Efficient electrocatalytic intramolecular anion radical cyclobutanation reactions

Article abstract of DOI:10.1016/j.tet.2004.08.088

Electrochemically initiated, intramolecular anion radical cyclobutanations of bis(enones) and related substrates are presented. The formation of novel anion radical Diels-Alder adducts in minor amounts is also observed. Total yields of pericyclic products, which include both cis- and trans-cyclobutanes and a single Diels-Alder adduct, are generally high (51-88%), with electrocatalytic factors in the range of 1.5-5. Mechanistically, strong evidence for the intervention of distonic anion radical intermediates as precursors of both types of pericyclic products is presented. The scope and limitations of these reactions are rather extensively explored and defined, and in particular the tendency, in some cases, for electrogenerated base-catalyzed reactions to compete with these anion radical pericyclic reactions.

Full text of DOI:10.1016/j.tet.2004.08.088

Tetrahedron 60 (2004) 10999–11010
Efficient electrocatalytic intramolecular anion radical
cyclobutanation reactions
*
Greg A. N. Felton and Nathan L. Bauld
Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, TX 78712, USA
Received 3 June 2004; revised 18 August 2004; accepted 26 August 2004
Available online 2 October 2004
Abstract—Electrochemically initiated, intramolecular anion radical cyclobutanations of bis(enones) and related substrates are presented.
The formation of novel anion radical Diels–Alder adducts in minor amounts is also observed. Total yields of pericyclic products, which
include both cis- and trans-cyclobutanes and a single Diels–Alder adduct, are generally high (51–88%), with electrocatalytic factors in the
range of 1.5–5. Mechanistically, strong evidence for the intervention of distonic anion radical intermediates as precursors of both types of
pericyclic products is presented. The scope and limitations of these reactions are rather extensively explored and defined, and in particular the
tendency, in some cases, for electrogenerated base-catalyzed reactions to compete with these anion radical pericyclic reactions.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction
potential utility. The present paper reports the development
of substantially more efficient conditions than those
previously reported for carrying out these electrocatalytic
intramolecular cycloaddition reactions in high yields, and
further extends the scope and defines the limitations of these
reactions. In particular, the use of tetraalkylammonium
tetrafluoroborates as electrolytes in acetonitrile solution is
developed as a particularly efficient method for these
cyclobutanations. From a mechanistic viewpoint, experi-
ments are described which strongly support the formulation
of the cycloaddition reaction as proceeding in a stepwise
fashion, via a distonic anion radical intermediate. Other new
mechanistic and theoretical aspects of the reaction are
clarified, including the requirement of at least one aroyl
group, and the preference for two aroyl moieties.
Cyclobutanation reactions of the cation radicals of alkenes
with neutral alkenes have by now become rather common-
place and are characterized by impressively high cyclo-
addition rates and low activation barriers, especially in
comparison to the corresponding thermal reactions.1,2 There
have been some recent indications that this extensive body
of cation radical cyclobutanation chemistry may have a
close counterpart in the domain of anion radical chemistry.
Specifically, the reduction of phenyl vinyl sulfone under
electrochemical conditions (mercury pool cathode) has been
reported to yield trans-1,2-bis(phenylsulfonyl)cyclo-
butane.3 Subsequently, the cyclodimerizations of a variety
vinylpyridines and vinylquinolines under similar conditions
have also been established.4 Still more recently, a few
intramolecular anion radical cyclobutanations of tethered
bis(enones) have been described from these laboratories.5,6
These anion radical reactions are of special interest because
they represent rare examples of intramolecular anion radical
cycloaddition, rather than the more common electrohydro-
cyclization/dimerization (EHC or EHD).7,8
2. Results and discussion
2.1. Electrolysis of 1a
The cyclobutanations of substrates 1a and 1g (Scheme 1)
have previously been reported from these laboratories.
Using lithium perchlorate (0.1 M) as the electrolyte, the
reaction of 1a was found to afford a total yield of pericyclic
products (2, 3) of just 45%, and the ratio of trans-2/cis-2/3
was found to be 2.4:2.0:1. The total yield of pericyclic
products obtained from 1g was even smaller (35%).5
Consequently, various electrolytes (and other procedural
changes) were investigated in the present work in order to
determine whether synthetically useful procedures might be
developed for carrying out these novel reactions.
The environmentally benign nature of these electrochemical
conversions, inherent in the simplified workup and the
consumption of electricity as the sole reagent (in catalytic
amounts), further add to their experimental appeal and
Keywords: Anion radical; Cyclobutanation; Catalytic; Electrochemical
reduction; Pericyclic; Electrogenerated base.
* Corresponding author. Tel.: C1-512-471-3017; fax: C1-512-471-8696;
e-mail addresses: bauld@mail.utexas.edu; greg_felton@hotmail.com
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.08.088

11000
G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
Scheme 1. Pericyclic products in the electrocatalytic, intramolecular anion radical cyclobutanation reactions of various bis(enones).
2.1.1. The anion radical mechanism. The anion radical
chain cycloaddition mechanism proposed for these reactions
is rather novel and is summarized in Scheme 2. According
to this mechanism, reduction of the substrate at the cathode
leads to a substrate anion radical, which then cyclizes to a
distonic anion radical intermediate. This intermediate
cyclizes to the anion radical of the cyclobutane product.
The anion radical moiety presumably resides upon one of
the benzoyl groups. Finally, exergonic electron transfer
(ET) from the product anion radical to a molecule of the
neutral substrate occurs, setting up the chain process and
affording the neutral cyclobutane product. The distonic
anion radical intermediate can also cyclize to a Diels–Alder
adduct anion radical, where the anion radical moiety again
resides upon a benzoyl moiety.
interest in that it represents the first documented instance of
an anion radical Diels–Alder reaction. Since complete
conversion of 1a to products is accomplished after a
maximum9 of 21% of the theoretical charge had flowed, the
reaction is mildly electrocatalytic, with a catalytic factor of
4.7 (representing a 0.21 F molK1 process).
2.2. Electrolysis of 1b
2.2.1. Effect of electrogenerated bases. The extension of
this chemistry to the ethereal substrate 1b was investigated
next. Under conditions similar to the earlier work (using
lithium perchlorate as the electrolyte)5 a total yield of 43%
of pericyclic products was isolated, with trans-2b being the
major product. Under both the present and former
conditions the catalytic factor was !3. When tetrabutyl-
ammonium tetrafluoroborate was used as the electrolyte,
pericyclics were isolated in 53% yield. Interestingly, this
reaction proceeded somewhat more efficiently in terms of
the catalytic factor (10.6) than any of the other reactions, but
the yield of pericyclic products was diminished in
comparison to the corresponding reduction of substrate 1a
when using the same electrolyte because of a competing
reaction of the substrate (Scheme 3). The formation of 4b
implies a base catalyzed deprotonation of 1b, followed by a
Michael type addition of the conjugate base via the alpha
carbon of the extended enolate to the beta carbon of the
enone moiety. Formation of this product in the reaction
using tetrabutylammonium tetrafluoroborate as the electro-
lyte, and not with any of the other electrolytes, is in
2.1.2. Tetraalkylammonium tetrafluoroborate electro-
lytes. The use of tetraalkylammonium tetrafluoroborates as
electrolytes was explored on the assumption that ion pairing
to the tetraalkylammonium cation should be much looser
than with the lithium ion, resulting in a possible increase in
reactivity of the anion radical intermediates. In fact, in the
case of 1a, a dramatic increase in yield to 88% was observed
under these conditions. As will be noted in Table 1, the yield
of the trans cyclobutane product (trans-2) is elevated to
59%. Since the cis isomer (obtained in 17% yield) is readily
isomerized to the trans isomer under acidic or basic
conditions, this reaction sequence makes the trans isomer
available in an overall 71% yield. The novel Diels–Alder
product (3), obtained in 13% yield, is also of inherent
Scheme 2. Proposed mechanism for the anion radical chain cyclobutanation reaction of 1,7-dibenzoyl-1,6-heptadiene.

