Spin-trapping evidence for the formation of alkyl, alkoxyl, and alkylperoxyl radicals in the reactions of dialkylzincs with oxygen

Article abstract of DOI:10.1002/chem.201002616

The reactions of dialkylzincs (Me₂Zn, Et₂Zn, and nBu₂Zn) with oxygen have been investigated by EPR spectroscopy using spin-trapping techniques. The use of 5-diethoxyphosphoryl-5-methyl-1-pyrrroline N-oxide (DEPMPO) as a spin-trap has allowed the involvement of alkyl, alkylperoxyl, and alkoxyl radicals in this process to be probed for the first time. The relative ratio of the corresponding spin-adducts depends strongly on the nature of the R group, which controls the C-Zn bond dissociation enthalpy, and on the experimental conditions (excess of spin-trap compared with R ₂Zn and vice versa). The results have demonstrated that Et ₂Zn and, to a lesser extent, nBu₂Zn are much better traps for oxygen-centered species than Me₂Zn. When the dialkylzincs were used in excess with respect to the spin-trap, the concentration of the oxygen-centered radical adducts of DEPMPO was much lower for Et₂Zn and nBu₂Zn than for Me₂Zn. A detailed reaction mechanism is discussed and C-Zn, O-Zn, and O-O bond dissociation enthalpies for the proposed reaction intermediates were calculated at the UB3LYP/6-311++G(3df,3pd)/ /UB3LYP/6-31G(d,p) level of theory to support the rationale. All trapped! The reactions of dialkylzincs with oxygen (see scheme) have been investigated with EPR spectroscopy by using spin-trapping techniques. The use of DEPMPO as a spin-trap has allowed the involvement of alkyl, alkylperoxyl, and alkoxyl radicals in this process to be probed for the first time. A detailed mechanism is discussed in the light of theoretical calculations of the C-Zn, O-Zn, and O-O bond dissociation enthalpies of the proposed reaction intermediates.

Full text of DOI:10.1002/chem.201002616

DOI: 10.1002/chem.201002616
Spin-Trapping Evidence for the Formation of Alkyl, Alkoxyl, and
Alkylperoxyl Radicals in the Reactions of Dialkylzincs with Oxygen
Julien Maury, Laurence Feray,* Samantha Bazin, Jean-Louis Clꢀment,
Sylvain R. A. Marque, Didier Siri, and Michꢁle P. Bertrand*[a]
Abstract: The reactions of dialkylzincs
(Me2Zn, Et2Zn, and nBu2Zn) with
oxygen have been investigated by EPR
spectroscopy using spin-trapping tech-
niques. The use of 5-diethoxyphosphor-
Zn bond dissociation enthalpy, and on
the experimental conditions (excess of
spin-trap compared with R2Zn and vice
versa). The results have demonstrated
that Et2Zn and, to a lesser extent,
nBu2Zn are much better traps for
oxygen-centered species than Me2Zn.
When the dialkylzincs were used in
excess with respect to the spin-trap, the
concentration of the oxygen-centered
radical adducts of DEPMPO was much
lower for Et2Zn and nBu2Zn than for
Me2Zn. A detailed reaction mechanism
ꢀ
ꢀ
ꢀ
yl-5-methyl-1-pyrrroline
N-oxide
is discussed and C Zn, O Zn, and O
O bond dissociation enthalpies for the
proposed reaction intermediates were
calculated at the UB3LYP/6-311++G-
(DEPMPO) as a spin-trap has allowed
the involvement of alkyl, alkylperoxyl,
and alkoxyl radicals in this process to
be probed for the first time. The rela-
tive ratio of the corresponding spin-ad-
ducts depends strongly on the nature
A
ACHTUNGTREN(NUNG d,p) level of
Keywords: EPR spectroscopy
·
oxygen · radical reactions · spin
traps · dialkylzinc
ꢀ
of the R group, which controls the C
Introduction
polar mechanism was adopted for the formation of
(ROO)2Zn and its subsequent reduction to the correspond-
ing zinc dialkoxide in the presence of an excess of dialkyl-
zinc.
At the very beginning of these mechanistic investigations
it was pointed out that diethyl- and dimethylzinc behaved
differently. In brief, Et2Zn, nBu2Zn, and iPr2Zn were shown
to lead to compounds formulated as ZnR2O or ZnR2O2, de-
pending on the experimental conditions. However, Me2Zn
The reactions of dialkylzincs with oxygen have been investi-
gated since their discovery in 1848.[1] Walling and Buckler
were the first to envisage a radical pathway involving an
electron-transfer step, but they opted instead for the forma-
tion of charge-transfer complexes.[2] In 1959, Hock and
Ernst suggested a cage recombination radical mechanism to
explain the formation of butyl hydroperoxide in excellent
yield from the oxidation of nBu2Zn.[3] Radical intermediates
were also proposed for the oxidation of allylzinc deriva-
tives.[4] Studies by Abraham showed that these oxidation re-
actions were not inhibited by good radical traps[5] and thus a
ꢀ
was found to oxidize more slowly. In this case only one Zn
C bond was broken, which led exclusively to the monoalk-
oxide MeZnOMe.[6] All these data can be summarized by
Equations (1) and (2).[1] Recent reports from inorganic
chemists have confirmed these characteristic features.[7]
R Zn O RZnOOR O ðROOÞ Zn R Zn 2 ðROÞ Zn
ð1Þ
ð2Þ
2
2
2
ƒ!
ƒ!
ƒƒ!
2
2
2
[a] J. Maury, Dr. L. Feray, Dr. S. Bazin, Dr. J.-L. Clꢀment,
Prof. S. R. A. Marque, Prof. D. Siri, Prof. M. P. Bertrand
Laboratoire Chimie Provence, UMR 6264
Me Zn O MeZnOOMe Me Zn MeZnOMe
2
2
ƒ!
ƒƒƒ!
2
Facultꢀ des Sciences et Techniques de Saint-Jꢀrꢁme
Avenue Normandie-Niemen, Aix-Marseille Universitꢀ
13397 Marseille Cedex 20 (France)
It was only in 1968 with the development of EPR tech-
niques that Davies and Roberts first probed the involvement
of radicals in these reactions and proposed a radical-chain
mechanism.[8] This mechanism, similar to the mechanism
suggested for the oxidation of other organometallic spe-
cies,[8,9] is now generally admitted by most synthetic chemists
E-mail: michele.bertrand@univ-cezanne.fr
laurence.feray@univ-cezanne.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem201002616.
1586
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1586 – 1595

FULL PAPER
using dialkylzinc in aerobic media.[10–14]
However the debate still continues. Recent studies by
[15]
´
Lewinsky and co-workers, which led to the first isolation
of four-coordinate zinc alkyl peroxide complexes,[15a,b] have
again addressed the question of the oxidation mechanism.
