Structural effects in radical clocks and mechanisms of grignard reagent formation: Special effect of a phenyl substituent in a radical clock when the crossroads of selectivity is at a metal/solution interface

Article abstract of DOI:10.1002/ejoc.200900096

A large class of radical clocks is based on the intramolecular trapping of a reactive radical by a suitably located, unsaturated system. Depending on the substituents present on this unsaturated system, the rate of cyclisation may vary drastically. This property has been repeatedly used, to diagnose the participation of very short-lived radicals in the mechanisms of a wide variety of reactions. For reactions occurring in homogeneous solution, a phenyl substituent capable of stabilizing the radical formed during the act of trapping has been one of the most widely used tools of this type. During study of the mechanisms of formation of Grignard reagents - reactions that occur at the interface of the metal and the solution - the phenyl substituent displayed a specific new behaviour pattern. Besides its stabilizing role, it was also able to play the role of mediator in redox catalysis of electron transfer. In this case, the first events on the pathway to the Grignard reagents involve a cascade of three (one intermolecular followed by two intramolecular) electron transfers. Introduction of a p-methoxy substituent on the phenyl ring, making the phenyl group a poorer electron acceptor, suppresses this specific second role. Applied to the mechanism of Grignard reagent formation, this p-methoxy effect is consistent with a triggering mechanistic act of electron transfer from the metal to the aryl halide rather than with a concerted oxidative addition. A similar change in selectivity is observed, when a p-methoxy group is introduced onto a phenyl group that also bears a halogen, but its origin is different: this effect is associated with the shortening of the lifetime of the radical anion formed by the triggering electron transfer. These observations reemphasise our earlier proposals to use concepts originating from, electrochemical kinetics to explain, the selectivities of reactions occurring at metal/solution interfaces. This conjecture could possibly hold for any interface where the diffusion of reactive species plays a role in the settling of selectivity. These concepts emphasise the necessity to consider, for each reactive species, their average distance of diffusion away from the metal/solution interface. Wiley-VCH Verlag GmbH & Co. KGaA.

Full text of DOI:10.1002/ejoc.200900096

FULL PAPER
DOI: 10.1002/ejoc.200900096
Structural Effects in Radical Clocks and Mechanisms of Grignard Reagent
Formation: Special Effect of a Phenyl Substituent in a Radical Clock when the
Crossroads of Selectivity is at a Metal/Solution Interface
Hassan Hazimeh,[a,b] Jean-Marc Mattalia,*[a] Caroline Marchi-Delapierre,[c]
Frédéric Kanoufi,[b] Catherine Combellas,[b] and Michel Chanon[a]
Keywords: Grignard reagent / Electron transfer / Diffusion and selectivity / Radicals / Radical probe / Substituent effects /
Reaction mechanisms
A large class of radical clocks is based on the intramolecular
trapping of a reactive radical by a suitably located unsatu-
rated system. Depending on the substituents present on this
unsaturated system, the rate of cyclisation may vary drasti-
cally. This property has been repeatedly used to diagnose the
participation of very short-lived radicals in the mechanisms
of a wide variety of reactions. For reactions occurring in
homogeneous solution, a phenyl substituent capable of stabi-
lizing the radical formed during the act of trapping has been
one of the most widely used tools of this type. During study
of the mechanisms of formation of Grignard reagents – reac-
tions that occur at the interface of the metal and the solu-
tion – the phenyl substituent displayed a specific new behav-
iour pattern. Besides its stabilizing role, it was also able to
play the role of mediator in redox catalysis of electron trans-
fer. In this case, the first events on the pathway to the Grig-
nard reagents involve a cascade of three (one intermolecular
followed by two intramolecular) electron transfers. Introduc-
tion of a p-methoxy substituent on the phenyl ring, making
the phenyl group a poorer electron acceptor, suppresses this
specific second role. Applied to the mechanism of Grignard
reagent formation, this p-methoxy effect is consistent with a
triggering mechanistic act of electron transfer from the metal
to the aryl halide rather than with a concerted oxidative ad-
dition. A similar change in selectivity is observed when a p-
methoxy group is introduced onto a phenyl group that also
bears a halogen, but its origin is different: this effect is associ-
ated with the shortening of the lifetime of the radical anion
formed by the triggering electron transfer. These observa-
tions reemphasise our earlier proposals to use concepts origi-
nating from electrochemical kinetics to explain the selectivi-
ties of reactions occurring at metal/solution interfaces. This
conjecture could possibly hold for any interface where the
diffusion of reactive species plays a role in the settling of
selectivity. These concepts emphasise the necessity to con-
sider, for each reactive species, their average distance of dif-
fusion away from the metal/solution interface.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
urated system. A review covering the use of these tools
would demand the citation of more than two thousand ref-
erences. Without going to this extreme, one may select refer-
ences that show the use of these mechanistic tools in or-
ganic,[6–8] organometallic,[9–11] and inorganic chemis-
try,[12–17] in electrochemistry,[18–23] in photochemistry[24–26]
and in enzymatic catalysis.[27–31] In this work we deal with
radical probes based on intramolecular additions of aryl
radicals to unsaturated systems suitably positioned in sub-
stituents situated in ortho positions to the aryl radicals. The
reaction mechanism investigated is the formation of a Grig-
nard reagent.
Curiously, although radical clocks based on intramolecu-
lar additions of fleeting radicals to unsaturated systems had
been frequently used since 1975 for corroborating the par-
ticipation of radicals during the formation of alkylMgX,
the world had to wait to 1998 to see this concept applied
to the study of arylMgX formation (Scheme 1, R1 = R2 =
R3 = H). Two teams independently used this type of tool
Introduction
Radical clocks or radical probes have been some of the
tools most widely used to demonstrate the participation of
radical species in reaction mechanisms.[1–4] As proposed in
a preceding work, radical clocks are radical probes for
which reactive intramolecular rearrangements have been ki-
netically measured.[5] “Rearrangement” covers a variety of
chemical acts occurring at the radical stage: racemisation,
β-scission, hydrogen atom transfer, or addition to an unsat-
[a] Université Paul Cézanne, iSm2, UMR CNRS 6263, Campus
Scientifique de Saint-Jérôme, Service 561,
13397 Marseille Cedex 20, France
Fax: +33-4-91288234
E-mail: jean-marc.mattalia@univ-cezanne.fr
[b] Laboratoire Environnement et Chimie Analytique, UMR
CNRS 7121, ESPCI,
10 rue Vauquelin, 75231 Paris Cedex 05, France
[c] Université Joseph Fourier, LCBM, CEA, iRTSV, CNRS UMR
5249,
38054 Grenoble, France
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2775

J.-M. Mattalia et al.
FULL PAPER
and observed the same surprising result: apparently, and
in contrast with their alkyl counterparts, aryl halides were
reacting with magnesium metal with almost no participa-
tion of radical species (here σ-type sp2-C-centred radi-
cals).[32,33] The unexpectedness of this observation was fur-
ther accentuated by the fact that the rate constants of intra-
molecular addition to unsaturated systems were known to
be higher for aryl radicals than for alkyl radicals (5ϫ108 s–1
vs. 2ϫ105 s–1, room temperature). In his quantitative
D model of reactivity, Garst was predicting that 64% of
cyclised aryl Grignard should be formed, whereas less than
1% was experimentally observed in THF.[33,34]
Scheme 1.
The two teams provided diverging explanations for the
same experimental observation. Garst’s team proposed that
the aryl halides, unlike their alkyl counterparts, were able
to accept a second electron from the magnesium. This in-
volves the intervention of dianion species or transition
states (route k1, k2, k4 dominant in Scheme 2, with weak
participation of k1, k3...).[33] Our group proposed that the
aryl radicals were indeed involved in the route to arylMgX
much as alkyl radicals are on the route to alkylMgX (route
k1, k3, k5 or k6 with no participation of k1, k2...). However,
being much more reactive intermolecularly than alkyl radi-
cals, the aryl radicals were far less readily trapped intramo-
lecularly by the double bond than the alkyl radicals.[32,35]
The logical conclusion of such a hypothesis, then, was to
synthesise aryl halides in which the side chain containing
the unsaturated system is specifically designed to increase
the rate constant of cyclisation. Radical clocks with higher
cyclisation rates should display higher yields of rearranged
products. Structural effects on the rate constants of radical
cyclisations have been intensively explored.[3,4,36,37] Among
the structural modifications that increase this rate, one of
the most widely recognised is the substitution of the unsatu-
rated system with a phenyl group (Scheme 1, R2 or R3 =
Ph). The phenyl group indeed stabilises the incipient formed
carbon-centred π radical formed by the addition of the radi-
cal species on the double bond.[38–42] This effect, could,
however, become less important for the very rapid radical
clocks.[43]
Scheme 2.
consequences of simple classical substituent effects as we
had postulated at the start.[35] A p-methoxy substituent
placed on such a phenyl group would be expected to in-
crease the quantity of cyclised product further.[44] Just the
opposite was observed. This result, combined with electro-
chemical studies of these substrates, hints at a dual role for
the phenyl group under the heterogeneous conditions of the
Grignard reagent formation. The first role is the widely rec-
ognised radical stabilizing effect, but it is completed by a
mediating effect. In this second role the phenyl substituent
of the styryl group accepts one electron from the metal sur-
face. As a consequence, the radical anion of the substrate
aryl halide formed by addition of this electron displays a
far longer lifetime than when the added electron directly
adds in the π antibonding orbital of the aryl halide group.
This leaves more time for the radical anion to diffuse away
from the metal surface. We have previously shown that the
participation of the radical route to the overall selectivity
increases under such conditions.[45–47]
Results and Discussion
The range of structural modifications is shown in
In this report we want to show that the introduction of Scheme 1. In another report we have shown that with a rad-
phenyl groups on the exo double bonds of our radical ical clock bearing an oxygen atom instead of a benzylic
clocks indeed leads to increases in production of cyclised CH2 group (CH2 replaced with O in (E)-2), high yields of
products in reactions with metallic magnesium. Further cyclised product are obtained in the formation of Grignard
studies, however, suggested that these increases were not the reagent.[35] However, the effect of the phenyl group on the
2776
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 2775–2787

Radical Clocks and Grignard Reagent Formation Mechanisms
oxygen probe cyclisation rate constant is minor[43] and this benzophenone was performed in THF at reflux and af-
more ambiguous[48] increase would deserve further investi- forded 3. The E and Z isomers of 2, 4, 5 and 6 were iden-
gation.
tified by 1H NMR spectroscopy, which showed the expected
Scheme 1 shows stereochemically pure radical probes. cis- and trans-vinyl coupling constants. Fortunately, the ste-
Actually, we will see below that our method of preparation reochemical composition does not depend on the methoxy
yielded mixtures of double bond isomers with the Z isomers group at the phenyl para position. Therefore, in Tables 1
in the majority, so radical probes 2, 4, 5 and 6 represent and 2, the reported results for radical clocks in which
these mixtures (Z/E ≈ 90:10). Radical probe (E)-2 repre- Z/E isomerism exists refer to reactions between mixtures of
sents the pure E isomer. In Scheme 2, 1c–6c represent the stereoisomers (Z/E ≈ 90:10) and magnesium turnings.
corresponding cyclised products (e.g. 1c: R1 = R2 = R3 =
Tables 1 and 2 summarise the experimental results ob-
H) and 1l–6l the corresponding linear products. The linear tained when structurally different radical clocks were
products of probes (E)-2, 2, 4, 5, 6 are formed in Z/E ratios treated with magnesium turnings in THF or diethyl ether
identical to those in the starting halogeno substrates [e.g. at room temperature under nitrogen (Schlenk techniques).
2l: R1 = H, R2 = Ph, R3 = H with Z/E ≈ 90:10, whereas This experimental specification is particularly important.
(E)-2l represents the pure E isomer].
Indeed, a recent report shows that cyclisations of RZnX,
Radical probes 2–6 were prepared as shown in Scheme 3. formerly viewed as carbanionic ones, were actually chain
Treatment of 7a or ether 7b with vinylmagnesium bromide radical cyclisations initiated by traces of dioxygen.[49] We
in the presence of CuI and 2,2Ј-bipyridyl yielded alkenes 8a ourselves have reported an example of a reduction with
and 8b, respectively. Treatment of 8a and 8b with HBr in LiAlH4 in which different products were obtained de-
toluene under radical conditions then gave the bromides 9a pending on the purity of argon used for the reaction.[50]
and 9b. The phosphonium salts 10a and 10b were prepared
The solvent effect reported by Garst when the Grignard
from the corresponding bromides and excess triphenylphos- reaction was performed on halide 1 in THF or diethyl ether
phane in xylenes at 120 °C. Radical traps 2, 4, 5 and 6 re- is found again with probes 2–6: yields of cyclised products
sulted from Wittig-type reactions between 10a–10b and are higher in diethyl ether than in THF. We will return later
benzaldehyde or p-anisaldehyde as electrophiles. These reac- to this point. With probes 2, 4, 5 and 6, low yields of
tions were carried out in THF at room temperature and cyclised products were obtained in THF (Ͻ 1% to 1.5%),
yielded mixtures of isomers (Z/E ≈ 90:10). The Wittig reac- so the effect of the methoxy group is discussed mainly for
tion between the phosphonium salt 10a and the less reactive diethyl ether.
Scheme 3.
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2777