G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
11001
Table 1. Yields and catalytic factors for the electrocatalytic intramolecular anion radical pericyclic reactions of various bis(enone) substrates (1)
Substrate
Procedurea
Yield of trans-2
Yield of cis-2
Yield of 3
Total pericyclic
yield
Other products
Catalytic factor
1a
1b
1b
1b
1c
1d
1d
1e
1f
A
A
B
C
A
BCC
D
A
C
A
C
D
B
59%
39
21
21
52
7
3
29
7
17%
11
20
39
11
33
0
14
21
0
13%
3
88%
53
43
88
75
56
3
51
53
12
9
—
4.7
10.6
1.6
1.8
5.1
1.3
7.3
5.2
1.9
1.4
!1
1.8
!1
4.5
4b; 11%
—
—
2.0
28
12
16
0
8
25
0
0
0
0
0
5c; 5%
6d; 6%
5d; 17%b
5e; 17%
5f; 13%, 7f; 7%
5f; 17%
7g; 30%c
—
1f
12
0
1g
1g
1i
9
0
0
0
19d
32e
0
19
32
0
—
5j; 42%
1j
D
a
The electrolytes used in the respective procedures were: AZ0.1 M tetrabutylammonium tetrafluoroborate; BZ0.1 M lithium perchlorate; CZ0.1 M mag-
nesium perchlorate; DZ0.1 M tetraethylammonium tetrafluoroborate.
Also obtained 35% of an unidentified polymer.
Two isomers.
Both trans isomers were observed: 12 and 7% (w2:1 ratio), X-ray of major isomer provided.
Both trans isomers were observed: 28 and 4% (7:1 ratio).
b
c
d
e
agreement with the postulate that the intermediate anionic
optimal efficiency, in the specific case of 1b the competing
reaction described above tended to lower the yields of the
pericyclic products. In view of that, it appeared worthwhile
to investigate the effect of using a more strongly ion-pairing
counterion than even the lithium ion, that is, the magnesium
ion. When the electrochemical reaction was carried out in
the presence of magnesium perchlorate as the electrolyte, an
88% yield of pericyclic products was obtained. The increase
in the cis/trans ratio in the cyclobutane product may also
reflect the stronger ion pairing between the magnesium ion
and the anion radical moiety in the transition state for the
final cyclization step (vide infra).
species are less tightly ion paired under these conditions,
and are therefore substantially more reactive (basic). Since a
product analogous to 4b was not observed in the electro-
chemical reaction of 1a, it is apparent that the substitution of
oxygen for carbon tends to acidify the adjacent C–H bond.
A similar competition has been reported in the EHC of
butenolides.10 In the case of substrate 1b, a product
corresponding to the isomerization of the double bond,
into conjugation with the carbonyl group, was not observed.
However, in several other instances products corresponding
to such a structure, 5,11 were obtained instead of 4. The
nature of the electrogenerated base (EGB)12,13 species
responsible for the deprotonation is unknown, but likely
candidates might be a substrate or product anion radical or a
dianionic intermediate produced by further reduction of the
proposed distonic anion radical intermediate.
2.3. Electrolysis of 1c
It had been found in the previously reported work5 that the
introduction of electron donating substituents, such as 4-
methoxy, onto the aryl ring of 1a sharply inhibits
cyclobutanation. It is presumed that this is the result, at
least in part, of the inability of the distonic anion radical
intermediate to cyclize to the product cyclobutane anion
radical, which would require the anion radical moiety to
2.2.2. Use of Mg(ClO4)2 electrolyte. Although the use of
tetraalkylammonium tetrafluoroborates as the electrolyte
appears to be the more general method of choice for
carrying out most of these anion radical cycloadditions with
Scheme 3. The electrogenerated base (EGB) catalyzed products (4 and 5) obtained in the electrochemical reduction of substrates 1, in the presence of
tetraalkylammonium tetrafluoroborate as the electrolyte.

11002
G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
Scheme 4. Formation of an aldol-type side product (6), via reduction of the distonic anion radical to an enolate dianion.
reside in a higher energy SOMO (that of a 4-methoxy-
benzoyl moiety) than in the case of the unsubstituted
substrate. Consequently, the facility of bis(enone) anion
radical cyclobutanation of substrates that would provide a
lower energy product anion radical SOMO was investi-
gated. Substrate 1c, which has electron withdrawing
4-chloro substituents on both of its benzoyl groups, was
found to undergo smooth anion radical cycloaddition under
the tetrabutylammonium tetrafluoroborate electrolyte con-
ditions, affording a 75% yield of total pericyclics, of which
52% was the trans cyclobutane. It is of special interest that
the chlorine substituent is retained in the product, even
though the 4-chlorobenzoyl anion radical moiety was
potentially susceptible to chloride ion loss. Presumably,
electron transfer from the product anion radical to substrate
is sufficiently rapid as to suppress the potential loss of
chloride ion.
subsequently is protonated and undergoes cyclization
(Scheme 4).
2.5. Electrolysis of 1e
The replacement of the phenyl group with a b-naphthyl or
4-biphenylyl group could also provide a more suitable
venue for the stabilization of the anion radical of the product
than in the case of 1a, but again the lowering of the SOMO
energy and the shifting of the SOMO density toward the
aroyl moiety could retard the first cyclization step. The
cyclobutanation of substrate 1e (which has b-naphthyl
substituents), using tetrabutylammonium tetrafluoroborate
as the electrolyte, was found to proceed, albeit in moderate
yield (51%), suggesting that the predominant effect of the
increased delocalization provided by the naphthyl group is
the lowering of the SOMO energy and the shift in density
away from the alkene linkage. However, the retardation of
the cyclization is evidently much less than in the case of the
dichloro substrate 1d. As in this latter case, the base-
catalyzed cyclization/isomerization product 5e was also
formed (17%; Table 2). It is worth noting that when the
reaction is not run to completion (w80% complete), the
yield of pericyclic products is relatively unchanged, but the
amount of the base catalyzed product is greatly reduced.
This may well be indicative of a build up of electro-
generated bases within the solution, accelerating the
cyclization to 5 in the latter stages of the reaction. This
build up of base could also provide an explanation for the
increase in the trans/cis ratio observed in the more complete
reaction (vide infra).
2.4. Electrolysis of 1d
The introduction of an even more strongly electron with-
drawing meta chloro substituent into the substrate, in
addition to the para chloro substituent, as in substrate 1d
would be expected to accelerate the rate of the second
cyclization step even further by providing a lower energy
SOMO for the product anion radical. However, the
considerably increased stabilization of the substrate anion
radical, and the expected shift in electron density in the
SOMO from the alkene moiety to the now rather strongly
electron-deficient aroyl ring could also be expected to have
an adverse effect upon the rate of the first cyclization step. In
accord with the latter idea, when 1d is electrolytically
reduced in the presence of the tetraethylammonium
tetrafluoroborate electrolyte, the yield of pericyclic products
falls precipitously to 3%, the main products consisting of a
17% yield of a base-catalyzed cyclization/isomerization
product (5d) and an uncharacterized polymer (35%).
However, in the presence of a mixed lithium perchlorate/
magnesium perchlorate electrolyte, the desired pericyclic
reactions are observed to occur with moderate efficiency,
but the yields of the pericyclic products are lower than in the
case of 1c, and an aldol-type product (6d) is also formed.
This latter product is considered to result from reduction of
the intermediate distonic anion radical to a dianion, which
2.6. Electrolysis of 1f
Substrate 1f, which contains 4-biphenylyl moieties,
responds in a manner rather similar to that of 1d. Limited
trans-2f formation is observed, along with greater formation
of the base-catalyzed cyclization/isomerization product 5f
(tetraalkylammonium electrolyte). However the use of
Mg(ClO4)2 afforded a moderate pericyclic yield of 53%.
Substrate 1f has limited solubility in acetonitrile, so that
these reductions were carried out in 1:1 THF/acetonitrile
solutions.
Table 2. The effect of the extent of reaction upon the yields of anion radical and base-catalyzed products
Substratea
Yield of trans-2
Yield of cis-2
Yield of 3
Total pericyclic
yield
Other products
Catalytic factor
1e
1eb
29
17
14
21
8
10
51
48
5e; 17%
5e; 6%
5.2
5.4
a
The electrolyte was 0.1 M tetrabutylammonium tetrafluoroborate.
Reaction run to 79% completion (based upon recovered 1e), yields and catalytic factor are corrected for 1e recovery.
b

