Chemistry of syn-o,o′-Dibenzene

Article abstract of DOI:10.1021/ja0023579

A new and improved synthesis cis,syn-o,o′-dibenzene 1 was developed to obtain 1 in larger amounts with improved purity. syn-Dibenzene 1 undergoes thermolysis to two molecules of benzene at a rate slower than that of the thermodynamically more stable anti-dibenzene 2. Kinetic analysis revealed that the higher thermal stability of 1 is due to the higher heat of activation in thermolysis. Photoelectron spectroscopy of 1 showed that the through-bond interaction between the two cyclohexadiene units in o,o′-dibenzenes is more important than their through-space interaction. A comparative study on the thermolyses of related syn-o,o′-arene:benzene dimers suggests that thermolyses of syn-o,o′-arene:benzene dimers proceed via their anti-isomers as an intermediate, syn-Dibenzene 1 also undergoes adiabatic photolysis to one molecule of excited benzene and one molecule of ground-state benzene in good efficiency. The mechanisms of these reactions are discussed.

Full text of DOI:10.1021/ja0023579

12098
J. Am. Chem. Soc. 2000, 122, 12098-12111
Chemistry of syn-o,o′-Dibenzene
Hong Gan,† M. Glenn Horner,† Bruce J. Hrnjez,† Thomas A. McCormack,† John L. King,†
Zbigniew Gasyna,† Grace Chen,§ Rolf Gleiter,*,§ and Nien-chu C. Yang*,†
Contribution from the Department of Chemistry, UniVersity of Chicago, Illinois 60637, and
Organisches-Chemisches Institut der UniVersita¨t, Heidelberg, D-69120 Heidelberg, Germany
ReceiVed June 29, 2000. ReVised Manuscript ReceiVed October 4, 2000
Abstract: A new and improved synthesis cis,syn-o,o′-dibenzene 1 was developed to obtain 1 in larger amounts
with improved purity. syn-Dibenzene 1 undergoes thermolysis to two molecules of benzene at a rate slower
than that of the thermodynamically more stable anti-dibenzene 2. Kinetic analysis revealed that the higher
thermal stability of 1 is due to the higher heat of activation in thermolysis. Photoelectron spectroscopy of 1
showed that the through-bond interaction between the two cyclohexadiene units in o,o′-dibenzenes is more
important than their through-space interaction. A comparative study on the thermolyses of related syn-o,o′-
arene:benzene dimers suggests that thermolyses of syn-o,o′-arene:benzene dimers proceed via their anti-isomers
as an intermediate. syn-Dibenzene 1 also undergoes adiabatic photolysis to one molecule of excited benzene
and one molecule of ground-state benzene in good efficiency. The mechanisms of these reactions are discussed.
Introduction
Due to the high internal energy of benzene cyclodimers, these
compounds are expected to possess unusual chemical and
physical properties of both experimental and theoretical sig-
nificance. The two o,o′-dimers, syn-o,o′-dibenzene 1 and anti-
o,o′-dibenzene 2, differ from each other only in their topology
and are of particular interest. Both compounds have now been
synthesized: anti-o,o′-dibenzene 2 by several different groups
since the 1960s1 and syn-o,o′-dibenzene 1 in our laboratory in
1987.2 Although 2 is now available in gram quantities,1a our
original synthesis of 1 only yielded the target compound in
minute quantities and of doubtful purity,2 which limited the
scope of our investigations. In this work, we wish to report an
improved synthesis of 1. The availability of both 1 and 2 enables
a parallel study of their highly exoergic thermolyses, formally
symmetry-forbidden 4n-retro-cycloadditions to two molecules
of benzene as governed by the Woodward-Hoffmann Rules.3
Because of their unique topology, dibenzenes 1 and 2 may be
an ideal system for the comparative study of the relative
importance of intramolecular through-space (π: π) and through-
bond (σ: π) interaction between two 1,3-cyclohexadiene systems.
Such a study may delineate these two types of interactions on
a quantitative basis. The synthesis of the p,p′-dibenzene 3
remains elusive at this moment.4 Recently, there is a renewed
interest in these systems, particularly on the effect of electron
transfers.5
Although the analogous syn- and anti-[2π + 2π] cyclobuta-
diene dimer systems, tricyclo[4.2.0.0]octa-2,6-diene 4 and 5,
have been prepared by the thermal dimerization of their
monomers,6 the same type of reaction cannot be applied to give
the dimers of benzene. Both the chemistry and the spectroscopy
of dienes 4 and 5 have been studied extensively. By analyzing
their photoelectron (PE) spectra, one of us (R.G.) has shown
that their through-bond interactions are much stronger than their
through-space interactions. As a result, the symmetric π-orbital
lies at higher energy than the antisymmetric π-orbital.7 This
interaction accounts for the observation that syn-octamethyltri-
cyclo[4.2.0.02,5]octa-3,7-diene and related systems yield only
to a very minor extent to the corresponding cubanes.8 The
importance of through-bond interaction has also been observed
in anti-dibenzene 2. The energy difference between the sym-
metric and antisymmetric π-orbitals amounts to 0.67 eV.9,10
While both the chemistry and the PE spectrum of a rigid cage-
* To whom the correspondence of this manuscript should be addressed.
† The University of Chicago.
§ Universita¨t Heidelberg.
(1) (a) Noh, T.; Gan, H.; Halfon, S.; Hrnjez, B. J.; Yang, N. C. J. Am.
Chem. Soc. 1997, 119, 7470-7482. (b) Berson, J. A.; Davis, R. F. J. Am.
Chem. Soc. 1972, 94, 3658-3659. (c) Schro¨der, G.; Martin, W.; Ro¨ttele,
H. Angew. Chem., Int. Ed. Engl. 1969, 8, 69-70. (d) Oth, J. F. M.; Ro¨ttele,
H.; Schro¨der, G. Tetrahedron Lett. 1970, 61-66. (e) Braun, D.; Grimme,
W.; Heinze, U. Unpublished results.
(2) Yang, N. C.; Hrnjez, B. J.; Horner, M. G. J. Am. Chem. Soc. 1987,
109, 3158-3159.
(3) Woodward, R. B.; Hoffmann, R. The ConserVation of Orbital
Symmetry; Academic Press: New York, 1970.
(5) Reddy, G. D.; Wiest, O.; Hurdlicky, T.; Schapiro, V.; Gonzales, D.
J. Org. Chem. 1999, 64, 2860-2863.
(6) Paquette, L. A.; Carmody, M. J. J. Am. Chem. Soc. 1976, 98, 8175-
8181.
(7) Gleiter, R.; Scha¨fer, W. Acc. Chem. Res. 1990, 23, 369-375.
(8) Gleiter, R.; Brand, S. Chem. Eur. J. 1998, 2532. Tetrahedron Lett.
1994, 35, 4969; 1997, 38, 2939.
(9) Gleiter, R.; Gubernator, K.; Grimme, W. J. Org. Chem. 1981, 46,
1247-1250.
(4) Gan, H.; King, J. L.; Yang, N. C. Tetrahedron Lett. 1989, 30, 1205-
1208. Yang, N. C.; Chen, M.; Chen, P. J. Am. Chem. Soc. 1984, 106, 7310-
7315.
10.1021/ja0023579 CCC: $19.00 © 2000 American Chemical Society
Published on Web 11/29/2000

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12099
like derivative of dibenzene 1, an intramolecular dimer 7 related
to pagodane, have been investigated and have led to many
interesting results,11 only a paltry amount of experimental data
is available due to the limited availability of 1. Furthermore,
molecular modeling of 1 reveals that the cyclobutyl ring in 1 is
flexible in contrast to the rigid skeletal framework of 7. It will
be interesting to compare the through-bond interaction of two
π-systems in dibenzenes 1, 2, and related systems with those
in 4 and 5 as well as to compare the through-bond interaction
of two π-systems in 1 with those in 7.
Scheme 1
Results and Discussion
Synthesis of 1. We were also interested in the preparation
of the other cyclodimer of benzene, p,p′-dibenzene 3, and have
synthesized the mixed cyclodimers related to 3,4 the naphthalene:
benzene dimer 812 and the anthracene:benzene dimer 9.13 The
key step in the syntheses of these compounds is the photocy-
cloaddition of a 1,2-disubstituted cyclohexadiene, a masked
benzene, to the arene, followed by the low-temperature elimina-
tion of the 1- and 2-substituents (eq 1). In contrast to the
11,12 this work led to the synthesis of a pentacyclic dimer 12
of benzene from 10 as a potential intermediate in the synthesis
of other interesting products.14 In an effort to purify 12 by
sublimation at 50-55 °C at 12 Torr, we were surprised to find
that the sublimate was contaminated with 1, that is, a symmetry-
forbidden intramolecular retrocycloaddition had occurred under
such a mild condition. Since both 1 and 12 are thermally
unstable and volatile hydrocarbons exhibiting similar chromato-
graphic properties, we were unable to separate these two
hydrocarbons by common chromatographic techniques. A
reasonably pure sample of 1 was prepared by refluxing a pentane
solution of 12 (bath temperature 40 °C) for 9.7 days until most
of 12 was decomposed. The only other detectable product was
benzene. Subsequent work indicated that 1 prepared by this
method was contaminated with 12, and we were unable to
reproduce the reported isolated yield of 16% of 1 in acceptable
purity from the thermolysis of 12.2 In contrast to 2, which may
be prepared in gram quantities in a matter of several days,1a
the preparation of 1 requires the intensive effort of a skilled
laboratory worker over a period of many weeks in order to yield
a few milligrams of the impure product. The overall yield of 1
from the dibromodiol 13 is about 0.01% (Scheme 1). Since we
were unable to synthesize the cis,syn-[2π + 2π] dimers of
derivatives of cis-3,5-cyclohexadien-1,2-diol, for example 14,
in detectable yield by either direct or sensitized irradiation of
their monomers (eq 3), the photocycloaddition of cyclohexa-
syntheses of 8 and 9 and due to the overlap in the ultraviolet
absorption spectra of benzene and cyclohexadienes, the photo-
cycloaddition of 1,3-cyclohexadiene derivatives to benzene led
to a complex mixture of products. These included polymeric
material insoluble in benzene, which precipitates out of the
solution and coats the light source, and often effectively disrupts
the photocycloaddition. These polymeric products may be
derived from both the excited singlet and from triplet cyclo-
hexadienes which result from the energy transfers from the
excited benzene. The yield of the 4π + 4π adduct 10 is usually
very poor, ranging from 0.2 to 2% (eq 2). All efforts to eliminate
the substituents from adduct 10 to form p,p′-dibenzene 3 failed.4
Since dimer 8 was synthesized from a pentacyclic intermediate
dienes to benzene remains the only viable alternative in
(12) (a) Mak, K. T.; Srinivasachar, K.; Yang, N. C. Chem. Commun.
1979, 1038-1040. (b) Gan, H.; King, J. L., Yang, N. C. Tetrahedron Lett.
1989, 30, 1205-1208. (c) King, J. L. Ph.D. Thesis, University of Chicago,
1988, p 44.
(13) Durnell, C. Unpublished results. See also: Durnell, C. Ph.D. Thesis,
University of Chicago, 1993, pp 275-280.
(14) Yang, N. C.; Horner, M. G. Tetrahedron Lett. 1986, 27, 543-546.
(10) Gleiter, R.; Zimmerman, H.; Fessner, W.-D.; Prinzbach, H. Chem.
Ber. 1985, 118, 3856-3860.
(11) (a) Sedelmeier, G.; Prinzbach, H.; Martin, H. D. Chimia 1979, 33,
329-332. (b) Fessner, W.-D.; Prinzbach, H.; Rihs, G. Tetrahedron Lett.
1983, 24, 5857-5860. (c) Fessner, W.-D.; Sedelmeier, G.; Spurr, P. R.;
Rihs, G.; Prinzbach, H. J. Am. Chem. Soc. 1987, 109, 4626-4642.