These authors, who first opted for a nonradical mecha-
nism,[15b] then reconsidered the mechanism and suggested
the activation of oxygen through electron transfer and the
formation of a caged radical pair.[15c] Eventually, having in-
vestigated the oxygenation of the Me2Zn/a-diimine complex
´
leading to a methyl peroxide cluster, Lewinsky and co-work-
ers proposed a rationale based on the homolysis of MeZ-
nOOMe [Eq. (3)]. This mechanism is consistent with the
production of both a methyl and a methoxyl radical and
with the observation that every zinc atom retains one
methyl group.[15d,e,f] An alternative pathway will be discussed
later.
Pioneering EPR studies have unambiguously proved the
formation of the ethyl radical from diethylzinc, initiated at
low temperature in the presence of di-tert-butyl peroxide.[8b]
Moreover, both methyl and ethyl radicals have been detect-
ed by spin-trapping using nitroso compounds[16] or nitrones
such as a-phenyl-N-tert-butylnitrone (PBN).[17]
Figure 1. Trapping of Et· by TTBNB. a) Et2Zn (0.4m in hexane,
0.04 mmol) and (tBu)3C6H2N=O (0.1m, 0.025 mmol) undegassed at
room temperature. b) Simulation.
However, in these reactions promoted by oxygen, no
spectroscopic study has ever demonstrated the involvement
of any oxygen-centered paramagnetic species, neither alkyl-
peroxyl nor alkoxyl radicals. We report in this article the
very first probe for the simultaneous formation of alkyl, al-
koxyl, and alkylperoxyl radicals during the oxidation of
Me2Zn, nBu2Zn, and Et2Zn in EPR experiments performed
in the presence of 5-diethoxyphosphoryl-5-methyl-1-pyrro-
line N-oxide (DEPMPO) as spin-trap. DFT calculations of
ꢀ
ꢀ
ꢀ
the C Zn, O Zn, and O O bond dissociation enthalpies
were performed to support the proposed mechanism.
Results and Discussion
Preliminary EPR experiments conducted with 2,4,6-tri-tert-
butylnitrosobenzene (TTBNB) enabled us to trap both ethyl
and methyl radicals from Et2Zn and Me2Zn, respectively. At
room temperature an undegassed hexane solution of dieth-
ylzinc containing TTBNB led to signals of a 68:32 mixture
of the nitroxyl and anilino radicals that were expected from
the trapping of an ethyl radical.[18,19] The same experiment,
carried out with dimethylzinc (commercially available 1m
solution in toluene), led similarly to superimposed signals of
nitroxyl and anilino radicals resulting from the addition of a
methyl radical to nitrogen and oxygen. These adducts were
formed in a 66:34 ratio, respectively (Figures 1 and 2).
Figure 2. Trapping of Me· by TTBNB. a) Me2Zn (0.4m in toluene,
0.04 mmol) and (tBu)3C6H2N=O (0.1m, 0.025 mmol) undegassed at
room temperature b) Simulation.
TTBNB is not recognized as being a good trap for radi-
cals other than alkyl radicals.[18] However, when oxygen was
bubbled at 253 K into a solution of dimethylzinc (1.6:1
Me2Zn/TTBNB), after immediate cooling to 78 K a three-
line spectrum was recorded that could be assigned to the
trapping of either the methoxyl or methylperoxyl radical
(aN =2.41 mT; Figure 3).[20,21]
Chem. Eur. J. 2011, 17, 1586 – 1595
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1587

L. Feray, M. P. Bertrand et al.
tra, are given in Table 1. Typical spectra are shown in Fig-
ures 4–6. The hyperfine coupling constants of the detected
adducts are reported in Table 2.
Due to the presence of the stereogenic center, DEPMPO
leads to mixtures of diastereomeric adducts except in the
case of the bulkiest radicals. The major isomers (M) are
likely to be trans owing to the bulk of the diethoxyphos-
phoryl group. These adducts are characterized by their hy-
perfine coupling constants, in good agreement with previ-
ously reported data.[26] Note that coupling constants can
vary depending on the reactions conditions, particularly on
the nature of the solvent and on the temperature.[27] The
main features are summarized as follows.
Figure 3. Trapping of oxygen-centered radicals MeO· or MeOO· by
TTBNB. ZnMe2 (0.4m in toluene, 0.04 mmol) and (tBu)3C6H2N=O
(0.1m, 0.025 mmol) (1.6:1 ratio), O2-bubbling at 253 K and then cooling
to 78 K
With regard to the trapping of the alkyl radical AR,[26] the
addition is not 100% stereoselective; a mixture of trans (M)
and cis (m) diastereoisomers was detected for the addition
of both Me· and nBu·. The aN values are typical of quasi-
planar nitroxyl groups. High values of aP indicate that the
EPR experiments conducted with 5,5-dimethyl-1-pyrroline
(DMPO) as radical trap led similarly to clear-cut trapping of
either alkyl- or oxygen-centered radicals depending on the
experimental conditions. However, these experiments did
not allow the unambiguous identification of the species re-
sulting from the trapping of oxygen-centered radicals be-
cause the hyperfine coupling constants of the alkoxyl and al-
kylperoxyl radical adducts are virtually identical, as already
reported in the literature.[22] Neither TTBNB nor DMPO al-
lowed the trapping of alkoxyl radicals to be differentiated
from the trapping of alkylperoxyl radicals.
One of the main problems concerning the trapping of al-
kylperoxyl radicals by nitrones resides in the very short half-
lives of the corresponding adducts. Moreover, at tempera-
tures higher than 250 K, reputedly only alkoxyl radical ad-
ducts are detected.[22,23]
ꢀ
mean dihedral angle between the C P bond and the pz orbi-
tal at the nitrogen atom is rather small.[28] This is consistent
with the phosphorus atom lying in a pseudo-axial mean posi-
tion. Note that aP trans is lower than aP cis. Values for aHb
are similarly high, which is consistent with a preferential
pseudo-equatorial mean position of the alkyl groups in both
diastereomeric adducts. Note that aHb trans is higher than
aHb cis.