J.-M. Mattalia et al.
FULL PAPER
Table 1. Reactions between radical probes and magnesium (turn- the large differences in behaviour for the cyclisations per-
ings) in THF at room temperature.
formed in diethyl ether and in THF.
Entry[a] Radical Reaction RMgX
Relative yields
(%)[d]
1c–6c
Comparison of Entries 1 and 2 in Tables 1 and 2 shows
that the introduction of a phenyl group on the exo double
bond indeed induces an increase in the formation of the
cyclised product. In a previous report we used this observa-
tion to suggest that this experimental fact is better ac-
counted for by considering the route that passes through an
aryl radical stage (k1, k3 in Scheme 2).[35] This suggestion
seems to be strengthened by consideration of Entries 7 in
both tables: double substitution by phenyl increases the
quantity of cyclised products still further. Such an effect has
also been reported with cyclisable radical probes in homo-
geneous solutions.[4] The rate constants, in s–1 at 20 °C, for
cyclisations of substituted hex-5-enyl radicals in which the
double bonds are either terminally unsubstituted or substi-
tuted with two phenyl groups are 2ϫ105 and 4ϫ107,
respectively.[4]
Garst, using radical probe 1, showed that the product
distributions in THF and in diethyl ether were not depend-
ent on the reaction times. Moreover, (indanylmethyl)magne-
sium halides do not ring-open, so the yields of linear and
cyclised Grignard products are kinetically controlled in
both solvents.[33] With radical probe (E)-2, we found similar
2l/2c ratios after reaction times of 3 h 15 or 1 h 40 in THF.
Preparation of the Grignard reagent in the presence of
tBuOH [2 equiv. relative to (E)-2], a proton donor, had not
changed this ratio after 19 h of stirring. In diethyl ether, the
same behaviour was found. The 2l/2c ratios were similar
after reaction times of 3 h 25 or 6 h. Addition of tBuOH
[2 equiv. relative to (E)-2] had not changed this ratio after
21 h of stirring. Under our reaction conditions, therefore,
addition of a phenyl group on the exo double bond does
not allow the cyclisation of the aromatic Grignard reagent.
Usually, five-membered ring cyclisations of alkenyl Grig-
nard reagents are very slow processes. Phenyl substitution
on the double bond appears modestly to accelerate the five-
membered cyclisation.[54,55] When the reaction mixtures ob-
tained from (E)-2, in THF or diethyl ether, were quenched
with D2O, more than 90% D incorporation was found in
probe[b] time
(%)[c]
1l–6l[e]
1
2
3
4
1
2 h 05
96
94
96
94
Ͻ 1
5
Ͻ 1
Ͻ 1
1.5
Ͻ 1
8[h]
Ͼ 99
95
Ͼ 99
Ͼ 99
98.5
Ͼ 99
92[h]
(E)-2
3 h 15
3 h 30
3 h 25
3 h 15
3 h 15
2 h 50
2
4
5
6
3
5
97
6
93
7[f]
84[g]
[a] [RX] = 0.15 , Mg/RX = 4.2–4.5, BrCH2CH2Br/RX = 0.38–
0.39. The conversion was 100%. For halide 1, see ref.[45]. [b] Radi-
cal probes 2, 4, 5, 6 were mixtures of stereoisomers: Z/E ≈ 90:10.
[c] Estimated by o-phenanthroline titration with 2-BuOH/xylene
(0.5 ) as titration reagent. [d] GC determination: c = cyclised and
l = linear. [e] Z and E linear products of probes 2, 4, 5, 6 are formed
in ratios identical to those in the starting halogeno substrates.
[f] Benzophenone is also formed as byproduct: NMR measurement
gives benzophenone/cy+lin ≈ 2:98. [g] Possibly approximate be-
cause the colour change during the titration was not sharp.
[h] Average of three experiments.
Table 2. Reactions between radical probes and magnesium (turn-
ings) in diethyl ether at room temperature.
Entry[a] Radical Reaction RMgX
Relative yields
(%)[d]
probe[b] time
(%)[c]
1c–6c
1l–6l[e]
1
2
3
4
1
2 h 05
84
73
82
78
9
91
80
83
98
(E)-2
3 h 25
3 h 30
3 h 20
3 h 10
3 h 10
3 h 30
20
17
2
2
4
5
6
3
5
85
11
1.5
25[h]
89
6
68[f]
81[h]
98.5
7[g]
75[h]
[a] [RX] = 0.15 , Mg/RX = 4.2–4.5, BrCH2CH2Br/RX = 0.38–
0.39. The conversion was 100%. For halide 1, see ref.[45]. [b] Radical
probes 2, 4, 5, 6 are mixtures of stereoisomers: Z/E ≈ 90:10. [c] Esti-
mated by o-phenanthroline titration with 2-BuOH/xylene (0.5 ) as
titration reagent. [d] GC determination: c = cyclised and l = linear.
[e] Z and E linear products of probes 2, 4, 5, 6 are formed in ratios
identical to those in the starting halogeno substrates. [f] The degree
of conversion was 90%. [g] Benzophenone is also formed as a by-
product: NMR measurement gives benzophenone/cy+lin
0.5:99.5. [h] Average of two experiments.
≈
1
the cyclic product 2c (–CHDPh, estimated from H NMR
500 MHz), demonstrating the presence of the cyclised Grig-
nard reagent in the mixture.
Entries 2–3 in Tables 1 and 2 show that the stereoisomers
Garst more recently discussed possible complications
display slight differences in terms of linear/cyclised selectiv- with probes such as (E)-2, proposing that in such probes,
ity, especially in THF. Kinetic studies on hex-5-en-1-yl radi- “the reduction of the styryl group is a plausible first step of
cal cyclisations have indicated that E and Z isomers cyclise the Grignard reaction…One can envision subsequent path-
at similar rates.[44,51] In a previous report we showed that ways to cyclic products…These considerations cast doubt
the linear/cyclised selectivity is best explained by use of the on the validity of these substrates as probes of mechanism
equations proposed by Saveant’s group to interpret the of Grignard reagent formation and other reductions. Iso-
selectivity observed in the yields of rearranged products in merisations in the corresponding Grignard reactions could
the vicinity of a cathode.[46,47,52] In these equations, the dif- be artefacts arising through other processes”.[48] We discuss
fusion coefficients associated with the reactive species play this proposition further later; starting from his first sen-
a definite role. Several reports have suggested that the dif- tence explains why we tried other structural modifications
fusion coefficients of isomers may slightly depend on the on the probes shown in Scheme 1.
shapes of these isomers (here Z vs. E isomers).[53] This ex-
Suppose that the substitution of the phenyl group on the
planation is, however, not sufficient to explain the larger exo double bond does indeed have a double effect on the
difference observed in THF. At this point we cannot explain linear/cyclised product selectivity. Firstly, this group in-
2778
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 2775–2787

Radical Clocks and Grignard Reagent Formation Mechanisms
creases the rate constant of radical cyclisation (k5 in fuse away from the metallic surface. The longer the overall
Scheme 2); secondly, it causes the first electron transfer set of three electron transfers takes to take place, the farther
from the metal to the substrate to occur on the styryl group from the metallic surface the σ aryl radical will be formed.
rather than on the aryl halide group. The formation of the For the set of substrates 1, 4 and 6, in which the redox
styryl-centred radical anion could, according to Garst’s catalysis plays a negligible role, the cascade is simplified to
proposition, open a route to species forming cyclised prod- two electron transfers (route a in Scheme 4). Such examples
ucts (deceptive ones with respect to actual radical probes). of intramolecular redox catalysis in the vicinity of a cathode
This author proposed that such cyclised products could have been explored.[71,72] This intramolecular redox cataly-
form through attack on the halogen group by the thus- sis is to be distinguished from the intermolecular redox ca-
formed styryl radical anion. Such an attack could possibly talysis involved in the arene catalysis of formation of
lead to final products with exactly the same structures as organolithium compounds.[73] The electrochemical treat-
those formed by step k5 in Scheme 2. In electron transfer ment of the cathodic reactivity of 9-chloroanthracene was
chemistry, deposition of an electron in an electrophoretic a simplified case of intramolecular redox catalysis: here the
group followed by an intramolecular electron transfer to the cascade consisted of only two electron transfers.[52] Such
actual reactive site has been repeatedly reported.[56–62]
cascades of intramolecular electron transfers driven by an
If such is the case, the introduction of a para-methoxy irreversible last reaction provide simplified models for the
group in the styryl component should change the situation. electron transport chains involved in photosynthetic reac-
Such substituents are well known to decrease the electron tion centres.
acceptor character of the styryl group. On the other hand,
available rate constants of cyclisation show that the radicals
formed from substrate 4 should cyclise slightly more rapidly
than those formed from substrate 2.[44] This is also in line
with the stabilisation conferred on a radical by a para-meth-
oxy substituent.[63–70]
Therefore, application of Garst’s line of thought to sub-
strate 4 should let smaller amounts of cyclised products
(mechanistically irrelevant ones) be expected, whereas the
simple line of thought “unbiased radical probe 4” would
have suggested an increase in amounts of cyclised products
with respect to (E)-2 (Scheme 2, higher k5). Entries 3 and 4
in Table 2 apparently converge to sustain Garst’s proposi-
tion: in diethyl ether, smaller amounts of cyclised products
are formed when a p-methoxy substituent diminishes the
electron acceptor ability of the styryl block in the fast radi-
cal probes.
Electrochemical studies and application of electrochemi-
cal treatment of selectivity at electrode surfaces, however,
offer an alternative explanation of this apparent conver-
gence (Scheme 4). This interpretation is based on the par-
ticipation of redox catalysis for the set of substrates (E)-2,
2, 3, and 5. This redox catalysis demands the presence of
electrophoretic groups on the exo double bonds; these elec-
trophoretic groups must be better electron acceptors than
the aryl halide part of the radical clock. The first electron
transfer then occurs on the styryl part (path b in Scheme 4).
In the anion radical thus formed, a reversible intramolecu-
lar electron transfer shuttle takes place (equilibrium h in
Scheme 4). Overall, the equilibrium should favour the form
Scheme 4.
in which the electron resides on the styryl group. Because
of the irreversible carbon halogen bond cleavage (step j) oc-
curring when the electron is in the aryl halide part, the equi-
For the fates of aryl radicals, we follow Bickelhaupt’s
librium is nevertheless driven toward the production of the group’s proposition. This group proposed that the best way
aryl radical. From the knowledge that the C–Br bond cleav- to explain their experimental results for aryl halides was to
age itself involves an intramolecular electron transfer (not assume that aryl radicals were reduced to carbanions.[74–76]
represented in Scheme 4), the overall route corresponds to This was an original proposition, because up to their work
a cascade of three electron transfers (one intermolecular fol- it had been believed that the route from radicals to RMgX
lowed by two intramolecular).[62] Meanwhile, the radical involved a direct coupling of radical R· with the paramag-
anion and its counterion, possibly MgBr+, have time to dif- netic MgX· species formed by reaction between Mg radical
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2779