12100 J. Am. Chem. Soc., Vol. 122, No. 49, 2000
Gan et al.
Scheme 2
experimental conditions, but from a side-reaction in the Mak-
Yang reaction. A plausible pathway was proposed previously.1a
Photoelectron, UV Absorption, and PMR Spectroscopy.
Photoelectron Spectra. The He(I) photoelectron (PE) spectrum
of 1 is shown in Figure 1, and the recorded first vertical
ionization energies are listed in Table 1. The PE spectrum of 1
shows four transitions (band 1-4) below 10 eV. The calculated
values were obtained using a 6-31G* basis.17 This comparison
is based on the assumption that Koopmans’ theorem18 is valid,
that is, the calculated orbital energies, ꢀj, can be correlated with
the recorded vertical ionization energies.
To understand the differences between 1 and 2 we consider
the correlation diagrams in Figure 2 a and b. On the left side of
each figure the two linear combinations (au, bg, and a2, b1,
respectively) are shown which result from the highest occupied
π MOs of two butadiene fragments. In the case of 1, the
corresponding orbital energies are split due to a through-space
interaction. This is not the case for 2 due to the large separation
of both π-units. At the right side of each figure are shown the
highest occupied σ-orbitals (bu, au, and b2, b1, respectively), the
so-called Walsh orbitals of the central four-membered ring
system.19 For the reason of symmetry only one linear combina-
tion (1:b1, 2:au) interacts with the corresponding Walsh orbital,
which yields a destablization of the π-combination as shown
in Figure 2.19
constructing the basic skeleton of 1 and 3. Although products
prepared from several runs had to be pooled to carry out the
next step of the synthesis, it was apparent that an improved
synthesis of 1 from the pentacyclic adduct 12 is needed to
achieve our study on the chemistry and spectroscopy of 1.
It is known that the [4π + 4π]-cycloadduct of benzene and
cyclopentadiene 15 undergoes a facile Cope rearrangement at
65 °C to the syn-[2π + 2π]-cycloadduct 16 (t1/2 ) 2 h), which
in turn undergoes a sequential rearrangement to the anti-isomer
17 at a much slower rate at the same temperature (t1/2 ) 24
h).13 The thermolysis of 17 to benzene and cyclopentadiene
requires an even more vigorous condition (t1/2 ) 35 h at 80 °C,
eq 4).15 On the basis of this knowledge, our experience in the
The smaller splitting for 2 (0.67 eV) is due to the through-
space interaction of the π-system in 1 which is not present in
2. The qualitative interaction diagram of Figure 2 is confirmed
by the MO calculation as shown in Table 1. These calculations
confirm the assignment given.
UV Absorption Spectra. Previously, we were puzzled by a
marked blue-shift (240 nm vs 266 nm)2 of the UV absorption
of 1 versus 2. The large difference in the observed electronic
transitions may be the result of an impurity, most probably 12,
in the original sample of 1. However, the pure sample of 1
obtained by HPLC exhibits a UV maximum at 260 nm which
is still blue-shifted from that of 2 despite the lower HOMO of
1. A closer examination of the orbital interactions in the
antibonding levels reveals that both through-bond and through-
space interactions similar to those in the bonding levels may
take place. The symmetric π* levels of cyclohexadienyl systems
in 1, π*+, which are lowered by favorable through-space
interaction, cannot undergo a favorable through-bond interaction
with the antibonding σ* orbital. The LUMO of cis-dibenzene
1 is thus π*-, the anti-symmetric π*-level. Since the through-
bond interaction in the bonding level is known to be appreciably
stronger than the through-space interaction, the LUMO of 2 will
be lower than that of 1. The electronic transitions of closely
related compounds may be correlated to the gaps between their
HOMOs and LUMOs. The orbital interactions that raise both
the HOMO and LUMO levels of 1 relative to those of 2 may
synthesis of 8 and 9, and the facile bis-dehydroxylation of 1,2-
diols to the corresponding olefins via the Mak-Yang reaction,16
an alternate synthesis of 1 was designed to circumvent the
thermolysis of 12 as a key step in the synthesis of 1 (Scheme
2). When the thermal Cope rearrangement of 10 was attempted,
we discovered that the rate of conversion of 10 to 18 and the
rate of subsequent rearrangement of 18 to 19 were much closer
to each other than those reported for related compounds in the
15 series (eq 5). Nevertheless, diol 18, contaminated with small
amounts of 10 and 19, was obtained in good yield by careful
manipulations and may be converted directly to dibenzene 1
containing biphenyl as the only detectable byproduct. Pure
dibenzene 1 was obtained from this mixture by HPLC, mp
54.0-54.5 °C, λmax(cyclohexane) 260 nm (ꢀ ) 4500). We wish
to point out that these values differ from those originally reported
from our laboratory, mp 45-46 °C, λmax(cyclohexane) 240 nm
(ꢀ ) 4900).2 The formation of biphenyl in the synthesis of 1
and in the synthesis of 2 via the Mak-Yang reaction suggests
that biphenyl was not derived from the dibenzenes under the
(17) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.;
Nanyakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, revision E2 and
Gaussian 98, revision A6; Gaussian, Inc.: Pittsburgh, PA, 1995 and 1998,
respectively.
(18) Koopmans, T. Physica 1934, 1, 104.
(15) Grimme, W.; Maure, W.; Reinhardt, G. Angew. Chem., Int. Ed. Engl.
1979, 18, 224-225.
(16) King, J. L.; Posner, B. A.; Mak, K. T.; Yang, N. C. Tetrahedron
Lett. 1987, 28, 3919-3922.
(19) (a) Walsh, A. T. Nature 1947, 159, 712; Trans. Faraday Soc. 1949,
45, 179. (b) Hoffmann, R.; Davidson, R. B. J. Am. Chem. Soc. 1971, 93,
5699. (c) Salem, L.; Wright, J. S. J. Am. Chem. Soc. 1969, 91, 5947; Gan,
H. Ph.D. Thesis, 1990, University of Chicago, p 138.

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12101
Figure 1. PE spectrum of syn-dibenzene 1.
Table 1. PE Spectra of o,o′-Dibenzenes, Values in eV, and
Assignments Are Given in Brackets9,10
band
1 exptl
calcd
2
7
1
2
3
4
7.88[b1(π+ - σ)]a 7.85b 7.85[au(π+ - σ)]8 7.6[b1(π+ - σ)]9
a degenerative rearrangement might lead to the equivalency of
all C-H’s in dibenzene 1.
8.34[a2(π-)]a
9.1[b2(σ)]a
8.17b 8.52[bg(π-)]8
11.02b 10.15[bu(σ-)]8
11.18b 10.4[au(σ)]8
8.1[a2(π-)]9
9.1[b2(σ)]9
9.65[a1(σ)]9
9.75[a1(σ)]a
Both syn-dibenzene 1 and anti-dibenzene 2 exhibit three sets
of proton resonance under common laboratory conditions
(Figure 3). The noticeable differences are that the cyclobutyl
proton resonance in dibenzene 1 is downfield-shifted from that
in 2 (from 3.26 to 3.60 ppm) and that the outer olefinic protons
are upfield-shifted (from 5.64 to 5.40 ppm). The inner olefinic
proton resonance remains at approximately the same position,
5.76 ppm for 1 and 5.70 ppm for 2. Due to the thermal instability
of dibenzene 1 with respect to its possible rearrangement to
cis-bicyclo[6.4.0]-2,4,6,9,11-pentaene 21 (eq 8), we found no
a Experimental values. b Values calculated from HF/6-31*G orbital
energies.
result in similar UV absorptions in two o,o′-dibenzenes. Ac-
cordingly, we also carried out SCF-CI calculations on the UV
absorptions of dibenzenes 1 and 2 using the Gaussian program
and a 6-31G set,17 assuming that 1 and 2 exist in a distorted C2
geometry. We obtained similar values for both isomers using
both Hartree-Fock (HF), 232 nm, and density functional
approaches (DFT), 244 nm.20 These results and the qualitative
estimate of HOMO-LUMO gaps of 1 and 2 give no basis for
the large difference in the UV maxima reported originally by
one of us using an impure sample.2
Proton Magnetic Resonance Spectrum. A C10H10 hydro-
carbon structurally related to dibenzene 1, hypostrophene 20,
is known to undergo degenerate Cope rearrangement (eq 6).21
experimental evidence to demonstrate the 5,5′-sigmatropic shift
in dibenzene 1. In the meantime, the synthetic complexity and
our limited resources prevented us from investigating the pos-
sible existence of this rearrangement in an appropriate isotopic-
labeled dibenzene 1, as in the case of hypostrophene.21 The
resonances of olefinic protons in dibenzene 1 resemble closely
those in caged dibenzene 7 (δ 5.30 and 5.73 ppm). The shifts
are presumably due to the through-space interaction between
two cyclohexadienyl rings. Similar shifts in PMR signals were
also observed in related 2πs + 2πs syn- and anti-naphthalene:
benzene cyclodimers 22 and 23. Although we failed to isolate
the metastable anti-dimer 23 in pure form, the experimental
results indicated that the cyclobutyl protons in syn-dimer 22
are downfield-shifted from those in anti-dimer 23, while the
Hypostrophene displays only two PMR signals at δ 3.2 ppm
(6H) and 6.1 ppm (4H) under common laboratory conditions.
Ab initio molecular orbital calculations suggest that a similar
symmetry-allowed 5,5′-suprafacial-sigmatropic shift may occur
in dibenzene 1 in the syn-configuration and cannot occur in
dibenzene 2 in the anti-configuration. It will be interesting to
examine the PMR spectrum of dibenzene 1 for this possible
degenerate symmetry-allowed shift (eq 7). Theoretically, such
(20) Hehre, W. J.; Radom, L.; P. v. R. Schleyer, P. Ab Initio Molecular
Orbital Theory; Wiley-Interscience: New York, 1986; p 34. Parr, R. G.;
Yang, W. Density Functional Theory of Atoms and Molecules; Oxford
University Press: Oxford, UK, 1989.
(21) McKennis, J. S.; Brener, L.; Ward, J. S.; Pettit, R. J. Am. Chem.
Soc. 1971, 93, 4957-4958.

