Electro-organic reactions. Part 54: Quinodimethane chemistry; Part 2 - Electrogeneration and reactivity of o-quinodimethanes

Article abstract of DOI:10.1039/b007476o

The electrochemical generation and characterisation of a variety of o-quinodimethanes (o-QDMs) are described together with the outcome of preparative experiments in which they are key intermediates. The quinodimethanes are conveniently formed, in DMF, by both direct and redox-catalysed electroreduction of 1,2-bis(halomethyl)arenes. Their predominant reaction is polymerisation to poly(o-xylylene) (o-PX) polymers. In the presence of dienophiles the electrogenerated o-QDMs may undergo efficient cycloaddition reaction and distinctions between the possible mechanisms have been attempted on the basis of voltammetric, preparative and stereochemical experiments. Contrary to the precedent of the corresponding methyl ester, diphenyl maleate radical-anion isomerises only slowly to the fumarate radical-anion, yet co-electrolysis of 2,3-bis(bromomethyl)-1,4-dimethoxybenzene and diphenyl maleate or diphenyl fumarate gives exclusively the corresponding trans-adduct. Co-electrolysis of dimethyl maleate with either 1,2-bis(bromomethyl)benzene (more easily reduced) or 2,3-bis(bromomethyl)-1,4-dimethoxybenzene (less easily reduced) gave only o-PX polymer. The results are rationalised in terms of a double nucleophilic substitution mechanism where electron transfer between dienophile radical-anion and dihalide is relatively slow. Where electron transfer from maleate or fumarate radical-anions is likely to be fast o-quinodimethanes are formed by redox-catalysis and they polymerise rather than undergo Diels-Alder reaction. Dimerisation of the dienophile radical-anions, with k₂ = 10⁴ to 10⁵ M^-1 s^-1, does not apparently compete with nucleophilic substitution or, where relevant, electron transfer.

Full text of DOI:10.1039/b007476o

View Article Online / Journal Homepage / Table of Contents for this issue
Electro-organic reactions. Part 54.† Quinodimethane chemistry.
Part 2.‡ Electrogeneration and reactivity of o-quinodimethanes§
James H. P. Utley,*a Shalini Ramesh,a Xavier Salvatella,a Sabine Szunerits,a Majid Motevallia
and Merete F. Nielsenb
2
a Department of Chemistry, Queen Mary and Westfield College (University of London),
Mile End Road, London, UK E1 4NS
b Department of Chemistry, University of Copenhagen, Symbion Science Park, Fruebjergvej 3,
DK-2100, Copenhagen, Denmark
Received (in Cambridge, UK) 14th September 2000, Accepted 15th November 2000
First published as an Advance Article on the web 3rd January 2001
The electrochemical generation and characterisation of a variety of o-quinodimethanes (o-QDMs) are described
together with the outcome of preparative experiments in which they are key intermediates. The quinodimethanes are
conveniently formed, in DMF, by both direct and redox-catalysed electroreduction of 1,2-bis(halomethyl)arenes.
Their predominant reaction is polymerisation to poly(o-xylylene) (o-PX) polymers. In the presence of dienophiles the
electrogenerated o-QDMs may undergo efficient cycloaddition reaction and distinctions between the possible
mechanisms have been attempted on the basis of voltammetric, preparative and stereochemical experiments.
Contrary to the precedent of the corresponding methyl ester, diphenyl maleate radical-anion isomerises only slowly
to the fumarate radical-anion, yet co-electrolysis of 2,3-bis(bromomethyl)-1,4-dimethoxybenzene and diphenyl
maleate or diphenyl fumarate gives exclusively the corresponding trans-adduct. Co-electrolysis of dimethyl maleate
with either 1,2-bis(bromomethyl)benzene (more easily reduced) or 2,3-bis(bromomethyl)-1,4-dimethoxybenzene (less
easily reduced) gave only o-PX polymer. The results are rationalised in terms of a double nucleophilic substitution
mechanism where electron transfer between dienophile radical-anion and dihalide is relatively slow. Where electron
transfer from maleate or fumarate radical-anions is likely to be fast o-quinodimethanes are formed by redox-catalysis
and they polymerise rather than undergo Diels–Alder reaction. Dimerisation of the dienophile radical-anions, with
k2 = 104 to 105 MϪ1 sϪ1, does not apparently compete with nucleophilic substitution or, where relevant, electron
transfer.
for reduction of 1,4-bis(chloromethyl)benzene and the corre-
sponding p-quinodimethanes. In these last cases p-QDMs
Introduction
o-Quinodimethanes (o-QDMs), derivatives of 5,6-bis(methyl-
could be generated by redox catalysis in the presence of suitable
ene)cyclohexa-1,3-diene 4a, are highly reactive intermediates.
organic mediators.7
The parent species 4a dimerises spontaneously1 above Ϫ80 ЊC
We have also explored the stereochemistry of cycloaddition
for reaction between fumarate and maleate dienophiles and
o-QDMs generated by either direct or mediated electrolysis.
Cava and Napier2 first suggested the participation of an
o-QDM in the conversion of 1,2-bis(dibromomethyl)benzene
to trans-dibromocyclobutabenzene, a reaction earlier reported
Radical-anions generated from dialkyl maleate esters are
by Finkelstein.3 Jensen and Coleman4 also proposed the inter-
known8,9 to isomerise to the fumarate species on the time
mediacy of an o-QDM in the preparation of 1,2-diphenyl-
benzocyclophane.
scale of cyclic voltammetric or rotating ring-disk electrode
experiments. On a preparative electrolysis time-scale such
isomerisation would, if significant, be expected to compete
with electron transfer to the 1,2-bis(halomethyl)arene and be
o-Quinodimethanes undergo Diels–Alder cycloaddition with
dienophiles. Recently it was shown5,6 that o-QDMs generated
by cathodic reduction of 1,2-bis(halomethyl)arenes may react
in situ with dienophiles to give good yields of adducts, even
reflected in the stereochemistry of the resulting adduct.
where the dienophile is sterically hindered. Furthermore rela-
tively long-lived radical-anions generated from the dienophiles
mediate the cathodic elimination from the 1,2-bis(halomethyl)-
arenes.
Results and discussion
Quinodimethane precursors
We here report on the reactions and characterisation of
o-QDMs formed by direct or mediated cathodic reduction
in the absence as well as in the presence of dienophiles. The
1,2-bis(chloromethyl)arenes 2a, b are more difficult to reduce
than the corresponding quinodimethanes, as was found7
o-Quinodimethanes were cathodically generated from 1,2-bis-
(halomethyl)arenes (1a–1d, 2a, 2b) that were mostly prepared
by modification of literature methods. 1-(α-Bromobenzyl)-2-(α-
bromo-α-phenylbenzyl)benzene 1c was prepared by treatment
of the known diol with hydrobromic acid and with hydrochloric
acid for the corresponding chloride 2b.
† For part 53, see ref. 27.
‡ For part 1, see ref. 7.
§ Further experimental details are available as supplementary data.
For direct electronic access see http://www.rsc.org/suppdata/p2/b0/
b007476o/
Characterisation of quinodimethanes by cyclic voltammetry
p-Quinodimethanes (p-xylylenes), intermediates in the form-
ation of poly(p-xylylene) (PPX) polymers, are relatively
DOI: 10.1039/b007476o
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163
This journal is © The Royal Society of Chemistry 2001
153

View Article Online
Table 1 Mediators for the cleavage of 1,2-bis(chloromethyl)benzene
2a
Mediator
ϪEЊ/V vs. SCEa
1.61
Dimethyl terephthalate
Diethyl terephthalate
Perylene
1.61
1.63
1.63
a Formal potentials measured by cyclic voltammetry at 0.3 V sϪ1 as 2
mM solutions in DMF–Bu4NPF6 (0.1 M) at an Hg/Pt electrode with
reference to SCE–calibrated against EЊ for anthracene—see Experi-
mental section for details.
long-lived on a voltammetric time-scale and have been detected
and characterised.7 The 1,4-bis(bromomethyl)arene precursors
are reducible at potentials less negative than those required for
reduction of the QDMs, which allows their formation and
detection by direct reduction. The corresponding chlorides
reduce, irreversibly, at more negative potentials and the
QDMs are rapidly further reduced following their generation
by reductive 1,4-elimination of halide. However, the p-QDMs
could be generated at less negative potentials through redox
catalysis by mediators and this enabled generation from
chloride precursors. Cathodic generation at lower temperature
(Ϫ15 ЊC), combined with relatively fast scan cyclic voltammetry
(160 V sϪ1), resulted in chemically reversible reduction for the
parent p-QDM and hence measurement, in DMF solution,
of its EЊ value (Ϫ1.94 V vs. SCE). Similar approaches are here
described for the detection and characterisation of o-QDMs.
