1.41-
1.10 (m, 4H), 1.02 (s, 3H, CH3-10a). Compound 1 not only is very
sensitive to acid but also rapidly turns yellow in the presence of air. It
can be stored for long periods of time at -85 °C.
Photodimerization of 1. A solution of 10 mg of 1 in 0.5 mL of
toluene-d8 was sealed in an NMR tube after three freeze-pump-thaw
cycles. The tube was irradiated for 2 h at -75 °C by a Hanovia medium-
pressure mercury arc lamp, and stored at -78 °C. The main product
4a was revealed by a 1H NMR (toluene-d8) spectrum taken at -42 °C:
6.16 (dd, 2H, J ) 9.9 and 2.0 Hz), 5.86 (s, 2H, 5.68 (m, 2H), 5.62 (s,
2H), 2.42-2.33 (m, 2H), 2.33-2.27 (m, 2H), 2.05-1.80 (m, 8H), 1.52-
1.47 (m, 2H), 1.42-1.24 (m, 8H), 1.11 (t, 2H, J ) 12.8 Hz), 1.05 (s,
6H).
After 15 min at room temperature, the 1H NMR spectrum indicated
a new product (16) and regenerated 1 in a ratio of 9:1. A similar
irradiation carried out at 0 °C for 2 h afforded only 16: 1H NMR
(toluene-d8) 5.46 (d, 2H, J ) 2 Hz), 5.29 (d, 2H, J ) 5.4 Hz), 5.22 (d,
2H, J ) 2.7 Hz), 2.39-2.30 (m, 2H), 2.26-2.24 (t, 2H, J ) 5.0 Hz),
2.14-1.90 (m, 8H), 1.78 (ddd, 2H, J ) 16.4, 5.1, and 1.9 Hz), 1.68-
1.53 (m, 8H), 1.17 (ddd, 2H, J ) 25.2, 12.6, and 5.1 Hz), 1.12 (s, 6H),
1.04 (t, 2H, J ) 11.9 Hz); MS m/z calcd for C32H40 424.3130 found
424.3112 (M+, base).
A sealed glass tube containing 1 without oxygen was heated at its
melting point (about 85 °C) for 10 min. A solution of the resulting
solid in toluene-d8 showed a trace of 16 by NMR. When 1 without
solvent was heated at 110 °C for 16 h, 2-3% 16 was detected. Heating
a 10% solution of 16 in toluene-d8 in a sealed NMR tube at 110 °C for
16 h gave identical results. Irradiation of a thin layer of solid 1 on the
glass wall of a sealed NMR tube at 0 °C for 2 h also gave 2-3% 16
along with a few percent of unidentified products.
1
10]: mp 105-106 °C; [R]25 -575° (methanol, c 1.967); H NMR
D
(400 MHz) 5.90 (s, 1H, H-9), 5.73 (s, 1H, H-1), 2.52 (m, 1H, H-3),
2.50 (m, 1H, H-4a), 2.41 (s, 1H, H-3′), 2.30 (m, 1H, H-8), 2.22 (1H,
H-8′), 2.00 (m, 1H, H-4), 1.88 (m, 1H, H-4′), 1.63 (m, 5H), 1.30 (m,
3H), 1.01 (s, 3H, CH3); 13C NMR (126 MHz) 200.1, 159.3, 158.6, 122.5,
122.3, 46.0, 42.0, 38.0, 36.6, 32.7, 31.8, 30.2, 27.7, 23.2, 21.9.
A solution of 1.16 g of (R)-10, 2.65 g of chloranil, and 5.9 mg of
p-toluenesulfonic acid in 35 mL of tert-butyl alcohol was boiled under
reflux for 15 h. After conventional workup, recrystallization of the
resulting brown oil from ether afforded 0.25 g (two crops) of (R)-8:
mp 99-100 °C; [R]25 -771° (methanol, c 0.744).
