Photoinduced electron-transfer cope rearrangements of 3,6-diaryl-2,6- octadienes and 2,5-diaryl-3,4-dimethyl-1,5-hexadienes: Stereospecificity and an unexpected formation of the bicyclo[2.2.0]hexane Derivatives

Article abstract of DOI:10.1021/jo981775h

Under the 9,10-dicyanoanthracene-sensitized photoinduced electron- transfer conditions, (Z,Z)-, (E,E)-, (E,Z)-3,6-diaryl-2,6-octadiene and (d,l),(meso)-2,5-diaryl-3,4-dimethyl-1,5-hexadiene stereospecifically undergo the Cope rearrangement to give a Cope

Full text of DOI:10.1021/jo981775h

1640
J . Org. Chem. 1999, 64, 1640-1649
P h otoin d u ced Electr on -Tr a n sfer Cop e Rea r r a n gem en ts of
3,6-Dia r yl-2,6-octa d ien es a n d
2,5-Dia r yl-3,4-d im eth yl-1,5-h exa d ien es: Ster eosp ecificity a n d a n
Un exp ected F or m a tion of th e Bicyclo[2.2.0]h exa n e Der iva tives
Hiroshi Ikeda,† Toshihiko Takasaki,† Yasutake Takahashi,† Akinori Konno,†
Masao Matsumoto,† Yosuke Hoshi,† Takashi Aoki,† Takanori Suzuki,†
J oshua L. Goodman,‡ and Tsutomu Miyashi*,†
Department of Chemistry, Graduate School of Science, Tohoku University, Sendai 980-8578, J apan and
Department of Chemistry, University of Rochester, Rochester, New York 14627
Received August 31, 1998
Under the 9,10-dicyanoanthracene-sensitized photoinduced electron-transfer conditions, (Z,Z)-,
(E,E)-, (E,Z)-3,6-diaryl-2,6-octadiene and (d,l),(meso)-2,5-diaryl-3,4-dimethyl-1,5-hexadiene ste-
reospecifically undergo the Cope rearrangement to give a Cope photostationary mixture. Remark-
ably, the photoinduced electron-transfer Cope rearrangements of the 4-methylphenyl derivatives
are concurrent with the formation of trans- or endo,cis-1,4-bis(4-methylphenyl)-2,3-dimethylbicyclo-
[2.2.0]hexane in a Cope photostationary mixture. Observed stereospecificity of the Cope rearrange-
ment and the formation of the bicyclo[2.2.0]hexane derivatives demonstrate the intermediacies of
both the chair and boat 1,4-diaryl-1,2-dimethylcyclohexane-1,4-diyl and cation radical intermediates
in a Cope rearrangement cycle. Photoreactions of trans- and exo,cis-1,4-diaryl-5,6-dimethyl-2,3-
diazabicyclo[2.2.2]oct-2-enes further support the interventions of the diyl intermediates in the Cope
rearrangement cycle. By photoacoustic analysis, a cation radical cyclization-diradical cleavage
mechanism is proposed for the photoinduced electron-transfer Cope rearrangement of the title
dienes.
Sch em e 1
We previously reported that the 9,10-dicyanoan-
thracene (DCA)-sensitized electron-transfer degenerate
Cope rearrangement of 2,5-diaryl-3,3,4,4-tetradeuterio-
1,5-hexadienes (d4-1) occurs in a cation radical cyclization
(CRCY)-diradical cleavage (DRCL) mechanism as shown
in Scheme 1.1 The important key process of this mech-
anism is the highly exothermic back electron transfer
(BET) from DCA•- to 1,4-diarylcyclohexane-1,4-diyl cat-
ion radical (d4-2•+) to form 1,4-diarylcyclohexane-1,4-diyl
(d4-2). Since the direct cleavage of d4-2•+ to d4-1•+ and
d4-1′•+ is highly endothermic, a competitive BET prefer-
entially occurs to give diyl d4-2, which, in turn, undergoes
cleavage to give a Cope photostationary mixture of d4-1
and d4-1′. If stereochemistry of d4-2•+ is conserved in d4-
2, the most important step to determine the stereochem-
ical course of the Cope rearrangement is the initial CRCY
step. Since the theoretical argument of Bauld on cycliza-
tion of the parent 1,5-hexadiene cation radical to cyclo-
hexane-1,4-diyl cation radical suggested that the energy
difference between the chair and boat conformations will
be quite small,2 it is of particular interest to know
experimentally which conformation is favorable in the
CRCY step in the photoinduced electron-transfer Cope
rearrangement.
For this purpose, the DCA-sensitized electron-transfer
photoreactions of 3,6-diaryl-2,6-octadienes (ZZ-3a -c, EE-
3a -c, and EZ-3a -c) and 3,4-dimethyl-2,5-diaryl-1,5-
hexadienes (dl-4a -c and meso-4a -c) (Chart 1) were
investigated under nitrogen and oxygen atmosphere.1b,4
The role of the boat 1,4-diaryl-2,3-dimethylcyclohexane-
1,4-diyl cation radical intermediates in the Cope rear-
rangement sequence was investigated by electron-
transfer photoreactions of 1,4-diaryl-2,3-dimethylbicyclo-
[2.2.0]hexanes (trans-5a -c, n,cis-5a -c, and x,cis-5a -c)
and 1,4-diaryl-5,6-dimethyl-2,3-diazabicyclo[2.2.2]oct-2-
enes (trans-6b-c and x,cis-6b-c) (Chart 1).
* To whom correspondence should be addressed.
† Tohoku University.
‡ University of Rochester.
While the thermal Cope rearrangements of dl-4a and
meso-4a are irreversible, the DCA-sensitized photoin-
(1) (a) Ikeda, H.; Minegishi, T.; Abe, H.; Konno, A.; Goodman, J . L.;
Miyashi, T. J . Am. Chem. Soc. 1998, 120, 87-95. (b) Miyashi, T.;
Konno, A.; Takahashi, Y. J . Am. Chem. Soc. 1988, 110, 3676-3677.
(c) Ikeda, H.; Minegishi, T.; Miyashi, T. J . Chem. Soc., Chem. Commun.
1994, 297-298. (d) Ikeda, H.; Minegishi, T.; Takahashi, Y.; Miyashi,
T. Tetrahedron Lett. 1996, 37, 4377-4380.
(3) Dunkin, I. R.; Andrews, L. Tetrahedron 1985, 41, 145-161.
(4) Ikeda, H.; Ishida, A.; Takasaki, T.; Tojo, S.; Takamuku, S.;
Miyashi, T. J . Chem. Soc., Perkin Trans. 2 1997, 849-850. A part of
this work was presented at XIIIth IUPAC International Symposium
on Photochemistry (UK, 1990).5
(2) Bauld, N. L.; Bellville, D. J .; Pabon, P.; Chelsky, R.; Green, G.
J . Am. Chem. Soc. 1983, 105, 2378-2382. See also refs 3 and 6f.
10.1021/jo981775h CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/18/1999

Photoinduced Electron-Transfer Cope Rearrangements
J . Org. Chem., Vol. 64, No. 5, 1999 1641
Ch a r t 1
Sch em e 3a
Sch em e 2a
ranged to EE-3. Diketone dl-8 was selectively converted
to the meso isomer (meso-8) by a continuous isomeriza-
tion-crystallization technique.8 The Wittig reaction of
meso-8 gave meso-4, which was quantitatively converted
to EZ-3 upon heating at 110 °C in toluene. The dl and
meso forms of 8 were confirmed by 1H NMR analyses
using a chiral shift reagent, tris[3-(2,2,2-trifluoro-1-
hydroxyethylidene)-d-camphorato] europium [Eu(tfc)3].