Cyclic voltammetry of 1,2-bis(bromomethyl)benzene 1a [Hg/
Pt cathode, DMF–Et4NBr (0.1 M)] gave a peak at Ep = Ϫ1.22 V
vs. SCE (1 V sϪ1) corresponding to halide cleavage with a
subsequent peak at Ep = Ϫ1.74 V vs. SCE that was chemically
irreversible at scan rates up to 100 V sϪ1. By analogy with the
results7 for the p-QDMs the second peak is assigned to reduc-
tion of the parent o-QDM (o-xylylene) 4a. The radical-anion
of o-QDM 4a, formed at ca. Ϫ1.74 V (1 V sϪ1), is, like that of
the p-QDM,7 too reactive at room temperature (ca. 22 ЊC) to
show chemically reversible reduction at up to 100 V sϪ1. In
contrast to the p-xylylene, reversible reduction of the o-QDM
4a was not achieved at a lower temperature (Ϫ15 ЊC).
Scheme 1
bromoindane 3. This precursor is a single isomer (sharp melting
point, clean H NMR), probably the more stable trans-isomer
1
based on the method of preparation (Grignard addition to
2,2-dimethylindane-1,3-dione followed by treatment of the
1,3-dihydroxy intermediate with HBr), although no assignment
was given in the literature.12 Cyclic voltammetry of 3 indicates a
small reversible peak at ca. Ϫ1.5 V vs. SCE followed by a large
irreversible peak at Ϫ1.99 V (1 V sϪ1). In the presence of a redox
mediator (2,3-dimethylmaleic anhydride, EЊ = Ϫ1.13 V) the
cyclic voltammogram of 5 developed and the formal redox
potential was measured (EЊ = Ϫ1.47 V). Thus the small peak
at Ϫ1.5 V is due to the QDM 5; either the first reduction
wave for 3 is ill-defined, as is common at Hg cathodes with
benzyl bromides, or the peak at Ϫ1.99 V corresponds to
direct reduction of 3, at an unexpectedly negative potential.
One explanation for such a difficult reduction would be that
dibromide 3 is rigid and highly hindered.
The redox behaviour of 1,2-bis(chloromethyl)benzene 2a is
similar to that reported for 1,4-bis(chloromethyl)benzene.7
Reduction is at more negative potentials than required for
reduction of the corresponding QDM. Consequently redox
mediators were sought to enable measurement of the
peak potential of the o-QDM 4a. Those mediators that were
effective are presented in Table 1; others that were tried but were
ineffective were 2,3-dimethylmaleic anhydride, fluorenone, 4-
ethoxycarbonylbenzophenone, and 1,4-diacetylbenzene. Both
dimethyl terephthalate and perylene were particularly effective
in allowing generation of 4a, and observation of its chemically
irreversible reduction peak and measurement of its Ep value
of Ϫ1.74 V, comfortingly identical to that measured by
direct reduction of the bromo precursor. The Ep and EЊ
values measured for the o-QDMs by voltammetry involving
generation by direct or mediated reduction are collected in
Table 2. The parent o-QDM, 4a, is more easily reduced than the
analogous p- compound.7
In the p-QDM cases stabilisation of the radical-anions,
and hence reversible reduction of the phenyl-substituted QDM,
was achieved by phenyl substitution at the methylene terminal
positions and shown7 to be the result of extended delocalis-
ation of the charge and spin. However, for reduction of 1,2-
bis(α-bromobenzyl)benzene 1b this stratagem fails. Cyclic vol-
tammetry gives an initial reduction wave (Ϫ1.46 V, 1 V sϪ1
)
followed by a broad irreversible peak at a very negative
potential (Ϫ2.11 V). Constant potential electrolysis of 1b gave
trans-1,2-diphenyl-1,2-dihydrocyclobutabenzene 8, and this
was shown by cyclic voltammetry to be responsible for the
peak at Ϫ2.11 V. The formation of 8 is rationalised in terms
of the known10,11 ring closure of the o-QDM when formed
photochemically (Scheme 1). The exclusive formation of the
trans-isomer, by conrotatory ring closure, is good evidence that
the first-formed intermediate is the more stable (E,E)-isomer of
the o-QDM 4b.
Substitution by phenyl groups at the termini of the methy-
lene functions, as in 1c, lowers the reversible reduction potential
of the corresponding QDM, 4c, relative to the peak potential
for irreversible reduction of 4a, by about 180 mV but in this
Ring closure is unlikely for the o-QDM 5, which is expected
to be formed by reduction of 1,3-diphenyl-2,2-dimethyl-1,3-di-
154
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163

View Article Online
Table 2 Reduction peak potentialsa of o-QDM precursors and the
corresponding QDMs
Controlled potential electrolysis
Direct electrolyses were typically carried out in DMF–Et4NBr
(0.1 M) on a 0.5–1 g scale in a divided cell with a mercury pool
cathode and an Ag wire reference (see Experimental section
for the relationship with SCE). For indirect electrolysis the
redox catalyst (mediator) was present in 10 mol%. The results
are summarised in Table 3.
1,2-Bis(chloromethyl)benzene 2a was electrolysed in the
presence of various redox mediators selected on the basis of the
data in Table 1. Constant potential electrolysis at the potential
of the mediator (see Table 1) gave poly(o-xylylene) 6. Using
diethyl terephthalate as mediator 5,6,11,12-tetrahydrodi-
benzo[a,e]cyclooctene 7 was formed as a side product. Perylene
is the most efficient mediator, as it gave good yields of polymer
[poly(o-xylylene) 6] with almost complete recovery of mediator.
ϪEЊ/V for
corresponding
Substrate
ϪEp(1)/Vb
QDM
ϪEЊ or Ep/V
p-QDM7
1a
2a
1b
1c
2b
3
1.22
2.08
4a
4a
(8)
4c
4c
5
1.74 (Ep)b
1.74c (Ep)b
2.11 (Ep)b
1.56 (EЊ)
1.94
1.94
—
1.56
1.56
—
1.46
0.88d
1.88
1.56c (EЊ)
1.47c (EЊ)
[1.99]e
a Potentials measured by cyclic voltammetry at 1–200 V sϪ1 as 2 mM
solutions in DMF–Bu4NPF6 (0.1 M) at an Hg/Pt electrode with
reference to SCE—calibrated against EЊ for anthracene—see Experi-
mental section for details. b Peak potential at 1 V sϪ1
.
c Determined
e Not clear
in the presence of mediator. d Peak potential at 0.3 V sϪ1
.
that this is the first reduction potential, see text.
Table 3 Products of controlled potential electrolysisa in the absence
and presence of a mediator
Product (%)
[% yield of
recovered
Substrate Mediator
Charge/Fb
mediator]
1a
2a
2.0
2.0
6 (33)
6 (24)
ortho-Xylene
(4.7)
7 (5.7)
2a (17)
2a
3.7c
2.0
6 (30)
o-Xylene (33)
7 (7)
6 (42) [24]
2a
Diethyl 9,10-dihydro-
9,10-ethenoanthracene-
11,12-dicarboxylate
Dimethyl terephthalate 2.0
6 (42) [62]
6 (29)
Diethyl terephthalate
2.0
o-Xylene(14)
[94]
6 (43) [90]
8 (74)
9 (37)
10 (7)
Perylene
2.0
2.0
2.0
1b
1c
2.0
1.6
11 (37)
13 (23)
11 (37)
14 (9)
3
3
2,3-Dimethylmaleic
anhydride
For comparison, direct electrolysis of 2a was performed at
Ered = Ϫ2.10 V vs. Ag, i.e. at the foot of the reduction wave.