J.-M. Mattalia et al.
FULL PAPER
cations and X anions.[77,78] In Bickelhaupt’s and Garst’s includes, for some substrates, the possibility that the aryl
representations, RMgX results from reactions of R carban- radicals generated by the radical anion cleavage might be
ions with MgX+[75] and MgX2,[48] respectively.
reduced in two different steps.[52] The first of these (step d
Solvent effects fit with this proposition. In diethyl ether, or n in Scheme 4) involves a reduction of this radical by
the cyclised/uncyclised ratio is higher than in THF. The vis- the metal surface. The second involves the reduction of this
cosity of diethyl ether (0.194 cP) is about half that of THF radical by radical anions migrating away from the metal
(0.389 cP).[34,79] Everything proceeds as if the radical anion surface. Its participation becomes non-negligible for radical
formed by heterogeneous electron transfer had more anions of sufficient lifetime. In the original electrochemical
chances to get away from the metal surface in diethyl ether, treatment this corresponds to substrates such as 9-chloro-
therefore giving birth to the aryl radical farther away from anthracene or 9-bromoanthracene.[52] Substrates such as
the surface. This aryl radical would therefore, on average, (E)-2, 2, 3 and 5 would obviously be those for which this
have a lower probability of undergoing reduction than when oversimplification would be the most damaging. The com-
the same succession of events occurs in THF. Bickelhaupt plete electrochemical treatment of intramolecular redox ca-
was the first author to propose the importance of viscosity talysis shows, however, that this neglect does not modify,
in the selectivity of formation of Grignard reagents.[79] overall, the interpretation of the experimental observations
Later, Garst returned to this point and laid stress on po- described in this report.
larity factors.[34,48] In this scheme of things, the polarity of
Cyclic voltammograms of radical clocks in DMF
the solvent plays a role in the value of k3 (Scheme 2): more (NBu4BF4 as electrolyte) make it evident that, at a glassy
polar solvents increase its value.[80] The higher polarity of carbon disk cathode, the first electron transfer occurs
THF therefore fits with the lower yield of cyclised products mainly on the halogenoaryl structural block only for the
obtained in this solvent: if the cleavage of the radical anion probes 1, 4 and 6. For (E)-2, 2, 3 and 5 it occurs mainly on
is faster, the aryl radical is created near the surface and has the styryl part.
good chances of being reduced more rapidly.
The voltammograms of the radical probes (RBr) display
Scheme 4 contains three oversimplifications. Firstly, al- either one or two peaks. One is connected with the re-
though the route going through step a suggests the con- duction of the aryl halide structural block, the second is
tinuous presence of the magnesium surface in close proxim- connected with the reduction of the substrate in which the
ity to the formed aryl radical, a better representation would bromine is replaced by hydrogen (reduction of the linear
be a short-distance shuttle of this radical between the met- product RH). The cyclised products are not seen in the vol-
allic surface and the solution.[81,82] Scheme 4 emphasises the tammograms because they are situated at too negative val-
average proximity of the metallic surface, to contrast this ues. Within this general pattern of reactivity, the observed
situation with that in which intramolecular redox catalysis peaks are rather close. The voltammograms of substrate 1
takes place.
show only the peak corresponding to the aryl halide group
Secondly, this scheme neglects the step of hydrogen (–2.58 V vs. aq. SCE). For this substrate the reduced prod-
atom transfer from the solvent to the aryl radical.[33,52] uct (RH) peak lies at a value too negative to be seen. It is
Ashby’s contribution concerning the role of radicals in the visible, however, for substrates 4 and 6 (4: peaks at –2.61 V
formation of alkylMgX clearly showed such participa- for RBr and –2.79 V for RH, 6: peaks at –2.65 V for RBr
tion.[78] Saveant’s report on the reactivity patterns of aryl and –2.80 V for RH). The set of radical probes in which the
halides suggests that the reduction of the σ aryl radicals electron transfer occurs predominantly on the styryl group
formed by the cleavage of radical anions is fast enough displays a distinct pattern of voltammograms. For (E)-2,
to overcome their dimerisation and hydrogen abstraction voltammograms show peaks at –2.46 V (RBr) and –2.63 V
reactivity.[52] The reduction (step d in Scheme 4) plays this (RH). For 2 (Z/E 64:36), peaks displayed were similar
predominant role because aryl radicals are better oxidiz- (–2.47 V and –2.64 V). For substrate 3 these peaks are at
ing agents than alkyl ones.[83–85] They are also better hy- –2.44 V (RBr) and –2.56 V (RH) and for substrate 5 they
drogen atom abstractors than their alkyl counterparts (k are seen at –2.50 V (RBr) and –2.64 V (RH). Under the
abstraction from THF is 6ϫ107 s–1 for aryl radicals at same experimental conditions, styrene itself displays a
25 °C vs. 7ϫ106 s–1 for cyclopropyl radical).[34,37,86] The unique peak at –2.52 V. Although the set of radical clocks
rates constants of cyclisation for the aryl-centred radical in which the first electron transfer takes place on the stryryl
clocks are about 5ϫ108 s–1,[37,87,88] or even higher when part of the molecule shows a strong similarity with styrene
the double bond is stabilised by a phenyl group.[38–43] For itself, the peak values (RBr) for substrates 1, 4 and 6 are not
the structures displayed in Tables 1 and 2, the H-abstrac- outstandingly distant from this styryl-type value (average
tion reaction could become less negligible for the set of difference of about 0.1 V). Competition between the two
substrates relevant to intramolecular redox catalysis [(E)- types of electrophores (aryl halide vs. styryl) is therefore
2, 2, 3 and 5]. If predominant, this reaction would lead to possible.
increased yields of linear products, in contrast with the
experimental trends for this set of substrates shown in by comparison of the RH peaks of substrates 2 and 4 (dif-
Tables 1 and 2. ference of 0.15 V), gives a quantitative measure of the de-
The effect of a p-methoxy group on the styryl part, seen
Thirdly, the complete electrochemical treatment of the crease in the electron acceptor ability of the styryl electro-
selectivity observed when aryl halides react at a cathode phoretic group para-substituted with a methoxy group.
2780
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 2775–2787

Radical Clocks and Grignard Reagent Formation Mechanisms
These electrochemical observations therefore suggest an Br bond of the radical anion formed from 5 were smaller
alternative to Garst’s proposition: in the set of radical than those related to radical anions from 2 and 3.
clocks (E)-2, 2, 3 and 5, intramolecular redox catalysis is
The shared structural feature of 4 and 6 in the set of
made possible by the presence of the phenyl group. In pre- substrates 1, 4 and 6 is the para substitution of the styryl
vious reports, we proposed that the reductive strength of group with a methoxy substituent. This substituent pre-
the magnesium surface was stronger than most of the vents, at least partly, the occurrence of the first electron
homogeneous reductants (including the solvated elec- transfer on the styryl part (induced difference in electropho-
tron).[46,47] As a consequence, an aryl radical created in retic activity of 0.15 V). As a consequence, the patterns of
close vicinity to this metallic surface should have less reactivity of these two substrates directly parallel that of
chance to cyclise because its rate of reduction into a carban- substrate 1 (in Tables 1 and 2, compare Entries 4 and 6 with
ion would be very high. In contrast, when the radical is Entry 1). In diethyl ether, the relative amount of cyclised
created at a greater distance from this metal surface, it compound for substrate 1 is even higher than that measured
would have better chances to cyclise. This situation will be for the other two radical clocks of the set. This effect was
met when the radical anions creating the aryl radical are not expected in terms of structural effects because the para-
long-lived enough to diffuse away from the metallic sur- methoxyphenyl substituent should increase the rate of
face.[89,90] Radical clocks (E)-2, 2, 3 and 5 illustrate this situ- cyclisation for the aryl radical.[44] It disappears, however, in
ation. For these radical clocks, the increase in the cyclised the electrochemical study. The data presented are not suf-
products has a double origin. Firstly, because of intramolec- ficient for discussion of its origin (counterion effects in the
ular redox catalysis the σ aryl radicals are, on average, diffusion coefficients of the radical anions?). In comparison
formed at a larger distance from the metallic surface. Sec- with the set of substrates (E)-2, 2, 3 and 5, this set is, over-
ondly, because of the stabilizing effect of the phenyl group, all, characterised by a shorter average distance of diffusion
the rate of radical cyclisation is increased.
before the cleavage of the C–Br bond of the radical anion
For substrate 3 one could have expected a relative yield formed from the starting substrate.
of cyclised product higher than those given in Tables 1 and
Garst’s proposal that the rapid radical clocks such as (E)-
2. The stabilisation of the radical formed during the cycli- 2, 2, 3 and 5 could provide deceptive results because the
sation is certainly higher when it is substituted with two initial electron transfer occurs on the styryl part of the
phenyl groups. These groups, however, could also hinder molecule to yield a radical anion capable of reacting intra-
the attack on the double bond during the radical cycli- molecularly as a nucleophile to substitute the halogen of
sation.[4,91,92]
the aryl halide part of the molecule must be discussed fur-
In the set of substrates in which intramolecular redox ther.[48] The best identified cases of ring-closure reactions
catalysis occurs, 5 is the member for which the relative triggered by an electron transfer and involving C–C bond
quantity of cyclised compound is the lowest in diethyl ether formation are intramolecular SRN1 reactions. These have
(Entry 5, Table 2). A possible reason for this observation recently been reviewed. The currently accepted mechanism
could be sought for in the effect of the p-MeO group on involves a key step in which an aryl radical formed by cleav-
the rate of cyclisation (k5, Scheme 2). If this group stabilises age of the radical anion reacts intramolecularly with a suit-
the σ aryl radical, this could induce a lower rate of cycli- ably situated carbanion in the same structure.[61] To the best
sation. Recent theoretical calculations suggest that this p- of our knowledge, there are no precedents for SNAr substi-
MeO group slightly destabilises this radical.[93] Another ex- tution in which the nucleophile is a radical anion. Usually,
planation for this experimental observation must therefore when an aromatic radical anion interacts with an aromatic
be provided. This pattern of reactivity may be qualitatively halide, the interaction involves electron transfer between the
understood in terms of average distance from the metallic interacting species driven by the cleavage of the C–X bond
surface when the cleavage of the C–Br bond occurs in the in the formed radical anion.[96] It might possibly be that
formed radical anion. Besides the intrinsic diffusion coeffi- future works will identify aromatic radical anions as nucleo-
cient of the species in the medium under consideration, this philes at sp2 carbon, as they have been on sp3 ones.[97–99] At
average distance depends on two parameters.[52,89,90] The the present state of knowledge, however, such an intramo-
first is the equilibrium constant of the first intramolecular lecular SNAr substitution performed by a radical anion in-
electron transfer in the formed radical anion (step h in troduces a layer of complexity for which there is no evi-
Scheme 4); for the radical anion formed from 5, this equi- dence.
librium should be shifted toward the species in which the
Theoretical and experimental approaches have given rise
electron stays in the styryl part. The second is the rate of to the hypothesis that the formation of the Grignard rea-
intramolecular electron transfer from the antibonding π gent, rather than being triggered by an electron transfer
cloud to the antibonding σ orbital of the C–Br bond in the from the metal, would be better understood in terms of a
aryl halide part. This second parameter has been repeatedly simple concerted oxidative addition of magnesium to the
studied: in the radical anion of 5 an increase in this rate C–X bond.[100,101] In the terms of such an interpretation, it
would be expected.[94–96] These two parameters play in op- seems that substituent effects further than six atoms away
posite directions; comparison of Entries 3 and 5 in Table 2 from the C–X reactive centre should play a negligible role
suggests that the second parameter dominates. Everything in the selectivity of the reaction. The results given in Table 1
takes place as if the average distance of cleavage for the C– and Table 2 display trends contrasting with this expectation.
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2781