Liquid chromatographic analyses using an optically
active column further confirmed the configuration of 4
and 8. Stereochemistries of 3 were determined by com-
duced electron-transfer Cope rearrangements of 3a and
4a are reversible and form a photostationary mixture.4
Remarkably, the Cope rearrangements of the 4-meth-
ylphenyl derivatives 3b and 4b proceed in concurrence
with the formation of the bicyclo[2.2.0]hexane derivative
(5b). Herein, we report characteristics of the photoin-
duced Cope rearrangement of 3 and 4 in detail.
1
parison of their H NMR with those of 4-substituted R,â-
dimethylstyrene and NOE analyses between C-2-H on
the benzene ring and methyl or olefinic proton.
Resu lts a n d Discu ssion
1,4-Diaryl-2,3-dimethylbicyclo[2.2.0]hexanes (trans-5
and x,cis-5) were synthesized as shown in Scheme 3.
Irradiation of 1,2-diarylcyclobutene (9) and dimethyl
fumarate gave trans-1,4-diaryl-2,3-dicarbomethoxybicyclo-
[2.2.0]hexane (10), which was converted to trans-5 through
trans-11 and 12. Three derivatives, trans-5a -c, were
prepared by this procedure. The exo adduct 13, an
exclusive product of the photoreaction of 9 and maleic
anhydride, was used as the starting material for x,cis-5.
Diol x,cis-11 was converted to 15 through 14. Reduction
of 15 followed by hydrolysis gave alcohol 16. Reaction of
16 with n-BuLi/(CF3SO2)2O followed by reduction with
LiEt3BH gave x,cis-5. Three derivatives, x,cis-5a -c, were
prepared by this procedure. Because the endo isomers of
13a -c were not formed in this photoreaction, n,cis-5 was
isolated from a photostationary mixture obtained in the
DCA-sensitized photoreaction of EZ-3 or meso-4 as noted
Syn th eses an d DCA-Flu or escen ce Qu en ch in g Rate
Con sta n ts of 3-6. The 4-methoxyphenyl, 4-methylphe-
nyl, and phenyl derivatives of 3,6-diaryl-2,6-octadienes
(ZZ-3a -c, EE-3a -c, and EZ-3a -c) and 2,5-diaryl-3,4-
dimethyl-1,5-hexadienes (dl-4a -c and meso-4a -c) were
prepared as illustrated in Scheme 2. The Wittig reaction
of 7 with ethyltriphenylphosphonium iodide gave ZZ-3
as a major product together with EE-3 and EZ-3. Because
yields of EE-3 and EZ-3 were low in this procedure, these
dienes as well as dl-4 and meso-4 were prepared by
different procedures. The key precursor, (d,l)-1,4-diaryl-
2,3-dimethyl-1,4-butadione (dl-8), was selectively pre-
pared by the base-catalyzed condensation7 of the substi-
tuted propiophenone and R-bromopropiophenone. The
subsequent Wittig reaction of dl-8 gives dl-4. Upon
heating at 110 °C in toluene, dl-4 quantitatively rear-
(5) Miyashi, T.; Ikeda, H.; Konno, A.; Okitsu, O.; Takahashi, Y. Pure
Appl. Chem. 1990, 62, 1531-1538. For other experimental studies on
the cation radical Cope rearrangement, see ref 6.
(6) (a) Lorenz, K.; Bauld, N. L. J . Catal. 1985, 95, 613-616. (b) Chen,
G.-F.; Williams, F. J . Chem. Soc., Chem. Commun. 1992, 670-672. (c)
Miyashi, T.; Takahashi, Y.; Ohaku, H.; Ikeda, H.; Morishima, S. Pure
Appl. Chem. 1991, 63, 223-230. (d) Ikeda, H.; Takasaki, T.; Takahashi,
Y.; Miyashi, T. J . Chem. Soc., Chem. Commun. 1993, 367-369. (e)
Ikeda, H.; Oikawa, T.; Miyashi, T. Tetrahedron Lett. 1993, 34, 2323-
2326. (f) Williams, F. J . Chem. Soc., Faraday Trans. 1994, 90, 1681-
1687. (g) Bauld, N. L. In Advances in Electron-Transfer Chemistry;
Mariano, P. S., Ed.; J AI: London, 1992; Vol. 2; pp 1-66. (h) Bauld, N.
L. In Radicals, Ion Radicals, and Triplets: The Spin Bearing Inter-
mediates of Organic Chemistry; Wiley-VCH: New York, 1997; pp 141-
180.
(7) Perry, C. W.; Kalnins, M. V.; Deitcher, K. H. J . Org. Chem. 1972,
37, 4371-4376.
(8) Biftu, T.; Hazra, B. G.; Stevenson, R. J . Chem Soc., Chem.
Commun. 1978, 491-492.

1642 J . Org. Chem., Vol. 64, No. 5, 1999
Ikeda et al.
Sch em e 4a
Ta ble 1. P h otosta tion a r y Ra tios of th e DCA-Sen sitized
Cop e Rea r r a n gem en t of 3a a n d 4a in CD2Cl2 a t 20 °Ca
photostationary ratiosb
sub
time (h) ZZ-3a EE-3a dl-4a EZ-3a meso-4a yields (%)b
ZZ-3a
EE-3a
dl-4a
EZ-3a
meso-4a
20
10
4
8
10
51
2
2
0
0
19
42
40
0
30
55
57
0
0
0
0
80
80
0
0
0
20
20
57
92
87
77
83
0
0
a
A 0.5 mL solution was irradiated with a 2 kW Xe lamp through
a cutoff filter (λ > 360 nm) at 20 °C. [sub] ) 0.1 M, [DCA] ) 5.3
b
× 10-4 M. Determined by 1H NMR analyses.
rangement (Scheme 1) the initial CRCY forms a chair
cation radical intermediate.3 The DCA-sensitized electron-
transfer photoreactions of 3a and 4a will provide the
direct information for the stereochemical course of the
Cope rearrangement.
Upon irradiation of DCA with EE-3a in dichlo-
romethane-d2 at 20 °C, EE-3a gradually diminishes,
forming dl-4a and a small amount of ZZ-3a . After 10 h
irradiation, a photostationary mixture of ZZ-3a , EE-3a ,
and dl-4a in the ratio of 2:42:55 is formed in 92% yield.
Similar photoreaction of dl-4a gives nearly the same
photostationary mixture in 87% yield after 4 h irradiation
as shown in Table 1. By contrast, the rearrangement of
ZZ-3a is apparently slow so that the reaction does not
reach a photostationary state even upon prolonged ir-
radiation, but a mixture of ZZ-3a , EE-3a , and dl-4a is
obtained in the ratio of 51:19:30 in 57% yield after 20 h
irradiation. Like EE-3a and dl-4a , EZ-3a and meso-4a
give nearly the same photostationary mixture under
similar DCA-sensitized photoconditions (Table 1). The
E-Z isomerization of ZZ-3a , EE-3a , or EZ-3a is not
observed under the DCA-sensitized electron-transfer
conditions. The dl-family dienes (ZZ-3a , EE-3a and dl-
4a ) and the meso-family dienes (EZ-3a and meso-4a )
interconvert within their own family through the Cope
rearrangement (Scheme 5). The observed stereochemical
results clearly demonstrate that both the CRCY and
DRCL processes take place stereospecifically in the chair
conformation14 as described by a CRCY-DRCL mecha-
nism in Schemes 6 and 7.
On the cation radical energy surface, EE-3a •+ and ZZ-
3a •+ initially form aaC-21a •+ and eeC-21a •+, respectively.