G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
11003
2.7. SOMO requirements for cyclization
unsaturated ester function. This higher energy anion radical
could then rapidly cyclize to the reactive benzoyl ene
function. The key factor here is that catalysis would
necessarily involve electron transfer from a product anion
radical to a substrate molecule, and this chemical electron
transfer undoubtedly is highly selective for formation of the
more stable, and apparently unreactive, anion radical
corresponding to the benzoyl enone function. Additionally,
this chemical electron transfer is likely to be irreversible in
nature, halting catalysis. The chemically reduced substrate
may prove reactive in non-cyclobutanation mechanisms,
which may account for the observed low yield.
2.7.1. Electrolysis of 1g and 1h. In contrast to the aroyl
groups considered above, acetyl groups provide a much less
extensively delocalized, higher energy SOMO for the
product cyclobutane or Diels–Alder anion radical. The
potential anion radical pericyclic chemistry of substrate 1h,
which has two acetyl substitutents, was therefore not
expected to be as efficient as when aroyl groups are present.
In accord with this supposition, no pericyclic products at all
could be detected in the electroreduction of this substrate.
However, since the anion radical of the pericyclic products
only requires (and perhaps can only utilize) a single aroyl
moiety, it was considered likely that pericyclic chemistry
might occur with unsymmetrical substrate 1g, which has
one benzoyl and one acetyl substituent. The previously
reported results, using lithium perchlorate as an electrolyte,
have already confirmed this conjecture, but further experi-
ments were carried out in the present work in connection
with magnesium perchlorate and tetraethylammonium
tetrafluoroborate as the electrolytes. The observed yields
in both cases are quite modest. However, an additional
feature of interest emerges when tetraethylammonium
tetrafluoroborate is used as the electrolyte, namely, that
both possible trans-2g isomers are formed, in a 2:1 ratio,
with the major cyclobutane having the benzoyl group syn to
the cyclopentane ring.
2.7.3. Diminished cyclization rates. Since the desired
pericylic reaction products provide the required low energy
SOMO associated with an aroyl function (in this case,
benzoyl), and the starting substrate provides a readily
reducible aroyl enone function, it appears likely that the
lower efficiency of the desired pericyclic chemistry in the
substrates which contain only a single aroyl function must
arise from diminished cyclization rates in either one or both
of the cyclization steps. It therefore seems reasonable to
propose that, in the transition states for both cyclization
steps, the SOMO is at least partially delocalized over both
enone moieties, as indicated in Scheme 5. The delocaliza-
tion is presumably greatest, and the SOMO energy the
lowest, when both keto functions are of the aroyl type.
2.8. Evidence for the distonic anion radical intermediate
2.7.2. Electrolysis of 1i. The reduction of 1i, which has one
benzoyl and one carboethoxy substituent, provides a still
further example of the relative inefficiency of pericyclic
chemistry when only a single benzoyl group is present.
When electrolyzed in the presence of lithium perchlorate, 1i
provides a 32% yield of two isomers of trans-2i. The major
isomer appears, on the basis of NMR comparisons, to be
structurally analogous to the major trans-2g isomer
obtained from 1g. This reduction is unusual in that it is
not catalytic. An intriguing possible interpretation of this
data is that while benzoyl reduction is occurring readily and
reversibly, the initial cyclization step is sharply retarded by
the ineffectiveness of the ester function at delocalizing and
stabilizing the SOMO in the transition state. This could
require cyclization to occur via the rarer reduction of the
In most cases the electrolyte systems (lithium/magnesium
perchlorate or tetraalkylammonium tetrafluoroborate) lead
to the formation of a mixture of the cis and trans
cyclobutane isomers. At the earliest point of the electro-
chemical reactions at which these products could be
detected, the trans cyclobutane was always formed in
modest excess over the cis isomer. As the reaction
progressed further toward completion, the trans/cis ratio
progressively increased. The values given in Table 1
correspond to reactions run essentially to completion. This
progressive change in the trans/cis ratio suggested the
possibility that a portion of the cis isomer was being
isomerized to the more stable trans isomer in the course of
the reaction. Although a base-catalyzed isomerization was
Scheme 5. Proposed delocalization of the SOMO over both carbonyl groups in the transition states for both cyclization steps.

11004
G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
Scheme 6. Mechanism of formation of 7, via protonation of distonic intermediate.
formally a possibility, it also appeared possible that anion
radicals of the cis-cyclobutane were being reformed during
the reaction and subsequently reverting to the distonic anion
radical, which could once again cyclize to give either
cyclobutane isomer or the Diels–Alder adduct. A distinction
between these two mechanistic possibilities is therefore
possible, based upon the predicted formation of small
amounts of the Diels–Alder adduct in the latter mechanism.
To test this possibility, isomers of 2b were isolated and used
in electrolysis reactions with lithium perchlorate as the
electrolyte. A quantity of cis-2b was reduced in isolation,
leading to the formation of trans-2b (32%), 3b (7%), an
aldol product 6b (11%), a dihydrocyclopentane product 7b
(20%; see Scheme 6), with a further 18% of unreacted cis-
2b. The attempted reduction of trans-2b, as expected, failed
to lead to any reaction. A degree of reactivity was seen
under extreme conditions (very negative potentials, large
amounts of charge). Although some trans-2b was still
returned, no other known products were obtained. This
clearly indicates that, while the trans isomer is stable toward
continued reduction at normal potentials, the cis isomer
readily reverts to the proposed distonic intermediate. This
not only allows for the formation of the trans and Diels–
Alder products, but also of 7b, which represents protonation
after the first cyclization, effectively ‘trapping’ the distonic
anion. In further accord with the postulate of cis–trans
isomerization via regeneration of the distonic anion radical
intermediate is the previously discussed absence of any
base-catalyzed products (e.g., 4 and 5) in any perchlorate
reaction system. Since the distonic anion radical inter-
mediate is evidently involved in the reversal of the
cycloaddition, an application of the law of microscopic
reversibility strongly suggests that the forward reaction also
involves the same distonic intermediate.
2.9. Trapping of the distonic anion radical
intermediate/inhibition of EGB pathway
It was thought that addition of a slight excess of a weak acid
would not only inhibit EGB pathways but would also trap
(protonate) the proposed distonic anion radical. Indeed this
was realized with a 1.6 M excess of acetic acid placed in the
solution from the beginning of electrolysis. The naphthoyl
substrate 1e was chosen as it leads to modest base-catalyzed
product formation (17% of 5e). The expected trapping is
clearly the dominant mechanism, represented by both 6e
and 7e formation (Table 3). The expected reduction in base-
catalysis (product 5e) is also seen, down to just 3%. Also, by
limiting the excess of acetic acid, we were still able to obtain
small amounts of our primary cyclobutanation product
(trans-2e), although formation of cis-2e and 3e were
reduced to levels below detection. The proposed mechanism
for the formation of 7 is given in Scheme 6.
2.10. Mediated electrolysis
The possibility of developing a mediated electrochemical
reduction method was also probed. The strategy adopted
was to provide a mediator which is (in the ideal case,
selectively) reduced at a less negative potential than the
substrate, and which forms a relatively long-lived anion
radical capable of mildly endergonic electron transfer to the
substrate molecules.14–16 This strategy should provide for a
low substrate anion radical concentration, which could
minimize anion radical to anion radical coupling as well as
over-reduction of the substrate (dianion formation). Initial
attempts utilized 1b as the substrate and magnesium or
lithium perchlorate as the electrolyte. Benzil (diphenyl
diketone), dypnone (E-3-phenyl-2-butenoylbenzene), and
benzophenone were investigated as mediators, but all three
Table 3. Effect of a weak acid in solution during electrolysis of 1e
Additive
% Yield of
trans-2e
% Yield of
cis-2e
% Yield of 3e
% Yield of 5e
% Yield of 6e
% Yield of 7e
% Yield of
products
None
1.6 M excess of
acetic acid
29
6
14
0
8
0
17
3
0
14
0
38
68
61