J.-M. Mattalia et al.
FULL PAPER
A possible origin of this discrepancy could be sought for
within the representation of Grignard reagent formation as
a form of organic corrosion.[34,48] In such a representation,
the kinetic measurements of metal consumption would have
to be connected with an act of anodic dissolution. The don-
icity of the solvent, the importance of which in the rate of
metal dissolution has been repeatedly shown by Maslenni-
kov’s team, could play its role at this anode.[102–104] The ki-
netic and selectivity results based on the fates of organic
substrates and products, would, on their part, be connected
to an act of cathodic reaction. The microanodes and micro-
cathodes would be spatially distinct, although by not more
than about five magnesium atoms.[34,48] Part of their reac-
tivity could be coupled.[105,106] More work is needed to ver-
ify this kind of hypothesis.
Experimental Section
General: THF (S.D.S., 99.7%) and diethyl ether (S.D.S., 99.7%)
were dried with sodium-benzophenone and distilled from purple
solutions prior to use. Butan-2-ol was dried with K2CO3 and then
distilled. Xylenes (mixture of isomers) was distilled from sodium/
benzophenone ketyl. All glassware and transfer needles were oven-
dried at 100 °C. 1,2-Dibromoethane (Acros, 99%), DMF (puriss
absolute anhydrous, Fluka) and magnesium turnings (Aldrich,
99.98%) were used as received. NBu4BF4 was synthesised from am-
monium tetrafluoroborate and tetrabutylammonium chloride
(Fluka), recrystallised from petroleum ether, and then dried over-
night before use.[107] Gas chromatography analysis was performed
on a Fisons GC 8000 instrument with use of a BPX5 capillary col-
umn (SGE, 25 m, 0.22 mm internal diameter), helium as carrier gas
and a flame ionisation detector (injector: 280 °C; detector: 250 °C).
The following program temperature was used: 150 °C (3 min) to
250 °C (15 min) at 5 °C min–1. Peak area integrations were per-
formed by electronic integration on a SP4600 Integrator (Spectra
Physics) and, if necessary, corrected by use of the ECN (effective
carbon number) concept.[108] Petroleum ether (boiling range 40–65
°C) was distilled prior to use for liquid phase chromatography.
Thin-layer chromatography (TLC) was performed with Merck
silica gel (60 F254) plates. Column chromatography was performed
with Merck silica gel 60 (230–400 mesh particle size). Separations
were monitored both by TLC and by gas chromatography (GC).
AgNO3-impregnated silica gel: AgNO3 (2.5 g) and CH3CN
(150 mL) were introduced into a flask containing a magnetic stirrer.
After the mixture had been stirred in the dark (10 min), silica gel
(50 g) was added. The mixture was stirred for 2 h in the dark, and
the solvent was then evaporated. Drying was achieved overnight in
a oven. The column chromatography was performed in the dark.
NMR spectra were run in CDCl3 and were generally recorded at
300 MHz (1H) and 75 MHz (13C). NMR of compounds 2, 4, 5 and
6 were each run with two different mixtures of isomers (Z/E ≈ 90:10
and ≈ 60:40). 13C NMR peak shifts are rounded off to the nearest
0.1 ppm except when greater precision is needed to distinguish
closely spaced peaks. For compounds 4, 5, 6, 3 and 4l, signals may
be blurred because of potential overlap. Gas chromatography/mass
spectrometry (GC-MS) was performed by electronic impact at
70 eV. 1-Bromo-2-(but-3-enyl)benzene (1)[33,109] 1-bromo-2-[(3E)-4-
phenylbut-3-enyl]benzene [(E)-2],[35] 4-bromo-3-(bromomethyl)-
phenyl methyl ether (7b),[110] authentic samples of but-3-enylben-
zene (1l)[33] and 1-methylindane (1c)[33,111] were prepared as de-
scribed in literature. The reaction between bromide 1 and Mg has
been described previously.[45]
Conclusions
This concomitant study of two complementary facets
(electrochemical and synthesis) of selectivity in a series of
structurally designed radical clocks sheds some fresh light
on the mechanism of Grignard reagent formation and on
the specificities of radical clocks used as mechanistic tools
in heterogeneous media.
Firstly, the consistent reactivity patterns associated with
structural variations when these substrates react to form
ArMgX or to accept an electron from a cathode sustain the
hypothesis of a single electron transfer from the magnesium
surface as the triggering act of Grignard reagent formation.
The proposition of a simple oxidative addition (magnesium
atom insertion) of magnesium atoms to ArX is difficult to
reconcile with the long-range substituent effects observed
in this work.
Secondly, under the heterogeneous reaction conditions
described here, the phenyl substituent effects classically en-
countered in radical reactions performed in homogeneous
solutions become more complex. The widely acknowledged
stabilisation of radicals by phenyl substituents has to be
completed. This substituent may behave as a distant electro-
phoretic group where the first electron transfer may occur
rather than going directly to the apparent centre. Under
homogeneous reaction conditions this event would be of
little importance in terms of selectivity because the electron
would in any case be finishing its travel on the apparent
centre. Spatially, at the end of the travel the molecule would
be reacting in the same microscopic environment as was
Reactions between Bromides 2–6 and Magnesium: Magnesium (4.2–
4.5 equiv.) was introduced into a Schlenk tube containing a mag-
netic bar. The Schlenk tube was successively degassed by three vac-
uum/nitrogen cycles, flame-dried in a flow of nitrogen and degassed
again. Bromide (0.29–0.30 mmol, 1 equiv.) was introduced into a
present when the electron was accepted. Under hetero- flask (10 mL) and this flask was flushed with nitrogen for about
10 min. Solvent (THF or diethyl ether, 2 mL) and 1,2-dibromoeth-
ane (0.38–0.39 equiv.) were added by syringe. The obtained solu-
tion was transferred by cannula to the Schlenk tube. After the allo-
cated time of stirring (Teflon-coated stirring bars were used) at
room temperature the Grignard reagent was titrated by Watson and
Eastham’s method.[112] A solution of o-phenanthroline (ca. 1 mg)
in the reaction solvent (ca. 1 mL) was added by syringe. If Grignard
reagent is present a purple or maroon colour develops. The mixture
was then titrated to the endpoint with a solution of butan-2-ol in
xylenes (0.5 ). After dilution with diethyl ether, the reaction mix-
ture was successively washed with ammonium chloride (10%) and
geneous conditions, though, the molecule also has time to
travel during the intramolecular electron transfer. This time,
at the end of the travel, the molecule reacts in a microscopic
environment that might be quite different from the one in
which the electron was accepted. In Grignard reagent for-
mation and in cathodic reduction, the heterogeneity is obvi-
ous. In some situations (particularly in biological media,
colloidal solutions, forming polymers) the heterogeneity of
the medium is less easy to recognise. The substituent effects
described in this report could help in recognition of this.
2782
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 2775–2787

Radical Clocks and Grignard Reagent Formation Mechanisms
water, dried (MgSO4) and filtered. The crude product was analysed
by GC and NMR spectroscopy. Components were identified by co-
injection of authentic samples, GC-MS and NMR comparisons.
solved in toluene (3 mL), and water was removed by decantation.
A solution of 8b (1.1 g, 4.84 mmol) and benzoyl peroxide (1 mL of
a 0.2  solution in toluene, 0.2 mmol) in toluene (10 mL) was pre-
pared under nitrogen and cooled to 0 °C. An excess of HBr was
bubbled through the mixture. HBr was prepared by addition of
Br2 (3.2 mL, 62.3 mmol) to tetraline (13 mL, 95.1 mmol) over 1 h
40 min. During this addition, two further portions of benzoyl per-
oxide (1 mL, ≈0.4 mmol) were added to the reaction mixture. With
the addition of Br2 finished, the reaction mixture was stirred for
4 h and allowed slowly to reach room temperature. The reaction
mixture was then washed with aqueous Na2SO3 (20%) and water,
dried (MgSO4) and concentrated. Column chromatography of the
residue (petroleum ether, petroleum ether/Et2O 99:1 and 98:2) af-
forded 9b (1.24 g, 83%) as a pale yellow liquid. 1H NMR
(300 MHz, CDCl3): δ = 2.18 (quint, J = 7.0 Hz, 2 H), 2.86 (t, J =
7.4 Hz, 2 H), 3.43 (t, J = 6.6 Hz, 2 H), 3.78 (s, 3 H), 6.65 (dd, J =
8.8, 3.0 Hz, 1 H), 6.81 (d, J = 3.0 Hz, 1 H), 7.41 (d, J = 8.8 Hz, 1
H) ppm. 13C NMR (75 MHz, CDCl3): δ = 32.5, 33.2, 34.8, 55.6,
113.7, 114.9, 116.4, 133.6, 141.0, 159.1 ppm. C10H12Br2O (308.01):
calcd. C 38.99, H 3.93; found C 39.29, H 3.79.
1-Bromo-2-(prop-2-enyl)benzene (8a) and 4-Bromo-3-(prop-2-enyl)-
phenyl Methyl Ether (8b): These compounds were prepared by a
modification of the method described by Knight and Parsons.[113]
1-Bromo-2-(prop-2-enyl)benzene (8a): A mixture of 2-bromobenzyl
bromide (7a, 5.0 g, 20.0 mmol), THF (35 mL), CuI (0.42 g,
2.20 mmol) and 2,2Ј-bipyridyl (0.325 g, 2.08 mmol) was cooled to
0 °C and placed under nitrogen. A solution of vinylmagnesium bro-
mide (20 mL of a 1  solution in THF, 20 mmol) was transferred
by cannula. The reaction mixture was stirred under nitrogen for
55 min at 0 °C and for 5 min without the ice bath. Solid NH4Cl
(8 g) was added, followed by Et2O (90 mL) and water (80 mL). The
solution was then transferred into a beaker and aqueous ammonia
(33%, 5 mL) was added. After stirring (20 min), the organic phase
was separated and the aqueous layer was extracted with Et2O. The
combined extracts were washed with aqueous HCl (10%) and satu-
rated aqueous NaHCO3, dried (MgSO4) and concentrated.
Chromatography on silica gel (petroleum ether) afforded 8a as a
colourless liquid (1.8 g, 46%). 1H NMR (300 MHz, CDCl3): δ =
3.51 (dt, J = 6.5, 1.5 Hz, 2 H), 5.04–5.14 (m, 2 H), 5.98 (ddt, J =
17.0, 10.2, 6.5 Hz, 1 H), 7.04–7.10 (m, 1 H), 7.21–7.30 (m, 2 H),
7.53–7.56 (m, 1 H) ppm. 1H NMR spectroscopic data obtained
agreed with published values.[113]
[3-(2-Bromophenyl)propyl](triphenyl)phosphonium Bromide (10a): A
solution of 9a (6.95 g, 25 mmol) and triphenylphosphane (13.1 g,
50 mmol) in xylenes (200 mL) was stirred under nitrogen at 120 °C
(bath temperature). During the reaction, triphenylphosphane
(60.3 g, 230 mmol) was added in five portions. The reaction time
was 30 h. A solid was formed, and after cooling to room tempera-
ture the mixture was filtered through a Büchner funnel. This solid
was then washed with xylenes and petroleum ether to afford pure
10a (10.4 g, 77%) as a white solid; m.p. 222–223 °C. 1H NMR
(300 MHz, CDCl3): δ = 1.93–2.06 (m, 2 H), 3.20 (t, J = 7.5 Hz, 2
H), 3.93–4.03 (m, 2 H), 7.05 (td, J = 7.7, 1.7 Hz, 1 H), 7.25 (td, J
= 7.7, 1.2 Hz, 1 H), 7.44 (dd, J = 7.7, 1.2 Hz, 1 H), 7.60 (dd, J =
4-Bromo-3-(prop-2-enyl)phenyl Methyl Ether (8b): This compound
was prepared by the same procedure with 7b (5.667 g, 20.2 mmol),
THF (38 mL), CuI (0.436 g, 2.29 mmol), 2,2Ј-bipyridyl (0.335 g,
2.15 mmol) and vinylmagnesium bromide (22 mL of a 1  solution
in THF, 22 mmol). Column chromatography on AgNO3-impreg-
nated silica gel (petroleum ether) afforded 8b as a colourless liquid
1
(2.28 g, 50%). H NMR (300 MHz, CDCl3): δ = 3.46 (dt, J = 6.6,
7.7, 1.7 Hz, 1 H), 7.63–7.70 (m, 6 H), 7.74–7.85 (m, 9 H) ppm. 13
C
1.4 Hz, 2 H), 3.77 (s, 3 H), 5.05–5.15. (m, 2 H), 5.96 (ddt, J = 16.8,
10.2, 6.6 Hz, 1 H), 6.65 (dd, J = 8.7, 3.0 Hz, 1 H), 6.78 (d, J =
3.0 Hz, 1 H), 7.42 (d, J = 8.7 Hz, 1 H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 40.5, 55.5, 113.6, 115.1, 116.2, 116.8, 133.3, 135.6,
140.6, 159.2 ppm. C10H11BrO (227.10): calcd. C 52.89, H 4.88, Br
35.18; found C 52.58, H 4.86, Br 33.96. The elemental analysis of
this product is the average of three experimental measurements,
except for Br, for which only one was performed.
NMR (75 MHz, CDCl3): δ = 22.1 (d, JC,P = 50.5 Hz), 22.6 (d, JC,P
= 3.3 Hz), 36.3 (d, JC,P = 17.6 Hz), 118.3 (d, JC,P = 86.2 Hz), 124.2,
128.0, 128.3, 130.5 (d, JC,P = 12.6 Hz), 132.3, 132.8, 133.7 (d, JC,P
= 9.9 Hz), 135.1 (d, JC,P = 3.3 Hz), 139.5 (d, JC,P = 1.1 Hz) ppm;
some chemical shifts and coupling constants were not definitely
assigned. 31P NMR (121 MHz, CDCl3): δ = 24.8 ppm. C27H25Br2P
(540.27): calcd. C 60.02, H 4.66; found C 59.92, H 4.50.
[3-(2-Bromo-5-methoxyphenyl)propyl](triphenyl)phosphonium Bro-
mide (10b): This compound was prepared by the same procedure
with 9b (0.889 g, 2.89 mmol) and xylenes (24 mL). Triphenylphos-
phane (6.195 g, 23.6 mmol) was added in three portions: 2.529 g
(initial time), 2.646 g (t = 7 h 30 min) and 1.020 g (t = 23 h).The
reaction time was 29 h. Workup afforded pure 10b (1.381 g, 84%)
1-Bromo-2-(3-bromopropyl)benzene (9a): A solution of 8a (1.747 g,
8.86 mmol) and 3-chloroperbenzoic acid (ca. 70%, ca. 0.09 g, ca.
0.36 mmol) in toluene (33 mL) was prepared under nitrogen and
cooled to 0 °C. An excess of HBr was bubbled through the mixture.
HBr was prepared from the addition of Br2 (3 mL, 58.4 mmol) to
tetraline (7.3 mL, 53.4 mmol) over 1 h 15 min. At the halfway point
of the addition, another portion of 3-chloroperbenzoic acid (ca.
70%, ca. 0.075 g, ca. 0.30 mmol) was added to the reaction mixture.
With the addition of Br2 finished, the reaction mixture was stirred
at 0 °C for 30 min and left under air for 20 min at room tempera-
ture to allow HBr evaporation. The reaction mixture was then
washed with aqueous Na2SO3 (20%) and water, dried (MgSO4) and
concentrated. Column chromatography of the residue (petroleum
1
as a white solid; m.p. 174–175 °C. H NMR (300 MHz, CDCl3): δ
= 1.93–2.07 (m, 2 H), 3.15 (t, J = 7.4 Hz, 2 H), 3.83 (s, 3 H), 3.93–
4.03 (m, 2 H), 6.63 (dd, J = 8.7, 3.2 Hz, 1 H), 7.23 (d, J = 3.2 Hz,
1 H), 7.30 (d, J = 8.7 Hz, 1 H), 7.64–7.70 (m, 6 H), 7.75–7.86 (m,
9 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 22.1 (d, JC,P
=
50.5 Hz), 22.5 (d, JC,P = 3.8 Hz), 36.6 (d, JC,P = 17.6 Hz), 56.1,
114.3, 115.1, 116.8, 118.3 (d, JC,P = 86.2 Hz), 130.5 (d, JC,P
=
1
12.6 Hz), 133.2, 133.7 (d, JC,P = 10.4 Hz), 135.1 (d, JC,P = 2.7 Hz),
140.5 (d, JC,P = 1.1 Hz), 159.2 ppm; some chemical shifts and coup-
ling constants were not definitely assigned. 31P NMR (121 MHz,
CDCl3): δ = 24.8 ppm. C28H27Br2OP (570.30): calcd. C 58.97, H
4.77; found C 58.14, H 4.59.
ether) afforded 9a (2.013 g, 82%) as a colourless liquid. H NMR
(300 MHz, CDCl3): δ = 2.14–2.23 (m, 2 H), 2.91 (t, J = 7.5 Hz, 2
H), 3.43 (t, J = 6.6 Hz, 2 H), 7.05–7.11 (m, 1 H), 7.21–7.27 (m, 2
H), 7.54 (d, J = 7.7 Hz, 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ
1
= 32.6, 33.1, 34.6, 124.6, 127.6, 128.1, 130.8, 133.1, 140.0 ppm. H
and 13C NMR spectroscopic data obtained agree with published
values.[114]
1-Bromo-2-(4-phenylbut-3-enyl)benzene (2): tBuOK (0.215 g,
1.92 mmol) was added to a suspension of phosphonium salt 10a
(1.007 g, 1.86 mmol) in THF (8 mL) under a flow of nitrogen. The
mixture was stirred under nitrogen for 15 min, and benzaldehyde
4-Bromo-3-(3-bromopropyl)phenyl Methyl Ether (9b): For this ex-
periment, benzoyl peroxide (ca. 75%, 0.2 g, ca. 0.62 mmol) was dis-
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2783