However, aaC-21a •+ is more stable than eeC-21a •+ be-
cause of steric hindrance between the methyl and aryl
groups in eeC-21a •+. Hence dl-4a •+ preferentially cyclizes
to aaC-21a •+, though dl-4a •+ may cyclize to both aaC-
21a •+ and eeC-21a •+. The facts that ZZ-3a is a minor
component in photostationary mixtures formed from EE-
3a and dl-4a and that the rearrangement of ZZ-3a is
slow suggest that aaC-21a •+ formed from EE-3a •+ and
dl-4a •+ is slow to interconvert to eeC-21a •+, and that eeC-
21a •+ formed from ZZ-3a •+ suffers BET from DCA•- in
in the Supporting Information. The structures of trans-
5c, x,cis-5c and n,cis-5c were directly confirmed by X-ray
analyses.
Diazenes trans-6b,c, and x,cis-6b,c were prepared as
shown in Scheme 4. trans-3,6-Diaryl-1,2-dicarbomethoxy-
1,4-cyclohexadiene prepared from dimethyl acetylenedi-
carboxylate and 1,4-diaryl-1,3-butadiene was converted
to trans-17.9 The isomeric cis-17 was obtained through
20.9 Both trans-18 and cis-18 were derived, respectively,
from trans-17 and cis-17 by reduction and mesylation
followed by reduction according to the known proce-
dure.9,10 The Diels-Alder reactions of trans-18 and cis-
18 with 4-methyl-1,2,4-triazoline-3,5-dione gave trans-
19 and x,cis-19, respectively. By usual procedures, trans-
6b,c and x,cis-6b,c were prepared in good yields.
Dienes (3a -c and 4a -c), bicyclo[2.2.0]hexanes (5a -
c), and diazenes (6b-c) are relatively good electron
donors, and their oxidation potentials (Eox1/2) are low
enough to quench the excited singlet of DCA exothermi-
cally. Free energy changes (∆G) associated with the
forward electron transfers are all negative as calculated
according to the Rehm-Weller equation.11 In accord with
calculated thermodynamics, the DCA fluorescence was
efficiently quenched by 3-6 in acetonitrile with large rate
constants (kq) close to the diffusion-control rate. In less
polar dichloromethane and nonpolar benzene, the DCA
fluorescence was also quenched, albeit less efficiently.
Values for Eox1/2, ∆G, and kq for each substrate are shown
in the Supporting Information.
Th e DCA-Sen sitized Electr on -Tr a n sfer P h otor e-
a ction s of 3a , 4a , a n d 5a . The thermal Cope rearrange-
ment generally occurs through the chair six-membered
cyclic transition state12 which is energetically more stable
than a boat counterpart.13 Thus, it is assumable that in
the photoinduced electron-transfer degenerate Cope rear-
(12) (a) Gajewski, J . J .; Conrad, N. D. J . Am. Chem. Soc. 1978, 100,
6269-6270. (b) Gajewski, J . J .; Conrad, N. D. J . Am. Chem. Soc. 1979,
101, 6693-6704. (c) Gajewski, J . J . Acc. Chem. Res. 1980, 13, 142-
148. (d) Gajewski, J . J . In Organic Chemistry, A Series of Monographs
44, Hydrocarbon Thermal Isomerizations.; Wasserman, H. H., Ed.;
Academic: New York, 1981; Vol. 5; pp 111-191.
(9) Courtot, P.; Rumin, R. Bull. Soc. Chim. Fr. 1969, 11, 3662-3665.
(10) Courtot, P.; Rumin, R. Bull. Soc. Chim. Fr. 1972, 14, 4238-
4250.
(13) Doering, W. von E.; Toscano, V. G.; Beasley, G. H. Tetrahedron
1971, 27, 5299-5306.
(11) (a) Rehm, D.; Weller, A. Isr. J . Chem. 1970, 8, 259-271. (b)
Weller, A. Z. Phys. Chem. (Munich) 1982, 133, 93-98. (c) Gould, I. R.;
Ege, D.; Moser, J . E.; Farid, S. J . Am. Chem. Soc. 1990, 112, 4290-
4301.
(14) The observed result does not always exclude a possibility that
the CRCY or DRCL step occurs in the twist-boat conformation. In this
paper, however, only chair and boat intermediates are discussed for
clarity and comparison with results of the theoretical calculation.2

Photoinduced Electron-Transfer Cope Rearrangements
J . Org. Chem., Vol. 64, No. 5, 1999 1643
Sch em e 5
Sch em e 7
Sch em e 6
Ta ble 2. Absor p tion Ma xim a (λm a x) a n d Ra te Con sta n ts
for Deca y (k0) a n d for MeOH Ad d ition (kMeOH) of a
Tr a n sien t In ter m ed ia te Gen er a ted fr om th e d l-F a m ily
Com p ou n d s (3-6) u n d er NMQ+BF 4--BP -Cosen sitized
a
Con d ition s in CH2Cl2
a : 4-MeOC6H4
k0,c kMeOH
b: 4-MeC6H4
k0,b kMeOH
c: C6H5
c
b
c
,
,
k0,c kMeOH
,
λmax
,
105
105
λmax
,
105
107
λmax
,
105
107
sub
nm s-1 M-1 s-1 nm s-1 M-1 s-1 nm s-1 M-1 s-1
ZZ-3
EE-3
dl-4
trans-5 523 1.0
trans-6
523 1.1
523 1.1
523 1.1
8.3
7.9
8.1
8.5
499 3.4
499 3.9
499 3.8
499 4.1
499 4.3
2.6
2.0
2.6
2.4
2.0
481 5.6
483 4.1
482 4.2
483 4.5
484 7.1
d
8.8
d
5.6
5.8
a
b
At 20 °C. An average of experimental error was about 10%.
c An average of experimental error was about 20%. No attempt.
d
Cope rearrangement sequence is that the DCA-sensitized
photoreaction of trans-5a gives a 39:61 photostationary
mixture of EE-3a and dl-4a , and no ZZ-3a . The boat
cation radical tB-21a •+ formed from trans-5a •+ and boat
diyl tB-21a formed from trans-5a •+ flip preferentially
to aaC-21a •+ and aaC-21a , respectively, as shown in
Scheme 6. Accordingly, trans-5a enters into a dl-EE
Cope manifold through tB-21a •+ and gives nearly the
same photostationary mixture as those from EE-3a and
dl-4a .
For the spectroscopic confirmation of the initial cy-
clizations of EE-3a •+ and dl-4a •+ as well as the prefer-
ential ring flip of tB-21a •+ to aaC-21a •+, EE-3a , dl-4a ,
and trans-5a were subjected to nanosecond laser flash
photolysis (LFP) under the N-methylquinolinium tet-
rafluoroborate (NMQ+BF4-)-biphenyl (BP)-cosensitized
conditions in aerated dichloromethane, minimizing BET
to form diyl intermediates.1a,15 As shown in Table 2, EE-
3a , dl-4a , and trans-5a yield the intense transient
preference to the ring flip to aaC-21a •+. Thus, the
concentration of eeC-21a •+ becomes low relative to that
of aaC-21a •+ in a Cope rearrangement cycle. In addition,
the ring cleavage of eeC-21a presumably gives ZZ-3a as
a major product. The BET from DCA•- to aaC-21a •+
similarly occurs to form aaC-21a which, in turn, under-
goes DRCL to form EE-3a and dl-4a . In a dl-EE Cope
manifold, EE-3a and dl-4a then repeat the Cope rear-
rangement to make a photostationary mixture, while ZZ-
3a fades gradually into a minor component of a Cope
photostationary mixture. Strong evidence to support the
(15) Ikeda, H.; Nakamura, T.; Miyashi, T.; Goodman, J . L.; Akiyama,
K.; Tero-Kubota, S.; Houmam, A.; Wayner, D. D. M. J . Am. Chem. Soc.
1998, 120, 5832-5833.

1644 J . Org. Chem., Vol. 64, No. 5, 1999
Ikeda et al.