With regard to the alkoxyl radical adducts BR,[29,30] the aN
values are lower for adducts B than for adducts A. This is
consistent with the electron-withdrawing effect of the
alkoxy group, which destabilizes the development of a posi-
tive charge on the nitrogen atom. The aP constants are high,
The discovery that the super-
Table 1. Relative ratios of the A, B, and C adducts depending on the experimental conditions[31]
.
oxide radical anion and hydrox-
yl radicals are involved in sever-
al human pathologies has
stimulated the development of
new selective spin-traps for the
detection of these transient re-
active
intermediates.[24]
DEPMPO, developed by Tordo
and co-workers,[25] was likely to
be ideally suited to probing the
formation of all three types of
radicals, that is, alkyl, alkoxyl,
Entry T [K]
AMe
BMe
CMe
AEt
BEt
CEt
ABu
BBu
CBu
G
E
G
E
G
R
G
E
ACHTUNGTRENNUNG
1
2
3
4
183[a]
213[a]
273[a]
290[b]
15 (15:–)
61 (61:–)
45 (42:3)
67 (67:–)
N
E
R
(1:–)
3
(3:–)
4
ACHTUNGTRENNUNG(32:–)
and
alkylperoxyl
radicals,
27 (27:–)
44 (44:–)
51 (46:5)
42 (42:–)
37 (37:–)
44 (43:1)
55 (49:6)
57 (54:3)
61 (55:6)
73 (61:12)
66 (66:–)
73 (67:6)
24
30
17
G
N
N
N
(24:–)
during the dialkylzinc oxidation
process.
8
(19:–)
5
(5:–)
–
16
(12:4)
(30:–)
EPR experiments were per-
formed with Me2Zn, Et2Zn, and
nBu2Zn in the presence of
DEPMPO used either in a 5:1
or 1:5 ratio with respect to the
dialkylzinc. The relative ratios
of the three types of nitroxyl
radicals (A, B, and C), deter-
10 (8:2)
G
E
(28:–)
–
9
ACHTUNGTRENNUNG
5
6
290[c,d] 100 (100:–)
290[c,e,f]
23 (15:8)
–
–
37
(27:10)
54 (49:5) 40 (38:2)
35 (31:4) 49 (49:-)
–
–
56 (45:11)
32
(32:–)
A
G
ACHTUNGTNER(NUNG 9:–)
[a] DEPMPO/R2Zn (5:1), the samples prepared at 183 K were warmed up to 273 K within 1 h. [b] DEPMPO/
R2Zn (5:1), the samples were prepared at 290 K and the spectra were directly recorded. [c] DEPMPO/R2Zn
(1:5). [d] In addition, 6% of radical D was detected when using nBu2Zn. [e] The same samples as in entry 5,
but after air bubbling. [f] In addition, 5, 22, and 16% of radical D were detected when using Me2Zn, Et2Zn,
mined from the simulated spec- and nBu2Zn, respectively.
1588
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1586 – 1595

Reaction of Dialkylzincs with Oxygen
FULL PAPER
reported for adducts B. The aP
values are similarly high.
Values observed for aHb are low,
but greater than those of ad-
ducts B. This is in agreement
with a preferential pseudo-axial
mean position of the alkylper-
oxy groups.
In summary, adducts A were
easily differentiated from both
adducts B and C on the basis of
their aN values. Furthermore,
the aHb constants are character-
istic of the patterns observed in
the spectra of adducts B and C.
According to the literature, for
all three types of radicals, aHg
ranged between 0–0.2 mT. Due
to the overlapping of several
species, the linewidth greater
than 0.2 mT did not allow these
coupling constants to be deter-
mined.
Figure 4. Trapping of Me· (D), MeO· (&), and MeOO· (*) by DEPMPO during the oxidation of Me2Zn
(Me2Zn/DEPMPO, 1:5). a) Experimental EPR spectra recorded at 183 and 213 K. b) Simulated spectra.
However, the relative ratio of
spin-adducts A, B, and C does not reflect the actual relative
concentrations of the trapped radicals because the rate of
trapping depends on the structure of the radical.[32] The
quantitative data reported in Table 1 give reliable compara-
tive information[33] because the stability of the alkyl and al-
koxyl radical adducts of DEPMPO are amongst the highest
known.[25a,30,34] Similarly, the alkylperoxyl radical adducts of
this particular trap have been reported to be stable for sev-
eral minutes at room temperature.[25a]
Table 2. Hyperfine coupling constants for the alkyl, alkoxyl, and alkyl-
peroxyl radical adducts of DEPMPO at 290 K.[a]
Diastereomeric Adducts
aN [mT]
aP [mT]
aHb [mT]
M
1.41
1.47
1.26
1.26
1.21
--
1.41
--
1.28
1.26
1.22
--
1.41
1.43
1.28
1.24
1.22
--
4.69
5.09
4.61
4.32
4.82
--
4.77
--
4.63
4.19
4.84
--
4.74
5.23
4.63
4.20
4.87
--
2.04
1.88
0.74
0.29
0.95
--
2.04
--
0.70
0.47
0.99
--
2.03
1.73
0.68
0.43
0.99
--
AMe
BMe
CMe
AEt
BEt
m
M
m
M
m
M
m
M
m
M
m
M
m
M
m
M
m
The data reported in Table 1 reflect the complex interplay
of the elementary steps proposed in Scheme 1.
The reactions performed in the presence of a 5:1 excess
of the spin-trap with respect to the dialkylzincs (footnote a,
entries 1–3) showed, in all cases, a very high ratio of trapped
oxygen-centered species (adducts B+C) to trapped alkyl
radicals (adducts A). At 183 K, this ratio was 85:15, 96:4,
and 99:1 in the cases of Me2Zn, Et2Zn, and nBu2Zn, respec-
tively. Because the rate of trapping of nucleophilic alkyl rad-
icals[35] is much higher than the rate of trapping of electro-
philic oxygen-centered radicals, these ratios are consistent
with the fact that the diffusion-limited step b[36] is much
faster than step l (Scheme 1). Furthermore, the radical addi-
tion to the hindered C=N bond of DEPMPO should be sen-
sitive to steric effects. The trapping of the methyl radical,
the least bulky of any primary alkyl radical, should be slight-
ly faster than the trapping of the latter. According to the rel-
ative amounts of detected adducts A, the rate constants for
the trapping by DEPMPO should decrease slightly in the
order Me>Et>nBu.[32]
CEt
ABu
BBu
CBu
[a] Experiments performed with an excess of the spin-trap (DEPMPO/
R2Zn=5:1).
which again is consistent with the diethoxyphosphoryl group
lying in a pseudo-axial mean position in both stereoisomers.
Note that aP trans is superior to aP cis. The values of the aHb
constants are low, which is in agreement with a preferential
pseudo-axial mean position of the alkoxy groups in both dia-
stereomeric adducts. This is likely to be the consequence of
the anomeric effect. A rather significant difference is ob-
served for the aHb values in the trans and cis adducts.
The second observation is that the ratio of trapped
oxygen-centered species to trapped alkyl radicals remains
quasi-unchanged on raising the temperature from 183 to
273 K in the case of Et2Zn and nBu2Zn (92:8 and 96:4 at
Finally, with regard to the alkylperoxyl radical adducts
CR,[25a,29] the aN values for the adducts C are similar to those
Chem. Eur. J. 2011, 17, 1586 – 1595
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1589

L. Feray, M. P. Bertrand et al.
trapping by DEPMPO increas-
es with temperature. Simultane-
ously, as the reaction evolves,
the amount of O2 is reduced.