12102 J. Am. Chem. Soc., Vol. 122, No. 49, 2000
Gan et al.
Figure 2. Schematic representation of orbital interaction in dibenzenes 1 and 2.
Figure 3. Comparative PMR spectra of dibenzenes 1 (peaks S) and 2 (peaks A), the signal at 7.24 ppm is due to chloroform.
corresponding outer olefinic protons are upfield-shifted (vide
infra). Since the degenerate rearrangement is unlikely in dimer
7 due to the steric restrain of its caged structure, and in dimer
22 due to the barrier imposed by the benzene resonance in the
naphthalene moiety of the system, we conclude that such a
rearrangement cannot be demonstrated at present.
Thermolyses of 1 and Related syn-o,o′-Arene:Benzene
Dimers 22 and 25. Although it has been reported that the
thermolysis of dibenzene 1 yields two molecules of benzene,2
the mechanistic aspects of this reaction have not been carefully
investigated due to the limited supply of this compound. We
were intrigued by our previous observation that syn-dibenzene
1 was more stable thermally than anti-dibenzene 2, and the
higher stability of 1 was the result of a large entropy of
activation in the thermolysis (∆Hq ) 22.5 ( 1.3 kcal/mol, ∆Sq
) -14.0 ( 2.2 eu).2 On the basis of this observation, we
proposed that thermolysis of syn-dibenzene 1 may proceed via
anti-dibenzene 2 as an intermediate: the rearrangement of 1
via a disrotatory ring opening to pentaene 21 followed by its
known rearrangement to anti-dibenzene 2 (eq 8).2 This conclu-
sion, although made on the basis of faulty data likely derived
from an impure sample, will be discussed in a later section.
Since we were also interested in the chemistry of p,p′-dibenzene
3, we had already synthesized the p,p′-dimer of naphthalene:
benzene 812 and a p,p′-dimer of anthracene:benzene 9.13 These
dimers underwent a smooth Cope rearrangement to the respec-
tive syn-o,o′-arene:benzene dimer 22 and 25 at a lower tem-
perature than their thermolyses, and these syn-dimers were
isolated and characterized.12,13 Since all syn-o,o′-arene:benzene
dimers may undergo thermolysis via a similar mechanism, we
initiated a dual probe on the mechanism of the thermolysis of
dibenzene 1, that is, parallel to the development of a more

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12103
Table 2. Activation Parameters for the Thermolysis of
syn-o,o′-Dibenzene 1
basis of transition-state resonance energies higher than the
calculated heats of formation of biradical intermediates in the
thermolysis of a series of cyclobutanes.23 If the thermolyses of
1, 2, and 7 all proceed via the same mechanistic pathway, we
must account for the large difference in the heat of activation
between 1 and 7. Alternatively, the thermolysis of 1 may
proceed via a pathway different from those of both 2 and 7, in
which case we must account for the anomaly on the thermolytic
pathway of 1.
Ea
∆Hq
(kcal/mol)
∆Sq
(eu)
∆Gq
(kcal/mol)
experiment (kcal/mol)
1
2
27.8 ( 0.6 27.2 ( 0.6
0.2 ( 1.6 27.2 ( 1.1
27.0 ( 0.5 26.4 ( 0.5 -2.0 ( 1.4 27.0 ( 0.9
27.4 ( 0.6 26.8 ( 0.6 -0.9 ( 1.6 27.1 ( 1.1
average
Table 3. Activation Parameters for the Thermolyses of
4nπ-Arene:Benzene Dimers
Birney and Berson found that there is a relationship between
kinetic and thermodynamic stabilities in cycloreversion reac-
tions.24 Grimme and co-workers noticed a linear dependence
on the heat of activation with the resonance energy gained in
the thermolysis of 4πs + 2πs arene:benzene dimers, for example,
26-28.25 A plot of the heat of activation of thermolysis of arene:
benzene adducts 26-28 versus the resonance energy gained in
the process yields a slope of -0.43. We have noticed a similar
structural dependence on the heat of activation of the thermoly-
ses of 4πs + 4πs arene:benzene dimers, 29-30 (Table 3),4
namely that the heat of activation decreases as the heat of
reaction becomes more favorable. Although the 4πs + 2πs
adducts tabulated by Grimme may undergo a symmetry-allowed
cycloreversion to their components, the thermolyses of 4n-
adducts are expected to proceed via a stepwise dissociation. It
is reasonable to assume that the thermolyses of a related series
of 2πs + 2πs arene:benzene dimers may bear a similar type of
relationship between heats of activation and the resonance
energies of arene products, anthracene, naphthalene, or benzene,
gained in the thermolyses.
Ea
∆Hq
∆Sq
∆Gq
dimer
(kcal/mol)
(kcal/mol)
(eu)
(kcal/mol)
ref
22
25
1
2
7
29
30
28.4
28.2
27.4
24.7
38.2
33.8
23.0
27.8
27.5
26.8
24.1
37.8
33.1
22.4
-1.8
-2.0
-0.9
-2.2
2.9
16.3
1.7
28.3
28.2
27.1
24.8
-
4
4
-
1a
8
4
4
27.4
21.9
efficient synthesis of 1 of high purity in order to reexamine its
kinetics of thermolysis. We also made a comparative thermolysis
study of the analogous syn-o,o′-arene:benzene dimers 22 and
25.
After we had achieved an improved synthesis of pure syn-
dibenzene 1, we reinvestigated its thermolysis in duplicate runs
in cyclohexane-d12, both at six different temperatures ranging
from 60 to 100 °C; progress of the reaction was monitored by
PMR spectroscopy. In both runs first-order kinetic behavior was
observed. We found that dimer 1 was kinetically more stable
than dimer 2, but the stability was the result of a higher
activation enthalpy in the thermolysis. The activation parameters,
obtained from transition state theory, are shown in Table 2. The
activation enthalpy of this thermolysis is 26.8 ( 0.6 kcal/mol,
and the reaction proceeds with little or no activation entropy
(∆Sq ) 0.9 ( 0.9 eu) as is usually observed in highly exoergic
processes. Our results clearly indicate that syn-o,o′-dibenzene
1 is kinetically more stable than its anti-isomer 2 (∆Hq ) 24.7
( 1.0 kcal/mol, ∆Sq ) -2.23 ( 2.29 eu), despite the possibility
that dimer 1 may be thermodynamically less stable than dimer
2 because of unfavorable overlapping interactions of the
cyclohexadiene rings.
It was suggested that the stability of 1 may be related to its
lower HOMO level.7 However, this cannot explain why 7 is
more stable than 2 even though its HOMO level lies much
higher in energy than that of 2 (∆E ) 0.25 eV). Therefore,
other factors must be more important in this consideration.
The activation parameters of the thermolyses of 2πs + 2πs
and other 4nπ arene- benzene dimers 1, 2, 7, 22, 25, 29, and
30 are listed in Table 3. The results indicate that, within
experimental error, the thermolyses of all three syn-o,o′-arene:
benzene dimers, 1, 22, and 25, proceed with similar activation
parameters. Dimer 1 is thermally more stable than dimer 2, while
dimer 7 is much more stable than both dimer 1 and 2. It has
been suggested that dimer 7 undergoes thermolysis via a
stepwise mechanism and a biradical intermediate.10 The forma-
tion of the biradical intermediate involves the cleavage of an
allylic bond with the release of the strain energy of a cyclobutane
ring of approximately 26.2 kcal/mol.22 The high value of the
heat of activation (37.8 kcal/mol) is in agreement with this
suggestion. Alternatively, Doering and his collaborators pos-
tulated a forbidden-concerted-[2πs + 2πa] cycloreversion mech-
anism for the thermolysis of cyclobutanes, including 7, on the
Since the resonance energy gained in the arene moiety in
the thermolyses of [2πs + 2πs] arene:benzene dimers 1, 22,
and 25 are 36, 26, and 22 kcal/mol respectively, the lack of
appreciable change in their heats of activation clearly indicates
that their thermolyses proceed via a different mechanistic
pathway from other arene:benzene dimers. Two possible mech-
anisms are proposed for the thermolysis of dimer 1.
Previously, it has been suggested that 1 may rearrange
thermally to 2 via a concerted disrotary ring-opening to cis-
bicyclo[6.4.0]dodeca-2,4,6,9,11-pentaene 21 on the basis of a
highly negative activation entropy in the thermolysis of 1 (eq
8).2 However, our efforts to repeat this kinetic measurement
were not successful. In the meantime, King and Durnell in our
laboratory have detected a transient minor product in the
thermolyses of related syn-o,o′-arene:benzene dimers 22 and
25.9,10 The result is illustrated by PMR spectroscopic analyses
of the thermolysis of dimer 22 in Figures 4 and 5. A minor
product 23 was detected which exhibits peaks in the PMR
attributable to the corresponding anti-isomer. Figure 4 is the
PMR of pure dimer 22, and Figure 5 is the PMR of dimer 22
after being heated at 40 °C for 21 days in cyclohexane-d12.
Several new peaks appeared, particularly those at the higher
field. These peaks disappeared either by heating at a higher
temperature or for a longer period of time. By comparing the
PMR spectrum of syn-dibenzene 1 to that of anti-dibenzene 2
in Figure 2, vide supra, we note that the cyclobutyl protons in
the anti-isomer 2 at 3.26 ppm are upfield-shifted from those in
1 at 3.60 ppm by approximately 0.3 ppm. In addition, the
olefinic protons are spaced further apart in 1 at 5.40 and 5.78
ppm than in 2 at 5.64 and 5.70 ppm. The minor transient product
(23) Doering, W. v. E.; Roth, W. R.; Breuckmann, R.; Figge, L.;
Lenmartz, H.-W.; Fessner, W.-D.; Prinzbach, H. Chem. Ber. 1988, 121,
1-9.
(24) Birney, D. M.; Berson, J. A. Tetrahedron 1985, 42, 1561-1570.
(25) Bertsch, A.; Grimme, W.; Reinhardt, G. Angew. Chem., Int. Ed.
Engl. 1986, 377-378.
(22) Benson, S. W. Thermochemical Kinetics, 2nd ed.; Wiley-Inter-
science: New York, 1976; pp 28 and 63.