Sch em e 8
absorption with λmax at 523 nm. They appear to be the
same transient species on the basis of the same observed
rate constants for decay (k0) and for methanol addition
(kMeOH). The observed transient species is assigned to
aaC-21a •+ by comparison with λmax of 2,5-bis(4-methoxy-
phenyl)hexane-2,5-diyl cation radical (λmax ) 520 nm in
1,2-dichloroethane)16 and 1,4-bis(4-methoxyphenyl)cyclo-
hexane-1,4-diylcation radical(λmax )508nm in acetonitrile).1a
If correct, results of LFP are consistent with experimental
conclusions that aaC-21a •+ is formed directly from EE-
3a •+, selectively from dl-4a •+, and stepwise from trans-
5a •+. The transient species due to aaC-21b•+ (λmax ) 499
nm) and aaC-21c•+ (λmax ) 483 nm) are similarly ob-
served from EE-3b,c, dl-4b,c, trans-5b,c, and trans-6b,c
as shown in Table 2.17
The stereochemical results observed in the electron-
transfer photoreactions of EZ-3a and meso-4a can be
similarly explained by a CRCY-DRCL mechanism
through common intermediates aeC-21a •+ and aeC-21a ,
as shown in Scheme 7. The masked intermediacy of n,cB-
21a •+ is suggested by the DCA-sensitized photoreaction
of n,cis-5a which gives a 82:18 photostationary mixture
of EZ-3a and meso-4a in 98% yield in dichloromethane-
d2. Interestingly, the result of similar electron-transfer
photoreaction of x,cis-5a is entirely different18 from that
of n,cis-5a , revealing that the isomeric exo intermediate,
x,cB-21a •+ shown in Scheme 8, does not participate in a
Cope rearrangement cycle of the meso-family dienes.
Sch em e 9
In conclusion, both chair and boat intermediates ap-
pear on the cation radical and diradical energy surfaces,
but boat intermediates never participate either in the
CRCY or in the DRCL step. However, as discussed later,
boat intermediates are immediate precursors of the
bicyclo[2.2.0]hexane derivatives in a Cope photostation-
ary mixture, supporting the Cope rearrangement se-
quences shown in Schemes 6 and 7.
Th e DCA-Sen sitized P h otooxygen a tion s of 3-5.
In a CRCY-DRCL mechanism, the initial CRCY occurs
in the chair conformation. Hence, when a cage-escaped
cation radical intermediate in a Cope rearrangement
cycle is captured by molecular oxygen, stereochemistry
gained at the initial cyclization step should be conserved
throughout oxygenation, no matter which conformational
isomer is captured by oxygen. For instance, as shown in
Scheme 8, trans-1,4-diaryl-5,6-dimethyl-2,3-dioxabicyclo-
[2.2.2]octane (trans-22) should be formed stereoselectively
(16) Tojo, S.; Toki, S.; Takamuku, S. J . Org. Chem. 1991, 56, 6240-
6243.
(17) Interestingly, similar transients with λmax ) 523, 499, and 481
nm were observed from ZZ-3a , ZZ-3b, and ZZ-3c, respectively, under
similar photoconditions (Table 2), suggesting an occurrence of the ring-
flip of eeC-21•+ to aaC-21•+. This is attributed to suppression of BET
in eeC-21a •+ under the NMQ+BF4--BP-cosensitized conditions. Thus,
free eeC-21a •+ formed under the cosensitized conditions readily flips
to aaC-21a •+ whereas in an ion radical pair eeC-21a •+ suffers rapid
BET as mentioned above. An alternative interpretation for this
through aaC-21•+, tB-21•+, or eeC-21•+ if EE-3•+ cyclizes
in the chair conformation, while cyclization in the boat
conformation should afford a mixture of the cyclic per-
oxides, x,cis-22 and n,cis-22. For the direct chemical
capture of the initially formed intermediate, the DCA-
sensitized photoreactions of 3-5 were investigated under
oxygen atmosphere.
The DCA-sensitized photoreaction of ZZ-3a in oxygen-
saturated acetonitrile results in a quantitative formation
of trans-22a (Scheme 9). Under similar conditions, EE-
3a and dl-4a give trans-22a in excellent yields as shown
in Table 3. On the other hand, both EZ-3a and meso-4a
observation may be accidental coincidence in
λmax, k0, and kMeOH
between eeC-21a •+ and aaC-21a •+. However, this is not likely because
eeC-21a •+ formed by γ-ray irradiation of ZZ-3a in BuCl matrixes at
77 K exhibits an absorption band with λmax at 525 nm while aaC-21a •+
from EE-3a or trans-5a has an absorption with λmax at 515 nm under
similar conditions.4 A possibility of the observation of tB-21•+ can be
also ruled out since a preliminary calculation using PM3 showed that
energy of tB-21b•+ is higher than that of aaC-21b•+. The detail of
calculation study will be given separately elsewhere.
(18) Under similar photoinduced electron-transfer conditions x,cis-
5a , a potential precursor of x,cB-21•+, gave not only the meso-family
but also the dl-family PS mixtures. The detail of the photoinduced
electron-transfer reaction of x,cis-5a -c will be given separately
elsewhere.

Photoinduced Electron-Transfer Cope Rearrangements
J . Org. Chem., Vol. 64, No. 5, 1999 1645
Ta ble 3. Th e DCA-Sen sitized P h otor ea ction of 3 a n d 4 u n d er Oxygen a
yields (%)b
sub
solvent
time (min)
convn (%)b
trans-22
x,cis-22
n,cis-22
3, 4, and 5
ZZ-3a
ZZ-3b
EE-3a
CH3CN
CH3CN
CH3CN
CH2Cl2
C6H6
CH3CN
CH2Cl2
C6H6
1
1
1
5
10
5
5
10
1
1
5
5
1
5
1
5
1
100
21
100
67
57
98
90
85
64
20
100
67
100
51
21
100
26
100
21
94
67
0
78
27
0
0
0
0
0
0
0
0
0
0
15
77
18
0
0
0
0
0
0
0
0
0
0
0
0
0
0
15 (dl-4a )
2 (dl-4b)
21 (dl-4b), 3 (trans-5b)
35 (dl-4b), 1 (trans-5b)
11 (dl-4c), 5 (trans-5c)
0
0
5 (dl-4b)
0
2 (EE-3b)
0
EE-3b
EE-3c
EZ-3a
CH3CN
CH3CN
13
0
0
4
19
trc
0
0
3
EZ-3b
dl-4a
dl-4b
CH3CN
CH3CN
CH3CN
CH3CN
0
90
44
0
0
0
0
meso-4a
12
72
13
17
0
meso-4b
CH3CN
trc
3 (n,cis-5b)
a
b
A 5 mL solution was irradiated with 2 kW Xe lamp through a cutoff filter (λ > 360 nm) at 20 °C. [sub] ) 0.01 M. Determined by 1H
NMR analyses. c Less than 2%.