All these parameters contribute
to an increase in the relative
amount of trapped methyl radi-
cal. The EPR spectra recorded
at 290 K clearly confirms this
rationale (footnote b, entry 4).
The second series of data
concerns spin-trapping experi-
ments in which R2Zn was used
in 5:1 excess with respect to
DEPMPO (footnote c, entries 5
and 6), that is, experimental
conditions similar to the condi-
tions generally used on a prepa-
rative scale when R2Zn is em-
ployed in large excess with re-
spect to the radical acceptor.[10]
At 290 K (entry 5), owing to
the smaller amount of oxygen
with respect to Me2Zn in excess
and the very slow oxidation of
the latter, only the methyl radi-
cal adduct was detected.
Figure 5. Trapping of Et· (D), EtO· (&), and EtOO· (*) by DEPMPO during the oxidation of Et2Zn (Et2Zn/
DEPMPO, 1:5). a) Experimental EPR spectra recorded at 183 and 213 K. (b) Simulated spectra.
In the case of Et2Zn, the ad-
ducts of Et·, EtO·, and EtOO·
were not detected. Thus, no
signal at all was detected in the
case of diethylzinc as if all ni-
troxyl radicals (A, B, and C)
formed in the medium were
transformed into spectroscopi-
cally “mute” nonparamagnetic
species through homolytic sub-
stitution at diethylzinc present
in excess (steps o, p, and q,
Scheme 1). These reactions
should be extremely slow in the
Figure 6. Trapping of nBu· (D), nBuO· (&), and nBuOO· (*) by DEPMPO during the oxidation of nBu2Zn case of dimethylzinc. The be-
(nBu2Zn/DEPMPO, 1:5). a) Experimental EPR spectra recorded at 183 and 213 K. b) Simulated spectra.
havior of nBu2Zn lies in be-
tween; a 54:40 ratio of trapped
alkyl radicals to trapped alkoxyl
273 K, respectively). This confirmed the stability of adducts
A, B, and even C, which led to negligible decay between 183
and 273 K.
radicals was detected.[37] This is in agreement with the rate
constant for the homolytic substitution at zinc by oxygen-
centered radicals varying in the order Et2Zn@nBu2Zn>
Me2Zn.
When the very same samples prepared under conditions c
were then submitted to air-bubbling at the same tempera-
ture (entry 6), a ratio of 23:56:16 of adducts A, B, and C
was detected with dimethylzinc, whereas this ratio was
32:37:9 with diethylzinc and 34:49:0 for dibutylzinc.
In all cases (entry 6) a new nitroxide appeared, that is,
radical D, which might result from the dialkylzincs behaving
as nucleophiles. The formation of D was first reported for
For comparison, this ratio decreased from 85:15 to 56:44
in the case of dimethylzinc (entries 1–3). This can be ration-
alized by assuming that the amount of oxygen present in the
reaction medium instantaneously consumes Et2Zn and
nBu2Zn. Conversely, in the case of Me2Zn, the oxidation
process being much slower, methyl radicals are continuously
produced when the medium is warmed. The rate for the
trapping of alkyl radicals by oxygen does not vary signifi-
cantly with temperature. In contrast, the rate of alkyl radical
1590
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1586 – 1595

Reaction of Dialkylzincs with Oxygen
FULL PAPER
bled through the reaction
medium, unconsumed R2Zn
starts to be oxidized and all
types of radicals are again pro-
duced in the medium (entry 6).
The latter can therefore be
trapped by any unreacted ni-
trone.
We assume that, in the case
of
diethylzinc,
was
all
the
DEPMPO
consumed
before starting air-bubbling
(entry 5). The A, B, and C radi-
cals detected in entry 6 origi-
nate from the “mute” species
that undergo homolytic substi-
tution at zinc by oxygen-cen-
tered radicals (steps r, s, and t).
Conversely, in the case of dime-
thylzinc, we speculate that,
Scheme 1. Mechanism for the reactions of dialkylzincs with oxygen.
owing to the slow rate of dime-
thylzinc oxidation, DEPMPO
the addition of methylmagnesium iodide to DEPMPO.[38]
Similarly, the nucleophilic addition of R2Zn would lead to F
(Scheme 2). A phosphorus-centered anionic species would
was not completely consumed before air-bubbling. Thus, ad-
dition of a large quantity of oxygen led to the slow con-
sumption of unreacted dimethylzinc, giving rise to the trap-
ping of all three radical species formed under these condi-
tions, that is, methyl, methoxyl, and methylperoxyl species.
The behavior of dibutylzinc lies in between because the ini-
tial ratio of (B+C)/A changes from 40:54 to 49:35, that is, a
final ratio very close to that observed for diethylzinc
(46:32).
Note that, after air-bubbling, the relative ratio of trapped
oxygen species to trapped alkyl radicals is higher in the case
of dimethylzinc than in any other case. It is close to 72:23.
This means that, when using dialkylzinc in excess (which is
the general protocol used for synthetic applications), more
oxygen-centered radicals are consumed by diethyl- and di-
butylzinc than by dimethylzinc and that a greater quantity
of oxygen-centered radicals is present in the reaction
medium when dimethylzinc is involved. This is likely to
have consequences for the reactivity of these species with
respect to their aptitude to abstract hydrogen atoms from
hydrogen atom donors such as ethers, acetals, or alkanes
used as solvents.[10b]
Scheme 2. Rationale for the formation of radical D.
result from the polar fragmentation of intermediate F and
the subsequent nucleophilic addition of this anion to
DEPMPO followed by oxidation would lead to D. Alterna-
tively, any stable nitroxide of types A–C might react further
with dialkylzinc to give F and consequently D through the
polar fragmentation step. The relative amount of D detect-
ed, which depends on the nature of R (5, 16, and 22% for
Me2Zn, nBu2Zn, and Et2Zn, respectively),[39] is in favor of
the second proposal. Radical D had never previously been
observed in any classical spin-trapping experiments involv-
ing DEPMPO except at high temperature.[40]
A series of qualitative experiments were performed with
tetramethylpiperidine N-oxyl (TEMPO) to support the as-
sumptions described above on the relative reactivity of dial-
AHCTUNGTREGkNNUN ylzincs depending on the nature of R. At room tempera-
ture (290 K), the signal of TEMPO instantaneously disap-
peared irrespective of its ratio with diethylzinc (1:50, 1:25,
1:5, and 1:1). This is consistent with the reactions shown in
Scheme 3. TEMPO can be consumed both by homolytic
substitution (SH2) at Et2Zn and by the trapping of the result-
ing ethyl radicals.[41] Homolytic substitution at zinc by
TEMPO, that is, path ii, competes with the direct trapping
of R· by the nitroxide (path i).[42] After the complete con-
sumption of TEMPO, when air was bubbled through the so-
The evolution of the EPR spectra (entries 5 and 6,
Table 1) can be explained as follows. When oxygen is bub-
Chem. Eur. J. 2011, 17, 1586 – 1595
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1591

L. Feray, M. P. Bertrand et al.
merous potentially reactive substrates in the medium this
coupling pathway should not be the main source of alkoxyl
radicals.