G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
Table 4. Anion radical versus base-catalyzed reactions in the mediated electrolysis of 1c
11005
Mediator
% Yield of trans-2c
% Yield of cis-2c
% Yield of 3c
% Yield of 5c
None
Benzophenone
Benzil
52
42
0
11
11
0
12
7
0
5
13
28a
a
Along with 42% unidentified polymer.
proved to be ineffective. Mediator reduction did occur at
less negative potentials (as evidenced by a temporary
solution color change) than that of the substrate, but
products did not form until the potential was reduced to
the usual value for reduction of the substrate.
theoretical consumption of electricity (i.e., electrocatalyti-
cally). The solvent/electrolyte combination acetonitrile/
tetraalkylammonium tetrafluoroborate is found to be an
especially effective one for producing high yields and large
catalytic factors. The formation of novel anion radical
Diels–Alder adducts in minor amounts is also verified. The
scope and limitations of these reactions are rather
extensively explored and defined. In particular, the reactions
have been found to have an absolute requirement for at least
one aroyl ketone moiety and a significant preference for
both ketonic moieties to be of the aroyl type. Theoretical
rationales for these requirements and preferences are
presented. Strongly electron-withdrawing substituents
(upon the aroyl moiety) tend to decrease reaction efficiency
by diminishing the rate of the first cyclization step, such that
a competition between anion radical mediated and electro-
generated base-catalyzed reactions is observed. Evidence
for a stepwise (as opposed to concerted) cycloaddition
mechanism involving a distonic anion radical intermediate
is presented, and the distonic anion intermediate has been
trapped.
Further studies were directed toward the use of tetra-
ethylammonium tetrafluoroborate as the electrolyte, this
time using 1c as the substrate. Although these conditions
failed to provide an increase in yield of pericyclic products,
they did provide important insights into the interplay
between anion radical cyclization and electrogenerated
base catalysis (Table 4). Benzophenone reduction was
found to occur at nearly the same potential as for 1c, when
benzil reduction occurred at a much less negative potential.
In the case of benzil, this occurred at a potential that did not
lead to substrate reduction, so no pericyclic products were
formed. On the other hand, in the absence of a mediator,
base-catalyzed products (such as 5c) were formed in a low
yield. When benzophenone was employed as the mediator,
both types of products were formed, indicating that both the
substrate and mediator are being reduced at the cathode.
Apparently, anion radical cyclobutanation occurs only when
the substrate is reduced, and reduction of the mediator leads
essentially only to the base-catalyzed reaction. These
observations indicate that the desired mildly endergonic
electron transfer is too slow to compete with the base-
catalyzed reactions.
4. Experimental
4.1. General electrolysis procedure
A typical experiment utilizes 100 mg of a given bis(enone)
substrate, which equates to 0.329 mmol for substrate 1a.
The substrate is dissolved in 22 mL of electrolyte solution,
giving a typical substrate concentration of 0.0150 M, and
added to the working electrode (WE) compartment of the
electrolysis cell. The electrolyte solution is 0.100 M (unless
stated) in either alkylammonium or perchlorate salt
(detailed below) in dry acetonitrile (unless stated). The
acetonitrile is distilled fresh for each electrolysis from a
reservoir containing phosphorus pentoxide. Electrolyte
solution (6 mL) is added to the counter electrode (CE)
compartment.
2.11. Efficient base-catalyzed reactions
The observations noted above immediately suggested the
possibility of employing an excess of benzil to engender a
highly catalytic method for selectively and efficiently
forming the base-catalyzed product. This possibility was
realized in the case of the electrochemical reduction of 1b in
the presence of a relatively large excess of benzil, which
gave only 4b, in 68% yield. The catalytic factor here was
rather modest at 2.8, although low concentrations of starting
material and a large excess of benzil made judging the end
of the reaction difficult. Comparison of this result with that
obtained for 1c (with benzil) suggests that the use of
mediators to engender these base-catalyzed cyclizations
may prove problematic, due to competition with
polymerization. The efficiency seen with formation of 4b
may be engendered by the greater acidity of the protons
(alpha to the bridging oxygen) in 1b.
Electrolysis of the substrate was carried out at specific
voltages (detailed below) with stirring under positive
nitrogen flow at room temperature. The voltage is
commonly increased through the course of an electrolysis
to help maintain the current (tracked by coulometer). This
increase is detailed in the description of each electrolysis.
Electrolysis voltages were versus a ‘pseudo-standard’ silver
wire (encased in porous vycor glass) reference electrode
(RE).17 The RE used is seen to have a calibration to SCE of
approximately C0.1 V, when in 0.1 M Et4NBF4 acetonitrile
solution. The CE and WE consisted of reticulated vitreous
carbon (25 mm!5 mm!5 mm), their corresponding com-
partments separated by a course frit. The RE was placed
within 0.5 cm of the WE. The reaction was stopped when
thin-layer chromatography (TLC) indicated that the starting
3. Conclusions
The present paper reports the development of conditions for
carrying out electrochemically initiated, intramolecular
anion radical cyclobutanations of bis(enones) and related
substrates in high yields and with substantially less than the

11006
G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
material had been consumed. The reactant solution (WE
compartment only) then underwent an aqueous workup with
sequential benzene washings (alkyl ammonium salt) or
dichloromethane (perchlorate salt). The organic phase was
retained and dried with Na2SO4. The benzene/dichloro-
methane was removed by rotary evaporation, with the crude
solution being purified by preparative TLC (1 mm thick,
elution with ethyl acetate/petroleum ether mixture, 1:9 ratio,
unless stated). Bands were identified, collected by scraping,
and extracted with dichloromethane. Filtering removed the
silica, with a rotary evaporator again employed to remove
the solvent, yielding the desired products.
13%) for a total of 98 mg of pericyclic products (88%). No
starting material was recovered, and no other products were
identified.
1
4.4.1. Compound cis-2a5,6. H NMR (300 MHz, CDCl3):
1.65 (2H, m), 1.84 (2H, m), 2.02 (2H, m), 3.20 (2H, br.s),
3.85 (2H, d, JZ4.2 Hz), 7.35 (4H, br.t, JZ7.2 Hz,), 7.44
(2H, br.t, JZ7.2 Hz), 7.75 (2H, br.d, JZ8.1 Hz).
4.4.2. Compound trans-2a5,6. 1H NMR (300 MHz, CDCl3):
1.42 (2H, m), 1.52 (1H, m), 1.85 (3H, br.m), 3.06 (1H, q,
JZ6.9 Hz), 3.24 (1H, m), 4.28 (1H, dd, JZ6.6, 7.5 Hz),
4.57 (1H, dd, JZ7.8, 10.5 Hz), 7.46 (4H, m), 7.55 (2H, m),
7.95 (2H, br.d, JZ6.9 Hz), 8.02 (2H, br.d, JZ7.2 Hz).
Notes. In cases when starting material is recovered, the
yields of products and the catalytic factors are corrected
accordingly. Characterization of products is in most cases
completed by comparison to the products obtained from 1b
electrolysis, and to results previously published (as
referenced). The products from 1b are more fully charac-
terized both by 500 MHz NMR (1H/13C) and HRMS. Their
ready crystallization also allowed for X-ray structural
determination (provided previously6). All novel products
are also characterized by HRMS or, in two cases, by X-ray
crystallography.
4.4.3. Compound 3a5,6. 1H NMR (300 MHz, CDCl3): 1.60
(2H, m), 1.84 (2H, m), 2.01 (2H, m), 2.69 (2H, m), 4.89 (1H,
d, JZ6.6 Hz), 5.59 (1H, d, JZ3.0 Hz), 7.26 (3H, m), 7.48
(4H, m), 8.08 (2H, d, JZ7.5 Hz).
4.5. Electrolysis of E,E-1,7-dibenzoyl-4-oxa-1,6-
heptadiene (1b)
Electrolysis of 100 mg (0.0149 M) of 1b with Mg(ClO4)2
electrolyte, at increasing voltages versus RE. The first 3.0 C
at K2.0 V, then 1.8 C at K3.0 V, followed by 1.3 C at
K3.5 V, and 6.0 C at K4.0 V. The reaction appeared
complete after 12.1 C, or 38.4% (of the required charge) had
passed through the cell. PTLC purification (1:4 ratio of EA/
PET) of the 156 mg of recovered crude yielded cis-2b
(26 mg, 39%), trans-2b (14 mg, 21%), and 3b (19 mg,
28%), giving a total of 59 mg of pericyclic products (88%).
While no other products were identified, 33 mg of starting
material was recovered. The above percent of required
charge is based upon the 100 mg of starting material being
used up; this becomes 54.8% based upon the 67 mg of
starting material used up (not recovered).
4.2. Analysis
1
Room temperature H NMR spectra were recorded on a
Varian UnityC300 as solutions in CDCl3. 13C NMR and
COSY spectra were recorded on a Varian Unity Inova 500
spectrometer. Chemical shifts (d) are relative to tetra-
methylsilane, and coupling constants (J) are given in Hertz
(Hz). Splitting patterns are designated as follows: s, singlet;
d, doublet; t, triplet; q, quartet; m, multiplet; dd, doublet of
doublets; br, broad. X-ray diffraction analyses were
conducted using a Nonius Kappa CCD diffractometer.
Low-resolution mass spectra (LRMS) were recorded on a
Finnigan MAT TSQ-70 mass spectrometer, with high-
resolution mass spectra (HRMS) recorded on a VGZAB-2E
mass spectrometer.
Electrolysis of 205 mg (0.0305 M) of 1b with LiClO4
electrolyte, at K1.3 V for 40.0 C (62.7% of required
charge). PTLC separation of the 170 mg of crude yielded
cis-2b (41 mg, 20%), trans-2b (42.2 mg, 21%), 3b (4 mg,
2%), along with 2.8 mg of unreacted 1b.
4.3. Equipment/reagents
A potentiostatic controller, the Electrosynthesis Company
(ESC) model 415, was used to control the applied potential.
The charge used was tracked by a digital coulometer, ESC
model 640. The applied potential was confirmed using a
digital multimeter (Wavetek DM7), operating as a potentio-
meter. The electrode material was Duocel 80 PPI reticulated
vitreous carbon. The substrates were kindly produced within
the Krische group or by Dr. Jingkui Yang.5,6 Reagent purity
was assayed by NMR/LRMS. Electrolytes were used as
purchased, from Alfa Aesar (98% purity).
Electrolysis of 105 mg (0.0156 M) of 1b with Bu4NBF4
electrolyte, at K1.6 V for 2.1 C and K2.0 V for 0.9 C. This
corresponds to 9.4% of required charge. PTLC separation of
the 108 mg of crude yielded cis-2b (12 mg, 11%), trans-2b
(41 mg, 39%), 3b (3 mg, 3%), and 4b (11 mg, 11%).
Electrolysis of 28 mg (0.0042 M) of 1b with Et4NBF4
electrolyte, and 157 mg (0.340 M) of benzil, at K1.4 V for
3.2 C. This corresponds to 36.2% of required charge, or a
catalytic factor of 2.76. PTLC separation proved
problematic due to the large benzil excess, however a
yield of 19.1 mg (68%) of 4b was obtained by NMR
integration of several mixed 4b/benzil PTLC bands. No
other products or unreacted starting material were observed.
4.4. Electrolysis of E,E-1,7-dibenzoyl-1,6-heptadiene
(1a)
Electrolysis of 111 mg (0.0166 M) of 1a with Bu4NBF4
electrolyte, at K2.0 V versus RE (first 3.0 C at K1.5 V).
The reaction appeared complete after 7.5 C, or 21.3% (of
required charge) had passed through the cell. PTLC
purification of the 105 mg of recovered crude yielded cis-
2a (19 mg, 17%), trans-2a (65 mg, 59%), and 3a (14 mg,
Electrolysis of 66 mg (0.0098 M) of cis-2b with LiClO4
electrolyte, at K2.5 V for 27.5 C, K3.0 V for 4.5 C, and
then at K3.5 V for a further 30.0 C. PTLC separation of the
crude yielded trans-2b (17.3 mg, 32%), 3b (4 mg, 7%), an