Ta ble 4. P h otosta tion a r y Ra tios of th e DCA-Sen sitized P h otor ea ction of ZZ-3b, EE-3a -c, a n d d l-4b in CD3CN, CD2Cl2,
a n d C6D6 a t 20 °Ca
CD3CN
CD2Cl2
C6D6
photostationary ratiosb
photostationary ratiosb
photostationary ratiosb
sub
ZZ-3 EE-3 dl-4 trans-5 yields (%)b ZZ-3 EE-3 dl-4 trans-5 yields (%)b ZZ-3 EE-3 dl-4 trans-5 yields (%)b
ZZ-3b
EE-3a
EE-3b
EE-3c
dl-4b
0
2
0
0
0
8
42
9
5
8
39
55
40
68
39
52
1
50
27
52
72
0
0
0
0
48
41
18
29
52
33
67
58
0
26
15
13
58
92
0
0
0
0
44
9
1
50
76
98
77
6
100
97d
100
95
61c
73c
64c
91d
98d
89d
15
tre
8
15
a
A 0.5 mL solution was irradiated with a 2 kW Xe lamp through a cutoff filter (λ > 360 nm) at 20 °C. [sub] ) 0.1 M, [DCA] ) 5.3 ×
10-4 M. Determined by H NMR analyses. c Not photostationary mixture. A small amount of the meso-family compounds was formed.
b
1
d
e Less than 2%.
give a 4:1 mixture of x,cis-22a and n,cis-22a . The
structure of x,cis-22a was determined by X-ray analysis1b
and those of trans-22a and n,cis-22a were determined
by comparisons of their NMR and NOE difference spectra
with those of x,cis-22a .
Solvent and substituent effects on oxygenation were
tested for EE-3a -c. As shown in Table 3, yields of trans-
22a from EE-3a and EE-3b significantly decrease in
dichloromethane. In nonpolar benzene, oxygenations of
EE-3a and EE-3b are suppressed. Interestingly, EE-3b
gives a small amount of trans-5b in dichloromethane and
benzene. The oxygenation efficiency in acetonitrile de-
creases in the order of a decrease in the electron-donating
ability of EE-3. The rearrangement of EE-3c to dl-4c and
the formation of trans-5c are observed for the less
electron-donating EE-3c.
ment sequences shown in Schemes 6 and 7 suggest that
trans-5 and n,cis-5 should be in an equilibrium with the
dl- and meso-family dienes, respectively, if trans-5 and
n,cis-5 are stable enough to survive under the DCA-
sensitized electron-transfer conditions. In fact, in photo-
oxygenation reactions in dichloromethane and benzene,
EE-3b and EE-3c unexpectedly gave trans-5b and trans-
5c, respectively, in low yields together with trans-22b,c
(Table 3). Because the Cope rearrangement accompanied
by the formation of the bicyclo[2.2.0]hexane derivative
is unprecedented, it was of particular interest to know
how 5 is formed. In addition, if the formation of 5 is
concomitant stereoselectively with the Cope rearrange-
ment, these results will provide direct evidence for the
masked intermediacy of boat intermediates in the Cope
rearrangement cycle. To confirm the mechanistic con-
nection between the Cope rearrangement and the forma-
tion of the bicyclo[2.2.0]hexane derivative, the DCA-
sensitized photoreactions of 3 and 4 were further
investigated in terms of solvent and substituent effects.
As mentioned before, while EE-3a gives a photosta-
tionary mixture of EE-3a and dl-4a in acetonitrile and
dichloromethane, EE-3a gives trans-5a in 6% yield
together with EE-3a and dl-4a in nonpolar benzene
(Table 4). Remarkably, a photostationary mixture from
the less electron-donating EE-3b includes a large amount
of trans-5b. The DCA-sensitized photoreaction of EE-3b
gives a photostationary mixture of EE-3b, dl-4b and
trans-5b in the ratio of 9:40:50 in dichloromethane
(Scheme 10). Similar electron-transfer photoreactions of
ZZ-3b and dl-4b in dichloromethane afford nearly the
same photostationary mixture. Similarly, EE-3c affords
a photostationary mixture of EE-3c, dl-4c, and trans-5c
The observed stereochemical results of oxygenation are
consistent with those of the Cope rearrangement, sup-
porting the initial CRCY in the chair conformation.14
Cation radicals ZZ-3•+ and EE-3•+ cyclize to eeC-21•+ and
aaC-21•+, respectively, and dl-4•+ preferentially cyclizes
to aaC-21•+. Molecular oxygen then captures either eeC-
21•+, aaC-21•+, or tB-21•+, giving rise to trans-22 via
trans-23•+. Both EZ-3•+ and meso-4•+ cyclize to aeC-21•+
,
and the sterically more favorable x,cis-22 is consequently
afforded as a major product together with n,cis-22 via
x,cis-23•+ and n,cis-23•+, respectively. Similar DCA-
sensitized photooxygenation of trans-5a gives trans-22a
quantitatively in acetonitrile, supporting the Cope rear-
rangement sequences shown in Scheme 6.
Th e P h otoin d u ced Electr on -Tr a n sfer Cop e Rea r -
r a n gem en t Accom p a n ied by th e F or m a tion of th e
Bicyclo[2.2.0]h exa n e Der iva tive. The Cope rearrange-

1646 J . Org. Chem., Vol. 64, No. 5, 1999
Ikeda et al.
Ta ble 5. P h otosta tion a r y Ra tios of th e DCA-sen sitized P h otor ea ction of EZ-3a -c a n d m eso-4a -c in CD3CN, CD2Cl2, a n d
C6D6 a t 20 °Ca
CD3CN
CD2Cl2
C6D6
photostationary ratiosb
photostationary ratiosb
photostationary ratiosb
sub
EZ-3
meso-4
n,cis-5
yields (%)b
64c
EZ-3
meso-4
n,cis-5
yields (%)b
EZ-3
meso-4
n,cis-5
yields (%)b
EZ-3a
EZ-3b
EZ-3c
meso-4a
meso-4b
meso-4c
80
53
20
14
0
33
77
90
5
5
78c
27
75
10
35
7
38
18
3
96c
97d
82c
80
67
19
20
18
34
0
15
45
83
94
92
4
5
89
81
7
14
84c
100
87
a
A 0.5 mL solution was irradiated with a 2 kW Xe lamp through a cutoff filter (λ > 360 nm) at 20 °C. [sub] ) 0.1 M, [DCA] ) 5.3 ×
10-4 M. Determined by 1H NMR analyses. c Not photostationary mixture. A small amount of the dl-family compounds was formed.
b
d
Ta ble 6. P h otosta tion a r y Ra tios of th e DCA-Sen sitized P h otor ea ction of tr a n s-5a -c a t 20 °Ca
yields (%)b
sub
solvent
time (h)
ZZ-3
EE-3
dl-4
trans-5
meso-family compounds
note
trans-5a
trans-5b
trans-5c
trans-5c
CD2Cl2
CD2Cl2
CD2Cl2
CD3CN
20
8
14
10
0
0
0
0
30
6
5
48
24
46
4
0
51
18
87
0
c
d
d
d
6 (EZ-3b), 8 (n,cis-5b)
2 (meso-4c), 7 (n,cis-5c)
0
2
a
A 0.5 mL solution was irradiated with a 2 kW Xe lamp through a cutoff filter (λ > 360 nm) at 20 °C. [sub] ) 0.1 M, [DCA] ) 5.3 ×
10-4 M. Determined by 1H NMR analyses. c Photostationary mixture. Not photostationary mixture.
b
d
Sch em e 10
and n,cis-5, respectively (vide supra). As shown in Tables
4 and 5, the solvent polarity and the electron-donating
ability of 3-5 are important for the formation of trans-5
and n,cis-5. In a Cope rearrangement cycle, these bicyclo-
[2.2.0]hexanes are formed and collapsed in the diradical
and cation radical energy surfaces, respectively, and
thereby the solvent polarity and the electron-donating
ability of substrates affect more significantly cleavage of
5•+ to 21•+ than closure of 21 to 5, because the ring
cleavage of nondistonic 5•+ to distonic 21•+ involves
localization of the unpaired electron and positive charge.