As in the para case the electrolysis was either stopped after 2 F
or when the background current was reached. The yield of the
poly(o-xylylene) 6 at 2 F was not as high as in the para case and
unreacted starting material was present (17.5%). Exhaustive
electrolysis (to 3.7 F) gives more polymer but also significant
further reduction, consistent with the relatively easy reduction
of o-xylylene 4a.
Constant potential electrolysis of 1,2-bis(α-bromobenzyl)-
benzene 1b resulted in the formation of trans-1,2-diphenyl-1,2-
dihydrocyclobutabenzene 8. Quinkert et al.10,11 investigated the
stereochemistry of the isomeric 1,2-diphenyl-1,2-dihydrocyclo-
butabenzene adducts that were obtained from the photo-
decarboxylation of 1,3-diphenylindanone (Scheme 1). The
reported 1H NMR spectrum of the trans-isomer 8 showed
a narrow absorption region for the aromatic signals and the
phenylmethine protons (PhCH) were at higher field than for the
corresponding cis-isomer. Comparing these data with those
obtained for the product isolated from the controlled potential
electrolysis of 1b confirmed formation of the trans-1,2-di-
phenyl-1,2-dihydrocyclobutabenzene, 8, the valence tautomer
of the o-QDM.
a Divided cell, Hg pool cathode, DMF–Et4NBr (0.1 M); potential used
corresponds to the reduction peak of the mediator (indirect electrolysis)
or the first reduction peak of the substrate (direct electrolysis)—see
Experimental section for details. b Based on the dihalide. c At foot of
reduction peak.
case the o-QDM and its p-analogue7 have identical EЊ
values (Ϫ1.56 V). For the p-QDMs, where it was also possible
to measure7 the relative effects of substitution by two and
three phenyl groups, it was found that the third phenyl sub-
stituent did not participate in delocalising charge and spin
because it is sterically prevented from achieving co-planarity
with the relevant π-system. It seems that the radical-anion of
o-xylylene, 4a, formed at Ϫ1.74 V, is more reactive than that
of p-xylylene, formed at Ϫ1.94 V. In cyclic voltammetry,
reversibility for the reduction of p-xylylene can be achieved
relatively easily. Reversibility in the reductions of both 4c and
of the analogous p-QDM is easily observed, so presumably
the radical-anions of these species are similarly unreactive by
virtue of a combination of steric hindrance and delocalisation
and there would seem therefore to be little intrinsic difference in
their reduction potentials.
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163
155

View Article Online
1-(α-Bromobenzyl)-2-(α-bromo-α-phenylbenzyl)benzene 1c
was electrolysed exhaustively at Ϫ1.2 V vs. Ag wire. A yellow
solid was isolated after ether extraction, which was neither a
polymer nor 1,1,2-triphenyl-1,2-dihydrocyclobutabenzene13 but
was found by TLC to be mixture of two products. Separation
by column chromatography on silica gel gave, as one product,
9,10-diphenylanthracene 9, identified by its melting point,
1H NMR and MS. This product has also been found14 in
photolysis of 1,1,3-triphenylindan-2-one. The photolytic
reaction is also believed14 to produce the o-QDM 4c which
decays by two competing cyclisation pathways leading to 9,10-
diphenylanthracene 9 and 1,1,2-triphenyl-1,2-dihydrocyclo-
butabenzene. The other product proved to be 2-(diphenyl-
methyl)benzophenone 10,
a compound produced in the
photolysis15 of phenyl-substituted o-QDM 4c. The o-QDM
cyclises to 9,10-diphenylanthracene 9, and 2-(diphenyl-
methyl)benzophenone 10 is probably a hydrolysis product of
the starting material, indicating low current efficiency in the
electrochemical reaction. A rationalisation of the results is
given in Scheme 2.
Scheme 3
the corresponding Diels–Alder adducts.5,6 Typically maleic
anhydride or maleimide derivatives have been used as dieno-
philes and they are more easily reducible than the 1,2-bis-
(bromomethyl)arenes with which they are co-electrolysed.
Consequently the mediated electrolysis of the halomethyl
compounds predominates, as has been demonstrated6 by
cyclic voltammetric experiments, which display characteristic
catalytic currents associated with redox catalysis and, in
coulometric experiments,17 a change from 1 F reduction of the
dienophiles to 2 F reduction in the presence of 1 equivalent of
the less easily reduced 1,2-bis(bromomethyl)arenes. We find17
that o-QDMs are similarly generated by mediated reduction
in aqueous electrolyte where reduction of the 1,2-bis(bromo-
methyl)arenes, followed by in situ Diels–Alder follow-up
reaction, completely suppressed the very fast electrohydro-
dimerisation of the dienophile (N-methylmaleimide) radical-
anion. In this case it is proposed that a pre-complex is formed
by “hydrophobic packing” between the 1,2-bis(bromomethyl)-
arene and the dienophile. Although the cases studied to date
involve redox catalysis by dienophile, o-QDMs can in principle
be generated by direct electrolysis and trapped as Diels–Alder
adducts by dienophiles less easily reduced than the 1,2-
bis(bromomethyl)arenes. In these cases the competing reaction
would be polymerisation.
We have attempted to demonstrate the direct reduction route,
and at the same time assess the relative reactivity of o-QDMs
towards cycloaddition vis à vis polymerisation. Co-electrolysis
of combinations of derivatives of 1,2-bis(bromomethyl)arenes
and of maleates and fumarates has been examined for cases
where substitution in either or both components is used
to manipulate reduction potentials. Thus it is possible to have
the maleates and fumarates more easily reduced than the 1,2-
bis(bromomethyl)arenes and vice versa. Maleate and fumarate
will be trapped by o-QDMs as cis- or trans-adducts respectively,
but should electron transfer to the dienophiles be the first step
then the trans-adduct may result irrespective of the dienophile
if there is fast isomerisation8,9 of maleate radical-anions to the
more stable fumarate radical-anions.
Scheme 2
For 1,3-diphenyl-2,2-dimethyl-1,3-dibromoindane 3, con-
stant potential electrolysis at Ep = Ϫ1.55 V vs. Ag wire resulted
in the formation of 1,3-diphenyl-2,2-dimethylindane 11 (37%),
which indicates that the initially formed quinodimethane is
rapidly reduced further at the electrolysis potential. In contrast,
mediated electrolysis of 1,3-diphenyl-2,2-dimethyl-1,3-di-
bromoindane 3 at Ep = Ϫ0.95 V vs. Ag wire, in the presence
of 2,3-dimethylmaleic anhydride, gave several products.
Separation by column chromatography and characterisation
showed one product to be 2-(2-methyl-1-phenylpropenyl)-
benzophenone 13 (23%) together with a small amount of
1,3-diphenylisobenzofuran 12, the hydrogenated product 11
(37%) and a Diels–Alder product 14 (9%). This last product was
also reported by Alder and Fremery12 and Johansson and
Skramstad.16 Part of the starting material (23%) is probably
converted chemically into the ring-opened ketone 13. This
side reaction can account for the consumption of only 1.6 F.
The formation of the 1,3-diphenylisobenzofuran 12 from 5
has been claimed16 in the context of the lability of 5, which
on exposure to air at ambient temperature changed colour
from orange to colourless. The pathways are summarised in
Scheme 3.