J.-M. Mattalia et al.
FULL PAPER
(0.19 mL, 1.87 mmol) was then slowly added by syringe. The mix-
ture was stirred for 22.5 h at room temperature. The reaction was
then quenched with HCl (1 ) and extracted with Et2O. The or-
1-Bromo-4-methoxy-2-[4-(4-methoxyphenyl)but-3-enyl]benzene (6):
tBuOK (0.132 g, 1.18 mmol) was added to a suspension of phos-
phonium salt 10b (0.646 g, 1.13 mmol) in THF (5 mL) under a flow
ganic layer was washed with water, dried (MgSO4) and concen- of nitrogen. The mixture was stirred under nitrogen for 10 min, and
trated. Column chromatography of the residue (petroleum ether)
afforded pure 2 (0.428 g, 80%, Z/E 92:8 from GC analysis) as a
colourless oil. 1H NMR (300 MHz, CDCl3): δ = 2.53 (br. q, J =
7.4 Hz, 2 H E), 2.61–2.70 (m, 2 H Z), 2.86–2.94 (m, 2 H Z and 2
H E), 5.72 (dt, J = 11.6, 7.2 Hz, 1 H Z), 6.27 (dt, J = 15.7, 6.7 Hz,
1 H E), 6.42 (br. d, J = 15.7 Hz, 1 H E), 6.47 (br. d, J = 11.6 Hz,
1 H Z), 7.02–7.08 (m, 1 H Z and 1 H E), 7.19–7.34 (m, 7 H Z and
7 H E), 7.51–7.56 (m, 1 H Z and 1 H E) ppm. 13C NMR (Z isomer,
75 MHz, CDCl3): δ = 28.9, 36.4, 124.7, 126.7, 127.5, 127.8, 128.3,
128.8, 129.9, 130.6, 131.5, 132.9, 137.6, 141.0 ppm.
C16H15Br (287.19): calcd. C 66.91, H 5.26; found C 66.49, H 5.26.
p-anisaldehyde (0.14 mL, 1.15 mmol) was slowly added by syringe.
The mixture was stirred for 27.5 h at room temperature. The reac-
tion was then quenched with HCl (1 ) and extracted with Et2O.
The organic layer was washed with water, dried (MgSO4) and con-
centrated. Column chromatography of the residue (petroleum ether,
petroleum ether/EtOAc 99:1 and 98:2) afforded 6 (0.26 g, 66%, Z/
E 90:10) as a colourless oil. 1H NMR (300 MHz, CDCl3): δ = 2.49
(br. q, J = 7.5 Hz, 2 H E), 2.64 (br. q, J = 7.6 Hz, 2 H Z), 2.81–
2.87 (m, 2 H Z and 2 H E), 3.75 (s, 3 H Z), 3.76 (s, 3 H E), 3.80
(s, 3 H Z and 3 H E), 5.62 (dt, J = 11.6, 7.0 Hz, 1 H Z), 6.12 (dt,
J = 15.7, 6.9 Hz, 1 H E), 6.37 (br. d, J = 15.7 Hz, 1 H E), 6.40 (br.
d, J = 11.6 Hz, 1 H Z), 6.62 (dd, J = 8.8, 3.0 Hz, 1 H Z), 6.63 (dd,
J = 8.8, 2.8 Hz, 1 H E), 6.76 (d, J = 3.0 Hz, 1 H Z), 6.79 (d, J =
2.8 Hz, 1 H E), 6.81–6.87 (m, 2 H Z and 2 H E), 7.17–7.30 (m, 2
H Z and 2 H E), 7.41 (d, J = 8.8 Hz, 1 H Z), 7.42 (d, J = 8.8 Hz,
1 H E) ppm. 13C NMR (75 MHz, CDCl3): δ = 28.82, 33.30, 36.53,
36.55, 55.30, 55.33, 55.43, 55.46, 113.30, 113.31, 113.68, 114.01,
115.00, 115.03, 116.18, 127.20, 127.33, 129.25, 129.79, 130.02,
130.07, 130.20, 130.56, 133.31, 133.33, 142.01, 142.13, 158.40,
158.88, 159.01 ppm. C18H19BrO2 (347.25): calcd. C 62.26, H 5.52,
Br 23.01; found C 62.17, H 5.54, Br 22.98.
4-[4-(2-Bromophenyl)but-1-enyl]phenyl Methyl Ether (4): The same
procedure with phosphonium salt 10a (1.008 g, 1.87 mmol), THF
(8 mL), tBuOK (0.215 g, 1.92 mmol) and p-anisaldehyde (0.23 mL,
1.89 mmol) afforded, after column chromatography (petroleum
ether and petroleum ether/Et2O 95:5), compound 4 as a colourless
oil (0.452 g, 76%), (Z/E 90:10 from GC analysis). 1H NMR
(300 MHz, CDCl3): δ = 2.50 (br. q, J = 7.4 Hz, 2 H E), 2.65 (br.
q, J = 7.7 Hz, 2 H Z), 2.86–2.92 (m, 2 H Z and 2 H E), 3.798 (s, 3
H E), 3.802 (s, 3 H Z), 5.62 (dt, J = 11.6, 7.1 Hz, 1 H Z), 6.12 (dt,
J = 15.8, 6.9 Hz, 1 H E), 6.36 (br. d, J = 15.8 Hz, 1 H E), 6.39 (br.
d, J = 11.6 Hz, 1 H Z), 6.82–6.86 (m, 2 H Z and 2 H E), 7.02–7.08
(m, 1 H Z and 1 H E), 7.18–7.29 (m, 4 H Z and 4 H E), 7.53 (d, J
= 7.7 Hz, 1 H Z), 7.54 (d, J = 7.9 Hz, 1 H E) ppm. 13C NMR
(75 MHz, CDCl3): δ = 28.89, 33.33, 36.36, 36.39, 55.32, 55.35,
113.70, 114.03, 124.59, 124.61, 127.22, 127.39, 127.49, 127.73,
127.76, 129.26, 129.85, 130.02, 130.07, 130.23, 130.52, 130.55,
130.59, 132.89, 132.91, 141.06, 141.15, 158.40, 158.89 ppm.
C17H17BrO (317.22): calcd. C 64.37, H 5.40, Br 25.19; found C
64.85, H 5.49, Br 24.84.
1-Bromo-2-(4,4-diphenylbut-3-enyl)benzene (3): tBuOK (0.487 g,
4.34 mmol) was added to a suspension of phosphonium salt 10a
(2.167 g, 4.01 mmol) in THF (16 mL) under a flow of nitrogen. The
mixture was stirred under nitrogen for 5 min, and benzophenone
(0.652 g, 3.58 mmol) was added. The mixture was heated at reflux
for 71.5 h. The reaction mixture was then quenched with HCl (1 )
and extracted with Et2O. The organic layer was washed with water,
dried (MgSO4) and concentrated. Column chromatography of the
residue on silica gel impregnated with AgNO3 (petroleum ether)
afforded 3 (1.125 g, 86%) as a yellow oil. 1H NMR (300 MHz,
CDCl3): δ = 2.43 (q, J = 7.6 Hz, 2 H), 2.86 (t, J = 7.6 Hz, 2 H),
6.13 (t, J = 7.6 Hz, 1 H), 7.00–7.36 (m, 13 H), 7.49 (dd, J = 7.9,
1.1 Hz, 1 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 30.14, 36.39,
124.68, 127.05, 127.33, 127.45, 127.74, 128.21, 128.30, 128.35,
129.95, 130.66, 132.89, 140.07, 141.01, 142.66, 142.70 ppm.
C22H19Br (363.29): calcd. C 72.73, H 5.27, Br 21.99; found C 72.81,
H 5.26, Br 21.71.
4-Bromo-3-(4-phenylbut-3-enyl)phenyl Methyl Ether (5): tBuOK
(0.070 g, 0.62 mmol) was added to a suspension of phosphonium
salt 10b (0.328 g, 0.57 mmol) in THF (2.6 mL) under a flow of
nitrogen. The mixture was stirred under nitrogen for 15 min. Benz-
aldehyde (0.049 g, 0.46 mmol) was then added under a flow of ni-
trogen, together with THF (0.4 mL) used for rinsing. The mixture
was stirred for 25 h. at room temperature. The reaction mixture
was then quenched with HCl (1 ) and extracted with Et2O. The
organic layer was washed with water, dried (MgSO4) and concen-
trated. To induce oxidation of residual benzaldehyde and to make
the purification easier, the crude product was exposed to air and
daylight for about 24 h. Two chromatographic columns were neces-
sary to obtain a pure product. A column of silica gel (petroleum
ether/EtOAc 97:3) followed by a separation on silica gel impreg-
nated with AgNO3 (petroleum ether/EtOAc 98:2) afforded 5 as a
colourless oil (0.09 g, 61%, Z/E 91:9). 1H NMR (300 MHz,
CDCl3): δ = 2.52 (br. q, J = 7.4 Hz, 2 H E), 2.61–2.69 (m, 2 H Z),
2.81–2.89 (m, 2 H Z and 2 H E), 3.74 (s, 3 H Z), 3.76 (s, 3 H E),
5.71 (dt, J = 11.5, 7.2 Hz, 1 H Z), 6.27 (dt, J = 15.8, 6.7 Hz, 1
H E), 6.43 (br. d, J = 15.8 Hz, 1 H E), 6.47 (br. d, J = 11.5 Hz, 1
H Z), 6.60–6.66 (m, 1 H Z and 1 H E), 6.75 (d, J = 3.2 Hz, 1 H Z),
6.80 (d, J = 3.0 Hz, 1 H E), 7.18–7.36 (m, 5 H Z and 5 H E), 7.40
(d, J = 8.7 Hz, 1 H Z), 7.42 (d, J = 8.7 Hz, 1 H E) ppm. 13C NMR
(75 MHz, CDCl3): δ = 28.82, 33.34, 36.43, 36.53, 55.50, 55.52,
113.36, 113.39, 115.04, 115.07, 116.21, 126.15, 126.75, 127.12,
128.28, 128.62, 128.85, 129.57, 129.89, 130.75, 131.41, 133.37,
133.40, 137.55, 137.76, 141.96, 142.06, 159.03, 159.05 ppm.
C17H17BrO (317.22): calcd. C 64.37, H 5.40, Br 25.19; found C
64.63, H 5.39, Br 25.14.
Authentic Products: The aim of the following experiments was to
isolate small amounts of pure samples of authentic products for
the identification of Grignard reaction mixtures. Authentic samples
of [(1E)-4-phenylbut-1-enyl]benzene [(E)-2l], (4-phenylbut-1-enyl)-
benzene (2l), (1,4-diphenylbut-1-enyl)benzene (3l), 1-(diphenyl-
methyl)-2,3-dihydro-1H-indene (3c), methyl 4-(4-phenylbut-1-enyl)-
phenyl ether (4l), methyl 3-(4-phenylbut-3-enyl)phenyl ether (5l)
and 1-methoxy-3-[4-(4-methoxyphenyl)but-3-enyl]benzene (6l) were
isolated (or identified) from Grignard reaction mixtures.
[(1E)-4-Phenylbut-1-enyl]benzene [(E)-2l]: Chromatography on sil-
ica gel (petroleum ether) afforded (E)-2l. 1H NMR (500 MHz,
CDCl3): δ = 2.54 (q, J = 7.9 Hz, 2 H), 2.80 (t, J = 7.9 Hz, 2 H),
6.26 (dt, J = 15.9, 6.7 Hz, 1 H), 6.42 (d, J = 15.9 Hz, 1 H), 7.19–
7.23 (m, 4 H), 7.28–7.34 (m, 6 H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 35.01, 36.02, 126.03, 126.13, 127.07, 128.50, 128.61,
128.63, 130.11, 130.51, 137.86, 141.90 ppm. GC-MS (70 eV): m/z
(%) = 208 (14), 117 (100), 115 (39), 91 (51). NMR and MS data
agree with published values.[115]
(4-Phenylbut-1-enyl)benzene (2l): This compound was directly iden-
tified from a crude reaction mixture (Z/E 92:8). 1H NMR
2784
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 2775–2787