Hence, the ring cleavage of 5•+ to 21•+ is facilitated much
more than closure of 21 to 5 with an increase in the
solvent polarity and the electron-donating ability of
substrates. Remarkable substituent effects on the bond
cleavage of bicumyl cation radical to the cumyl cation and
cumyl radical19 is a suitable example in favor of this
assumption. Thus, under the DCA-sensitized electron-
transfer conditions in acetonitrile, the highly electron-
in the ratio of 5:68:27 in dichloromethane, while in
acetonitrile EE-3c does not afford a clear photostationary
mixture.
donating trans-5a (Eox ) +1.00 V vs SCE in acetoni-
1/2
trile) and n,cis-5a (+0.87 V) cannot survive in a Cope
rearrangement cycle. In contrast, the ring cleavages of
trans-5b,c•+ and n,cis-5b,c•+ become relatively slow, and
thus 3b,c and 4b,c form less electron-donating trans-5b,c
(+1.38 V and +1.54 V) and n,cis-5b,c, (+1.19 V and +1.31
V), respectively, even in acetonitrile. In fact, the DCA-
sensitized photoreaction of trans-5a gives a photosta-
tionary mixture of EE-3a and dl-4a , but those of trans-
5b,c give photostationary mixtures of EE-3b,c, dl-4b,c,
and trans-5b,c (Table 6). Similar results are observed for
the DCA-sensitized photoreactions of n,cis-5a-c as shown
in Table 7.
Th e DCA-Sen sitized Electr on -Tr a n sfer P h otor e-
a ction s of tr a n s-6b a n d x,cis-6b. To gain further
insight into the formation of trans-5b and n,cis-5b in the
Cope photostationary mixture, the reactivities of the boat
intermediates tB-21b•+ and n,cB-21b•+ were investigated
by the independent generation from trans-6b and x,cis-
6b under the DCA-sensitized electron-transfer conditions.
As shown in Table 8, the DCA-sensitized electron-
transfer photoreaction of trans-6b gives trans-5b, EE-
A photostationary mixture from EZ-3a or meso-4a does
not include either n,cis-5a or x,cis-5a in dichloromethane,
but in nonpolar benzene meso-4a gives n,cis-5a (Scheme
10). As shown in Table 5, in dichloromethane EZ-3b gives
a 67:18:15 photostationary mixture of EZ-3b, meso-4b,
and n,cis-5b. The fact that the exo isomers, x,cis-5a -c,
are not formed from EZ-3a -c and meso-4a -c is consis-
tent with the assumption that the exo intermediate, x,cB-
21•+ shown in Scheme 8, does not participate in the Cope
rearrangements of EZ-3 and meso-4 as mentioned previ-
ously.18
Since bicyclo[2.2.0]hexanes are thermally labile, the
formation of bicyclo[2.2.0]hexanes has never been ob-
served in thermal reactions of 1,5-hexadienes. Conse-
quently, the formation of bicyclo[2.2.0]hexanes is the
most unique characteristic of the photoinduced electron-
transfer Cope rearrangement. The formation of bicyclo-
[2.2.0]hexanes can be explained either by the cation
radical closure of tB-21•+ and n,cB-21•+ or by the diradical
closure of tB-21 and n,cB-21 as shown in Schemes 6 and
7. The diradical closure mechanism is favorable, and in
fact direct irradiation of trans-6 and x,cis-6 give trans-5
(19) Maslak, P.; Asel, S. L. J . Am. Chem. Soc. 1988, 110, 8260-
8261.

Photoinduced Electron-Transfer Cope Rearrangements
J . Org. Chem., Vol. 64, No. 5, 1999 1647
Ta ble 7. P h otosta tion a r y Ra tios of th e DCA-Sen sitized
Sch em e 11
P h otor ea ction of n ,cis-5a in CD2Cl2 a t 20 °Ca
yields (%)b
sub
time (h) EZ-3 meso-4 n,cis-5 dl-family compounds note
n,cis-5a
n,cis-5b
n,cis-5c
10
3
8
81
59
21
19
16
32
0
21
38
0
0
c
d
d
7 (dl-4c), 3 (ZZ-3c)
a
A 0.5 mL solution was irradiated with a 2 kW Xe lamp through
a cutoff filter (λ > 360 nm) at 20 °C. [sub] ) 0.1 M, [DCA] ) 5.3
b
× 10-4 M. Determined by 1H NMR analyses. c Photostationary
d
mixture. Not photostationary mixture.
Ta ble 8. Dea zeta tion of tr a n s-6b (1) a n d x,cis-6b (2)
u n d er P h otoin d u ced Electr on -Tr a n sfer a n d Dir ect
Ir r a d ia tion Con d ition s a t 20 °Ca
(1)
relative yields (%)b
sub
conditions
convn (%)b
EE-3
dl-4
trans-5
trans-6b
hνsens/DCAc
hνCT/TCNBc
hν (direct)d
9
21
21
37
42
42
25
26
30
38
32
28
(2)
relative yields (%)b
convn (%)b EZ-3 meso-4 n,cis-5
sub
conditions
x,cis-6b
hνsens/DCAc
hνCT/TCNBc
hν (direct)d
24
17
18
0
0
0
0
0
0
100
100
100
a
A 0.5 mL of CD2Cl2 solution was irradiated with a 2 kW Xe
lamp through a cutoff filter. [6b] ) 0.01 M. Determined by 1H
b
NMR analyses. c λ > 410 nm. λ > 360 nm.
d
3b, and dl-4b in the ratio of 38:37:25 at 9% conversion
in dichloromethane-d2 as well as photoexcitation of the
charge-transfer complex of x,cis-6b and 1,2,4,5-tetracy-
anobenzene (TCNB). Upon prolonged irradiation, a pho-
tostationary mixture of these three compounds is formed
in nearly the same ratio as those from EE-3b, dl-4b, and
trans-5b (Table 4). Interestingly, direct irradiation of
trans-6b in dichloromethane-d2 gives trans-5b together
with EE-3b and dl-4b in the ratio of 28:42:30 at 21%
conversion (Table 8). Thus, under the photosensitized
electron-transfer conditions, both trans-5b and dienes
(EE-3b, dl-4b) are probably formed from tB-21b formed
by BET from DCA•- to tB-21b•+. A ring-closure of tB-21b
affords the former whereas the latter is given by ring flip
of tB-21b followed by cleavage of aaC-21b. Alternatively,
dienes (EE-3b, dl-4b) may be formed by ring flip of tB-
21b•+ followed by BET and cleavage as shown in Scheme
11. However, results of the reactions under the photo-
sensitized and directly irradiated conditions and calcu-
lated rate constant, kbet ) 2 × 1010 s-1, in dichlo-
romethane,20,23 suggest that the rapid BET takes place
in [tB-21b•+/DCA•-] to afford tB-21b in preference to the
ring flip.
In contrast to photoreaction of trans-6b, n,cis-5b is
exclusively formed in similar DCA-sensitized photoreac-
tion of x,cis-6b and photoexcitation of the charge-transfer
complex of x,cis-6b and TCNB as shown in Table 8. Thus,
under the photoinduced electron-transfer conditions the
BET from DCA•- to n,cB-21b•+ followed by closure of
n,cB-21b to x,cis-5b occur much faster than flip of n,cB-
21b•+ to aeC-21b•+. The observed result is consistent with
the Cope rearrangement sequence shown in Scheme 7.
(20) Rate constant, kbet, of the BET in [21•+/DCA•-] at 20 °C in
dichloromethane was estimated by using the following eqs. (1, 2,21 and
3) and parameters reported by Kikuchi.22
(21) Miller, J . R.; Beitz, J . V.; Huddleston, R. K. J . Am. Chem. Soc.
1984, 106, 5057-5068. Siders, P.; Marcus, R. A. J . Am. Chem. Soc.
1981, 103, 741-747, 748-752. Van Duyne, R. P.; Fischer, S. F. Chem.
Phys. 1974, 5, 183-197. Ulstrup, J .; J ortner, J . J . Chem. Phys. 1975,
63, 4358-4368.