G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
11007
aldol product 6b (5.7 mg, 11%), a dihydro product 7b
(11 mg, 20%), with a further 12 mg (18%) of unreacted cis-
2b.
7.8 Hz), 7.58 (2H, br.t, JZ7.8 Hz), 7.94 (4H, m); HRMS
(CIC): calcd; 309.149070. Found; 309.149110.
4.6. Electrolysis of E,E-1,7-bis(4-chlorobenzoyl)-1,6-
heptadiene (1c)
Electrolysis of 52 mg (0.0077 M) of trans-2b with LiClO4
electrolyte, at K3.0 V for 11 C, K3.5 V for 130 C, and
K4.0 V for 20 C. No product spots were observed until after
w40 C, indicating that no reaction was occurring, hence the
continued flow of charge. PTLC of the 68 mg of crude
yielded 8 mg of impure trans-2b, along with 19 mg of
unidentified product (single benzoyl moiety). None of the
characterized products (such as cis-2b) were observed.
Electrolysis of 99 mg (0.012 M) of 1c with Bu4NBF4
electrolyte, at K1.2 V for 2.1 C, K1.6 V for 2.2 C, and
K1.8 V for 0.8 C. The reaction appeared complete after
5.0 C, or 19.5% (of required charge) had passed through the
cell. PTLC purification of the 124 mg of recovered crude
yielded cis-2c (11 mg, 11%), trans-2c (51.5 mg, 52%), and
3c (11.5 mg, 12%), for a total yield of 74 mg of pericyclic
products (75%). No starting material was recovered.
However, a further 5 mg (5%) of 5c was recovered.
4.5.1. Compound cis-2b.6 1H NMR (500 MHz, CDCl3):
3.42 (2H, m), 3.63 (2H, d.m, JZ10.2 Hz), 4.14 (4H, m),
7.34 (4H, br.t, JZ7.6 Hz), 7.45 (2H, br.t, JZ7.3 Hz), 7.72
(4H, br.d, JZ8.2 Hz); 13C NMR (500 MHz in CDCl3):
39.72, 47.70, 73.46, 127.81, 128.59, 132.8, 136.05, 198.08;
NMR COSY (500 MHz), and X-ray crystallography
confirm structure;6 HRMS (CIC): calcd; 307.133420.
Found; 307.132575.
Electrolysis of 55 mg (0.0067 M) of 1c with 32 mg
(0.0080 M) of benzophenone and Et4NBF4 electrolyte, at
K1.8 V for 1.5 C (10.5% of required charge). 1c/benzo-
phenoneZ1:1.19. PTLC separation yielded cis-2c (6 mg,
11%), trans-2c (23 mg, 42%), and 3c (4 mg, 7%), for a total
yield of 33 mg of pericyclic products (60%). A further 7 mg
(13%) of 5c was recovered.
4.5.2. Compound trans-2b.6 1H NMR (500 MHz, CDCl3):
3.36 (3H, br.m), 3.46 (1H, dd, JZ4.6, 9.6 Hz), 3.68 (1H, d,
JZ9.8 Hz), 4.06 (1H, d, JZ9.6 Hz), 4.41 (1H, m), 4.53
(1H, m), 7.45 (4H, m), 7.54 (2H, br.m), 7.90 (2H, m), and
8.01 (2H, m); 13C NMR (500 MHz in CDCl3): 39.71, 40.74,
42.99, 43.01, 69.27, 72.55, 128.29, 128.71, 128.76, 128.91,
133.41, 133.43, 135.27, 135.61, 196.70, 199.70; NMR
COSY (500 MHz), and X-ray crystallography confirms the
structure;6 HRMS (CIC): calcd; 307.133420. Found;
307.133048.
Electrolysis of 99 mg (0.012 M) of 1c with 165 mg
(0.036 M) of benzil and Et4NBF4 electrolyte, at K1.3 V
for 2.5 C (9.7% of required charge). 1c/benzilZ1:1.2.96.
PTLC separation yielded 28 mg (28%) of 5c along with
42 mg (42%) of an unidentified polymer.
4.6.1. Compound cis-2c.6 1H NMR (300 MHz, CDCl3)
1.68 (2H, m), 1.82 (2H, dd, JZ9.9, 2.4 Hz), 2.00 (2H, m),
3.16 (2H, m), 3.77 (2H, d, JZ3.9 Hz), 7.32 (4H, d, JZ
8.4 Hz), 7.66 (4H, d, JZ9.0 Hz); HRMS (CIC): calcd;
373.076210. Found 373.076164.
4.5.3. Compound 3b.6 1H NMR (500 MHz, CDCl3): 2.99
(1H, m), 3.06 (1H, m), 3.58 (1H, m), 3.75 (1H, dd, JZ4.4,
9.6 Hz), 4.17 (2H, m), 4.98 (1H, d, JZ8.6 Hz), 5.58 (1H, d,
JZ4.2 Hz), 7.26 (2H, m), 7.48 (5H, m), 7.61 (1H, m), 8.08
(2H, m); 13C NMR (500 MHz in CDCl3): 35.54, 36.94,
70.30, 74.03, 76.30, 98.89, 124.65, 128.24, 128.44, 128.62,
129.51, 133.70, 134.44, 135.41, 151.80, 196.01; NMR
COSY (500 MHz) and X-ray crystallography confirms the
structure;6 HRMS (CIC): calcd; 307.133420. Found;
307.132957.
4.6.2. Compound trans-2c.6 1H NMR (300 MHz, CDCl3):
1.38 (2H, m), 1.58 (1H, m), 1.82 (3H, m), 3.04 (1H, dd, JZ
6.9, 13.2 Hz), 3.22 (1H, m), 4.20 (1H, dd, JZ0.9, 7.5 Hz),
4.49 (1H, dd, JZ2.4, 10.2 Hz), 7.44 (4H, dd, JZ2.7,
8.4 Hz), 7.88 (2H, br. d, JZ8.4 Hz), 7.95 (2H, br.d, JZ
9.0 Hz):); HRMS (CIC): calcd; 373.076210. Found;
373.075821.
4.6.3. Compound 3c.6 1H NMR (300 MHz, CDCl3): 1.41
(1H, m), 1.59 (2H, m), 1.78 (1H, m), 2.01 (2H, m), 2.68 (2H,
m), 4.80 (1H, d, JZ6.9 Hz), 5.52 (1H, d, J!1.5 Hz), 7.40
(2H, m), 7.47 (4H, m), 8.01 (2H, d, JZ8.4 Hz); HRMS
(CIC): calcd; 373.076210. Found; 373.076898.
1
4.5.4. Compound 4b. H NMR (300 MHz, CDCl3): 3.08
(1H, m), 3.17 (1H, d, JZ5.7 Hz), 3.20 (1H, d, JZ7.5 Hz),
3.85 (1H, m), 3.98 (1H, dd, JZ3.3, 10.8 Hz), 4.23 (1H, dd,
JZ2.7, 10.8 Hz), 4.72 (1H, td, J!1.5, 6.0 Hz), 6.52 (1H,
dd, JZ1.8, 6.0 Hz), 7.48 (4H, br.t, JZ7.5 Hz), 7.58 (2H,
m), 8.00 (4H, m). X-ray crystallography confirms structure
(CCDC 240331).
1
4.6.4. Compound 5c. H NMR (300 MHz, CDCl3): 1.64
(1H, m), 1.72 (3H, m), 2.30 (2H, br.m), 2.79 (1H, dd, JZ
10.2, 14.7 Hz), 3.34 (1H, dd, JZ3.6, 14.7 Hz), 3.43 (1H,
br.s), 6.61 (1H, td, JZ3.9, 1.2 Hz), 7.43 (3H, m), 7.51 (1H,
m), 7.62 (2H, m), 7.99 (2H, m); HRMS (CIC): calcd;
373.076210. Found; 373.076307.
4.5.5. Compound 6b.6 1H NMR (300 MHz, CDCl3): 1.98
(1H, ddd, JZ1.5, 8.1, 12.9 Hz), 2.45 (1H, dd, JZ8.1,
12.9 Hz), 3.22 (2H, m), 3.64 (2H, dd, JZ6.3, 9.0 Hz), 3.85
(2H, t, JZ9.9 Hz), 4.12 (1H, d, JZ9.6 Hz), 5.35 (1H, d, JZ
2.1 Hz), 7.14 (1H, m), 7.24 (2H, m), 7.40 (2H, m), 7.52 (2H,
m), 7.59 (1H, m), 7.80 (2H, dd, JZ1.5, 8.7 Hz).
4.7. Electrolysis of E,E-1,7-bis(3,4-dichlorobenzoyl)-1,6-
heptadiene (1d)
4.5.6. Compound cis-7b. 1H NMR (300 MHz, CDCl3):
3.04 (4H, br.m), 3.16 (2H, dd, JZ3.0, 14.1 Hz), 3.60 (2H,
dd, JZ4.2, 8.7 Hz), 4.12 (2H, m), 7.47 (4H, br.t, JZ
Electrolysis of 102 mg (0.0105 M) of 1d with LiClO4 and
Mg(ClO4)2 electrolyte (both 0.10 M), at K4.0 V (first 1.5 C
at K3.0 V). The reaction appeared complete after 17.0 C, or