Radical Clocks and Grignard Reagent Formation Mechanisms
(300 MHz, CDCl3, Z isomer): δ = 2.61–2.70 (m, 2 H), 2.75–2.80
(m, 2 H), 5.70 (dt, J = 11.7, 7.0 Hz, 1 H), 6.44 (br. d, J = 11.7 Hz,
1 H), 7.16–7.34 (m, 10 H) ppm. GC-MS (70 eV): m/z (%) = 208
(19), 117 (100), 115 (38), 91 (46). 1H NMR and MS data agree with
published values.[115]
115 (15), 91 (27), 78 (12); E isomer: m/z (%) = 268 (6), 121 (13),
147 (100), 115 (19), 91 (34), 78 (15).
1-Phenylmethyl-2,3-dihydro-1H-indene (2c): This compound was
prepared by treatment of bromide (E)-2 with Bu3SnH in toluene at
reflux. 1H NMR (300 MHz, CDCl3): δ = 1.70–1.82 (m, 1 H), 2.08–
2.19 (m, 1 H), 2.68 (dd, J = 13.6, 9.3 Hz, 1 H), 2.75–2.92 (m, 2 H),
3.14 (dd, J = 13.6, 5.7 Hz, 1 H), 3.44 (quint, J = 7.3 Hz, 1 H),
7.11–7.31 (m, 9 H) ppm. GC-MS (70 eV): m/z (%) = 208 (5), 117
(100), 116 (25), 115 (31), 91 (22). 1H NMR spectroscopic data agree
with published values.[118]
(1,4-Diphenylbut-1-enyl)benzene (3l) and 1-(Diphenylmethyl)-2,3-di-
hydro-1H-indene (3c): Chromatography on AgNO3-impregnated
silica gel (petroleum ether) afforded 3l and 3c as colourless oils.
Compound 3l: 1H NMR (300 MHz, CDCl3): δ = 2.43 (q, J =
7.6 Hz, 2 H), 2.74 (t, J = 7.6 Hz, 2 H), 6.10 (t, J = 7.6 Hz, 1 H),
7.06–7.36 (m, 15 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 31.77,
36.30, 125.97, 127.00, 127.03, 127.34, 128.20, 128.27, 128.42,
128.65, 128.93, 129.94, 140.19, 141.79, 142.37, 142.78 ppm. 1H
NMR spectroscopic data obtained agree with published values.[116]
GC-MS (70 eV): m/z (%) = 284 (8), 193 (100), 115 (85), 91 (65).
Cyclised products 4c, 5c and 6c were also prepared by Bu3SnH-
induced radical cyclisation of the corresponding bromides: a solu-
tion of bromide 4, 5 or 6 (0.22–0.26 mmol), AIBN (1.3–1.7 mg)
and Bu3SnH (0.14–0.15 mL) in THF (2.4–3.0 mL) was heated at
reflux under nitrogen for 20–21 h. After concentration, the residue
was purified by column chromatography.
1
Compound 3c: H NMR (300 MHz, CDCl3): δ = 1.73–1.84 (m, 1
4-(2,3-Dihydro-1H-inden-1-ylmethyl)phenyl Methyl Ether (4c):
Chromatography on silica gel (petroleum ether, petroleum ether/
EtOAc 99:1) afforded a colourless oil that crystallised on standing
in a freezer. Recrystallisation from EtOH afforded a white solid;
m.p. 44.5–46.5 °C (lit. 46–47 °C).[119] 1H NMR (300 MHz, CDCl3):
δ = 1.69–1.81 (m, 1 H), 2.08–2.19 (m, 1 H), 2.64 (dd, J = 13.7,
9.1 Hz, 1 H), 2.73–2.93 (m, 2 H), 3.07 (dd, J = 13.7, 5.8 Hz, 1 H),
3.40 (quint, J = 7.4 Hz, 1 H), 3.80 (s, 3 H), 6.81–6.86 (m, 2 H),
7.10–7.23 (m, 6 H) ppm. 13C NMR (75 MHz, CDCl3): δ = 31.3,
32.0, 40.6, 46.7, 55.4, 113.8, 124.0, 124.6, 126.1, 126.6, 130.1, 133.1,
144.3, 147.1, 158.0 ppm.
H), 2.06–2.17 (m, 1 H), 2.70–2.87 (m, 2 H), 3.94 (d, J = 10.8 Hz,
1 H), 4.05–4.13 (m, 1 H), 6.37 (d, J = 7.6 Hz, 1 H), 6.88 (brt, J =
7.6 Hz, 1 H), 7.09 (brt, J = 7.3 Hz, 1 H), 7.13–7.33 (m, 11 H) ppm.
13C NMR (75 MHz, CDCl3): δ = 31.0, 31.9, 48.8, 57.2, 124.5,
125.4, 125.7, 126.3, 126.5, 126.7, 128.3, 128.5, 128.6, 128.7, 144.6,
144.7, 144.8, 145.9 ppm. GC-MS (70 eV): m/z (%) = 167 (23), 117
(100), 115 (20).
Methyl 4-(4-Phenylbut-1-enyl)phenyl Ether (4l): This compound
was directly identified from a crude reaction mixture (Z/E 90:10).
1H NMR (300 MHz, CDCl3): δ = 2.50 (br. q, J = 7.2 Hz, 2 H E),
2.60–2.69 (m, 2 H Z), 2.73–2.80 (m, 2 H Z and 2 H E), 3.79 (s, 3
H E), 3.80 (s, 3 H Z), 5.61 (dt, J = 11.7, 6.9 Hz, 1 H Z), 6.11 (dt,
J = 15.7, 6.8 Hz, 1 H E), 6.36 (br. d, J = 15.7 Hz, 1 H E), 6.37 (br.
d, J = 11.7 Hz, 1 H Z), 6.82–6.87 (m, 2 H Z and 2 H E), 7.16–7.32
(m, 7 H Z and 7 H E) ppm. 13C NMR (Z isomer, 75 MHz,
CDCl3): δ = 30.6, 36.3, 55.4, 113.7, 126.0, 128.5, 128.6, 128.9,
130.0, 130.4, 141.9, 158.4 ppm. GC-MS (70 eV): Z isomer m/z (%)
= 238 (4), 147 (100), 115 (22), 103 (12), 91 (46); E isomer m/z (%)
= 238 (2), 147 (100), 115 (32), 91 (24). 1H NMR and MS data agree
with published values.[117]
Methyl 1-Phenylmethyl-2,3-dihydro-1H-inden-5-yl Ether (5c):
Chromatography on silica gel (petroleum ether, petroleum ether/
EtOAc 99:1) afforded an oil that crystallised, in the presence of
EtOH, on standing in a freezer. This white solid was washed with
cold EtOH; m.p. 39–40 °C. 1H NMR (300 MHz, CDCl3): δ = 1.71–
1.82 (m, 1 H), 2.09–2.20 (m, 1 H), 2.67 (dd, J = 13.6, 9.1 Hz, 1 H),
2.71–2.90 (m, 2 H), 3.08 (dd, J = 13.6, 5.9 Hz, 1 H), 3.38 (quint, J
= 7.2 Hz, 1 H), 3.78 (s, 3 H), 6.68 (dd, J = 8.2, 2.5 Hz, 1 H), 6.77
(d, J = 2.5 Hz, 1 H), 6.98 (d, J = 8.2 Hz, 1 H), 7.18–7.32 (m, 5
H) ppm. 13C NMR (75 MHz, CDCl3): δ = 31.5, 32.5, 41.9, 45.8,
55.5, 110.1, 112.0, 124.4, 126.0, 128.4, 129.2, 139.2, 141.1, 145.8,
159.0 ppm. GC-MS (70 eV): m/z (%) = 238 (2), 147 (100), 115 (13),
91 (29).
Methyl 3-(4-Phenylbut-3-enyl)phenyl Ether (5l): Chromatography
on silica gel (petroleum ether/EtOAc 99:1) followed by chromatog-
raphy on AgNO3-impregnated silica gel (petroleum ether) afforded
5l (Z/E 91:9). 1H NMR (300 MHz, CDCl3): δ = 2.53 (br. q, J =
7.4 Hz, 2 H E), 2.61–2.69 (m, 2 H Z), 2.72–2.80 (m, 2 H Z and 2
H E), 3.78 (s, 3 H Z), 3.80 (s, 3 H E), 5.70 (dt, J = 11.7, 6.8 Hz, 1
H Z), 6.25 (dt, J = 15.9, 6.6 Hz, 1 H E), 6.42 (br. d, J = 15.9 Hz,
1 H E), 6.44 (br. d, J = 11.7 Hz, 1 H Z), 6.72–6.83 (m, 3 H Z and
3 H E), 7.17–7.34 (m, 6 H Z and 6 H E) ppm. 13C NMR (Z isomer,
75 MHz, CDCl3): δ = 30.4, 36.2, 55.2, 111.3, 114.4, 121.0, 126.7,
128.3, 128.9, 129.4, 129.6, 131.9, 137.7, 143.4, 159.8 ppm. GC-MS
(70 eV): Z isomer: m/z (%) = 238 (7), 121 (25), 117 (100), 115 (32),
91 (32), 78 (12); E isomer: m/z (%) = 238 (4), 121 (23), 117 (100),
115 (32), 91 (33), 78 (14).
5-Methoxy-1-[(4-methoxyphenyl)methyl]-2,3-dihydro-1H-indene (6c):
Chromatography on silica gel (petroleum ether, petroleum ether/
EtOAc 95:5) afforded an oil that crystallised on standing at room
temperature. Recrystallisation from EtOH afforded a white solid;
m.p. 69–70 °C (lit.[120] 68–70 °C). 1H NMR (300 MHz, CDCl3): δ =
1.69–1.81 (m, 1 H), 2.09–2.20 (m, 1 H), 2.62 (dd, J = 13.7, 8.8 Hz, 1
H), 2.70–2.89 (m, 2 H), 3.01 (dd, J = 13.7, 5.9 Hz, 1 H), 3.34 (quint,
J = 7.3 Hz, 1 H) 3.78 (s, 3 H), 3.80 (s, 3 H), 6.68 (dd, J = 8.3,
2.3 Hz, 1 H), 6.77 (d, J = 2.3 Hz, 1 H), 6.81–6.86 (m, 2 H), 6.97
(d, J = 8.3 Hz, 1 H), 7.08–7.13 (m, 2 H) ppm. 13C NMR (75 MHz,
CDCl3): δ = 31.5, 32.4, 41.0, 45.9, 55.4, 55.5, 110.0, 112.0, 113.8,
124.5, 130.1, 133.2, 139.3, 145.9, 158.0, 158.9 ppm. GC-MS
(70 eV): m/z (%) = 268 (6), 121 (21), 147 (100).
1-Methoxy-3-[4-(4-methoxyphenyl)but-3-enyl]benzene (6l): Column
chromatography of the residue (petroleum ether/EtOAc 99:1) af-
1
forded 6l (Z/E 90:10). H NMR (300 MHz, CDCl3): δ = 2.50 (br.
Cyclic Voltammetry Experiments: The working electrode was a 3-
mm diameter glassy carbon disk. It was carefully polished before
each voltammogram. The counter electrode was a platinum wire
and the reference electrode an aqueous SCE electrode. Cyclic vol-
tammograms were recorded with a CHI600B potentiostat (CH In-
struments, IJ Cambria Scientific, Burry Port, UK).
q, J = 7.4 Hz, 2 H E), 2.60–2.69 (m, 2 H Z), 2.71–2.78 (m, 2 H Z
and 2 H E), 3.78 (s, 3 H Z), 3.79 (s, 6 H E), 3.80 (s, 3 H Z), 5.60
(dt, J = 11.6, 6.9 Hz, 1 H Z), 6.11 (dt, J = 15.8, 6.8 Hz, 1 H E),
6.34–6.39 (m, 1 H E and 1 H Z), 6.72–6.87 (m, 5 H Z and 5 H E),
7.16–7.28 (m, 3 H Z and 3 H E) ppm. 13C NMR (Z isomer,
75 MHz, CDCl3): δ = 30.5, 36.3, 55.3, 55.4, 111.3, 113.7, 114.4,
121.0, 129.0, 129.4, 130.1, 130.3, 130.4, 143.6, 158.4, 159.8 ppm.
GC-MS (70 eV): Z isomer: m/z (%) = 268 (7), 121 (8), 147 (100),
[1] D. Griller, K. U. Ingold, Acc. Chem. Res. 1980, 13, 317–323.
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2785