It is of course possible that electron transfer from maleate
radical-anion, to produce the o-QDMs with subsequent
Diels–Alder reaction, is faster than isomerisation of the
maleate radical-anion although this competition also depends
on substrate concentration. We therefore also examined the
stereochemistry of adducts formed in cases where the maleates
are more easily reduced than are the 1,2-bis(bromomethyl)-
arene precursors. However, even if formed initially the cis-
Diels–Alder adducts might epimerise under the usually basic
conditions of cathodic reduction and this possibility was also
tested.
A comparison of Diels–Alder reactions between dienophiles and
o-QDMs generated either by redox catalysis or by direct
reduction
The possible pathways for cathodic reaction of the fumarate
and maleate esters alone and in the presence of 1,2-
bis(bromomethyl)arene are outlined in Scheme 4, in which the
particular substrates used in this study are labelled.
The electrogeneration of o-QDMs in the presence of dieno-
philes has been extensively used as a preparative route to
156
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163

View Article Online
Scheme 4
the voltammograms for dimethyl maleate and dimethyl
fumarate (Fig. 1a).
The reactivity of maleate radical-anions
The radical-anions of dialkyl maleates are known8,9 to iso-
merise to the corresponding fumarate species and, further-
more, they undergo dimerisation, eventually to yield18 tetra-
alkylbutane-1,2,3,4-tetracarboxylate. For bimolecular dimeris-
ation in DMF at ambient temperature (ca. 24 ЊC) of dimethyl
maleate radical-anions Yeh and Bard9 reported k2 = 1.9 ×
105 MϪ1 sϪ1 and for unimolecular isomerisation a value of
k1 = 2.2 sϪ1, both values estimated on the basis of simulation
of rotated ring-disk electrode experiments. The corresponding
fumarate radical-anion was reported to dimerise9 with k2 =
Contrary to expectation we find that under comparable
conditions the radical-anion of diphenyl maleate does not
isomerise rapidly to the more stable fumarate species. Over a
scan rate range of 0.1 to 50 V sϪ1 no re-oxidation of diphenyl
fumarate radical-anion was seen clearly (Fig. 1b(iv)), whereas
for the reduction of dimethyl maleate the re-oxidation of the
corresponding fumarate radical-anion is prominent (Fig. 1a).
The voltammogram in Fig. 1b(iv) reveals a small impurity
peak, probably due to diphenyl fumarate, that distorts com-
parison of the peak heights. We have determined rates of
dimerisation for the radical-anions of diphenyl maleate and
diphenyl fumarate and, for comparison, have re-examined
the corresponding kinetics of the dimethyl esters. The com-
pounds used in this part of the study, together with relevant
voltammetric and kinetic data, are included in Table 4. We
find that under our reaction conditions the radical-anion of
diphenyl maleate dimerises about an order of magnitude faster
than does the diphenyl fumarate radical-anion. The radical-
anions of both diphenyl maleate and fumarate are considerably
more reactive than the corresponding dimethyl species; this
is not evident from Fig. 1 because of the different intervals
between peak potential and switching potential in the two
examples. Our result for the dimerisation of the radical-anion
110 MϪ1 Ϫ1. The greater rate for reductive dimerisation of the
s
maleate esters is the reason for their choice in the large-scale
production18 of butane-1,2,3,4-tetracarboxylic acid.
In cyclic voltammetry, at low scan rates, the reduction of
dimethyl maleate (2–5 mM) gives a single peak that is
chemically irreversible because of the dimerisation follow-up
reactions. However, at 10 V sϪ1 (Fig. 1a and 1b) the appearance
of the voltammograms is changed; the dimerisation is partly
outrun but isomerisation to the less reactive dimethyl fumarate
radical-anion is evident in the re-oxidation peak at ca. 0.3 V less
negative than the forward reduction peak. This has been
shown8 previously to correspond to oxidation of the dimethyl
fumarate radical-anion, a result reproduced here by overlaying
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163
157

View Article Online
Table 4 Components of co-electrolyses, of reduction potentialsa and kinetic data (Scheme 4)
d
Scan rate/
Ϫ1 sϪ1
Concn./
mM
10Ϫ4 k2
Ϫ1 sϪ1
/
Compoundb
ϪEp(1)c
ϪEЊ
V
M
k1/sϪ1
[2.2]9
15a, R1 = H, R2 = Me
—
1.59
30
40
2
5
2
5
1.22
1.30
1.21
1.29
1.26 0.05
[1.9 × 105]9
12.0
12.2
11.7
12.2
12.0 0.2
0.109
15b, R1 = H, R2 = Ph
—
—
1.24
20
30
0.88e
1.76
0.88
1.76
16a, R1 = H, R2 = Me
16b, R1 = H, R2 = Ph
1.30
1.07
0.8
2
5
5
0.118
0.114 0.005
[1.1 × 102]9
1.55
1.52
1.53
10
20
1
2
1
2
1.52
1.53 0.001
1a, X = H
1b, X = OMe
1.22
1.90
—
—
a Hg/Pt cathode, DMF–Bu4PF6 (0.1 M), 0.8–40 V sϪ1, substrate 1–5 mM [second-order rate constants acceptably independent of substrate
concentration]. b See Scheme 4. c Potentials wrt. SCE, calibrated against EЊ for anthracene—see Experimental section for details. d By cyclic
voltammetry with digital simulation (BioAnalytical Systems DigiSim v2.1), details of simulations, including parameters used, are available as
supplementary data. e After allowance for an impurity, probably diphenyl fumarate, see Fig. 1b.
dimerisation reactions. Examination of the voltammograms in
Fig. 1a shows that the forward reduction peak for dimethyl
maleate has, by comparison with the reversible peak for
dimethyl fumarate, n > 1. It seems that, in addition to hydro-
dimerisation, another reaction is taking place, probably hydro-
genation (n = 2) involving trace acidic impurities in the DMF.
This explanation does not account, however, for our observa-
tion of a lower rate of dimerisation of dimethyl maleate
radical-anions. However, the diphenyl esters show no similar
abnormality in cyclic voltammetry in the same medium. At
30 and 40 V sϪ1 the isomerisation of the maleate radical-anion
to the fumarate radical-anion is almost completely outrun
although there may still be unquantified competition from
protonation in the methyl maleate case. In deriving the rate
constants for dimerisation by simulation both possibly
competing reactions were ignored.
The dialkyl maleates are known18 efficiently to hydrodimerise
under preparative conditions and we also find that controlled
potential electrolysis of diphenyl maleate (Hg cathode, DMF–
Bu4NPF6 (0.1 M)) proceeds smoothly to 1 F. Aqueous work-up
is made difficult by the rapid hydrolysis of the phenyl ester
1
products, but H NMR examination of the crude residue after
solvent removal indicates that a substantial product (ca. 50%)
is a mixture of meso- and ( )-isomers of the expected linear
hydrodimer (see Experimental section). A second product has
not yet been isolated or identified.
Because of the uncertainty over follow-up reactions other
Fig. 1 Cyclic voltammetry in DMF–n-Bu4NPF6 (0.1 M); Hg/Pt
cathode, 10 V sϪ1, substrate 2 mM, of dimethyl fumarate [1a(i)] and
dimethyl maleate [1a(ii)] and of diphenyl fumarate [1b(iii)] and diphenyl
maleate [1b(iv)]—see Table 4 for data and the text for details.
than dimerisation, it was not possible to simulate the cyclic
voltammogram in Fig. 1a, with a view to determining the rate
constant for isomerisation of dimethyl maleate radical-anion
to dimethyl fumarate radical-anion. The voltammograms for
dimethyl fumarate, and diphenyl maleate and fumarate are not
substantially affected by adventitious impurity in the DMF.