11008
G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
76.3% (of required charge) had passed through the cell.
PTLC purification of the 87 mg of recovered crude yielded
cis-2d (34 mg, 33%), trans-2d (7 mg, 7%), and 3d (16 mg,
16%), for a total yield of 57 mg of pericyclic products
(56%). A further 6 mg (6%) of a product was isolated and
identified as 6d.
Electrolysis of 104 mg (0.0117 M) of 1e, with Et4NBF4
electrolyte and 25 mg of acetic acid, giving a 1.6:1 excess of
acetic acid. Initially at K2.5 V for 15.0 C and then at
K3.0 V for 32.0 C. PTLC purification of the crude yielded
trans-2e (6.0 mg, 6%), 5e (2.8 mg, 3%), 6e (14.1 mg, 14%),
and 7e (39.2 mg, 35%).
Electrolysis of 100 mg (0.0103 M) of 1d with Et4NBF4
electrolyte, at K3.2 V for 3.0 C (first 0.75 C at K2.5 V,
then 0.75 C at K3.0 V). This corresponds to 13.7% of
required charge. PTLC purification of the 150 mg of crude
yielded 3 mg (3%) of trans-2d, 17 mg (17%) of 5d, and
35 mg (35%) of an unidentified polymer.
Electrolysis of 98 mg (0.0110 M) of 1e with Bu4NBF4
electrolyte, at K1.5 V. The reaction was stopped after
3.4 C, or 14.5% (of required charge, corrected to 18.4%) had
passed through the cell. PTLC purification of the 122 mg of
recovered crude yielded cis-2e (16.6 mg, 21%), trans-2e
(13 mg, 17%), and 3e (8 mg, 10%), for a total yield of
37.6 mg of pericyclic products (48%). 20.4 mg (21%) of
starting material was recovered, along with 5 mg (6%) of 5e.
Reaction is 79% complete based upon recovered 1e.
4.7.1. Compound cis-2d.6 1H NMR (300 MHz, CDCl3):
1.74 (2H, m), 1.85 (2H, dd, JZ11.7, 3.3 Hz), 2.05 (2H, m),
3.19 (2H, d, JZ2.4 Hz), 3.77 (2H, d, JZ3.6 Hz), 7.46 (2H,
d, JZ8.4 Hz), 7.57 (2H, dd, JZ8.7, 2.1 Hz), 7.81 (2H, d,
JZ2.1 Hz); HRMS (CIC): calcd; 440.998266. Found;
440.997859.
4.8.1. Compound cis-2e.6 1H NMR (300 MHz, CDCl3)
1.74 (2H, m), 1.93 (2H, dd, JZ14.1, 5.4 Hz), 2.15 (2H, m),
3.29 (2H, d, JZ2.4 Hz), 4.08 (2H, d, JZ4.2 Hz), 7.45 (4H,
m), 7.74 (4H, m); 7.81 (4H, m), 8.22 (2H, s); HRMS (CIC):
calcd; 405.185455. Found 405.185760.
1
4.7.2. Compound trans-2d. H NMR (300 MHz, CDCl3):
1.38 (2H, m), 1.55 (1H, m), 1.83 (3H, m), 3.04 (1H, m), 3.24
(1H, m), 4.15 (1H, m), 4.47 (1H, dd, JZ2.1, 10.2 Hz), 7.55
(1H, d, JZ3.3 Hz), 7.57 (1H, d, JZ3.0 Hz), 7.76 (1H, br.d),
7.81 (1H, br.d), 8.02 (1H, d, JZ2.1 Hz) 8.08 (1H, d, JZ
1.8 Hz); HRMS (CIC): calcd; 440.998266. Found;
440.997697.
4.8.2. Compound trans-2e.6 1H NMR (300 MHz, CDCl3):
1.45 (1H, m), 1.62 (1H, m), 1.85 (2H, m), 1.99 (2H, m), 3.16
(1H, m), 3.35 (1H, m), 4.50 (1H, m), 4.79 (1H, dd, JZ2.4,
10.2 Hz), 7.58 (4H, m), 7.89 (4H, m), 8.03 (4H, br.m), 8.48
(1H, br.s), 8.60 (1H, br.s); HRMS (CIC): calcd;
405.185455. Found; 405.186095.
1
4.7.3. Compound 3d. H NMR (300 MHz, CDCl3): 1.60
4.8.3. Compound 3e.6 1H NMR (300 MHz, CDCl3): 1.60
(2H, m), 1.85 (2H, m), 2.07 (2H, m), 2.83 (2H, m), 5.09 (1H,
d, JZ6.9 Hz), 5.73 (1H, d, J!3 Hz), 7.39 (2H, m), 7.62
(3H, m), 7.74 (3H, m), 7.95 (4H, m), 8.15 (1H, m), 8.74 (1H,
br.s); HRMS (CIC): calcd; 405.185455. Found;
405.186005.
(2H, m), 1.85 (2H, m), 2.07 (2H, m), 2.83 (2H, m), 5.09 (1H,
d, JZ6.9 Hz), 5.73 (1H, d, J!3 Hz), 7.39 (2H, m), 7.62
(3H, m), 7.74 (3H, m), 7.95 (4H, m), 8.15 (1H, m), 8.74 (1H,
br.s); HRMS (CIC): calcd; 440.998266. Found;
440.997720.
1
4.7.4. Compound 5d. H NMR (300 MHz, CDCl3): 1.73
4.8.4. Compound 5e.11 1H NMR (300 MHz, CDCl3): 1.71
(1H, m), 1.87 (3H, m), 2.35 (2H, br.m), 3.00 (1H, dd, JZ
11.4, 15.6 Hz), 3.66 (1H, dd, JZ3.6, 15.6 Hz), 3.77 (1H,
br.s), 6.75 (1H, br.t), 7.60 (4H, m), 7.92 (5H, m), 7.99 (1H,
m), 8.05 (1H, m), 8.14 (1H, m), 8.22 (1H, s), 8.71 (1H, s).
(4H, m), 2.31 (2H, br.m), 2.84 (1H, dd, JZ9.6, 15.0 Hz),
3.66 (1H, dd, JZ3.6, 15.0 Hz), 3.40 (1H, br.s), 6.66 (1H, t,
JZ3.3 Hz), 7.51 (2H, m), 7.56 (1H, d, JZ8.4 Hz), 7.74
(1H, d, JZ1.8 Hz), 7.90 (1H, dd, JZ2.1, 8.4 Hz), 8.12 (1H,
d, JZ2.1 Hz); HRMS (CIC): calcd; 440.998266. Found;
440.998110.
4.8.5. Compound 6e.6 1H NMR (300 MHz, CDCl3). Partial
only: 1.60 (m), 1.91 (m), 3.20 (2H, m), 5.74 (1H, d, JZ
1.8 Hz), 7.41 (2H, m), 7.59 (6H, br.m), 7.76 (2H, m), 7.87
(2H, m), 8.00 (2H, m).
1
4.7.5. Compound 6d. H NMR (300 MHz, CDCl3): 1.71
(3H, m), 1.98 (1H, m), 2.17 (1H, m), 2.34 (1H, m), 2.96 (3H,
m), 3.17 (1H, br.d, JZ17.4 Hz), 3.65 (1H, d, JZ9.0 Hz),
5.24 (1H, d, J!1.5 Hz), 7.31 (1H, m), 7.54 (2H, m), 7.76
(1H, br.d), 7.81 (1H, d, JZ2.1 Hz), 8.00 (1H, m); HRMS
(CIC): calcd; 443.013916. Found; 443.012405.
1
4.8.6. Compound trans-7e. H NMR (300 MHz, CDCl3):
1.35 (1H, m), 1.67 (2H, m), 1.88 (1H, m), 2.04 (2H, m), 2.31
(1H, m), 2.78 (1H, m), 2.95 (1H, dd, JZ8.1, 15.6 Hz), 3.10
(1H, dd, JZ8.1, 16.2 Hz), 3.27 (1H, dd, JZ5.7, 15.3 Hz),
3.38 (1H, dd, JZ4.8, 16.2 Hz), 7.56 (4H, m), 7.88 (4H, m),
7.95 (2H, br.d, JZ7.8 Hz), 8.02 (2H, m), 8.47 (2H, br.d, JZ
6.0 Hz); HRMS (CIC): calcd; 407.201105. Found;
407.202010.
4.8. Electrolysis of E,E-1,7-di-1-naphthoyl-1,6-
heptadiene (1e)
Electrolysis of 97 mg (0.0109 M) of 1e with Bu4NBF4
electrolyte, at K1.8 V (first 3.0 C at K1.5 V). The reaction
appeared complete after 4.5 C, or 19.4% (of required
charge) had passed through the cell. PTLC purification of
the 123 mg of recovered crude yielded cis-2e (14 mg, 14%),
trans-2e (28 mg, 29%), and 3e (8 mg, 8%), for a total yield
of 50 mg of pericyclic products (51%). No starting material
was recovered. A further 16 mg (17%) of 5e was recovered.
4.9. Electrolysis of E,E-1,7-bis(4-phenylbenzoyl)-1,6-
heptadiene (1f)
Electrolysis of 76 mg (0.0076 M) of 1f with Bu4NBF4
electrolyte (in a 1:1 THF/acetonitrile solution), at K3.5 V
(first 1.8 C at K2.5 V, then 1.1 C at K3.0 V). The reaction