In addition, the involvement of the spontaneous homoly-
[15d]
´
sis of the peroxidic bond proposed by Lewinsky et al.
cannot be ruled out (step k). However, compared with the
steps previously envisaged, it should be endothermic.
In EPR experiments, alkoxyl radicals may also result
from the decomposition of adducts C.[23b]
Scheme 3. Reaction of R2Zn with TEMPO.
With regard to the possibility of a nucleophilic addition of
ROZnX species to DEPMPO, it must be argued that at
least Et2Zn and nBu2Zn should behave similarly, which is
not the case in the experiments described in Table 1,
entry 5. Furthermore, the synthesis of authentic samples of
the adduct resulting from the addition of MeO· to
DEPMPO from MeOH necessitates the presence of an oxi-
lution, the signal of TEMPO was again detected, which con-
firmed that oxygen-centered radicals are able to displace ni-
troxyl radicals from the alkoxyaminoalkylzinc that would
act temporarily as “mute” species (path iii).
Conversely, in the presence of dimethylzinc, whatever the
ratio (even in a 50:1 excess of dimethylzinc) the signal of
TEMPO was persistent. It slightly decayed overnight due to
the slow formation of methyl radicals and their subsequent
coupling to TEMPO. However, no increase of the nitroxide
signal could be observed when oxygen was afterwards bub-
bled through the solution. This suggests that SH2 at dime-
thylzinc by TEMPO should be an extremely slow process.
These comments also apply to experiments carried out on
the laboratory scale, in which the presence of TEMPO
could be qualitatively monitored by the intensity of the
characteristic red color of the solution.[43]
In summary, we propose the following pathways as the
major elementary steps involved in the dialkylzinc oxidation
mechanism (Scheme 1). Step a leads to the formation of R·,
which reacts with oxygen to give ROO· (step b). The reac-
tion of the alkylperoxyl radical with R2Zn leads to the mo-
noperoxidic ZnII species that reacts with ROO· to give
(ROO)2Zn (only when R=Et, nBu; step e).
ACTHNUTRGNEUNG
dant like FeIII or Pb(OAc)4.[30,44]
Computational data were obtained with a view to sup-
porting experimental observations through the calculation
of bond dissociation enthalpies. The bond dissociation ener-
ꢀ
gies (BDEs) of O O bonds in RZnOOR species were calcu-
lated at the UB3LYP/6-311++G
A
AHCTUNGTREGUN(NN d,p) level of theory. Values of 151.8, 150.4, and
147.6 kJmolꢀ1 at 298 K were obtained for MeZnOOMe,
EtZnOOEt, and nBuZnOOnBu, respectively. This means
that the spontaneous homolysis of these species at room
temperature would have rate constants ranging from 2.5ꢃ
10ꢀ12 to 1.3ꢃ10ꢀ11 sꢀ1 (assuming that the value of the pre-ex-
ponential factor A[45] is 1015 sꢀ1) and therefore is an unlikely
ꢀ
source of alkoxyl radicals. Note that the O O BDE is lower
for EtOOZnOOEt (144.4 kJmolꢀ1 at 298 K) than for EtZ-
nOOEt but its spontaneous cleavage remains unlikely.
ꢀ
There is little data in the literature relative to the C Zn
Several pathways can be envisaged for the formation of
alkoxyl radicals. They may be formed by the reaction of R·
to give the zinc monoalkoxide by homolytic substitution at
the peroxidic oxygen (step d). Similarly, (ROO)2Zn may
react with R· to form the mixed alkoxyperoxyzinc species
according to step f. Both steps d and f generate RO· that
may further react with R2Zn to form again the zinc mono-
alkoxide species (step g). Reaction of RZnOR (R=Et,
nBu) with RO· then leads to the zinc dialkoxide (step h),
whereas its reaction with ROO· gives rise to mixed alkoxy-
peroxyzinc species (step i). The latter would therefore be a
possible source of zinc dialkoxide by reaction with R·
(step j). Note that on a laboratory scale the oxidation pro-
cess can be selectively driven towards the formation of hy-
droperoxides when oxygen is used in excess with respect to
dialkylzincs. This clearly implies that if any of the reactions
involving homolytic substitution at the peroxidic oxygen
atom, that is, steps d, j, and f, are effectively involved, they
must be slow compared not only to step b, which is obvious,
but also to step e.
bond strength in dialkylzincs[39,46] and no data are available
ꢀ
for the C Zn bond strength in mixed compounds like
RZnOR or RZnOOR. DFT calculations using hybrid meth-
ods were performed. Geometry optimization was conducted
without any constraint by using the UB3LYP/6-31G
method and single-point energies were calculated at the
UB3LYP/6-311++G(3df,3pd) level of theory. Zero-point
energies and temperature corrections were calculated using
the specified scaling factor (0.9923). The calculated values
are given in Table 3.
ACHTUNGTRENNUNG(d,p)
AHCTUNGTRENNUNG
These data, similar to the most recommended experimen-
tal values given for dialkylzincs,[46c] are consistent with the
ꢀ
Table 3. Bond dissociation enthalpies calculated at 298 K for the C Zn
and Zn O bonds in R2Zn, RZnOOR, RZnOR, and (ROO)2Zn deriva-
tives at the UB3LYP/6-311++G
ꢀ
N
ACHTUNGTREN(NUNG d,p) level of
theory.
]
R
R2Zn
Exp[46c]
RZnOOR
Calcd
RZnOR
Calcd
ACHTUNGTRENNUNG(ROO)2Zn
Calcd
Calcd
An alternative pathway, often encountered, is the possibil-
ity of forming tetraoxides through the dimerization of alkyl-
peroxyl radicals that would then decompose to release al-
koxyl radicals and oxygen.[32a,b] Due to the presence of nu-
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
C Zn
C Zn
278.2
249.9
247.1
O Zn
255.3
--
C Zn
O Zn
228.6
--
Me
nBu
Et
275.4
241.0
237.9
266.5ꢁ6.3
285.0
256.0
251.6
--
219.1ꢁ8.4
249.6
223.6
1592
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1586 – 1595

Reaction of Dialkylzincs with Oxygen
FULL PAPER
ꢀ
fact that the reactivity depends on the nature of R. The C
Zn BDE increases slightly on replacing the ethyl group by
[1] D. Seyferth, Organometallics 2001, 20, 2940–2955, and references
therein.