D
(b) In an alternate procedure, (R)-5 was first dehydrogenated by
chloranil to (R)-(-)-4,4a,5,6-tetrahydro-4aâ-methyl-2(3H)-naphthale-
none [(R)-6] following a procedure of Banerjee et al.,4,27 modified by
further purification of the oily product by two recrystallizations from
1
pentane (-10 to -20 °C): 99% purity; ee > 99% (GLC); H NMR
6.21 (m, 1H), 6.13 (dd, 1H, J ) 9.3, 0.6 Hz), 5.65 (s, 1H), 2.66-2.58
(m, 1H), 2.43-2.35 (m, 2H), 2.30-2.24 (m, 1H), 1.92-1.82 (m, 1H),
1.79-1.74 (m, 1H), 1.62-1.50 (m, 2H), 1.16 (s, 3H); IR 3027, 2967,
2920, 1660, 1619, 1586, 1253, 1211, 872, 622.
A Michael addition of (R)-6 to methyl vinyl ketone then gave the
diketone (R)-7, (R)-(-)-4,4a,5,6-tetrahydro-3-[4′-(2′-ketobutyl)]-4aâ-
methyl-2(3H)-naphthalenone. A solution of 5.0 g (30.8 mmol) of (R)-6
in 10 mL of anhydrous THF was added dropwise to a solution of LDA
(from 19.25 mL of a 1.6 M solution of n-BuLi and 3.43 g of
diisopropylamine) at -78 °C under nitrogen. To this solution, after
being stirred for 10 min, was added at -78 °C over a 10-min period
a solution of 2.15 g (30.7 mmol) of freshly distilled MVK in 10 mL of
anhydrous THF. After 35 min, the reaction mixture was quenched with
10 mL of 5% aqueous NaHCO3, and extracted with ether (3 × 50 mL).
The combined organic layers were washed with water and saturated
brine, dried over Na2SO4, and concentrated in vacuo to a yellow oil,
which was chromatographed on silica gel (elution with hexane/ethyl
acetate (10:1 and 4:1)) to give recovered (R)-6 (1.69 g, 34%) and
diketone (R)-7 (3.31 g, 66%): 1H NMR 6.20 (m, 1H), 6.12 (dd, 1H, J
) 9.7, 2.5 Hz), 5.62 (s, 1H), 2.65-2.53 (m, 3H), 2.43-2.34 (m, 1H),
2.30-2.21 (m, 1H), 2.14 (s, 3H), 2.09-2.00 (m, 1H), 1.78-1.48 (m,
5H), 1.13 (s, 3H); IR 2968, 2920, 1713, 1660, 1621, 1361, 1225, 1204,
1179, 880, 622.
Kinetics of Cycloreversion of 16 to 1. The rates of fragmentation
of 16 to 1 as a function of temperature were determined in the manner
described previously.28 Samples of 16 in toluene-d8 were prepared from
1 by irradiation in sealed NMR tubes at room temperature, and then
heated in the vapors of solvents of appropriate boiling points.
Kinetics of Stereomutation and Cycloreversion of 4a. A solution
of ∼10 mg of 1 in 0.5 mL of toluene-d8 in an NMR tube was degassed
by three freeze-pump-thaw cycles and sealed under vacuum (10-5
mmHg). The tube was irradiated with a medium-pressure Hg lamp for
2 h at -75 °C (cooled with dry ice/isopropyl alcohol) and then stored
at -85 °C. Kinetic measurements at low temperature in the range of
-42.3 to -7.8 °C were conducted in a Bruker AM 500N instrument
equipped with a variable-temperature unit. After the temperature was
set, about 30 min was allowed for equilibrium to be reached. The sample
was then moved from the dry ice/isopropyl alcohol bath to the NMR
probe, 10 min being allowed for returning to equilibrium before the
measurements were begun. Because 1 precipitates below -45 °C, the
lowest temperature employed for kinetic studies was -42.3 °C.