G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
11009
appeared complete after 11.7 C, or 69.6% (of required
charge) had passed through the cell. PTLC purification of
the 107 mg of recovered crude yielded trans-2f (9 mg,
12%). Also recovered was 13 mg (17%) of 5f.
PTLC separation yielded two isomers of trans-2g, 9 mg
(12%) of the isomer where the benzoyl group is syn to the
cyclopentane ring, and 5 mg (7%) with an anti benzoyl
group.
4.10.1. Compound cis-2g.6 1H NMR (300 MHz, CDCl3):
1.59 (2H, m), 1.68 (4H, br.m), 2.03 (3H, s), 3.00 (2H, m),
3.11 (1H, m), 3.79 (1H, dd, JZ4.5, 9.9 Hz), 7.45 (2H, br.t,
JZ7.2 Hz), 7.52 (1H, br.t, JZ7.2 Hz), 7.83 (2H, br.d, JZ
8.1 Hz); HRMS (CIC): calcd; 243.138505. Found;
243.138743.
Electrolysis of 55 mg (0.0055 M) of 1f with Mg(ClO4)2
(0.1 M in 1:1 THF/acetonitrile) at K2.0 V for 5.0 C
(corrected for recovered 1f to 53.7% of required charge),
PTLC purification of the 61 mg of crude yielded cis-2f
(9 mg, 21%), trans-2f (3.1 mg, 7%), and 3f (11 mg, 25%),
for a total yield of 23.1 mg of pericyclic products (53%).
11 mg of unreacted 1f was recovered, allowing the corrected
yields given above. Two additional products were isolated;
5.8 mg (13%) of 6f; and 3.2 mg (7%) of 7f.
4.10.2. Compound trans-2g (benzoyl group syn to the
cyclopentane ring). 1H NMR (300 MHz, CDCl3): 1.36 (2H,
m), 1.74 (4H, m), 2.13 (3H, s), 2.93 (1H, m), 3.13 (1H, m),
3.49 (1H, t, JZ8.1 Hz), 4.32 (1H, dd, JZ2.1, 10.5 Hz), 7.47
(2H, br.t, JZ7.2 Hz), 7.57 (1H, br.t, JZ6.9 Hz), 7.93 (1H,
br.d, JZ7.2 Hz). X-ray crystallography confirms structure
(CCDC240332).
4.9.1. Compound cis-2f.6 1H NMR (300 MHz, CDCl3):
1.84 (4H, m), 2.06 (2H, m), 3.27 (2H, m), 3.92 (2H, m), 7.37
(6H, m), 7.55 (8H, m), 7.84 (4H, d, JZ8.7 Hz).
1
4.9.2. Compound trans-2f. H NMR (300 MHz, CDCl3):
4.10.3. Compound trans-2g.6 (benzoyl group anti to the
cyclopentane ring) 1H NMR (300 MHz, CDCl3): 1.58 (2H,
m), 1.80 (4H, m), 2.11 (3H, s), 2.94 (1H, m), 3.07 (1H, m),
3.38 (1H, m), 3.86 (1H, dd, JZ2.4, 10.5 Hz), 3.99 (1H, t,
JZ6.3 Hz), 7.45 (2H, br.t, JZ7.2 Hz), 7.55 (1H, br.t, JZ
7.2 Hz), 7.95 (1H, br.d, JZ7.2 Hz).
1.47 (2H, m), 1.66 (2H, m), 1.91 (2H, m), 3.13 (1H, m), 3.30
(1H, m), 4.33 (1H, m), 4.61 (1H, m), 7.45 (6H, m), 7.65 (2H,
m), 7.70 (2H, d, JZ8.4 Hz), 8.04 (2H, d, JZ9.0 Hz), 8.11
(2H, d, JZ8.7 Hz); HRMS (CIC): calcd; 457.216755.
Found; 457.215380.
1
4.9.3. Compound 3f. H NMR (300 MHz, CDCl3): 1.56
(3H, m), 1.78 (1H, m), 2.05 (2H, m), 2.75 (2H, m), 4.93 (1H,
m), 5.61 (1H, m), 7.40–7.72 (12H, m), 8.19 (2H, m): calcd;
457.216755. Found; 457.218026.
1
4.10.4. Compound 7g. H NMR (300 MHz, CDCl3). 1.23
(2H, m), 1.61 (2H, m), 1.98 (4H, br.m), 2.13 (3H, s), 2.40
(1H, dd, JZ7.8, 16.8 Hz), 2.67 (1H, dd, JZ4.2, 16.8 Hz),
2.91 (1H, dd, JZ7.8, 16.5 Hz), 3.14 (1H, dd, JZ4.2,
16.8 Hz), 7.46 (2H, br.t, JZ7.5 Hz), 7.56 (1H, br.t, JZ
7.2 Hz), 7.95 (1H, br.d, JZ7.2 Hz); HRMS (CIC): calcd;
245.154155. Found; 245.154390.
1
4.9.4. Compound 5f. H NMR (300 MHz, CDCl3): 1.67
(1H, m), 1.78 (3H, m), 2.31 (2H, br.m), 2.85 (1H, dd, JZ
10.5, 14.7 Hz), 3.47 (1H, dd, JZ6.6, 13.8 Hz), 3.54 (1H,
br.s), 6.70 (1H, m), 7.43 (6H, m), 7.66 (6H, m), 7.79 (2H, d,
JZ8.4 Hz), 8.03 (2H, d, JZ8.7 Hz), 8.16 (2H, d, JZ
8.7 Hz); HRMS (CIC): calcd; 457.216755. Found;
457.216312.
4.11. Electrolysis of E,E-1,7-diacetyl-1,6-heptadiene (1h)
Electrolysis of 97 mg (0.025 M) of 1h with Et4NBF4
electrolyte, at K2.5 V for 10.0 C (initial 4.0 C at
K2.0 V). H NMR (300 MHz, CDCl3) of the 121 mg of
1
4.9.5. Compound 6f.6 1H NMR (300 MHz, CDCl3). Partial
only: 3.94 (1H, d, JZ8.7 Hz), 5.63 (1H, d, Jw1.5 Hz).
recovered crude showed only starting material present (plus
electrolyte).
4.9.6. Compound cis-7f. 1H NMR (300 MHz, CDCl3): 1.34
(2H, m), 1.67 (2H, m), 2.02 (2H, m), 2.24 (2H, m), 2.98 (2H,
dd, JZ7.2, 17.4 Hz), 3.26 (2H, dd, JZ4.5, 16.2 Hz), 7.44
(6H, m), 7.65 (8H, m), 8.03 (4H, d, JZ8.7 Hz); HRMS
(CIC): calcd; 459.232406. Found; 459.231734.
4.12. Electrolysis of E,E-7-ethoxy-1-benzoyl-1,6-
heptadiene (1i)
Electrolysis of 117 mg (0.020 M) of 1i with LiClO4
electrolyte (0.3 M), at K2.0 V for 10.0 C, K2.2 V for
20.0 C, and K3.0 V for 10.0 C (136% of required charge).
PTLC separation yielded 23 mg (28%) of one isomer of
trans-2i. Thought to be the same isomer as the major isomer
seen in the 1g electrolysis (based upon NMR comparison).
A further 3 mg (4%) of the alternate trans-2i isomer was
also obtained. 34 mg of starting material 1i was recovered.
4.10. Electrolysis of E,E-1-acetyl-7-benzoyl-1,6-
heptadiene (1g)
Electrolysis of 98 mg (0.018 M) of 1g with Mg(ClO4)2
electrolyte, at K4.5 V for 150 C (initial 4.0 C at K4.0 V,
last 16 C at K5.0 V), corresponding to 459% of required
charge. PTLC separation yielded cis-2g (7 mg, 9%), and 7g
(14 mg, 17%), along with 11 mg (13%) of an unidentified
product, tentatively described as an isomer of 7g (LRMS
(CIC): 245, 227). 16 mg of unreacted 1g was also
recovered.
Electrolysis of 136 mg (0.023 M) of 1i with Et4NBF4
electrolyte, at K2.0 V for 10.5 C. Yielding only unidenti-
fied polymers
4.12.1. Compound trans-2i (benzoyl group syn to the
cyclopentane ring). 1H NMR (300 MHz, CDCl3): 0.98 (3H,
t, JZ7.2 Hz), 1.63 (2H, m), 1.72 (2H, m), 1.91 (2H, m), 3.04
(2H, m), 3.22 (1H, m), 3.69 (1H, dd, JZ5.1, 9.9 Hz), 3.87
Electrolysis of 73 mg (0.014 M) of 1g with Et4NBF4
electrolyte, at K2.0 V for 16.0 C (initial 2.2 C at
K1.6 V), corresponding to 55.0% of required charge.