12104 J. Am. Chem. Soc., Vol. 122, No. 49, 2000
Gan et al.
Figure 4. PMR spectrum of syn-o,o′-naphthalene:benzene dimer 22.
formed in this reaction displays the cyclobutyl signals at higher
fields, approximately 0.3 ppm higher than those in dimer 22,
no olefinic signals in the higher-field region from 5 to 5.5 ppm,
and the olefinic signals are closer together in the 5.6-5.8 ppm
region. These changes are closely analogous to the changes from
syn-dibenzene 1 to anti-dibenzene 2. We thus assigned the
structure of this minor transient product to anti-o,o′-naphthalene:
benzene dimer 23 (eq 9). However, we were unable to increase
rearrangement to be 25.5 and 25.1 kcal/mol, with a small entropy
of activation of -2 and -1 eu, respectively. These values are
quite close to the activation parameters of the thermolyses of
syn-o,o′-arene:benzene dimers given in Table 3. It is possible
that the similarity of the activation parameters between the
thermolysis of syn-o,o′-dimers (1, 22, and 25) and the rear-
rangement of the bicyclooctadiene to cyclooctatriene may be
coincidental. However, this observation, coupled with the
detection of a transient intermediate in the thermolyses of 22
and 25 and the marked stability of 7 (which cannot undergo
disrotatory ring opening because of its cagelike structure),
strongly suggests this reaction pathway for the thermolysis of
dibenzene 1. The thermolysis of dibenzene 1, therefore, may
proceed through dibenzene 2 prior to its dissociation to two
benzenes, as is shown in eq 8, above. In this mechanism, the
disrotatory opening of the cyclobutane ring to cis-pentaene 21a
is the rate-determining step (eq 12). Pentaene 21a undergoes a
the yield of 23 formed in the thermolysis (the data given
represents the optimized effort), and the amount of 23 formed
(8%) in the thermolysis was too small for its isolation and
additional characterization. Since such a rearrangement is not
possible in dimer 7 due to its cage-like structure, this observation
provides a rational basis for the difference between the ther-
molytic behavior of the rigid cage-like dimer 7 and those of
the “more flexible” dimers 1, 22, and 25. In addition, this
explanation is also in agreement with the kinetics of the thermal
rearrangement of bicyclo[4.2.0]octa-2,4-diene 31 to cycloocta-
1,3,5-triene 32 via an analogous disrotatory ring opening, which
was investigated by Huisgen, Winstein, and their co-workers
(eq 10).26,27 They found the enthalpy of activation of this
(26) Huisgen, R.; Boche, G.; Dahmen, A.; Fechtl. W. Tetrahedron Lett.
1968, 5215-5219.
(27) Glass, D. S.; Watthey, J. W. H.; Winstein, S. Tetrahedron Lett. 1965,
conformational change to pentaene 21b which is known to
377.
undergo a disrotary ring closure with an activation energy of

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12105
Figure 5. PMR spectrum of dimer 22 after being heated at 40 °C for 21 days, peaks marked with B are attributed to dimer 23.
23.3 kcal/mol to 2,2d and 2 is known to dissociate to two
molecules of benzene much faster than 1 under the same
experimental conditions. Since the thermal dissociation of anti-
dimer 23 is expected to be slower than that of dimer 2, the
failure to observe intermediates 2 and 21 in the thermolysis of
dimer 1 is not unexpected.
There is considerable interest in the thermolysis of cyclobu-
tanes to two molecules of ethylene, a subject explored by
Doering and his collaborators23 and recently reviewed by
Berson.24,28 The discussion concerns whether the biradical
formed in the stepwise cleavage of cyclobutanes represents the
transition state or an intermediate on the potential surface in
the thermolysis (Figure 6). Zewail and co-workers detected a
transient by using femto-second laser spectroscopy coupled with
Alternatively, syn-dibenzene 1 may also rearrange directly
to anti-dimer 2 via a symmetry-allowed 1,5-sigmatropic shift.
For other syn-o,o′-dimers 22 and 25, their rearrangements to
the corresponding anti-isomers may be achieved by a similar
mass spectrometry in a molecular beam.29 Their observations
provide experimental evidence for the existence of tetrameth-
ylene biradical and its tetramethyl derivative as intermediates
in chemical reactions. Therefore, a biradical in thermolysis of
1,5-sigmatropic shift of the cyclohexadienyl ring. However, this
a cyclobutane is in an energy well following the transition state.
direct pathway is judged to be not preferred on the theoretical
To contribute to the discussion on the mechanism of the
ground, vide infra.
A Theoretical Treatment of the Chemistry of Dibenzenes.
If the rigid dimer 7 dissociates into its monomeric form via a
stepwise mechanism and a biradical intermediate 33, and the
thermolyses of dimer 1 and 2 proceed via different reaction
pathways from that of dimer 7, an important experimental
thermolyses of 1 and 2 to benzene we carried out quantum
mechanical calculations using the Gaussian 94 and 98 pro-
grams.17 The ranges of variation of energy values using different
sets in calculations have been noted previously by Shriver and
Gerson.30 Geometry optimizations were carried out using density
observation will have to be accounted for, that is, the substantial
difference between kinetic parameters in the thermolysis of 7
from its flexible analogues 1 and 2 (eqs 13, 14, and 15).
(28) Berson, J. A. Science 1994, 266, 1338-1339.
(29) Pederson, S.; Herek, J. L.; Zewail, A. H. Science 1994, 266, 1359-
1364.