(λs + ∆Gbet + ωhν)2
4λskbT
1/2
∞
4π3
e-sSω
2
kbet
)
|V|
exp
-
(1)
∑
(
)
(
) {
}
h2λskbT
ω!
ω)0
(22) Niwa, T.; Kikuchi, K.; Matsushita, N.; Hayashi, M.; Katagiri,
T.; Takahashi, Y.; Miyashi, T. J . Phys. Chem. 1993, 97, 11960-11964.
(23) These calculations have been done on the assumption that
Eox of tB-21b and aaC-21a are comparable with Eox of the 4-
S ) λv/hν
(2)
(3)
1/2
1/2
methylcumyl radical24 (+0.03 V vs SCE in acetonitrile) and 4-meth-
∆Gbet (eV) ) -[Eox1/2(21) - Ered1/2(DCA) - e2/ꢀr]
oxycumyl radical25 (-0.14 V), respectively, and that Eox in dichlo-
1/2
2
where parameters |V| , λs, λv, ν, and ∆Gbet are respectively an electronic
matrix element squared, solvent reorganization energy, vibrational
reorganization energy, single average frequency, and free energy
change for BET process.23 Ered1/2(DCA) is -0.89 and -0.95 V vs SCE
in dichloromethane and acetonitrile, respectively. The Coulombic term
(e2/ꢀr) is 0.23 eV in dichloromethane while this term is ignored in
acetonitrile.11c
romethane are generally more positive by ca. 0.1 V than those in
acetonitrile. These estimations give higher limits of ∆Gbet because Eox
1/2
of 21 may be more positive. This was suggested from a significant
electronic coupling between the substituted cumyl cation and cumyl
radical parts in a cyclohexane-1,4-diyl cation radical.1a,d Unfortunately,
however, it is difficult to evaluate this electronic coupling effects upon
redox potentials.

1648 J . Org. Chem., Vol. 64, No. 5, 1999
Ikeda et al.
Ta ble 9. Decon volu tion F ittin g P a r a m eter sa of P AC
An a lyses for Deter m in a tion of En er gy of th e Ion Ra d ica l
P a ir , ∆Hir p ([a a C-21a •+/DCA•-])
∆Hirp([aaC-21a •+
/
sub
R1
R2
τ2 (ns)
DCA•-]) (kcal/mol)b
ZZ-3a
EE-3a
dl-4a
0.25 ( 0.02 0.24 ( 0.02 205 ( 25
0.26 ( 0.03 0.22 ( 0.02 160 ( 53
0.25 ( 0.01 0.26 ( 0.00 191 ( 16
44.9 ( 2.1
46.3 ( 1.5
43.0 ( 0.4
29.3 ( 1.5
trans-5a 0.28 ( 0.00 0.40 ( 0.01 219 ( 11
a
b
The errors are 1σ. Face values relative to each substrate and
DCA.
Upon prolonged irradiation n,cis-5b enters into a Cope
rearrangement cycle to give a Cope photostationary
mixture as shown in Table 7. The fact that direct
irradiation of x,cis-6b forms n,cis-5b exclusively supple-
ments a diradical closure mechanism shown in Scheme
7.
P h otoa cou stic Ca lor im etr ic An a lysis a n d En er -
getics of th e P h otoin d u ced Electr on -Tr a n sfer Cop e
Rea r r a n gem en ts of EE-3a a n d d l-4a . Nanosecond
time-resolved photoacoustic calorimetry (PAC)26 allows
for the simultaneous determination of dynamics and
energetics of various photoinduced electron-transfer
reactions.1a,15,27 In fact, our recent PAC experiments
demonstrated that a CRCY-DRCL mechanism (Scheme
1) is energetically most favorable for the photoinduced
electron-transfer degenerate Cope rearrangement of d4-
1.1a For further confirmation of this mechanism, we have
used PAC to investigate the Cope rearrangement of EE-
3a and dl-4a .
F igu r e 1. Potential energy diagram for the DCA-sensitized
PET Cope rearrangement of EE-3a and dl-4a . Relative energy
was represented in kcal/mol.
statistic value, 47.3 kcal/mol is used for ∆Hirp([aaC-21a •+
/
DCA•-]).
By using redox potentials of EE-3a , dl-4a , and DCA,
the energies of ion pair [EE-3a •+/DCA•-] and [dl-4a •+
/
DCA•-] are calculated to be 58.1 and 47.5 kcal/mol,
respectively. The CRCY of dl-4a •+ to aaC-21a •+ is thus
about 11 kcal/mol exothermic. In contrast, the energy of
[EE-3a •+/DCA•-] is close to that of [aaC-21a •+/DCA•-],
and hence the direct cleavage of aaC-21a •+ to EE-3a •+
seems to be energetically feasible but perhaps slower.
However, BET to form diyl aaC-21a is assumed to occur
faster than cleavage to EE-3a •+. To judge experimentally
whether cleavage of aaC-21a •+ to EE-3a •+ is operative,
aaC-21a •+ was independently generated from trans-5a ,
and temperature effects on product ratios from aaC-21a •+
were examined at low conversion in the temperature
range between 60 and -80 °C. The product ratios, dl-
4a :EE-3a , at low conversion does not depend on temper-
ature, and the average ratio is near 32:68.4 This obser-
vation suggests that cleavage of aaC-21a •+ to EE-3a •+ is
not thermally accelerated and thus unlikely to occur in
a Cope rearrangement sequence. Thus, aaC-21a •+ gener-
ated from EE-3a •+ and dl-4a •+ preferentially suffers BET
to give aaC-21a , through which EE-3a and dl-4a are
formed. Using ∆Gbet for the BET from DCA•- to aaC-21a •+
to be about 19 kcal/mol,20,23 aaC-21a is about 29 kcal/
mol higher in energy than EE-3a .
Experiments were done under DCA-BP-cosensitized
conditions in acetonitrile.1a,15,27e The enthalpy of forma-
tion of [aaC-21a •+/DCA•-] can be expressed by ∆Hirp([aaC-
21a •+/DCA•-]) ) hν(1 - R1 - R2)/φ and φ ) hν(1 - R1)/
E([BP•+/DCA•-]), where hν, R, φ, and E([BP•+/DCA•-]) are
the photon energy (415 nm, 68.9 kcal/mol), the deconvo-
lution parameters, the quantum yield to form [BP•+
/
DCA•-], and the energy (66.2 kcal/mol) of [BP•+/DCA•-]
determined from redox potentials of BP and DCA,
respectively. ∆Hirp was determined from several experi-
ments for EE-3a and dl-4a , and shown in Table 9
together with another deconvolution parameter, τ2. The
energy of ion radical pair, ∆Hirp([aaC-21a •+/DCA•-]),
relative to dl-4a and DCA is determined to be 43.0 ( 0.4
kcal/mol. By adding this value to the difference in heat
of formation between dl-4a and EE-3a , 6.0 ( 1.5 kcal/
mol, determined by differential scanning calorimetry,
∆Hirp([aaC-21a •+/DCA•-]) relative to EE-3a and DCA is
estimated to be 49.2 ( 1.9 kcal/mol which corresponds
to the experimental value, 46.3 ( 1.5 kcal/mol, obtained
from EE-3a within experimental errors. In Figure 1, the
Con clu sion . Observed stereospecificity and energetics
based on PAC support a CRCY-DRCL mechanism of the
Cope rearrangement for the title dienes. The initial
cyclization of the diene cation radicals occurs in the chair
conformation to give the chair 1,4-diaryl-2,3-dimethyl-
cyclohexane-1,4-diyl cation radical intermediates. Sub-
sequent cleavage of the chair 1,4-diaryl-2,3-dimethylcy-
clohexane-1,4-diyl intermediates regenerates the neutral
dienes. These two processes are connected by the highly
exothermic BET from DCA•- to the 1,4-diaryl-2,3-di-
methylcyclohexane-1,4-diyl cation radicals. A CRCY-
DRCL mechanism is thus a general rearrangement
sequence for the photoinduced electron-transfer Cope
rearrangement of the 2,5-diaryl-1,5-hexadiene system.