11010
G. A. N. Felton, N. L. Bauld / Tetrahedron 60 (2004) 10999–11010
(2H, q, JZ7.2 Hz), 7.43 (2H, br.t, JZ8.1 Hz), 7.53 (1H,
br.t, JZ6.9 Hz), 7.84 (1H, br.d, JZ7.2 Hz). HRMS (CIC):
calcd; 273.149070. Found; 273.148638.
Acknowledgements
The authors acknowledge the assistance of Michael J.
Krische, Jingkui Yang, and thank the Robert A. Welch
Foundation (F-149) for support of this research.
4.12.2. Compound trans-2i (benzoyl group anti to the
cyclopentane ring). 1H NMR (300 MHz, CDCl3): 0.98 (3H,
t, JZ7.2 Hz), 1.69 (4H, m), 2.01 (2H, m), 2.99 (1H, m), 3.21
(1H, m), 3.38 (1H, m), 4.14 (2H, q, JZ7.2 Hz), 4.34 (1H,
dd, JZ2.7, 10.5 Hz), 7.47 (2H, br.t, JZ6.3 Hz), 7.56 (1H,
br.t, JZ6.9 Hz), 7.93 (1H, br.d, JZ6.6 Hz). HRMS (CIC):
calcd; 273.149070. Found; 273.148198.
References and notes
1. Bauld, N. L. Tetrahedron 1989, 45, 5307–5363.
2. Bauld, N. L. In Electron Transfer in Chemistry; Balzani, V.,
Ed.; Wiley-VCH: Weinheim, 2001; Vol. 2, pp 133–205.
3. Delaunay, J.; Mabon, G.; Orliac, A.; Simonet, J. Tetrahedron
Lett. 1990, 31, 667–668.
4.13. Electrolysis of E,E-1,7-dicarbethoxy-1,6-
heptadiene (1j)
4. Jannsen, R.; Motevalli, M.; Utley, J. H. P. J. Chem. Soc.,
Chem. Commun. 1998, 539–540.
Electrolysis of 112 mg (0.021 M) of 1j with Et4NBF4
electrolyte, at K2.5 V for 10.0 C, corresponding to 22.2%
of required charge. PTLC purification of the 112 mg of
recovered crude yielded 47 mg (42%) of 5j. No starting
material was recovered.
5. Roh, Y.; Jang, H.-Y.; Lynch, V.; Bauld, N. L.; Krische,
M. J. Org. Lett. 2002, 4, 611–613.
6. Yang, J.; Felton, G. A. N.; Bauld, N. L.; Krische, M. J. J. Am.
Chem. Soc. 2004, 126(6), 1634–1635.
7. Fry, A. J.; Little, R. D.; Leonetti, J. J. Org. Chem. 1994, 59,
5017–5026.
Electrolysis of 96 mg (0.018 M) of 1j with LiClO4
electrolyte, at K3.5 V for 4.5 C, K3.5 V for 7.9 C, and
K4.0 V for 55 C. 1H NMR (300 MHz, CDCl3) of the
recovered crude showed only starting material present.
8. Little, R. D. Chem. Rev. 1996, 96(1), 108–114.
9. The reactions were monitored for completion in a step-wise
fashion, such that the exact amount of charge used to give
completion is not known (see experimental data). Therefore,
the catalytic factors may be greater than that stated. Catalytic
factors are found using 96,485 C molK1, see Hamman, C. H.;
Hamnett, A.; Vielstich, W. Electrochemistry; Wiley-VCH:
Weinheim, 1998; pp 7–9.
1
4.13.1. Compound 5j. H NMR (500 MHz, CDCl3): 1.26
(3H, t, JZ7.5 Hz), 1.29 (3H, t, JZ7.0 Hz), 1.64 (4H, m),
2.18 (2H, br.m), 2.27 (1H, dd, JZ10.5, 15.0 Hz), 2.64 (1H,
dd, JZ3.5, 15.0 Hz), 3.09 (1H, br.s), 4.16 (4H, m), 7.03
(1H, t, JZ4.0 Hz). 13C NMR (500 MHz in CDCl3): 14.21
(2C), 17.11, 25.77, 26.52, 29.92, 38.40, 60.20, 60.26,
132.79, 140.99, 166.86, 172.53; NMR COSY (500 MHz)
(consistent with structure). HRMS (CIC): calcd;
241.143984. Found; 241.142945.
10. Amputch, M. A.; Little, R. D. Tetrahedron 1991, 47(3),
383–402.
11. Wang, L.; Luis, A. L.; Agapiou, K.; Jang, H.; Krische, M. J.
J. Am. Chem. Soc. 2002, 124(11), 2402–2403.
12. Utley, J. H. P.; Nielsen, M. F. Electrogenerated Bases; Lund,
H., Hammerich, O., Eds. 4th ed.; Organic Electrochemistry;
Marcel Dekker: New York, 2001. Chapter 30.
13. Felton, G. A. N.; Bauld, N. L. Tetrahedron Lett. 2004, 45(25),
4841–4845.
4.14. X-ray data
Crystallographic data (excluding structure factors) for 4b
and the primary trans-2g isomer, have been deposited with
the Cambridge Crystallographic Data Centre as supplemen-
tary publication numbers CCDC 240331 and CCDC
240332, respectively. Copies of the data can be obtained,
free of charge, via www.ccdc.cam.ac.uk/data_request/cif or
on application to CCDC, 12 Union Road, Cambridge CB2
1EZ, UK (fax: C44(0)-1223-336033 or e-mail:
deposit@ccdc.cam.ac.uk).
14. Ihara, M.; Katsumata, A.; Setsu, F.; Tokunaga, Y.; Fukumoto,
K. J. Org. Chem. 1996, 61, 677–684.
15. Utley, J. Chem. Soc. Rev. 1997, 26, 157–167.
16. Inokuchi, T.; Kusumoto, M.; Torii, S. Electroorganic
Synthesis; Little, R. D., Weinberg, N. L., Eds.; Marcel Dekker:
New York, 1991; pp 233–239.
17. Becker, J. Y.; Koch, T. A. Electrochim. Acta 1994, 39(13),
2067–2071.