Consequently, digital simulation provides kinetic data for
dimerisation of the radical-anions of diphenyl maleate and
diphenyl fumarate. We did not find conditions in the scan rate
range 1–50 V sϪ1 where a fumarate re-oxidation peak was
significant on reduction of diphenyl maleate. A comparison
of Figs. 1a and 1b clearly indicates that isomerisation of
the diphenyl maleate radical-anion is much slower than that
of dimethyl fumarate is not in good accord with that found
earlier by Yeh and Bard9 who used a different technique
(experiments at rotating ring-disk electrodes) and who took
steps rigorously to dry the solvent (DMF). In our case the
electrolyte was dried in a heated vacuum pistol and the DMF
was freshly distilled and both solvent and electrolyte were
introduced into the cell after passage through a column of
activated alumina. However, it is likely that the solutions still
contained adventitious water, known19 to accelerate hydro-
158
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163

View Article Online
Table 5 Co-electrolysis experimentsa
QDM
precursor
Dienophile
Method
Charge/F
∆Eb
Product
Yield (%)
1d
1d
1d
1a
1d
Diphenyl fumarate (16b)
Diphenyl maleate (15b)
Diphenyl maleate (15a)
Diphenyl maleate (15a)
Diphenyl maleate (15b)
Cathode
Cathode
Cathode
Cathode
Zn dust
3.7
3.2
2.3
2.5
—
0.86
0.66
trans-17
97
85
—
—
ca. 60
ca. 6
67
trans-17
0.31
o-PX polymer
o-PX polymer
trans-17
Ϫ0.37c
—
cis-17
1d
1a
Diphenyl fumarate (16b)
Diphenyl fumarate (16b)
Zn dust
Zn dust
—
—
—
—
trans-17
trans-17
46
a Hg pool cathode, divided cell, DMF–Et4NBr (0.1 M), controlled potential (see Experimental for details), substrate concentrations 25–50 mM,
dienophile in two-fold excess, charge based on the dihalide. b Difference between Ep for dihalide and EЊ for dienophile. c i.e. Dienophile less easily
reduced by 0.37 V.
Fig. 3 X-Ray crystallographic structure of trans-17.
Fig. 2 Cyclic voltammetry of 15b with incremental addition of 1d;
Hg/Pt cathode, DMF–Et4NBr (0.1 M), 0.5 V sϪ1
.
catalysis is usually observed (∆E ≤ 0.5 V) and electron transfer
is likely to be relatively slow. Thus, although it may be that
the reduction of 1d is mediated by the radical-anion of 15b,
a two-fold increase in current would also be consistent with a
double nucleophilic attack of radical-anion on the dibromide
and this possibility is included in Scheme 4.
of the corresponding dimethyl species, although we cannot say
by how much.
On a preparative electrolysis time-scale the dimerisation
of dienophile radical-anions could in principle compete with
electron transfer to the QDM precursors (mediated reduction).
The problem of competing dimerisation must also be con-
sidered in the use of maleic anhydride derivatives as redox
mediators. They were found5,6 to be very effective as dual
mediators/dienophiles, but the anhydrides give fairly persistent
radical-anions in DMF. In the case of diphenyl maleate,
because isomerisation to fumarate is slow, rapid electron trans-
fer to give an o-QDM, followed by cycloaddition, would pre-
serve in 17 the cis-stereochemistry of the maleate dienophile,
providing that cis-17 is stable under the reaction conditions.
The reduction potentials of maleate and fumarate esters,
and of the 1,2-bis(bromomethyl)arenes 1a and 1d, are found
to be very sensitive to substitution (Table 4). In particular, the
diphenyl esters 15b and 16b are more easily reduced than
the corresponding methyl esters 15a and 16a. In the presence
of increasing amounts of the dibromide, 1d, the cyclic
voltammetric reduction current for diphenyl maleate (15b) is
seen to increase by about two-fold (Fig. 2a). At the scan rate
used the reduction of 15b is not reversible and the current
increase probably relates to a switch from 1 F dimerisation to
2 F reductive cleavage, i.e. it is not clearly a catalytic current
associated with redox catalysis. The manifestation of redox
catalysis is typically found for maleic anhydride derivatives as
dienophiles/mediators.6 In these earlier cases where the catalytic
current is much greater than a two-fold increase the differences
between reduction potentials of the mediator and dibromide
[∆E = EЊ (mediator) Ϫ Ep (dibromide)] were much smaller
(typically5,6 ca. 0.3 V) than for the case in Fig. 2a, where
∆E = 0.67 V. This is at the limit of separations at which redox
Co-electrolyses where the dienophile is the more easily reduced
component. Co-electrolysis experiments, at controlled potential,
were designed in which either the dienophile or the 1,2-bis-
(bromomethyl)arene could be the more easily reduced compon-
ent. The results are summarised in Table 5. Preparative-scale
exhaustive co-electrolysis of diphenyl maleate, 15b, and 2,3-
bis(bromomethyl)-1,4-dimethoxybenzene, 1d, with the maleate
in two-fold excess, proceeded to 3.2 F (based on the dibromide)
and gave trans-17 (85%, R1 = H, R2 = Ph) with no trace of the
cis-isomer. The assignment of trans-stereochemistry was made
by X-ray crystallography (Experimental section and Fig. 3).
Similar co-electrolysis of 1d with diphenyl fumarate, 16b, con-
sumed 3.7 F and gave a near-quantitative yield of the adduct,
trans-17.
Isolation of the trans-adduct only is consistent with: (a) prior
generation of radical-anions of the dienophiles with cis–trans
isomerisation of the maleate radical-anion being faster than
electron transfer to the QDM precursor; (b) formation of the
cis-adduct from maleate with rapid epimerisation by electro-
generated base; and (c) double nucleophilic attack by the
radical-anions, with equilibrium control. We have shown that
the diphenyl maleate radical-anion isomerises rather slowly
so possibility (a) is unlikely given that either electron transfer
or nucleophilic reaction between maleate radical-anion and
1,2-bis(bromomethyl)arene is sufficiently fast to outrun the
competing dimerisation (k2 = 105 MϪ1 Ϫ1). Route (c) for reac-
s
tion between 1d and diphenyl maleate (15b) is supported by the
large ∆E value (Table 5) that presages slow electron transfer;
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163
159

View Article Online
the lack of a convincing catalytic current (Fig. 2) also favours
this explanation.
radical-anions and the dihalides is slow (∆E ≥ 0.5 V) the
double nucleophilic substitution route comes into play. This
also accounts for the exclusive formation of trans-cycloadducts.
The expected isomerisation of diphenyl maleate radical-anions
to the fumarate species is surprisingly slow but even had the
expected cis-adduct been formed initially by the Diels–Alder
route it would have rapidly to epimerised in the presence of
electrogenerated base, probably the dianion formed by rapid
dimerisation of the maleate radical-anions.
Epimerisation of cis-17 in the presence of diphenyl maleate
radical-anion. The cis-isomer was available for comparison,
having been prepared by unambiguous synthesis (see Scheme 5)
Experiments have also been carried out using chemical
generation (Zn dust reduction) of the o-quinodimethanes
from the 1,2-bis(bromomethyl)arenes in the presence of the
maleate and fumarate dienophiles. Generation of 4d by
zinc-dust reduction, in DMF, and in the presence of diphenyl
maleate, 15b, gave a mixture of the adducts 17 (ca. 66%) with
a trans:cis ratio of 10:1. The formation of trans-isomer in
this case is probably also a result of partial epimerisation of
the cis-isomer through enolisation of cis-adduct catalysed by
ZnBr2, a product of the reaction. In further experiments, using
the Zn-dust method, the corresponding trans-17 adducts were
formed in moderate yields from combinations of 1d or 1a with
diphenyl fumarate. The electrochemical method is cleaner and
more efficient.
Scheme 5
It is unlikely that these reactions proceed by the double
nucleophilic substitution route, yet if o-quinodimethanes are
generated they apparently react by Diels–Alder cycloaddition
rather than polymerisation—the preservation of some cis-
stereochemistry supports this. This contention contrasts with
the findings for the electrochemical method. The probable
reason is that the zinc-dust reductions take place over a
considerable period (5–10 hours). Consequently the o-QDMs
are formed slowly and in low concentration relative to the
dienophile. Diels–Alder reaction may therefore compete
favourably against polymerisation.
and the corresponding trans-isomer was the major product of
preparative electrolysis. The adducts also contain convenient
aromatic chromophores for detection by UV absorption.