J.-M. Mattalia et al.
FULL PAPER
[2]
[38] E. A. Hill, C. L. Harder, R. Wagner, D. Meh, R. P. Bowman,
J. Organomet. Chem. 1986, 302, 5–17.
[39] D. P. Curran, Synthesis 1988, 417–439.
[40] D. G. Peters, in: Organic Electrochemistry: An Introduction and
a Guide (Eds.: M. M. Baizer, H. Lund), 3rd edition, M. Dekker,
New York, 1991, pp. 361–400.
J.-M. Surzur, in: Reactive Intermediates (Ed.: R. A. Abramov-
itch), Plenum Press, New York, 1982, vol. 2, 121–295.
M. Newcomb, Tetrahedron 1993, 49, 1151–1176.
M. Newcomb, in: Radicals in Organic Synthesis (Eds.: P. Re-
naud, M. P. Sibi), Wiley-VCH, Weinheim, 2001, vol. 1,
pp. 317–336.
[3]
[4]
[41] T. A. Halgren, J. D. Roberts, J. H. Horner, F. N. Martinez, C.
Tronche, M. Newcomb, J. Am. Chem. Soc. 2000, 122, 2988–
2994.
[42] E. Martinez II, M. Newcomb, J. Org. Chem. 2006, 71, 557–561.
[43] A. Annunziata, C. Galli, M. Marinelli, T. Pau, Eur. J. Org.
Chem. 2001, 1323–1329.
[44] S. U. Park, S. K. Chung, M. Newcomb, J. Am. Chem. Soc.
1986, 108, 240–244.
[45] H. Hazimeh, J. M. Mattalia, C. Marchi-Delapierre, R. Barone,
N. S. Nudelman, M. Chanon, J. Phys. Org. Chem. 2005, 18,
1145–1160.
[46] H. Hazimeh, J.-M. Mattalia, M. Attolini, N. Bodineau, K.
Handoo, C. Marchi-Delapierre, E. Péralez, M. Chanon, Indian
J. Chem. Sect. B 2006, 45, 2270–2280.
[47] H. Hazimeh, F. Kanoufi, C. Combellas, J.-M. Mattalia, C.
Marchi-Delapierre, M. Chanon, J. Phys. Chem. C 2008, 112,
2545–2557.
[5]
[6]
[7]
K. E. Liu, C. C. Johnson, M. Newcomb, S. J. Lippard, J. Am.
Chem. Soc. 1993, 115, 939–947.
E. W. Della, C. H. Schiesser, D. K. Taylor, J. C. Walton, J.
Chem. Soc. Perkin Trans. 2 1991, 1329–1333.
P. E. Eaton, Y. C. Yip, J. Am. Chem. Soc. 1991, 113, 7692–
7697.
M. Journet, M. Malacria, J. Org. Chem. 1992, 57, 3085–3093.
R. R. Conry, J. M. Mayer, Organometallics 1991, 10, 3160–
3166.
W. E. Piers, D. J. Parks, L. R. Macgillivray, M. J. Zaworotko,
Organometallics 1994, 13, 4547–4558.
J. A. Franz, J. C. Linehan, J. C. Birnbaum, K. W. Hicks, M. S.
Alnajjar, J. Am. Chem. Soc. 1999, 121, 9824–9830.
D. Sattari, C. L. Hill, J. Am. Chem. Soc. 1993, 115, 4649–4657.
E. I. Karasevich, V. S. Kulikova, A. E. Shilov, A. A. Shtein-
man, Russ. Chem. Rev. 1998, 67, 376–398.
W. Adam, F. G. Gelalcha, C. R. Saha-Moller, V. R. Stegmann,
J. Org. Chem. 2000, 65, 1915–1918.
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[48] J. F. Garst, M. R. Soriaga, Coord. Chem. Rev. 2004, 248, 623–
652.
M. Costas, K. Chen, L. Que, Coord. Chem. Rev. 2000, 200,
517–544.
[49] T. Cohen, H. Gibney, R. Ivanov, E. A. H. Yeh, I. Marek, D. P.
Curran, J. Am. Chem. Soc. 2007, 129, 15405–15409.
[50] E. Péralez, J.-C. Négrel, M. Chanon, Tetrahedron Lett. 1995,
36, 6457–6460.
[51] A. L. J. Beckwith, I. A. Blair, G. Phillipou, Tetrahedron Lett.
1974, 15, 2251–2254.
[16]
[17]
R. H. Crabtree, J. Chem. Soc., Dalton Trans. 2001, 2437–2450.
F. Alonso, I. P. Beletskaya, M. Yus, Chem. Rev. 2002, 102,
4009–4091.
M. L. Andersen, W. Long, D. D. M. Wayner, J. Am. Chem. Soc.
1997, 119, 6590–6595.
[18]
[19]
[20]
[21]
[22]
[23]
[52] F. M’Halla, J. Pinson, J. M. Savéant, J. Am. Chem. Soc. 1980,
J. M. Tanko, J. P. Phillips, J. Am. Chem. Soc. 1999, 121, 6078–
6079.
102, 4120–4127.
[53] A. Leahy-Dios, A. Firoozabadi, J. Phys. Chem. B 2007, 111,
191–198.
D. M. Fang, D. G. Peters, M. S. Mubarak, J. Electrochem. Soc.
2001, 148, E464–E467.
H. Jensen, H. S. Sorensen, S. U. Pedersen, K. Daasbjerg, J.
Chem. Soc. Perkin Trans. 2 2002, 1423–1428.
J. P. Stevenson, W. F. Jackson, J. M. Tanko, J. Am. Chem. Soc.
2002, 124, 4271–4281.
T. Kjaersbo, K. Daasbjerg, S. U. Pedersen, Electrochem. Acta
2003, 48, 1807–1816.
W. Adam, R. Finzel, J. Am. Chem. Soc. 1992, 114, 4563–4568.
R. M. Wilson, K. A. Schnapp, Chem. Rev. 1993, 93, 223–249.
C. Chatgilialoglu, A. Altieri, H. Fischer, J. Am. Chem. Soc.
2002, 124, 12816–12823.
[54] H. G. Richey Jr., H. S. Veale, Tetrahedron Lett. 1975, 16, 615–
618.
[55] E. A. Hill, in: Advances in Organometallic Chemistry (Eds.:
F. G. A. Stone, R. West), Academic Press, New York, 1977, vol.
16, pp. 131–165.
[56] M. Julliard, J. P. Scagliarini, M. Rajzmann, M. Chanon, Chi-
mia 1986, 40, 16–19.
[57] H. Morrison, K. Muthuramu, G. Pandey, D. Severance, B.
Bigot, J. Org. Chem. 1986, 51, 3358–3363.
[24]
[25]
[26]
[58] M. Chanon, M. Rajzmann, F. Chanon, Tetrahedron 1990, 46,
6193–6299.
[27]
[28]
E. J. Corey, R. Nagata, J. Am. Chem. Soc. 1987, 109, 8107–
8108.
M. Sono, M. P. Roach, E. D. Coulter, J. H. Dawson, Chem.
Rev. 1996, 96, 2841–2887.
[59] J. S. Duca, M. H. Gallego, A. B. Pierini, R. A. Rossi, J. Org.
Chem. 1999, 64, 2626–2629.
[60] D. Laage, I. Burghardt, T. Sommerfeld, J. T. Hynes, J. Phys.
Chem. A 2003, 107, 11271–11291.
[29]
[30]
M. Newcomb, P. H. Toy, Acc. Chem. Res. 2000, 33, 449–455.
M. Merkx, D. A. Kopp, M. H. Sazinsky, J. L. Blazyk, J. Muller,
S. J. Lippard, Angew. Chem. Int. Ed. 2001, 40, 2782–2807.
M. Newcomb, M. B. Manek, A. G. Glenn, J. Am. Chem. Soc.
1991, 113, 949–958.
M. Chanon, J.-C. Négrel, N. Bodineau, J.-M. Mattalia, E.
Péralez, A. Goursot, Macromol. Symp. 1998, 134, 13–28.
J. F. Garst, J. R. Boone, L. Webb, K. E. Lawrence, J. T. Baxter,
F. Ungváry, Inorg. Chim. Acta 1999, 296, 52–66.
J. F. Garst, F. Ungváry, in: Grignard Reagents: New Develop-
ments (Ed.: H. G. Richey Jr.,), Wiley, Chichester, 2000,
pp. 185–275.
[61] R. A. Rossi, A. B. Pierini, A. B. Penenory, Chem. Rev. 2003,
103, 71–167.
[62] N. Takeda, P. V. Poliakov, A. R. Cook, J. R. Miller, J. Am.
Chem. Soc. 2004, 126, 4301–4309.
[31]
[32]
[33]
[34]
[63]
[64]
[65]
[66]
J. M. Dust, D. R. Arnold, J. Am. Chem. Soc. 1983, 105, 1221–
1227.
A. M. de P. Nicholas, D. R. Arnold, Can. J. Chem. 1986, 64,
270–276.
R. A. Jackson, M. Sharifi, J. Chem. Soc. Perkin Trans. 2 1996,
775–778.
J. P. Cheng, B. Liu, Y. Y. Zhao, Z. Wen, Y. K. Sun, J. Am.
Chem. Soc. 2000, 122, 9987–9992.
[35]
[36]
[67]
[68]
N. Bodineau, J.-M. Mattalia, V. Timokhin, K. Handoo, J.-C.
Négrel, M. Chanon, Org. Lett. 2000, 2, 2303–2306.
A. L. J. Beckwith, K. U. Ingold, in: Rearrangements in Ground
X. Zhang, W. M. Nau, J. Phys. Org. Chem. 2000, 13, 634–639.
Z. Wen, Z. Li, Z. Shang, J.-P. Cheng, J. Org. Chem. 2001, 66,
1466–1472.
and Excited States (Ed.: P. de Mayo), Academic Press, New [69]
York, 1980, vol. 1, pp. 161–310.
L. W. Menapace, M. B. Loewenthal, J. Koscielecki, L. Tucker,
L. C. Passaro, R. Montalbano, A. J. Frank, J. Marrantino, J.
Brunner, Organometallics 2002, 21, 3066–3068.
N. K. Singh, P. L. A. Popelier, P. J. O’Malley, Chem. Phys. Lett.
2006, 426, 219–221.
[37]
L. J. Johnston, J. Lusztyk, D. D. M. Wayner, A. N. Abeywick-
reyma, A. L. J. Beckwith, J. C. Scaiano, K. U. Ingold, J. Am.
Chem. Soc. 1985, 107, 4594–4596.
[70]
2786
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 2775–2787