[2] C. Walling, S. A. Buckler, J. Am. Chem. Soc. 1955, 77, 6032–6038.
[3] H. Hock, F. Ernst, Chem. Ber. 1959, 92, 2716–2723.
ꢀ
the n-butyl group. The C Zn bond is stronger by
37.5 kJmolꢀ1 in Me2Zn than in Et2Zn.
[4] a) S. Czernecki, C. Georgoulis, B. Gross, C. Prevost, C. R. Acad. Sc.
(C) 1968, 266, 1617–1619; b) S. Czernecki, C. Georgoulis, E. Michel,
J. Organomet. Chem. 1977, 140, 127–131.
[5] M. H. Abraham, J. Chem. Soc. 1960, 4130–4135.
[6] C. H. Bamford, D. M. Newitt, J. Chem. Soc. Abstr. 1946, 688–695,
and references therein.
Replacing an alkyl group by an alkylperoxy group in-
ꢀ
creases the polarity of the C Zn bond and therefore the cor-
responding BDE. This effect is even more pronounced for
alkoxy than for alkylperoxy derivatives.[47]
SH2 reactions are sensitive to enthalpic factors. To make a
[7] S. Jana, R. J. Berger, R. Frçhlich, T. Pape, N. W. Mitzel, Inorg.
Chem. 2007, 46, 4293–4297.
[8] a) A. G. Davies, B. P. Roberts, J. Chem. Soc. B 1968, 1074–1078;
b) A. G. Davies, D. Griller, B. P. Roberts, J. Chem. Soc. B 1971,
1823–1829; c) A. G. Davies, B. P. Roberts, Acc. Chem. Res. 1972, 5,
387–392.
[9] a) A. G. Davies, K. U. Ingold, B. P. Roberts, R. Tudor, J. Chem. Soc.
B 1971, 698–712; b) P. J. Krusic, J. K. Kochi, J. Am. Chem. Soc.
1969, 91, 3942–3944; c) P. J. Krusic, J. K. Kochi, J. Am. Chem. Soc.
1969, 91, 3944–3946.
[10] For reviews, see: a) S. Bazin, L. Feray, M. P. Bertrand, Chimia 2006,
60, 260–265; b) T. Akindele, K.-I. Yamada, K. Tomioka, Acc. Chem.
Res. 2009, 42, 345–355.
ꢀ
complete analysis of enthalpic factors, the Zn O bond disso-
ciation energies were also calculated for the ethyl and
methyl derivatives. The results are given in Table 3. Because
ꢀ
the strength of the O Zn bond varies only slightly with the
nature of the R group in the peroxidic species, the rate of
the SH2 reaction is likely to be influenced mainly by the
ꢀ
strength of the C Zn bond.
All these data converge to account for the greater difficul-
ty of displacing one of the two methyl groups from
dimethylACHTUNGTRENNUNGzinc than one of the two primary alkyl groups from
either diethyl- or dibutylzinc. They also account for the
quasi-impossibility of displacing the second methyl group of
either MeZnOOMe or MeZnOMe.
[11] a) M. P. Bertrand, L. Feray, R. Nouguier, P. Perfetti, Synlett 1999,
1148–1150; b) M. P. Bertrand, L. Feray, R. Nouguier, P. Perfetti, J.
Org. Chem. 1999, 64, 9189–9193; c) M. P. Bertrand, S. Coantic, L.
Feray, R. Nouguier, P. Perfetti, Tetrahedron 2000, 56, 3951–3961;
d) S. Bazin, L. Feray, D. Siri, J.-V. Naubron, M. P. Bertrand, Chem.
Commun. 2002, 2506–2507; e) F. Cougnon, S. Bazin, L. Feray, M. P.
Bertrand, Tetrahedron 2007, 63, 11959–11964; f) S. Bazin, L. Feray,
N. Vanthuyne, D. Siri, M. P. Bertrand, Tetrahedron 2007, 63, 77–95;
g) L. Feray, M. P. Bertrand, Eur. J. Org. Chem. 2008, 3164–3170.
[12] T. Cohen, H. Gibney, R. Ivanov, E. A.-H. Yeh, I. Marek, D. P.
Curran, J. Am. Chem. Soc. 2007, 129, 15405–15409.
[13] a) F. Denes, F. Chemla, J. F. Normant, Angew. Chem. 2003, 115,
4177–4180; Angew. Chem. Int. Ed. 2003, 42, 4043–4046; b) F.
Denes, S. Cutri, A. Perez-Luna, F. Chemla, Chem. Eur. J. 2006, 12,
6506–6513; c) S. Giboulot, A. Pꢀrez-Luna, C. Botuha, F. Ferreira, F.
Chemla, Tetrahedron Lett. 2008, 49, 3963–3966; d) A. Perez-Luna,
C. Botuha, F. Ferreira, F. Chemla, Chem. Eur. J. 2008, 14, 8784–
8788; e) K.-I. Yamada, Y. Yamamoto, M. Maekawa, T. Akindele, H.
Umeki, K. Tomioka, Org. Lett. 2006, 8, 87–89; f) T. Akindele, Y.
Mamamoto, M. Maekawa, H. Umeki, K.-I. Yamada, K. Tomioka,
Org. Lett. 2006, 8, 5729–5732, and references therein; g) I. Ryu, F.
Araki, S. Minakata, M. Komatsu, Tetrahedron Lett. 1998, 39, 6335–
6336.
[14] For the proposal of the dual mechanism depending on the nature of
the R group, also see the pioneering work of Van Koten and co-
workers: a) J. M. Klerks, J. T. B. H. Jastrzebski, G. Van Koten, K.
Vrieze, J. Organomet. Chem. 1982, 224, 107–119; b) G. Van Koten,
J. T. B. H. Jastrzebski, K. Vrieze, J. Organomet. Chem. 1983, 250,
49–61; c) M. Kaupp, H. Stoll, H. Preuss, W. Kaim, T. Stahl, G.
Van Koten, E. Wissing, W. J. J. Smeets, A. L. Spek, J. Am. Chem.
Soc. 1991, 113, 5606–5618; d) E. Wissing, R. W. A. Havenith, J.
Boersma, G. Van Koten, Tetrahedron Lett. 1992, 33, 7933–7936;
e) E. Wissing, M. Kaupp, J. Boersma, A. L. Spek, G. Van Koten, Or-
ganometallics 1994, 13, 2349–2356; f) E. Wissing, S. Van der Linden,
E. Rijnberg, J. Boersma, W. J. J. Smeets, A. L. Spek, G. Van Koten,
Organometallics 1994, 13, 2602–2608; g) E. Wissing, E. Rijnberg,
P. A. Van der Schaaf, K. Van Gorp, J. Boersma, G. Van Koten, Orga-
nometallics 1994, 13, 2609–2615.