Consequently conditions were established for the base-line
separation in HPLC of a mixture of cis- and trans-17 together
with diphenyl maleate (15b). The extinction coefficients of the
cis- and trans-isomers proved, conveniently, to be identical.
An equimolar mixture of diphenyl maleate (15b) and the
electrochemically inactive cis-17 was electrolysed in DMF–
Et4NBr (0.1 M) at an Hg cathode at the first reduction potential
of 15b. Analysis at the start of the electrolysis and after the
consumption of 0.56 F showed that cis to trans epimerisation
had occurred to the extent of 53%. This is compelling evidence
for epimerisation pro rata to the formation of diphenyl maleate
radical-anions.
Experimental
Materials and general procedures
General-purpose reagents grade solvents (GPR) were used
unless stated otherwise. For preparative electrolyses DMF
(Aldrich) was used without further purification. For cyclic
voltammetry and controlled potential coulometry DMF
(Aldrich, HPLC grade) was used and stored over freshly baked
(150 ЊC) 4 Å molecular sieve. Supporting electrolytes were
tetraethylammonium bromide (Aldrich) and tetraethyl-
ammonium hexafluorophosphate (Aldrich), without any
further purification.
Co-electrolyses where the dienophile is the less easily reduced
component. The o-QDM precursor 1a is more easily reduced
than dimethyl maleate, 15a (Table 4). Consequently, direct
reduction at the potential for conversion of 1a in the presence
of unreduced dimethyl maleate should result in formation of
cis-17, (R1 = H, R2 = Me, X = H). Co-electrolysis to just over 2 F
yielded copious amounts of the corresponding o-PX polymers
with no detectable Diels–Alder adduct. The order of reduction
is reversed for 1d vis à vis 15a and mediated reduction of 1d
(∆E = 0.39 V) would be expected. The corresponding o-PX was
also the only detectable product of this co-electrolysis.
Melting points were determined on an Electrothermal
capillary tube melting point apparatus and were uncorrected.
For polymers the melting points were obtained from DSC
experiments using a Perkin-Elmer PC series DSC7 calorimeter.
Infrared spectra were recorded on KBr discs using a Perkin-
Mechanistic conclusions. A rationalisation must crucially
account for (i) the reaction between the radical-anions of
diphenyl maleate and fumarate and 2,3-bis(bromomethyl)-1,4-
dimethoxybenzene (1d) giving exclusively the trans-cyclo-
adducts, despite the fact that the separations of peak potentials
are >0.5 V suggesting relatively slow electron transfer and
therefore inefficient redox catalysis; and (ii) that where redox
catalysis to form the quinodimethane is likely to be favoured,
such as in co-electrolysis of 1d and dimethyl maleate (∆E =
0.31 V), cycloaddition is not found and polymerisation of
the QDM predominates.
The simplest explanation is that when o-QDMs are formed
by redox catalysis involving the maleate and fumarate dieno-
philes Diels–Alder reaction is slower than polymerisation. We
do know from the cyclic voltammetric studies (see earlier) that
the parent o-xylylene does not give reversible redox behaviour
at up to 160 V sϪ1, i.e. it reacts (polymerises) relatively quickly.
For those cases where electron transfer between the dienophile
1
Elmer 1600 series FTIR spectrophotometer. H NMR spectra
and 13C NMR (63 MHz) on a Bruker AM 250 (250 MHz)
spectrometer. Solid-state 13C NMR spectra were recorded at
75 MHz on a Bruker MSL-300 multinuclear NMR spectro-
meter of the University of London Intercollegiate Research
Service (ULIRS). Direct pyrolysis mass spectra were recorded
on a Kratos MS5ORT mass spectrometer.
HPLC experiments were carried out using a Hewlett Packard
HP1100 Series Liquid Chromatograph with UV detection at
270 nm. The instrument was equipped with a Zorbax Sil
column (5 µm, 4.6 mm ID × 250 mm) and guard, run iso-
cratically with EtOAc–hexane (30:70, Merck HPLC grade)
eluent at 1.0 ml minϪ1 and an injection volume of 10 µl.
Crystallography
The intensity data were collected on an Enraf Nonius CAD-4
diffractometer using Mo-Kα radiation (λ = 0.71069 Å) with an
160
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163

View Article Online
ω-2θ scan at 120 K. The unit cell parameters were determined
by least-squares refinement on diffractometer angles 6.4 ≤
θ ≤ 11.5Њ for 25 automatically centred reflections.20 All data
were corrected for absorption by semi-empirical methods
(ψ scan)21 and for Lorentz-polarization effects by XCAD4.22
The structure was solved by direct methods using SHELXS-
97,23 and refined anisotropically (non-hydrogen atoms) by full-
matrix least-squares on F2 using the SHELXL-97 program.22
The program ORTEP-324 was used for drawing the molecules.
WINGX25 was used to prepare the material for publication.
1-(ꢀ-Bromobenzyl)-2-(ꢀ-bromo-ꢀ-phenylbenzyl)benzene (1c).
1-(α-Hydroxybenzyl)-2-(α-hydroxy-α-phenylbenzyl)benzene
(1.0 g, 1.12 mmol) was dissolved in glacial acetic acid (20 ml)
and the mixture heated gently until all the diol was dissolved
when HBr (5 ml in HOAc (45%)) was added. The mixture was
heated under reflux at 120 ЊC (1 hour), cooled to room tem-
perature, diluted with water (50 ml) and the product extracted
into chloroform. The organic layer was washed (H2O), dried
(MgSO4) and the solvent evaporated. Crystals precipitated on
cooling (0.91 g, 68%); mp 109 ЊC; 1H NMR δH (250 MHz,
CDCl3) 6.40 (1H, s, PhCH(Br)Ar), 7.35–8.05 (19H, m, ArH),
IR (KBr)/cmϪ1 3058 (m, νaromؒC–H), 3029 (s, νsatؒC–H), 1596
(s, benzene ring), 1482 (s, νC–H), 738 (m, δAr–H, 2 adjacent
H), 700 (m, δC–H); m/z 491.9590, C26H20Br2 requires 491.957;
elemental analysis, found (calcd): C 63.1% (63.4), H 4.1 (4.09),
Br 32.8 (32.52).
Crystal data. The structure of trans-17, (R1 = H, R2 = Ph,
X = OMe) (Fig. 3) was determined by M. Motevalli (QMW
Crystallographic Service): C26H24O6, M = 432.45, orthorhom-
bic, a = 18.582(5), b = 10.594(4), c = 22.431(8) Å, U = 4416(3)
Å3, T = 291 K, space group Pbca, Z = 8, µ = 0.092 mmϪ1
.
The final R indices were R1 = 0.0656, wR2 = 0.0515. CCDC ref-
erence number 188/281. See http://www.rsc.org/suppdata/p2/
b0/b007476o/ for crystallographic files in .cif format.
1-(Chlorobenzyl)-2-(phenylchlorobenzyl)benzene (2b). 1-(α-
Hydroxybenzyl)-2-(α-hydroxy-α-phenylbenzyl)benzene (1.0 g,
1.12 mmol) was dissolved in glacial acetic acid (20 ml). Con-
centrated HCl (10 ml) was added and refluxed (120 ЊC) for
30 minutes. The solution was cooled to room temperature,
extracted with ether, washed with water, dried (MgSO4) and the
solvent evaporated. A yellow oil was obtained (0.40 g, 36%);
1H NMR δH (250 MHz, CDCl3) 6.14 (1H, s, PhCH(Cl)Ar),
7.15–7.95 (19H, m, ArH); IR (KBr)/cmϪ1 3059 (m, νaromؒC–H),
3030 (s, νsatؒC–H), 1597 (s, benzene ring), 1482 (s, νC–H),
738 (m, δAr–H, 2 adjacent H), 699 (m, δC–H); m/z 402.0949,
C26H20Cl2 requires 402.0941; elemental analysis, found (calcd):
C 77.3% (77.59), H 5.1 (5.01), Cl 17.3 (17.39).