12106 J. Am. Chem. Soc., Vol. 122, No. 49, 2000
Gan et al.
Figure 6. Schematic representation of thermolyses of dibenzenes 1 and 2 via biradical intermediates.
Table 4. Calculated Enthalpies of Formation of o,o′-Dibenzenes
and Related Molecules Relative to Two Molecules of Benzene (in
kcal/mol)
undergo disrotatory ring opening to form an octatriene as 1 and
2 and must dissociate by a stepwise mechanism and a biradical
intermediate 33, 38.2 kcal/mol. Dibenzene 1 may also rearrange
directly to dibenzene 2 via a suprafacial 1,5-sigmatropic shift
in which the 1,1′-σ-bond migrates to the 5′-position by a 60°
rotation of the 6,6′-bond (eq 11). The distance between the
migrating 1-carbon to both 1′-C origin and the 5′-C terminus is
estimated to be over 2.9 Å at the transition state, which is too
far for an appreciable stablization of a 1,5-sigmatropic shift.32
No meaningful results were obtained in our calculations via this
pathway, and this direct pathway is thus not preferred.
method
1
33
2
34
28 reference
Gaussian 86 or 88 with 60.3
RHF 3-21G set
Gaussian 86 or 88 with 61.0
-
56.0
-
46.0 29
-
57.5
-
54.3 29
RHF 6-31G* set
Gaussian 94 or 98 with 60.8 78.9 57.4 72.8 54.9 this work
UB3LYP/6-31G* set
function theory (DFT)20 with the (U)B3LYP/6-31G* functional31
for both dibenzenes and the biradical intermediates 34 and 35.
At this level of theory, the energies of 1 and 2 are 60.8 and
57.4 kcal/mol, respectively higher than two molecules of
benzene, in good agreement with energy values in the litera-
ture.30 These values are listed in Table 4. We note that the values
for dibenzenes from different laboratories are in general
agreement with each other. The energy difference between 1
and biradical 34 is 18.1 kcal/mol, which is higher than that
between 2 and biradical 35, 15.4 kcal/mol. If the stability of
biradicals were involved in determining the rate of thermolyses
of flexible dibenzenes, this could offer a qualitative explanation
of the higher ∆Hq for thermolysis of 1 than that of 2.
Accordingly, we also performed ab initio calculations on the
activation energies for the formation of biradicals 34 and 35
from dibenzenes 1 and 2 (Figure 6) as well as the pericyclic
rearrangement of 1 to 2 via the pentaene intermediates 21a and
21b (eq 12, vide supra). We explore the activation energies of
these pathways using Gaussian 9417 with the (U)B3LYP/6-31G*
level of DFT.20 The values were 39.3 kcal/mol for the formation
of 34 from 1 and 41.3 kcal/mol for the formation of 35 from 2.
These values far exceed the experimental values of 27.4 kcal/
mol and 24.7 kcal/mol for the thermolyses of dimers 1 and 2,
and indicate clearly that biradicals do not represent the transition
states in the thermolyses of flexible dimers 1 and 2. However,
the value for the syn-biradical 34 formation from 1 of 39.3 kcal/
mol is in reasonable agreement with the experimental value for
the thermolysis of rigid dimer 7. Note that dimer 7 cannot
Since the thermolysis of syn-2π + 2π naphthalene:benzene
dimer 22 may proceed via the tricyclic-24 and anti-isomer 23
as the intermediate (eq 9), a similar process may occur in the
thermolysis of dibenzenes 1 and 2 via a pentaene intermediate
21. The cyclooctatriene moiety of 21 may exist in either the
cis- or the trans-conformations, 21a and 21b, and we propose
that dibenzene 1 may open to pentaene 21a (equation 12, first
step). We thus carried out ab initio calculations on the thermal
rearrangments of dibenzene 1 to pentaene 21a, the relative
stabilities between 21a and 21b, and the energy barrier of
conformational changes between 21a and 21b. The results
(Figure 7) indicate that 21a is destablized by 2.23 kcal/mol from
21b and the activation energy barrier for conformational change
of 21a to 21b is 5.13 kcal/mol. The calculations also show that
the barrier of cyclization of pentaene 21b to 2 is 22.09 kcal/
mol which is in good agreement with the experimental value
of 23.3 kcal/mol, vide supra.2 The experimental value for the
thermolysis of dibenzene 1 is 27.2 kcal/mol. The theoretical
calculation, ∆H ) 27.95 kcal/mol, thus supports that dibenzene
1 may rearrange via pentaene 21 to dibenzene 2, which dissoci-
ates to two molecules of benzene under experimental conditions.
At this time, we conclude that the mechanism of thermolysis
of 2 plays a key role in the understanding of the chemistry of
4n-dimers of benzene. We are now exploring the possibility
that dibenzene 2 may thermolyze to two benzenes via a
concerted-[2πs + 2πa] cycloreversion23 as well as via other
reaction pathways. These symmetry-allowed pathways might
have an activation energy barrier in better agreement with our
experimental results.
(30) Shriver, G. W.; Gerson, D. J. J. Am. Chem. Soc. 1990, 112, 4723-
4728.
(31) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785; Becke,
A. D. J. Chem. Phys. 1993, 98; Stephens, P. J.; Dedlin, F. J.; Chabalvrsky,
C. F.; Frisch, M. J. Phys. Chem. 1985, 82, 299.
(32) Wiest, O.; Houk, K. N. Top. Curr. Chem. 1996, 183, 1-22; Kless,
A.; Nendel, M.; Wilsey, S.; Houk, K. N. J. Am. Chem. Soc. 1999, 121,
4523-4524.

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12107
Figure 7. Various reaction pathways calculated at the (U)B3LYP/6-31G* level.
Adiabatic Photodecomposition of Dibenzenes.33 Dewar
Table 5. A Summary on Photochemical Decompositions of
o,o′-Dibenzenes
benzenes and 2πs + 2πs dimers of benzene are among the most
energetic derivatives of benzene. Both types of compounds will
yield over 80 kcal/mol of energy in the form of enthalpy and
enthalpy of activation.34 In a previous communication, we
demonstrated that thermolysis of dibenzene 2, under optimal
sensitized conditions, yields an excited polynuclear aromatic
sensitizer in extremely low yield.1a Since energetic arene
precursors, in the form of either dimers or Dewar isomers, will
undergo adiabatic photochemical reactions,35 the photochemistry
of dibenzenes 1 and 2 was thus investigated. It has been reported
that the photolysis of dibenzene 2 yields two molecules of
benzene, but the nature of their excited state is not known.2a
Using light at 345 nm at the red-edge of dibenzene’s absorption
(equivalent to 85 kcal/einstein) where benzene has no detectable
absorption, there is 148 kcal/mol of enthalpy of reaction
available for the photochemical dissociation of dibenzene 1.
Since the energy of the first excited state of benzene is 110
kcal/mol and the adiabatic photochemical reaction of 1 to one
molecule each of excited- and ground-state benzene is a
symmetry-allowed process, we expected this reaction to occur
efficiently for both dibenzenes 1 and 2. We found that the only
emission observed in the irradiation of dibenzenes at 300-345
nm is benzene fluorescence (Figure 1S).33 This was confirmed
by the absence of emission from a benzene sample of the same
concentration and at the same wavelength, as well as by the
excitation spectrum of benzene fluorescence from the irradiation
of 1, which possesses a peak shape identical to the UV
absorption of 1 (Figure 2S). The quantum efficiency of
photolysis of 1 and 2 (φ[-cpd]) was determined with the aid of
a potassium ferrioxalate actinometer.36 The quantum efficiency
of the formation of excited benzene from 1 and 2 (φ[benz*]), 0.41
anti-o,o′-dibenzene
syn-o,o′-dibenzene
φ[-cpd]
φ[benz*]
1.04 ( 0.09
0.32 ( 0.03
1.01 ( 0.14
0.41 ( 0.03
and 0.32, respectively, was determined by comparing the
integrated emission spectrum with that of benzene (Table 5).
Adiabatic photochemical reactions generally occur with light
more energetic than that needed for the excitation of the
product.34 Contributions by Turro and co-workers demonstrated
the “red-light to blue-light uphill conversions” in the generation
of n,π* excited states of acetone from the direct or sensitized
photolysis of tetramethyl-dioxetane,37 but such upconversions
have a propensity to occur in the triplet mani-fold. Our finding
in the detection of benzene fluorescence, the formation of
excited benzene in its singlet excited state in the photolysis of
o,o′-dibenzenes, demonstrates that such upconversions can also
occur efficiently in the singlet manifold.
Experimental Methods.
General Preparative Irradiation Procedure. All preparative ir-
radiations were performed with a Hanovia medium-pressure mercury
lamp (450 W) irradiated through a Vycor, Pyrex, or Uranyl glass filter
which was placed in a quartz immersion well cooled with chilled water.
The immersion well was inserted in a Pyrex jacket equipped with a
water-cooled condenser and an inert gas (nitrogen or argon) dispersion
inlet tube. The reaction solution was placed in the outer jacket and
deaerated for at least 20 min before the lamp was activated. During
the irradiation, a gentle stream of inert gas was passed through the
(36) Murov, S. L.; Carmichael, I.; Hog, G. L. Handbook of Photochem-
istry, 2nd. Ed.; M. Dekker: New York, 1993; 299-230.
(37) Turro, N. J.; Lechtken, P. Tetrahedron Lett. 1973, 565-568. Turro,
N. J.; Brewer, D.; Farneth, W.; Ramamurthy, V. NouV. J. Chem. 1978, 2,
85-89.
(38) VanRheenen, V.; Kelly, R. C.; Cha. D. Y. Tetrahedron Lett. 1976,
1973-1976.
(33) Yang, N. C.; Noh, T.; Gan, H.; Halfon, S.; Hrnjez, B. J. J. Am.
Chem. Soc. 1988, 110, 5919-5920.
(34) Turro, N. J.; McVey, J.; Ramamurthy, V.; Lechtken, P. Angew.
Chem., Int. Ed. Engl. 1979,18, 572-586.
(35) Michl, J. Pure Appl. Chem. 1975, 41, 507-534.