Conclusion
The use of DEPMPO as a radical trap has enabled the for-
mation of all three types of radicals, that is, alkyl, alkoxyl,
and alkylperoxyl radicals, during the reactions of dialkyl-
zincs with oxygen to be probed for the first time. The rela-
tive concentration of these species is greatly influenced by
ꢀ
the nature of the alkyl group, which controls the C Zn
bond dissociation enthalpy. These data, supported by theo-
retical calculations, clearly demonstrate that diethylzinc and,
to a lesser extent, dibutylzinc are much better traps for
oxygen-centered species than dimethylzinc. The concentra-
tion of oxygen-centered radical adducts of DEPMPO is
much lower for diethyl- and dibutylzinc than for dimethyl-
zinc when dialkylzincs are used in excess, which is typical of
radical reactions performed on a preparative scale. This has
a high impact on the radical reactivity of these organometal-
lic species. The low concentration of oxygen-centered spe-
cies in the reaction medium containing an excess of diethyl-
zinc is probably responsible for the poor ability of the latter
to promote the addition of solvents like THF, ethers, or al-
kanes to unsaturated substrates like imines by hydrogen
atom abstraction, in contrast to dimethylzinc.[48]
Experimental procedures and computational data are in-
cluded in the Supporting Information
´
[15] a) J. Lewinski, Z. Ochal, E. Bojarski, E. Tratkiewicz, I. Justyniak, J.
Lipkowski, Angew. Chem. 2003, 115, 4791–4794; Angew. Chem. Int.
´
Ed. 2003, 42, 4643–4646; b) J. Lewinski, W. Marciniac, J. Lipkowski,
Acknowledgements
´
I. Justyniak, J. Am. Chem. Soc. 2003, 125, 12698–12699; c) J. Lewin-
ski, W. Sliwinski, M. Dranka, I. Justyniak, J. Lipkowski, Angew.
Chem. 2006, 118, 4944–4947; Angew. Chem. Int. Ed. 2006, 45, 4826–
4829; d) J. Lewinski, K. Suwala, M. Kubisiak, Z. Ochal, I. Justyniak,
The authors gratefully acknowledge the ANR (D2R2; ANR-09-JCJC-
0081) for financial support.
´
Chem. Eur. J. 2011, 17, 1586 – 1595
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1593

L. Feray, M. P. Bertrand et al.
J. Lipkowski, Angew. Chem. 2008, 120, 8006–8009; Angew. Chem.
tion of oxygen (c) in various organic solvents, see: I. B. Golovanov,
S. M. Zhenodarova, Russ. J. Gen. Chem. 2005, 75, 1795–1797.
´
Int. Ed. 2008, 47, 7888–7789; e) J. Lewinski, K. Suwala, T. Kaczor-
owski, M. Galezowski, D. T. Gryko, I. Justyniak, J. Lipkowski,
[32] No kinetic data could be referred to for the rate constants of radical
addition to DEPMPO. The rate constants should depend on the nu-
cleophilic character of the added radical. Rate constants as different
as 30 and>103 mꢀ1 sꢀ1 have been reported in the literature for the
´
Chem. Commun. 2009, 215–217; f) J. Lewinski, M. Koscielski, K.
Suwala, I. Justyniak, Angew. Chem. 2009, 121, 7151–7154; Angew.
Chem. Int. Ed. 2009, 48, 7017–7020.
C
[16] a) S. Bazin, PhD Thesis, University of Marseille (France), 2005;
b) M. Rouge, R. Spitz, A. Guyot, J. Polym. Sc. Polym. Lett. Ed.
1974, 12, 407–410; c) M. Rouge, R. Spitz, A. Guyot, J. Chim. Phys.
1975, 72, 611–617; d) D. F. Grishin, V. K. Cherkasov, V. A. Dodo-
nov, Russ. Chem. Bull. 1994, 43, 862–866.
trapping of tBuOO by DMPO, see: a) C. M. Jones, M. J. Burkitt, J.
Chem. Soc. Perkin Trans. 2 2002, 2044–2051 and b) J. D. Honeywill,
B. Mile, J. Chem. Soc. Perkin Trans. 2 2002, 569–575. For compari-
son, rate constants of 1.4ꢃ107 and 1.6ꢃ107 mꢀ1 sꢀ1 were reported for
the trapping of the methyl and ethyl radicals by DMPO, respective-
ly, see: c) H. Taniguchi, K. P. Madden, J. Am. Chem. Soc. 1999, 121,
11875–11879.
[17] E. Mileo, F. Benfatti, P. G. Cozzi, M. Lucarini, Chem. Commun.
2009, 469–470.
[18] For the trapping of the ethyl radical, the hyperfine coupling con-
stants in the nitroxide splitting pattern (aN =1.32 mT, aH =1.75 mT
(methylene protons), and aH =0.08 mT (aromatic protons)) are in
good agreement with those reported in the literature. The anilino
radical is characterized by aN =1.16 mT and aH =0.21 mT (methylene
protons) and aH =0.19 mT (aromatic protons), see: a) S. Terabe, R.
Konaka, J. Chem. Soc. Perkin Trans. 2 1973, 369–374; b) P. G. Me-
karbane, B. J. Tabner, Magn. Reson. Chem. 1998, 36, 826–832;
c) P. G. Mekarbane, B. J. Tabner, Magn. Reson. Chem. 2000, 38,
845–832.
[19] For the trapping of the methyl radical, the hyperfine coupling con-
stants in the nitroxide: aN =1.23 mT, aH =1.23 mT (methylene pro-
tons), and aH =0,08 mT (aromatic protons) are in good agreement
with those reported in the literature, see ref. [18]. The anilino radi-
cal is characterized by aN =1.21 mT, aH =0.24 mT (methyl protons),
and aH =0.18 mT (aromatic protons).
[33] It must be underlined that all experiments reported in Table 1 led to
reproducible data. All samples were prepared according to the pro-
tocol described in the Supporting Information.
[34] K. Stolze, N. Udilova, H. Nohl, Free Radical Biol. Med. 2000, 29,
1005–1014.
[35] For a discussion of the factors controlling the reactivity of the
methyl radical, see: a) M. Wah Wong, A. Pross, L. Radom, J. Am.
Chem. Soc. 1993, 115, 11050–11051; b) M. Wah Wong, A. Pross, L.
Radom, J. Am. Chem. Soc. 1994, 116, 6284–6292; c) T. Zytowski, H.
Fischer, J. Am. Chem. Soc. 1997, 119, 12869–12878; d) R. Arnaud,
N. Bugaud, V. Vetere, V. Barone, J. Am. Chem. Soc. 1998, 120,
5733–5740.