Electrochemical experiments
Cyclic voltammetric experiments (CV) were performed using
a VersaStat EG&G Princeton Applied Research potentiostat
with Model 270/250 Research Electrochemistry Software
v 4.00. For faster scan rates (1–100 V sϪ1) either an assembly of
a digital oscilloscope (Nic-310, Nicolet) connected to a wave-
form generator (PP R1, Hi-Tek Instruments, England), a
DT 2101 (Hi-Tek Instruments, England) potentiostat and a
XYt-recorder (PM 8271, Philips) or an EG&G Princeton
Applied Research potentiostat Model 263A with appropriate
software control were used.
trans-1,2,3,4-Tetrahydro-5,8-dimethoxynaphthalene-2,3-di-
carboxylic acid diphenyl ester (trans-17, R1 ؍ H, R2 ؍ Ph,
X ؍ OMe). Activated zinc dust (0.5 g) was suspended in a
solution of 2,3-bis(bromomethyl)-1,4-dimethoxybenzene 1d
(0.2 g, 0.0006 mol) and diphenyl fumurate 16b (0.5 g, 0.002 mol)
in DMF (5 ml), under nitrogen. The reaction was followed
by TLC (CH2Cl2) and after 5 hours the mixture was filtered
through Celite and the solvent evaporated. The oil obtained
was suspended in ethyl acetate (50 ml) and the solution washed
with brine (150 ml). After drying (MgSO4) and solvent removal
flash chromatography (CH2Cl2 85%, petroleum ether 15%) gave
the trans product, (0.178 g, 67%); mp 128–130 ЊC; m/z 432.1573,
C26H24O6 requires 432.1566. Full NMR data (1H and 13C) are
deposited as supplementary data, and for the relevant X-ray
crystallographic structure, see Fig. 3.
Glass cells for cyclic voltammetry were undivided and
equipped with a mercury coated platinum disk (0.1–0.6 mm)
working electrode (cathode), Ag wire reference electrode and
platinum coil counter electrode (anode). The experiments were
carried out in 0.1 M Et4NBr–DMF or 0.1 M Bu4NPF6 DMF
solutions. The concentration of the substrate was 2–3 mM and
the scan rate varied from 0.1–200 V sϪ1. Because the potential
of the Ag wire reference electrode can vary over a relatively
short time-scale, the EЊ value for anthracene was measured
at the beginning and at the end of a series of measure-
ments, against the Ag wire electrode, so that reduction poten-
tials could be consistently referred to the SCE electrode (EЊ for
anthracene = Ϫ1.92 V vs. SCE in DMF26).
Preparative scale electrochemical reductions were performed
using conventional glass cells, with the anode and cathode com-
partments separated by a glass sinter. The cells were equipped
with a mercury pool working electrode (cathode), a Ag wire
reference electrode and a carbon rod counter electrode (anode).
Reduction potentials for controlled potential electrolyses were
determined by cyclic voltammetry on the solutions immediately
prior to electrolysis, i.e. at greater than ideal concentrations.
Consequently the peak potentials so determined, against
the relatively unstable Ag wire reference, serve to enable reduc-
tion at the first reduction wave in an electrolysis immediately
following measurement. However, actual values fluctuate
from day-to-day and cannot be compared with peak potentials
measured at low concentration with reference to SCE (see
above). The reactions were kept under inert atmosphere by
the slow bubbling of dry nitrogen through the electrolytes. The
catholyte was mechanically stirred.
cis-1,2,3,4-Tetrahydro-5,8-dimethoxynaphthalene-2,3-di-
carboxylic acid diphenyl ester, cis-17 (R1 ؍ H, R2 ؍ Ph,
X ؍ OMe). The Zn-dust reduction was carried out as above
in the presence of diphenyl fumarate, except that reaction took
10 hours. Flash chromatography (CH2Cl2 90%, petroleum
ether 10%) gave a mixture of cis and trans-17 (R1 = H, R2 = Ph,
X = OMe), approximately 1:10.
Unambiguous preparation of cis-1,2,3,4-tetrahydro-5,8-di-
methoxynaphthalene-2,3-dicarboxylic acid diphenyl ester, cis-17
(R1 ؍ H, R2 ؍ Ph, X ؍ OMe)
The cis-adduct expected for cycloaddition involving diphenyl
maleate was prepared by first reacting the appropriate QDM
with maleic anhydride and the ensuing cis-adduct was then
converted into the corresponding diphenyl ester.
Starting materials
cis-5,8-Dimethoxy-1,2,3,4-tetrahydronaphthalene-2,3-di-
carboxylic anhydride. 2,3-Bis(bromomethyl)-1,4-dimethoxy-
benzene 1d (0.2 g, 0.0006 mol) and maleic anhydride (0.3 g,
0.003 mol) in DMF (15 ml), under nitrogen, were reacted
with Zn dust (0.4 g, 0.006 mol). Work-up as above gave cis-
5,8-dimethoxy-1,2,3,4-tetrahydronaphthalene-2,3-dicarboxylic
Many of these are known compounds that were prepared by
modification of literature methods. Consequently, descriptions
of their preparation, and physical and spectroscopic data are
deposited as supplementary data. Only new compounds are
described in this section.
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163
161

View Article Online
anhydride as a white solid (0.105 g, 65%); mp 209–210 ЊC
(lit.29 210–211 ЊC); δH(250 MHz, CDCl3) 2.62 (2H, m, CHax-
Heq), 3.34 (2H, m, CHaxHeq), 3.68 (2H, m, CHCO), 3.72 (6H, s,
OCH3), 6.82 (2H, s, ArH).
of the phenyl esters and consequent formation of the tetra-
alkylammonium salt, which is soluble in both organic and
aqueous phases. Examination of the 1H NMR spectrum of
the precipitated mixture of electrolyte and organic product was
problematical because signals from the large excess of electro-
lyte obscured important parts of the spectrum. However,
according to accessible parts of the spectrum, and relative
integration, two products appear to be present, compound A
(55%), proposed as the meso- and ( )-isomers (in the ratio
6:1, or 1:6) of [RO2CCH2CH(CO2R)CH(CO2R)CH2CO2R],
the linear hydrodimer, and an as yet unidentified compound
B (45%). Only two signals are visible for compound B, two
separate double doublets in the ratio 7:1. Compound A (major
isomer, meso or ( )) δH (600 MHz, CDCl3) 3.50 (1H, dd,
J 16.48, 8.92 Hz, CH2CH), 2.89 (1H, dd, J 16.30, 9.42 Hz,
CH2CH); (minor isomer, ( ) or meso) 3.65 (1H, dd, J 19.60,
8.71 Hz). Compound B (major isomer) 4.42 (1H, d, J 7.32 Hz)
(minor isomer) 4.47 (1H, d, J 8.45 Hz).
cis-5,8-Dimethoxy-1,2,3,4-tetrahydronaphthalene-2,3-di-
carboxylic acid diphenyl ester, cis-17 (R1 ؍ H, R2 ؍ Ph,
X ؍ OMe). To a cooled (0 ЊC) solution of cis-5,8-dimethoxy-
1,2,3,4-tetrahydronaphthalene-2,3-dicarboxylic anhydride (see
above) (0.1 g, 3.8 × 10Ϫ4 mol) and phenol (0.01 g, 0.001 mol)
in CH2Cl2 was added 1,3-dicyclohexylcarbodiimide (DCC,
0.17 g, 8 × 10Ϫ4 mol) and 4-(dimethylamino)pyridine (DMAP,
0.01 g, 4 × 10Ϫ4 mol). The reaction was run at 0 ЊC and followed
by TLC. After 4 hours the dicyclohexylurea was filtered off
and the product isolated by flash chromatography (CH2Cl2–
petroleum ether, 3:1) to give cis-5,8-dimethoxy-1,2,3,4-tetra-
hydronaphthalene-2,3-dicarboxylic acid diphenyl ester (0.103 g,
62%); mp 132–133 ЊC; m/z 432.1585, C26H24O6 requires
432.1590; NMR data (1H and 13C) are deposited as supplemen-
tary data.