Radical Clocks and Grignard Reagent Formation Mechanisms
[71]
[72]
[95] J. M. Savéant, in: Advances in Physical Organic Chemistry (Ed.:
T. T. Tidwell), Academic Press, New York, 2000, vol. 35, pp.
117–192.
[96] J. M. Savéant, Elements of Molecular and Biomolecular Electro-
chemistry. An Electrochemical Approach to Electron Transfer
Chemistry, Wiley, Hoboken, 2006.
[97] B. A. Selivanov, I. I. Bilkis, V. L. Varand, V. D. Shteingarts, Zh.
Org. Khim. 1985, 21, 952–965.
[98] S. Fukuzumi, T. Suenobu, T. Hirasaka, R. Arakawa, K. M.
Kadish, J. Am. Chem. Soc. 1998, 120, 9220–9227.
[99] S. Fukuzumi, D. M. Guldi, in: Electron Transfer in Chemistry
(Ed.: V. Balzani), Wiley-VCH, Weinheim, 2001, vol. 2, pp. 270–
337.
S. Antonello, F. Maran, J. Am. Chem. Soc. 1998, 120, 5713–
5722.
S. Antonello, M. Crisma, F. Formaggio, A. Moretto, F. Taddei,
C. Toniolo, F. Maran, J. Am. Chem. Soc. 2002, 124, 11503–
11513.
N. C. L. Wood, C. A. Paddon, F. L. Bhatti, T. J. Donohoe,
R. G. Compton, J. Phys. Org. Chem. 2007, 20, 732–742.
H. J. R. de Boer, O. S. Akkerman, F. Bickelhaupt, Angew.
Chem. Int. Ed. Engl. 1988, 27, 687–689.
G. P. M. van Klink, H. J. R. de Boer, G. Schat, O. S. Akker-
man, F. Bickelhaupt, A. L. Spek, Organometallics 2002, 21,
2119–2135.
[73]
[74]
[75]
[100] B. J. Beals, Z. I. Bello, K. P. Cuddihy, E. M. Healy, S. E.
Koon-Church, J. M. Owens, C. E. Teerlinck, W. J. Bowyer, J.
Phys. Chem. A 2002, 106, 498–503.
[101] P. G. Jasien, J. A. Abbondondola, Theochem, J. Mol. Struct.
2004, 671, 111–118.
[76]
[77]
F. Bickelhaupt, M. Newcomb, C. B. DeZutter, H. J. R. de Boer,
Eur. J. Org. Chem. 2008, 6225–6231.
F. Kanoufi, C. Combellas, H. Hazimeh, J. M. Mattalia, C.
Marchi-Delapierre, M. Chanon, J. Phys. Org. Chem. 2006, 19,
847–866.
E. C. Ashby, J. Oswald, J. Org. Chem. 1988, 53, 6068–6076.
H. W. H. J. Bodewitz, C. Blomberg, F. Bickelhaupt, Tetrahe-
dron 1975, 31, 1053–1063.
N. Kimura, S. Takamuku, Radiat. Phys. Chem. 1987, 29, 179–
183.
[102] S. V. Maslennikov, E. N. Kirillov, N. V. Berezina, I. V. Spirina,
A. V. Kondin, Russ. J. Gen. Chem. 1997, 67, 667–671.
[103] S. V. Maslennikov, A. V. Piskunov, L. N. Bochkarev, I. V. Spi-
rina, Russ. J. Gen. Chem. 1998, 68, 907–909.
[104] S. V. Maslennikov, I. V. Spirina, A. V. Piskunov, S. N. Maslen-
nikova, Russ. J. Gen. Chem. 2001, 71, 1741–1743.
[105] M. Spiro, in: Modern Aspects of Electrochemistry (Eds.:
J. O. M. Bockris, B. E. Conway, R. E. White), Kluwer Aca-
demic, New York, 2001, vol. 34, pp. 1–11.
[106] R. Akid, M. Garma, Electrochem. Acta 2004, 49, 2871–2879.
[107] C. Amatore, M. Azzabi, P. Calas, A. Jutand, C. Lefrou, Y.
Rollin, J. Electroanal. Chem. 1990, 288, 45–63.
[108] J. T. Scanlon, D. E. Willis, J. Chromatogr. Sci. 1985, 23, 333–
340.
[78]
[79]
[80]
[81]
[82]
J. F. Garst, B. L. Swift, J. Am. Chem. Soc. 1989, 111, 241–250.
J. F. Garst, Acc. Chem. Res. 1991, 24, 95–97.
[83] S. Antonello, M. Musumeci, D. D. M. Wayner, F. Maran, J.
Am. Chem. Soc. 1997, 119, 9541–9549.
[84] C. P. Andrieux, J. Pinson, J. Am. Chem. Soc. 2003, 125, 14801–
14806.
[85] Y. Fu, L. Liu, H.-Z. Yu, Y.-M. Wang, Q.-X. Guo, J. Am. Chem.
Soc. 2005, 127, 7227–7234.
[86] B. Branchi, C. Galli, P. Gentili, Eur. J. Org. Chem. 2002, 2844–
2854.
[87] A. N. Abeywickrema, A. L. J. Beckwith, J. Chem. Soc., Chem.
Commun. 1986, 464–465.
[88] S. J. Garden, D. V. Avila, A. L. J. Beckwith, V. W. Bowry, K. U.
Ingold, J. Lusztyk, J. Org. Chem. 1996, 61, 805–809.
[89] C. Amatore, in: Organic Electrochemistry: An Introduction and
a Guide (Eds.: M. M. Baizer, H. Lund), 3rd edition, M. Dekker,
New York, 1991, pp. 207–232.
[90] C. Amatore, in: Organic Electrochemistry (Eds.: H. Lund, O.
Hammerich), Fourth Edition, M. Dekker, New York, 2001,
pp. 183–205.
[91] C. Ha, J. H. Horner, M. Newcomb, T. R. Varick, B. R. Arnold,
J. Lusztyk, J. Org. Chem. 1993, 58, 1194–1198.
[92] M. Newcomb, S.-Y. Choi, J. H. Horner, J. Org. Chem. 1999,
64, 1225–1231.
[93] F. De Vleeschouwer, V. Van Speybroeck, M. Waroquier, P.
Geerlings, F. De Proft, J. Org. Chem. 2008, 73, 9109–9120.
[94] C. P. Andrieux, J. M. Savéant, D. Zann, New J. Chem. 1984, 8,
107–116.
[109] G. F. Meijs, J. F. Bunnett, A. L. J. Beckwith, J. Am. Chem.
Soc. 1986, 108, 4899–4904.
[110] T. R. Hoye, L. Mi, J. Org. Chem. 1997, 62, 8586–8588.
[111] M. Adamczyk, D. S. Watt, D. A. Netzel, J. Org. Chem. 1984,
49, 4226–4237.
[112] S. C. Watson, J. F. Eastham, J. Organomet. Chem. 1967, 9,
165–168.
[113] J. Knight, P. J. Parsons, J. Chem. Soc. Perkin Trans. 1 1989,
979–984.
[114] M. Oestreich, U. K. Schmid, G. Auer, M. Keller, Synthesis
2003, 2725–2739.
[115] Y. S. Hon, L. Lu, R. C. Chang, S. W. Lin, P. P. Sun, C. F. Lee,
Tetrahedron 2000, 56, 9269–9279.
[116] H. Uno, K. Sakamoto, F. Semba, H. Suzuki, Bull. Chem. Soc.
Jpn. 1992, 65, 210–217.
[117] T. Satoh, N. Hanaki, N. Yamada, T. Asano, Tetrahedron
2000, 56, 6223–6234.
[118] M. Sindler-Kulyk, W. H. Laarhoven, J. Am. Chem. Soc. 1978,
100, 3819–3830.
[119] C. Courtot, Ann. Chim. 1916, 9, 52–108.
[120] D. M. Lynch, W. Cole, J. Med. Chem. 1968, 11, 291–295.
Received: January 28, 2009
Published Online: April 29, 2009
Eur. J. Org. Chem. 2009, 2775–2787
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2787