12108 J. Am. Chem. Soc., Vol. 122, No. 49, 2000
Gan et al.
solution to provide sufficient agitation. The reactions were followed
by PMR until the maximum amount of desired products was ac-
cumulated.
reduced pressure to give a deep red liquid (2.1 g) which was
chromatographed on silica gel (70-230 mesh, active III), eluting with
dichloromethane-ethyl acetate solvent mixture. The enriched [4 +
4]-diol 10 was collected by elution with 10% ethyl acetate, and it was
further purified via recrystallization from dichloromethane-hexane.
Colorless crystals (10) were obtained (0.082 g, 1.0% based on the diene
orthoformate): mp 124-125 °C. PMR (CDCl3) δ 6.41 (dd, 2H, -CHd
Synthesis of Diol 10 from Dibromodiol 13 (Scheme 1). In a 50-
mL round-bottom flask a solution of dibromodiol 13 (16.0 g, 5.84 ×
10-2 mol) and a catalytic amount of p-toluenesulfonic acid monohydrate
(16.9 mg) in trimethylorthoformate (20 mL) was stirred at room
temperature for 4 h. The solvent was then removed under reduced
pressure, and the oily residue was dissolved in 100 mL of diethyl ether
and transferred to a 250-mL separatory funnel. The ether solution was
washed with saturated aqueous sodium carbonate solution (2 × 100
mL), water (100 mL), and brine (100 mL). The aqueous layers were
further extracted twice with diethyl ether (2 × 100 mL). The combined
organic layers were dried over anhydrous magnesium sulfate, the
solution filtered, and the solvent evaporated under reduced pressure.
The resulting colorless crystals were dried overnight under vacuum
(0.05 mmHg) to yield 18.3 g (99%) of orthoformate of 13. PMR
indicated that the reaction affords two stereoisomers, a and b. The ratio
between these two isomers had been shown by further investigation to
be dependent on the reaction time:
CH-, JAA′ ) 4.74, JAX ) 3.57), 6.27 (dd, 2H, -CHdCH-, JAA′
)
4.81, JAX ) 3.38), 5.98 (dd, 2H, -CHdCH-, JAA′ ) 5.19, JAX ) 3.31),
4.56 [br d, 2H, -CH(OR)-CH(OR)-, J ) 4.84], 3.21 (m, 2H, -Cd
C-CH-CdC-), 3.06 [m, 2H, -CdC-CH-C(OR)-], 2.27 (br d,
2H, -OH, J ) 5.73); PMR (CD3OD) δ 6.37 (dd, 2H, -CHdCH-,
JAA′ ) 4.77, JAX ) 3.58), 6.20 (dd, 2H, -CHdCH-, JAA′ ) 4.84, JAX
) 3.38), 5.86 (dd, 2H, -CHdCH-, JAA′ ) 5.31, JAX ) 3.23), 4.52 [s,
2H, -CH(OR)-CH(OR)-], 3.16 (m, 2H, -CdC-CH-CdC-), 2.98
[m, 2H, -CdC-CH-C(OR)-]; CMR (CDCl3) δ 138.3, 136.1, 133.9,
69.5, 51.6, 39.3.
Instead of chromatography, a simplified procedure was developed
later to purify [4 + 4]-diol 10. This procedure involved the following
steps. After the reaction solution was neutralized, the solvent was
concentrated under reduced pressure to remove most of the tetrahy-
drofuran. The residue was then transferred to a 1000-mL separatory
funnel and extracted with chloroform (2 × 100 mL). The organic phases
were combined and extracted with 20% silver nitrate twice (2 × 50
mL). To the combined aqueous solution was added 400 mL of
concentrated ammonium hydroxide (29.4%). The resulting dark solution
was extracted with chloroform (2 × 150 mL). The organic phase was
then washed with water (200 mL), dried over anhydrous magnesium
sulfate, and filtered, and the solvent was evaporated. A yellow solid
resulted which was recrystallized from dichloromethane-hexane to
afford colorless crystals (10). With this procedure 89 mg of diol 10
(0.63% based on diene used) was obtained starting from 11.4 g diene
orthoformate used in the photoaddition.
Synthesis of Phenylacetal of Diol 10 and 10PA. Diol 10 (30 mg,
0.158 mmol) was dissolved in 0.5 mL (3.33 mmol) of benzaldehyde
dimethyl acetal and 0.5 mL of CHCl3. To this solution 3 mg of
p-toluensulfonic acid was added. The mixture was stirred at room
temperature for 3 h, after which the reaction mixture was transferred
in 50 mL of ether to a separatory funnel. The organic layer was washed
twice with 0.1 M KOH, once with distilled water, and once with brine.
The aqueous layers were back-washed with 50 mL of ether, and the
organic layers were combined, dried over sodium sulfate, filtered, and
concentrated under reduced pressure. The crude product, containing
cyclic acetal 10PA, benzaldehyde, and benzaldehyde dimethyl acetal,
was chromatographed with centrifugal assistance on silica gel, on a
Chromatotron with hexanes-ethyl acetate gradient elution. The cyclic
phenyl acetal 10PA eluted with 19:1 hexane/ethyl acetate, with a yield
of 35 mg (80%). The producet was recrystallized from dichloromethane/
hexane to give white needles, mp. 143-143.5 °C.
IR (CCl4) 3044, 2948, 2919, 2830 (w), 1457, 1401, 1310, 1216,
1088, 1059 (s), 1026, 1018; mass spectrum (CI, isobutane) 279, 200,
173, 155, 154, 107, 95, 94, 91, 79, 78, 67. HRMS: calcd for
C19H19O2: 279.1385. Found: 279.1306. PMR (CDCl3) 7.43 (m, 2H,
ArH), 7.29 (m, 3H, ArH), 6.52 (br s, 2H, -CHdCH-), 6.32 (br s,
2H,-CHdCH-), 6.00 (br s, 2H, -CHdCH-), 5.64 (s, 1H, acetal H),
4.99 (br s, SH, -CH(OR)-CH(OR)-), 3.16 (br s, 4H, bridgehead H);
CMR (CD2Cl2) 139.3, 137.6, 134.4, 133.7, 129.9, 128.6, 128.0, 104.6,
79.8, 48.9, 39.7; UV (hexanes) 266 (66), 262 (110), 256 (130), 250
(120), 207 (4900). Analysis: calcd for C19H18O2: C, 81.98; H, 6.52.
Found: C, 81.52; H, 6.60.
time (hours)
4
6
14
ratio (a:b)
3.3:1.0
2.0:1.0
1.3:1.0
The major isomer exhibits PMR (CDCl3) δ 5.77 (s, 1H), 4.45 (q, 1H),
4.38 (m, 1H), 4.36 (q, 1H), 4.19 (q, 1H), 3.34 (s, 3H), 2.79 (dt, 1H),
2.71 (m, 1H), 2.37 (dt, 1H), 2.25 (m, 1H). The characteristic signals
of the minor isomer in PMR (CDCl3) are δ 5.68 (s, 1H) and 3.42 (s,
3H). Separation of the two isomers was not attempted. The product
was used in the next reaction without further purification.
The dibromocyclohexanediol orthoformate was then converted to
the cyclohexadiendiol by dehydrobromination. A solution of 60.6 g
(0.192 mol) of the orthoformate of 13 in 400 mL of benzene was stirred
in a 1000-mL round-bottom flask. DBU (61.0 g, 0.401 mol) was added
to this in one portion. The solution was then stirred and refluxed under
nitrogen for 6 h. During the reaction a white solid (DBU hydrobromide)
precipitated from the solution. Upon completion of the reaction, 300
mL of 5% potassium hydroxide was added to dissolve the precipitate,
and the reaction solution was cooled to room temperature and
transferred to a 1000-mL separatory funnel. After discarding the aqueous
phase, the organic phase was washed once again with 300 mL of 5%
potassium hydroxide, once with 400 mL of water, and once with 400
mL of brine. The water phases were back-extracted with benzene (2 ×
100 mL). The combined organic phases were then dried over anhydrous
magnesium sulfate, filtered, and evaporated under reduced pressure to
afford a yellow liquid which was distilled under reduced pressure, bp
61-62 °C (0.45 mmHg) to yield diene orthoformate as a colorless liquid
(14.7 g, 49%).
Diene orthoformate (6.43 g, 4.17 × 10-2 mol, in 750 mL of dry
benzene) obtained in the previous experiment was added to the
photoreaction apparatus. The solution was deaerated with argon for 30
min and cooled in a constant-temperature bath to 3-5 °C. The solution
was then irradiated with a 450-W medium-pressure mercury lamp
through a Vycor filter for 20 h. The reaction progress was monitored
by PMR. When more than 80% of the diene had been consumed, the
solvent was evaporated under reduced pressure to give a brown liquid.
Isolation of the [4 + 4] adduct from the reaction mixture was not
attempted. Instead, the crude product was used in the conversion of
diol 10.
In a 250-mL round-bottom flask, the crude reaction mixture obtained
above was dissolved in 50 mL of tetrahydrofuran; to this was added
50 mL of 0.2 N hydrochloric acid. The solution was stirred at room
temperature for 3 h, and then a concentrated solution of potassium
hydroxide was added dropwise until all of the acid had been neutralized.
The solution was transferred into a 1000-mL separatory funnel, and
250 mL of 0.2 N potassium hydroxide was added. This was then
extracted with ethyl acetate (3 × 150 mL), and the aqueous phases
were discarded. The combined organic phases were washed once with
0.2 N potassium hydroxide (250 mL), once with water (500 mL), and
once with brine (250 mL). After drying over anhydrous sodium
carbonate, the solution was filtered and the solvent removed under
Photocyclization of Acetal 10PA. Phenylacetal PA (35 mg, 0.126
mmol) and xanthone (13.3 mg, 0.075 mmol) were dissolved in 2 mL
of CH2Cl2 in a 10-mm NMR tube and deaerated for 10 min with argon.
The solution was immersed in a constant-temperature bath at -17 °C
and irradiated through a Pyrex filter with a 450-W Hanovia medium-
pressure Hg lamp for 2 h. During this time the temperature of the bath
rose to -12 °C. The reaction mixture was concentrated under reduced
pressure and chromatographed on silica with gradient elution with
hexanes-ethyl acetate solvent mixture. Unreacted triene 10PA (5 mg)
eluted with 2% ethyl acetate, followed by cage acetal 12PA (25 mg,
71%) and xanthone (13 mg) which eluted with 5% ethyl acetate. The