Co-electrolysis
of
cis-5,8-dimethoxy-1,2,3,4-tetrahydro-
naphthalene-2,3-dicarboxylic acid diphenyl ester (cis-17) with
diphenyl maleate, 15b. Compound cis-17 (0.114 g, 0.25 mmol)
and diphenyl maleate (0.069 g, 0.25 mmol) were co-electrolysed
in DMF–Et4NBr (0.1 M) until 0.6 F had passed. The catholyte
was evaporated to dryness, water added and the product
extracted into CH2Cl2. The combined organic extract was dried
(MgSO4) and evaporated to dryness. The syrup so obtained
was analysed by HPLC at room temperature [column, Zorbax
sil (4.6 mm ID × 25 cm), isocratic, EtOAc–hexane (30:70),
detector wavelength of 270 nm].
Electrolysis
a. ortho-Quinodimethane formation with polymerisation.
General procedure. The substrate (0.5–2.0 g for 80 ml cell 0.2–
0.5 g for 30 ml cell) and mediator, if appropriate, were dis-
solved in the previously de-oxygenated (N2 bubbling) electrolyte
in the cathode compartment of a divided cell. After electrolysis
at controlled potential reaction was considered complete after
the current had decayed to the background level (direct elec-
trolysis) or when 2 F had passed through the cell (mediated
electrolysis). Insoluble (polymeric) products were then filtered
off and soluble products were obtained by adding distilled
water to the reaction mixture. Any precipitate formed was fil-
tered off and washed with water. The remaining solution was
extracted three times with 20 ml diethyl ether and then washed
at least three times with 30 ml water, the ether extract was dried
(MgSO4) and the solvent evaporated. The next purification step
was crystallisation or column chromatography. Physical and
spectroscopic data (FTIR, NMR and MS) for products
obtained in these experiments are available as supplementary
data.
Acknowledgements
We are grateful to QMW for the award of a Graduate Teaching
Studentship (to S. S.) and to the EPSRC for a studentship (to
S. R.). The collaboration between the London and Copenhagen
groups, and a stipend to X. S., was supported under the EU
Human Capital and Mobility Programme No. ERBCHRXCT
920073. We also thank the University of London Intercol-
legiate Research Services (ULIRS) for provision of the High
Field NMR Service at QMW and the solid-state NMR facilities
at University College London.
References
b. ortho-Quinodimethane formation with cycloaddition. The
following example may be taken to represent the typical
procedure.
1 St. K Pollack, B. C. Raine and W. J. Hehre, J. Am. Chem. Soc., 1981,
103, 6308.
2 M. P. Cava and D. R. Napier, J. Am. Chem. Soc., 1957, 79, 1701.
3 H. Finkelstein, Chem. Ber., 1910, 43, 1528.
4 F. R. Jensen and W. E. Coleman, J. Am. Chem. Soc., 1958, 80,
6149.
5 E. Eru, G. E. Hawkes, J. H. P. Utley and P. B. Wyatt, Tetrahedron,
1995, 52, 3033.
6 E. Oguntoye (née Eru), S. Szunerits, J. H. P. Utley and P. B. Wyatt,
Tetrahedron, 1996, 52, 7771.
7 S. Szunerits, J. H. P. Utley and M. F. Nielsen, J. Chem. Soc., Perkin
Trans. 2, 2000, 669.
8 A. J. Bard, V. J. Puglisi, J. V. Kenkel and A. Lomax, Faraday
Discuss. Chem. Soc., 1973, 56, 353.
9 L.-S. R. Yeh and A. J. Bard, J. Electrochem. Soc., 1977, 124, 189.
10 G. Quinkert, K. Opitz, W. W. Wiersdorff and M. Finke, Liebigs Ann.
Chem., 1965, 44, 693.
11 G. Quinkert, Pure Appl. Chem., 1964, 9, 607.
12 K. Alder and M. Fremery, Tetrahedron, 1961, 14, 190.
13 L. A. Carpin, J. Org. Chem., 1969, 34, 461.
14 J. C. Netto-Ferreira, V. Wintgens and J. C. Scaiano, Tetrahedron
Lett., 1989, 49, 6851.
Co-electrolysis of 2,3-bis(bromomethyl)-1,4-dimethoxy-
benzene, 1d, with diphenyl maleate, 15b. 2,3-Bis(bromomethyl)-
1,4-dimethoxybenzene (1d, 0.45 g, 0.0014 mol) and diphenyl
maleate (15b, 0.75 g, 0.0028 mol) were added to the catholyte
(ca. 80 ml DMF–Et4NBr (0.1 M)) in a divided cell equipped
with a stirred mercury pool cathode, a silver wire reference
electrode and a carbon rod anode. After electrolysis at Ϫ1.1 V
the current fell from 130 mA to 8 mA and 3.7 F had been
consumed. The catholyte was removed and DMF evaporated
under reduced pressure. The residue was dissolved in ethyl
acetate (70 ml) and washed with brine (3 × 50 ml). The ethyl
acetate layer was dried (MgSO4) and after solvent removal flash
chromatography (CH2Cl2–petroleum ether, 3:1) gave (trans-17,
R1 = H, R2 = Ph, X = OMe) as a white solid (0.580 g, 97%).
Controlled potential reduction of diphenyl maleate, 15b.
Diphenyl maleate (0.268 g, 0.001 mol) was electrolysed at
Ϫ1.3 V vs. Ag wire in DMF–Bu4NPF6 (0.1 M), according to the
general procedure, until the background current was reached
(0.94 F). The catholyte was evaporated to dryness and the
water (50 ml) added. The beige precipitate was filtered and
dried (1.07 g, containing ca. 90 mg of diphenyl maleate-derived
product, according to 1H NMR). Attempts at normal aqueous
work-up and extraction failed, probably due to rapid hydrolysis
15 T. Paul, M. A. Hassan, H. Korth and R. Sustmann, J. Org. Chem.,
1996, 61, 6835.
16 E. Johansson and J. Skramstad, J. Org. Chem., 1981, 46, 3752.
17 J. H. P Utley, E. Oguntoye (née Eru), C. Z. Smith and P. B. Wyatt,
Tetrahedron Lett., 2000, 41, 7249.
18 E. A. Casanova, M. C. Dutton, D. J. Kalota and J. H. Wagenknecht,
J. Electrochem. Soc., 1993, 140, 2565.
162
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163

View Article Online
of crystal structures, University of Göttingen, Göttingen, Germany,
1997.
19 I. Fussing, M. Gullu, O. Hammerich, A. Hussain, M. F. Nielsen
and J. H. P. Utley, J. Chem. Soc., Perkin Trans. 2, 1996, 649.
20 Enraf-Nonius CAD-4/PC Software, Enraf-Nonius, Delft, The
Netherlands, 1992, ver. Version 1. 5c.
21 A. C. T. North, D. C. Phillips and F. S. Mathews, Acta Crystallogr.,
Sect. A, 1968, 42, 351.
22 K. Harms, XCAD4, Program for data reduction, Philipps-
Universität, Marburg, Germany, 1996.
23 G. M. Sheldrick, SHELX-97, program for solution and refinement
24 L. J. Farrugia, J. Appl. Crystallogr., 1997, 30, 565.
25 L. J. Farrugia, WinGX - a Windows program for crystal structure
analysis, University of Glasgow, Glasgow, 1998.
26 A. J. Bard and L. R. Faulkner, Electrochemical Methods,
Fundamentals and Applications, Wiley, New York, 1980.
27 J. H. P. Utley and C. Z. Smith, J. Mater. Chem., 2000, 10,
2642.
J. Chem. Soc., Perkin Trans. 2, 2001, 153–163
163