The chair intermediates interconvert with the boat
(24) Eox1/2 of the 4-methylcumyl radical is estimated by interpolation
from the correlation between the known Eox of 4-substituted cumyl
1/2
radicals25 and the substituent constants σ+.
(25) Sim, B. A.; Milne, P. H.; Griller, D.; Wayner, D. D. M. J . Am.
Chem. Soc. 1990, 112, 6635-6638.
(26) Rudzki, J . E.; Goodman, J . L.; Peters, K. S. J . Am. Chem. Soc.
1985, 107, 7849-7854. Herman, M. S.; Goodman, J . L. J . Am. Chem.
Soc. 1989, 111, 1849-1854. Peters, K. S. In Kinetics and Spectroscopy
of Carbenes and Biradicals; Platz, M. S., Ed.; Plenum: New York, 1990;
pp 37-49. Griller, D.; Wayner, D. D. M. Pure Appl. Chem. 1989, 61,
717-724.
(27) (a) Rothberg, L. J .; Simon, J . D.; Bernstein, M.; Peters, K. S. J .
Am. Chem. Soc. 1983, 105, 3464-3468. (b) Goodman, J . L.; Peters, K.
S. J . Am. Chem. Soc. 1986, 108, 1700-1701. (c) Ci, X.; Silva, R. S. da;
Goodman, J . L.; Nicodem, D. E.; Whitten, D. G. J . Am. Chem. Soc.
1988, 110, 8548-8550. (d) LaVilla, J . A.; Goodman, J . L. J . Am. Chem.
Soc. 1989, 111, 712-714. (e) Zona, T. A.; Goodman, J . L. Tetrahedron
Lett. 1992, 33, 6093-6096.

Photoinduced Electron-Transfer Cope Rearrangements
J . Org. Chem., Vol. 64, No. 5, 1999 1649
410 nm) for 6 under N2 at 20 ( 1 °C. Time-dependent changes
of product ratios were determined by 200 MHz 1H NMR
analyses.
intermediates on the cation radical and diradical energy
surfaces in a Cope rearrangement cycle, but the boat diyl
intermediates formed by BET never participate in the
cleavage step, perhaps because of poor orbital overlap.
The boat intermediates, however, play an important role
in an unusual Cope rearrangement yielding the bicyclo-
[2.2.0]hexane derivative in the photostationary mixture.
The concurrent formation of bicyclo[2.2.0]hexane deriva-
tives with the Cope rearrangement is thus the most
remarkable characteristic of the photoinduced electron-
transfer Cope rearrangement of the title dienes.
DCA-Sen sitized P h otor ea ction s of 3-5 u n d er O2 in
Va r iou s Solven ts. A general procedure: A 5 mL solution
containing 0.05 mmol of substrate (3-5) (0.01 M) and 2-3 mg
of DCA in CH3CN, CH2Cl2, or C6H6 was irradiated with a 2
kW Xe lamp through a Toshiba cutoff filter L-39 (λ > 360 nm)
under O2 at 20 ( 1 °C. Removal of solvent and PTLC followed
by recrystallization afforded cyclic peroxide 22. Physical data
of 22 are shown in Supporting Information.
P h otoexcita tion of th e Ch a r ge-Tr a n sfer Com p lex of
6b a n d TCNB. A 0.5 mL CH2Cl2 solution containing 15.9 mg
(0.05 mmol) of 6b and ca. 1 mg (ca. 0.005 mmol) of TCNB was
irradiated with a 2 kW Xe lamp through a Toshiba cutoff filter
Y-44 (λ > 410 nm) under N2 at 20 ( 1 °C. The product ratios
at low conversion were determined by 200 MHz 1H NMR
analyses.
Exp er im en ta l Section
Gen er a l Meth od . See the Supporting Information.
Syn th eses of Su bstr a tes Sh ow n in Ch a r t 1. See the
Supporting Information.
Dir ect Ir r a d ia tion of 6b. A 0.5 mL CD2Cl2 solution
containing 15.9 mg (0.05 mmol) of 6b in CD2Cl2 was irradiated
with a 2 kW Xe lamp through a Toshiba cutoff filter L-39 (λ >
360 nm) under N2 at 20 ( 1 °C. The product ratios at low
conversion were determined by 200 MHz 1H NMR analyses.
En er gy Deter m in a tion by Tim e-Resolved P AC. The
detail of PAC experiment has been described previously.1a,26
A chart of the PAC waveforms for EE-3a -DCA-BP system and
lists of deconvolution fitting parameters for experimental
waveforms in each PAC experiment are shown in the Sup-
porting Information. Values were obtained by at least two
separate runs and errors are with in 1σ.
X-r a y Str u ctu r a l An a lyses. All of data collections of trans-
5c, n,cis-5c, and x,cis-5c were performed on an AFC-5R
automated four-circle diffractometer (45 kV, 200 mA) equipped
with a rotating anode (Mo KR radiation, λ ) 0.71049 Å or Cu
KR radiation, λ ) 1.5418 Å) at the Instrumental Analyses
Center for Chemistry, Graduate School of Science, Tohoku
University. These structure were solved by direct method using
the RANTAN8128 program with some modification. Atomic
parameters were refined by the block-diagonal least-squares
methods by applying anisotropic temperature factors for non-
hydrogen atoms. At the final stage, hydrogen atoms were
included in the refinement with isotropic temperature factors.
All the calculations were carried out on a ACOS 2020 and
ACOS 3900 computers at Tohoku University by using the
applied library program of the UNICS III system.29 Details of
X-ray structural analyses, ORTEP diagrams, final atomic
coordinates and thermal parameters, bond length and angles,
structure factors, and thermal ellipsoids with atom numbering
systems are given in the Supporting Information.
Ack n ow led gm en t. We gratefully acknowledge fi-
nancial support from the Ministry of Education, Science,
Sports and Culture (Grant-in-Aid for Scientific Research
Nos. 03303001, 03403005, 05740439, and 08740560).
An a lyses of Tim e-Dep en d en t Ch a n ge of P r od u ct Ra -
tios for th e DCA-Sen sitized P h otor ea ction s of 3-6. A
typical procedure: A 0.1 M solution (0.5 mL) containing 0.05
mmol of substrate (3-6) and 2-3 mg of DCA in CD3CN, CD2-
Cl2, or C6D6 was irradiated with a 2 kW Xe lamp through a
Toshiba cutoff filter L-39 (λ > 360 nm) for 3-5 or Y-44 (λ >
Su p p or tin g In for m a tion Ava ila ble: Syntheses of sub-
strates shown in Chart 1, details of X-ray structural analyses
of trans-5c, n,cis-5c, and x,cis-5c, a table of Eox1/2, ∆G, and kq
for 3-6, physical data of trans-22a ,b, x,cis-22a ,b, and n,cis-
22a , a chart of the PAC waveforms (EE-3a -DCA-BP system
in CH3CN), tables of deconvolution fitting parameters for
experimental waveforms in each PAC experiment, and a table
of differential scanning calorimetric analysis. This material
is available free of charge via the Internet at http://pubs.acs.org.
(28) (a) J ia-Xing, Y. Acta Crystallogr. Sect. A 1983, 39, 35-37. (b)
J ia-Xing, Y. Acta Crystallogr. Sect. A 1981, 37, 642-644.
(29) Sakurai, T.; Kobayashi, K. Rikagaku Kenkyusyo Hokoku 1979,
55, 69-77.
J O981775H