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12109
product acetal was recrystallized from CH2Cl2/hexane, to give white
needles, mp 173-174 °C.
If the compound was kept in chloroform-d for 3 h, it isomerized to
the stereoisomer 10B. PMR (CDCl3) δ 6.48 (m, 2H, -CHdCH-),
6.30 (m, 2H, -CHdCH-), 5.94 (m, 2H, -CHdCH-), 5.29 [s, 1H,
CH(OR)2(NR2)], 5.01 [s, 2H, -CH(OR)-CH(OR)-], 3.15 [m, 4H,
-CdC-CH-CdC- and -CdC-CH-C(OR)-], 2.23 (s, 6H,
-NCH3). The ratio between the stereoisomers 10A:10B was 1:1.24.
Separation of the two was not attempted. CMR (CDCl3) of the mixture
was obtained: δ 138.51 (138.48), 136.85 (136.68), 133.76 (132.98),
114.66 (112.49), 79.61 (76.34), 48.84 (48.21), 39.05 (38.83), 37.52
(37.42). The mixture was used in subsequent reactions without further
purification.
IR (CCl4) 3028, 2964, 2892, 1458, 1401, 1306, 1216, 1088, 1063,
1027, 1007, 983; mass spectrum (Cl, isobutane) 279 (2.75), 277 (2.98),
201 (3.48), 200 (4.51), 172 (15.4), 171 (10.8), 155 (28.6), 154 (46.4),
143 (11.3),128 (10.9), 107 (48.8), 105 (32.4), 96 (11.1), 95 (100), 94
(50.8), 91 (18.1), 79 (20.2), 78 (51.8). HRMS (CI, isobutane) calcd
for C19H19O2: 279.1385. Found: 279.1348. PMR (CDCl3) 7.32-7.35
(m, 5H, ArH), 6.30 (m, 2H, JAA′ ) 9.6, JAX ) 7.2, JAX′ ) 0.3, -CHd
CH-), 5.60 (s, 1H, acetal H) 4.33 (br s, 2H, -CH(OR)-CH(OR)-),
3.49 (m, 4H, bridgehead H), 3.07 (m, 2H, cyclobutyl H), 2.85 (m, 2H,
cyclobutyl H); CMR (CDCl3) 137.8, 132.3 130.1, 129.0, 127.6, 102.3,
73.0, 44.0, 33.1, 32.1, 31.6; UV (hexane) 266 (92), 262 (260), 256
(310), 250 (280), 207 (3770). Analysis: calcd for C19H18O2: C, 81.98;
H, 6.52. Found: C, 81.57; H, 6.59.
Synthesis of Caged Diene 12. Phenylacetal 12PA (24.9 mg, mmol)
was dissolved in 0.5 mL of dry THF under N2 and stirred and cooled
in an ice bath. To the solution, 150 µL of tert-butyl lithium solution
(1.6 M in pentane) was added dropwise over about 1 min. The solution
immediately turned deep brown. The solution was allowed to warm
slowly to 20 °C, and after 1 h the solution was light brown. Another
120 µL of tert-butyl lithium solution was added, and the reaction was
continued for an additional 1.5 h. After this time, 200 µL of H2O was
added, and the reaction mixture was transferred in ether to a separatory
funnel. The organic layer was washed once with 0.1 M KOH, once
with distilled H2O, and once with brine. The aqueous layers were all
back-washed with fresh ether, and the organic layers were combined
and dried over Na2SO4. The solution was filtered and concentrated under
reduced pressure to yield 13.1 mg of diene 12 (94%). An analytical
sample could be prepared by dissolving the crude product in several
drops of methanol at room temperature and by precipitation of the diene
by rapid cooling to -25 °C. Alternatively, the sample could be sublimed
at 35-40 °C at 10 Torr to give white plates, mp 73-76°.
Attempted Elimination of 2-Dimethylamino-1,3-Dioxolane De-
rivative of 10. In a 5-mL round-bottom flask, 25.7 mg (0.105 mmol)
of 2-(dimethylamino)-1,3-dioxolane derivative (mixture of stereoiso-
mers) was dissolved in 2.5 mL of dry dichloromethane. To this was
added 33 µL (0.19 mmol) of N,N-diisopropylethylamine. The solution
was stirred under nitrogen while the temperature was maintained at 0
°C with an ice-water bath. A solution of trifluoromethanesulfonic
anhydride (16 µL, 0.095 mmol) in dichloromethane (1.0 mL) was then
added dropwise via syringe over a period of 3 min. The solution was
allowed to react at 0 °C under nitrogen for 2 min before adding 0.016
mL (0.11 mmol) of DBU to quench the remaining acid. The reaction
mixture was then partitioned between dichloromethane (25 mL) and
saturated sodium carbonate solution (25 mL). The aqueous layer was
back-extracted with 25 mL of dichloromethane. The organic layers were
then washed once more with saturated aqueous sodium carbonate
solution, dried over anhydrous sodium carbonate, and filtered, and the
solvent was evaporated under reduced pressure, leaving a yellow oil.
PMR and TLC showed that there was none of the desired p,p′-dibenzene
3 or its Cope rearrangement product, syn-o,o′-dibenzene 1 produced.
The reaction mixture was purified via preparative TLC with
dichloromethane as eluent. A pure white solid was obtained which was
determined to be the carbonate of diol 10 (5.3 mg): mp 147.5-148.5
°C; PMR (CDCl3) δ 6.45 (br s, 2H, -CHdCH-), 6.30 (br d, 2H,
-CHdCH-, J ) 0.82), 5.95 (br d, 2H, -CHdCH-, J ) 1.98), 5.24
[s, 2H, -CH(OCOO)CH-], 3.27 [br d, 4H, -CdC-CH-CdC- and
-CdC-CH-C(OR)-, J ) 2.19]; CMR (CDCl3) δ 154.79, 138.18,
136.18, 132.86, 78.41, 47.46, 38.33. IR (CCl4) 1837, 1819, 1808, 1356,
IR (KBr) 3024, 2954, 2931, 2852, 1436, 872, 804; mass spectrum
(CI, isobutane) 157, 156, 155, 91, 79, 78. HRMS (CI, isobutane) calcd
for C12H11: 155.0861. Found: 155. PMR (CDCl3) 6.17 (m, 4H, JAA′
) 9.7, JAX ) 7.4, JAX′ ) 0.0, -CHdCH-), 3.38 (br m, 4H, C-CH-
C)), 2.90 (m, 4H, -CH-C-C)); CMR (CDCl3) 130.88 (d, JCH
157.5), 41.07 (d, JCH ) 138.0), 32.90 (d, JCH ) 148.9).
)
1167, 1056 cm-1
.
When sublimation of diene 12 was attempted at 55-60 °C at 12
Torr, dibenzene 1 appeared as an oil which was more volatile than
diene 12. PMR (CHCl3) 5.79 (dd, 4H, J ) 8.5, J′ ) 2.3, dCH-CH)),
5.43 (br d, 4H, J ) 8.5, dCH-C-), 3.62 (s, 4H, cyclobutyl H).
This oil was treated with phenylurazole and lead tetraacetate at 0
°C for 1 h and worked up by transferring the reaction mixture in
CH2Cl2 to a separatory funnel and washing the organic layer once with
0.1 M HCl, once with 0.1 M KOH, and once with brine. The organic
layers were all back-washed with fresh CH2Cl2, and the combined
organic layers were dried over MgSO4. The solution was filtered and
concentrated under reduced pressure to yield a small amount of product
whose NMR spectrum indicated the presence of unreacted phenylura-
zole. PMR (CHCl3) δ 6.30 (m), 4.42 (m), 3.05 (m), 2.76 (m), 2.41
(m), 2.30 (m) ppm.
Synthesis of Dibenzene 1 via the Thermolysis of Diene 12. Since
we were unable to separate 1 and 12 by common chromatographic
techniques and the photorearrangement of 12 to 1 is faster than the
thermolysis of 1 to two molecules of benzene, we were able to heat 12
in refluxing pentane (34-5 °C) to affect the its rearrangement to 1
prior to its complete thermolysis.
Preparation of 2-Dimethylamino-1,3-Dioxolane Derivative of Diol
10. In a 10-mL round-bottom flask, 21.8 mg (0.11 mmol) of diol 10
was dissolved in 1.0 mL of dichloromethane. An excess of neat N,N-
dimethylformamide dimethyl acetal (DMF acetal) (2.0 mL) was added,
and the reaction solution was stirred at room temperature for 1 h. The
solvent and the excess DMF acetal were then evaporated, leaving 28.8
mg of the white solid 10A. PMR (CDCl3) indicated only one
stereoisomer had been formed: δ 6.44 (dd, 2H, -CHdCH-), 6.28
(dd, 2H, -CHdCH-), 5.87 (dd, 2H, -CHdCH-), 5.27 [s, 1H,
CH(OR)2(NR2)], 4.79 [s, 2H, -CH(OR)-CH(OR)-], 3.13 (m, 2H,
-CdC-CH-CdC-), 3.05 [m, 2H, -CdC-CH-C(OR)-], 2.30 [s,
6H, -N(CH3)2].
Further elution gave the monoformate ester of diol 10 as a colorless
liquid (16 mg). PMR (CDCl3) δ 8.01 (s, 1H, HCO-), 6.48 (m, 2H,
-CHdCH-), 6.27 (m, 2H, -CHdCH-), 6.02 (m, 2H, -CHdCH-
), 5.73 [d, 1H, -C-CH-C(OCOR)-, J ) 7.21], 4.67 [br t, 1H, -C-
CH-C(OR)-, J ) 7.72], 3.23 (m, 2H), 3.09 [m, 1H, -CH-
C(OCOR)-], 3.00 [m, 1H, -CH-C(OR)-], 1.71 (d, 1H, -OH, J )
9.69); IR (CCl4) 3576, 3045, 2952, 2931, 1732, 1371, 1162 cm-1
.
Repeating the reaction at the refluxing temperature gave the same
results.
Cope Rearrangement of Diol 10 to Diol 18. In a 250-mL round-
bottom flask equipped with a condenser, a solution of 22 mg (1.16
mmol) of diol 10 in 100 mL of absolute ethanol was stirred while
heating to reflux for 3 h. After the reaction was completed, the solvent
was removed under reduced pressure. The residue was dissolved in 20
mL of dichloromethane and 5 mL of n-hexane. The solvent was then
evaporated slowly on a Rotavap until only 5-7 mL of solution
remained, at which point a white crystalline precipitate diol 10 formed
on the wall of the flask (7.1 mg). The remaining solution was decanted
from the precipitate, and the solvent was evaporated to give a mixture
of diol 10, syn-[2 + 2]-diol 18 and a trace of anti-[2 + 2]-diol 19.
This mixture was separated via preparative TLC on silica gel, eluting
with ethyl acetate/n-hexane (1/1) and developing the plate three times.
Two strong bands (Rf ) 0.65/3 and Rf ) 0.50/3) were observed by UV
light (254 nm). Upon recovery of these two bands from the TLC plate,
the faster moving one was identified as 10 (72.5 mg recovered), while
the slower moving band was diol 18 as a waxy solid (75.7 mg, 34%).
PMR (CDCl3) δ 6.05 (m, 1H), 5.79 (m, 2H), 5.69 (m, 1H), 5.48 (dd,
1H, J ) 9.55, J′ ) 5.47), 5.36 (dd, 1H, J ) 9.80, J′ ) 3.66), 4.42 (dd,
1H, J ) 9.19, J′ ) 3.37), 4.32 (dd, 1H, J ) 5.53, J′ ) 3.61), 3.68 (m,
1H), 3.51 (br t, 1H), 3.42 (m, 1H), 2.64 (ddd, 1H, J ) 17.68, J′ )
8.81, J′′ ) 1.77), 2.43 (br s, 1H), 2.22 (br s, 1H); CMR (CDCl3) δ
131.29, 128.43, 127.83, 124.07, 123.88, 123.39, 69.26, 65.93, 44.34,