The temperature unit of the NMR spectrophotometer was calibrated
by a digital thermometer with a J-type thermocouple placed inside the
NMR probe. A further calibration was obtained by measurements at
several temperatures of the NMR spectrum of methanol in a sealed
capillary placed in the NMR tube. The temperature was monitored
periodically during all runs by such an internal capillary tube.
A stirred methanol (50 mL) solution of (R)-7 was purged of oxygen
by a nitrogen stream for 15 min before the addition of 50 mg of sodium
methoxide in 5 mL of methanol. After being stirred at 60 °C under
nitrogen for 16 h, the cooled solution was quenched with acetic acid
to pH 2-3. Removal of methanol in vacuo gave a residue which was
dissolved in 100 mL of CHCl3, washed with aqueous NaHCO3 (2 ×
20 mL), water, and brine, then dried over Na2SO4, and chromatographed
on silica gel (hexane/ethyl acetate (6:1)) to yield (R)-(-)-4,4aâ,5,6,-
10,10a-hexahydro-10aâ-methyl-2(3H)-anthracenone [(R)-8] (2.81 g,
92%). Recystallization from ethyl acetate afforded a slightly yellow
material (2.30 g, 75%): mp 105 °C; [R]20 -83° (ethanol, c 0.109);
D
1H NMR (toluene-d8) 6.11 (dd, 2H, J ) 9.9, 2.5 Hz), 6.07-6.03 (m,
1H), 5.92 (s, 1H), 5.79 (s, 1H), 2.86-2.78 (m, 2H), 2.56-2.51 (m,
1H), 2.49-2.32 (m, 2H), 2.26-2.21 (m, 2H), 2.06-2.02 (m, 1H), 1.78-
1.55 (m, 4H), 1.49-1.46 (m, 1H), 1.35 (t, 1H, J ) 12.9 Hz), 1.13 (s,
3H); IR 2920, 2852, 1647, 1578, 1561, 1323, 1201, 907, 651; UV-
vis 327 (4.56); MS m/z calcd for C15H18O 214.1358, found 214.1335
(M+, base).
(R)-(-)-4,4aâ,5,6,10,10a-Hexahydro-10aâ-methyl-2(3H)-meth-
yleneanthracene (1). This compound was prepared by a Wittig reaction
as described above. From 200 mg of trienone (R)-8 and 680 mg of
methyltriphenylphosphonium bromide, a reaction mixture (THF) was
diluted with pentane, freed of a suspension by being passed through a
(26) Revial, G.; Pfau, M. Org. Synth. 1991, 70, 35-45.
(27) Banerjee, D. K.; Angadi, V. B. J. Org. Chem. 1961, 26, 2988-
2989.
(28) Doering, W. von E.; Mastrocola, A. R. Tetrahedron 1981, 37 (Suppl.
1), 329-344.

Isomerization of the [2 + 2] Photodimer of a Tetraene
J. Am. Chem. Soc., Vol. 123, No. 37, 2001 9161
Quantitative 1H NMR spectra were measured allowing an interpulse
time delay (relaxation delay and saturation period of at least 5 times
the longest spin-lattice relaxation time (T1) at each temperature). The
signals used for analysis follow (in parts per million): 4a, 5.68; 4b,
5.91; 4c, 5.94; 16, 5.22; 1, 4.90.
edged. We thank the Norman Fund in Organic Chemistry,
Harvard University, in memory of Ruth Alice Norman Weil
Halsband, for its partial support of this work.
Supporting Information Available: Untreated data from
17 kinetic runs between -42.3 and -7.8 °C for the system 4a,
4b, 4c, 16, and 1. This material is available free of charge via
the Internet at http://pubs.acs.org.
Acknowledgment. We are indebted to the late Professor
Wolfgang Roth for all the calculations employing his MMEVBH
program. Partial support by the National Science Foundation,
Grants CHE-88 16186 and 91 23207, is gratefully acknowl-
JA011135N