12110 J. Am. Chem. Soc., Vol. 122, No. 49, 2000
Gan et al.
43.08, 37.25, 32.44; IR (CCl4) 3640, 3569, 3033, 2926, 1586, 1379,
1073, 1047, 1023 cm-1; UV (CH3OH) λmax (ꢀ) ) 278.3 nm (2054);
MS (Cl+) m/e 155 (M + 1 - H2O), 112 (M - 78). A trace of anti-[2
+ 2]-diol 19 was also detected in the PMR spectrum of the reaction
mixture. If the reaction solution was allowed to reflux for 8 h or longer,
substantial amounts of 19 could be isolated via a Chromatotron, eluting
with 5% diethyl ether in dichloromethane. PMR (CDCl3) δ 5.96 (m,
1H), 5.76 (m, 3H), 5.65 (m, 2H), 4.26 (br s, 1H), 3.91 (br t, 1H), 3.12
(m, 1H), 2.89 (m. 1H), 2.74 (br s, 2H), 2.30 (br s, 1H), 2.05 (br s, 1H).
Preparation of 2-Dimethylamino-1,3-Dioxolane Derivatives of
syn-[2 + 2]-Diol 18. syn-[2 + 2]-Diol 18 (0.139 g, 7.32 × 10-4 mol)
obtained from Cope rearrangement of 10 was dissolved in excess N,N-
dimethylformamide dimethyl acetal (2.5 mL) and stirred at room
temperature for 2 h. The solution was then evaporated to dryness,
affording a brown oily residue which was further dried under vacuum
at room temperature for 2 h. This crude product was shown by PMR
to be a mixture of two stereoisomers and was used for the subsequent
elimination step without further purification. The reaction is quantitative.
Preparation of syn-o,o′-Dibenzene 1. The above 2-(dimethylamino)-
1,3-dioxolane derivative of syn-[2 + 2]-diol 18 and 0.24 mL of N,N-
diisopropylamine (1.38 × 10-3 mol) were dissolved in 12 mL of dry
dichloromethane. The solution was stirred at 0 °C (ice-water bath),
and 0.11 mL of trifluoromethanesulfonic anhydride (6.54 × 10-4 mol)
was added dropwise over a period of 3 min via syringe. The reaction
solution was stirred for another 7 min and then filtered through a short-
column of silica gel (3 cm × 3 cm) with the aid of dichloromethane
(100 mL). The filtrate was concentrated under reduced pressure,
affording a pale yellow residue which was dissolved in pentane (5 mL)
and filtered through a short column of silica gel (3 cm × 3 cm) with
pentane as eluent (150 mL). The solvent was removed under reduced
pressure to give 0.0654 g of a colorless crystalline material. PMR
spectra indicated the material was a mixture of syn-o,o′-dibenzene 1
(51.8% from 18) and biphenyl (10.3% from 18).
Purification of syn-o,o′-Dibenzene 1 by High-Pressure Liquid
Chromatography (HPLC). An analytical sample of syn-o,o′-dibenzene
1 was separated from biphenyl by high-pressure liquid chromatography
using a Water Associates HPLC system, vide supra. Semipreparative
scale HPLC was performed with a Dupont Instruments 9.4 mm × 25
cm Zorbax silica gel column, using hexane as eluent at a flow rate of
2.0 mL of per min. The retention time for syn-o,o′-dibenzene 1 and
biphenyl were 9.29 and 10.65 min, respectively. A solution of 20.5
mg of the above mixture in 3 mL of hexane was prepared and injected
repeatedly in 0.015 mL aliquots. The collected syn-o,o′-dibenzene 1
solution was concentrated and transferred into a small sublimator (20
mL). The solvent was removed under reduced pressure to give a
colorless crystalline residue which was sublimed at 0 °C under pressure
of 10 mmHg. Colorless needles of 1 condensed on the coldfinger was
collected: mp 54.3-54.6 °C. PMR (CDCl3) δ 5.76 (dd, 4H, J ) 7.9,
J′ ) 2.6 Hz), 5.40 (br d, 4H, J ) 8.5 Hz), and 3.60 ppm (s, 4H); UV
(cyclohexane) λmax (ꢀ) ) 260 nm (4500).
The temperature 40 °C was then chosen for our investigation, and the
result is displayed in Figure 4.
Activation Parameters for Thermal Decompositions of Dibenzene
1 and Dimer 22, and for the Thermal Rearrangement of Dimer 8
to Dimer 22. The activation parameters of these processes were
obtained from their rates of decomposition in cyclohexane-d12 by PMR
spectroscopy. The thermal decomposition of 1 was determined in
duplicate runs, that of 22 in both cyclohexane-d12 and chloroform-d.
The experimental details are given in the Supporting Information.
Activation Parameters (298 K) for Decompositions of
Dimers 1, 22, and 8
Ea,
(kcal/mol) (kcal/mol)
∆Hq
∆Sq
(eu)
∆Gq
(kcal/mol)
dimer
1
22
27.4 ( 0.8 26.8 ( 0.8 -0.9 ( 1.6 27.1 ( 1.1
28.4 ( 0.2 27.5 ( 0.2 -1.7 ( 0.6 28.2 ( 0.2
22 (CDCl3)a
-
28.8 ( 1.1
7.6 ( 2.9
-
8
23.2 ( 0.4 22.8 ( 0.4
1.4 ( 1.3 23.2 ( 0.8
a A trial run.
Conclusions
A pure sample of syn-o,o′-dibenzene 1 was obtained for the
first time. The chemistry and spectroscopy of 1 were studied in
comparison to those of the anti-isomer 2, particularly with
respect to the influence of through-bond and through-space
interactions between the two 1,3-cyclohexadienyl systems.
Photoelectron spectroscopy of 1 and 2 reveals that the through-
bond interaction between the two cyclohexadienyl systems is
more important than the through-space interaction. syn-Diben-
zene 1 undergoes thermolysis to two molecules of benzene at
a slower rate than thermodynamically more stable anti-dibenzene
2. Kinetic analysis reveals that the higher thermal stability of 1
is due to the higher heat of activation in thermolysis. Molecular
orbital calculations of dibenzenes, the corresponding bis-
cyclohexadienyl diradicals, and the transition states of ther-
molyses of dibenzenes 1 and 2 suggest that diradicals are not
intermediates in the thermolyses. A comparative study on the
thermolyses of syn-o,o′-arene:benzene dimers suggests that the
thermolyses of syn-o,o′-arene:benzene dimers proceed via their
anti-isomers as intermediates. syn-Dibenzene 1 also undergoes
adiabatic photolysis to one molecule of excited benzene and
one molecule of ground-state benzene in good efficiciency. Since
the adiabatic photochemical formation of excited benzene from
dibenzene 1 may be achieved with light longer than 320 nm in
wavelength, the process represents an upconversion of a less
energetic photon to one of higher energy. The mechanisms of
these reactions were discussed.
Synthesis of syn,o,o′-Naphthalene:Benzene Dimer 22. The dimer
22 was synthesized from the thermal Cope rearrangement of a 0.1 M
solution of the p,p′-dimer 8 in chloroform or cyclohexane at 40-50
°C. The rearrangement was complete after 4 h. The product was purified
by semipreparative HPLC on a 0.94 cm × 25 cm Zorbax silica gel
column using hexane as the eluent. The retention time of dimer 22
was 25.61 min. syn-o,o′-Dimer 22 displays: mp 60-61 °C; PMR-
(cyclohexane-d12) (Figure 3) δ 6.92 (m, 2H, ArH), 6.78 (m, 1H, ArH),
6.67 (m, 1H, ArH), 6.19 (dd, 1H, J10,9 ) 10.0, J10,8 ) 1.4, H10), 5.71
(dd, 1H J9,10 ) 9.9, J9,8 ) 3.9, H9), 5.61 (dd, 1H, J5,6 ) 9.7, J5,4 ) 5.5,
H5), 5.39 (dd, 1H, J4,3 ) 9.8, J4,5 ) 5.5, H4), 5.26 (dd, 1H, J4,3 ) 9.7,
J3,2 ) 5.3, H3), 5.04 (dd, 1H, J6,5 ) 9.9, J6,7 ) 4.0, H6) 4.11 (t, 1H,
J ) 9.5, H1), 3.46-3.60 (m, 3H, H2, H7, and H8); CMR (cyclohexane-
d12) δ 133.8, 132.4, 129.0, 128.6, 128.4, 127.7, 127.7, 127.7, 127.3,
124.8, 124.4, 45.2, 44.2, 42.7, 40.7; IR (CS2) 3056, 3029, 3016, 2918,
Acknowledgment. We thank the National Science Founda-
tion and the National Institute of General Medical Sciences,
Deutsche Forschungsgemeinschaft, and Fonds der Chemischen
Industrie for financial support of this work. G.C. is grateful to
the Humboldt Foundation for a stipend. T.A.McM. is grateful
to the College of the University of Chicago for a Richter
Undergraduate award. We also thank Professor Jerry Berson
for the spectral data of adduct of dibenzene and N-phenylmale-
imide and many valuable discussions, Drs. Fred Arnold and
Michael Green for fruitful conversations regarding the ab initio
calculations, Ms. Sabine Bethke for her assistance in the
calculations on the thermal rearrangements of dibenzenes, Dr.
Christopher Durnell for the use of his experimental results on
p,p′-dibenzenes prior to their publication, Dr. Taehee Noh and
Professor Hansen Shou for their assistance in the adiabatic
photochemical experiments, and Professor Luping Yu for
reading this manuscript and his valuable comments.
2880, 1489, 1451, 811, 788, 755, 738, 700 cm-1
.
Attempts to Detect Metastable Intermediates in the Thermolysis
of Dimer 22. In our attempts to detect a metastable intermediate, the
anti-dimer 23, in the thermolysis of dimer 22, we found that the reaction
was too slow at temperatures below 40 °C. Although the reaction was
faster at 55 °C, the relative amount of 23 formed was not improved.

Chemistry of syn-o,o′-Dibenzene
J. Am. Chem. Soc., Vol. 122, No. 49, 2000 12111
Supporting Information Available: The chemiluminescent
and its excitation spectra of dibenzene 1, the general information
concerning sources of chemicals and instrumentation used, an
improved synthesis of 4,5-dibromo-1,2-cyclohexanediol 13,
methodology used in determining the kinetic parameters of the
thermolyses of dibenzenes, and results obtained (PDF). This
material is available free of charge via the Internet at http://
pubs.acs.org.
JA0023579