[36] In fluid solutions, the rate constant for the trapping of the alkyl radi-
cal by oxygen is in the range (1.6–4.9)ꢃ109 mꢀ1 sꢀ1, see: a) A. March-
aj, D. G. Kelley, A. Bakac, J. H. Espenson, J. Phys. Chem. 1991, 95,
4440–4441; b) B. Maillard, K. U. Ingold, J. C. Scaiano, J. Am. Chem.
Soc. 1983, 105, 5095–5099.
[20] This is consistent with the value reported in the literature for the
trapping of tert-butoxyl or tert-butylperoxyl radicals in toluene (~
2.5 mT), see ref. [18b].
[37] In the case of dibutylzinc, 6% of radical D was detected. The possi-
ble origin of this nitroxide will be discussed later.
[21] A three-line spectrum exhibiting a coupling constant of aN ꢂ1.5 mT
was observed under the same conditions with diethylzinc. The cou-
pling constant is inconsistent with the trapping of an oxygen-cen-
tered radical species. At the moment we have no rationale for this
experimental observation.
[38] This nitroxide was previously detected during the addition of orga-
nomagnesium compounds to DEPMPO, see: J.-L. Clꢀment, DEA,
Marseille, 1993.
ꢀ
[39] The polarity of the C Zn bond decreases in the order Me2Zn>
Et2Zn, see: A. Haaland, J. C. Gree, A. J. McGrady, A. J. Downs, E.
Gullo, M. J. Lyall, J. Timberlake, A. V. Tutukin, H. V. Volden, K.-A.
Ostby, Dalton Trans. 2003, 4356–4366, and this article (the Support-
ing Information).
[22] For assignments through 17O-labelled radicals, see: S. I. Dikalov,
R. P. Mason, Free Radical Biol. Med. 1999, 27, 864–872.
[23] a) J. A. Howard, J. C. Tait, Can. J. Chem. 1978, 56, 2163–2166;
b) E. G. Janzen, P. H. Krygsman, D. A. Lindsay, D. L. Haire, J. Am.
Chem. Soc. 1990, 112, 8279–8284; c) M. V. Merritt, R. A. Johnson, J.
Am. Chem. Soc. 1977, 99, 3713–3719.
[40] At a high temperature (401 K) in chlorobenzene, the formation of
D from DEPMPO was observed. According to the author (P. G. Me-
karbane, Ph. D. Thesis, Lancaster University (UK), 1998) D would
result from a radical fragmentation process leading to the diethoxy-
phosphoryl radical that would be trapped by DEPMPO.
[41] For the trapping of the benzyl radical from the reaction of benzyl-
zinc iodide with TEMPO, see: T. Nagashima, D. P. Curran, Synlett
1996, 330–332.
[24] a) C. Nsanzumuhire, J.-L. Clement, O. Ouari, H. Karoui, J.-P. Finet,
P. Tordo, Tetrahedron Lett. 2004, 45, 6385–6389, and references
therein; b) K. Stolze, N. Rohr-Udilova, T. Rosenau, A. Hofinger, D.
Kolarich, H. Nohl, Bioorg. Med. Chem. 2006, 14, 3368–3376, and
references therein; c) S. I. Dikalov, R. P. Mason, Free Radical Biol.
ˇ
´
ˇ ´
´
Med. 2001, 30, 187–197; d) G. Bacic, I. Spasojevic, B. Secerov, M.
[42] For the sake of clarity, this reaction was neglected in the previous
discussion.
´
Mojovic, Spectrochim. Acta, Part A, 2008, 69, 1354–1366; e) F. A.
Villamena, J. L. Zweier, Antioxid. Redox Signaling 2004, 6, 619–629.
[25] a) C. Frejaville, H. Karoui, B. Tuccio, F. Le Moigne, M. Culcasi, S.
Pietri, R. Lauricella, P. Tordo, J. Med. Chem. 1995, 38, 258–265;
b) J.-L. Clement, J.-P. Finet, C. Frejaville, P. Tordo, Org. Biomol.
Chem. 2003, 1, 1591–1597.
[26] S. Barbati, PhD Thesis, University of Marseille (France), 1997.
[27] B. R. Knauer, J. J. Napier, J. Am. Chem. Soc. 1976, 98, 4395–4400.
[28] L. Dembrowski, J.-P. Finet, C. Freville, F. Le Moigne, R. Maurin, A.
Mercier, P. Pages, P. Stipa, P. Tordo, Free Radical Res. Commun.
1993, 19, S23–S32.
[43] Commercially available solutions of diethyl-, dimethyl-, and dibutyl-
zinc were mixed with a 0.2m solution of TEMPO in dichlorome-
thane. For a 1.5:1 molar ratio of R2Zn to TEMPO, in the presence
of Et2Zn the color of the solution faded within 10 s (after having left
the solution under an air atmosphere for 48 h TEMPO was quantita-
tively recovered), no decoloration was recorded with dimethylzinc
after 22 days, and nBu2Zn behaved exactly as Me2Zn. For a 5:1
molar ratio of R2Zn to TEMPO, in the presence of Me2Zn no decol-
oration was recorded after 21 days, however, the solution containing
nBu2Zn faded completely after 5 min. Again, the behavior of dibu-
tylzinc was in between, but was much closer to that of diethylzinc
than that of dimethylzinc.
[29] For the conformational analysis of the cis and trans isomers, see
ref. [26].
[30] For the characterization of the methoxyl radical adducts in metha-
nol, see: S. Dikalov, P. Tordo, A. Motten, R. P. Mason, Free Radical
Res. 2003, 37, 705–712.
[44] When ZnCl2 was used as a Lewis acid to promote the addition of
MeOH to DEPMPO in the presence of air, no signal was detected
even after 10 min.
[31] Based on the estimated solubility of oxygen in CH2Cl2, oxygen is
20% molar with respect to dialkylzinc in experiments reported in
entries 1–4. Oxygen is 2% molar in the experiments reported in
entry 5. The solubility of oxygen was evaluated from the mole frac-
[45] P. Molyneux, Tetrahedron 1966, 22, 2929–2943.
[46] a) G. P. Smith, R. Patrick, Int. J. Chem. Kinet. 1983, 15, 167–185;
b) R. Georgiadis, P. B. Armentrout, J. Am. Chem. Soc. 1986, 108,
2119–2126; c) R. L. Jackson, Chem. Phys. Lett. 1989, 163, 315–322.
1594
www.chemeurj.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1586 – 1595

Reaction of Dialkylzincs with Oxygen
FULL PAPER
[47] The calculated NBO charges and their repercussion on the bond
lengths are given in the Supporting Information.
moto, M. Maekawa, K. Tomioka, J. Org. Chem. 2004, 69, 1531–
1534.
[48] a) K.-I. Yamada, H. Fujihara, Y. Yamamoto, Y. Miwa, T. Taga, K.
Tomioka, Org. Lett. 2002, 4, 3509–3511; b) K.-I. Yamada, Y. Yama-
Received: September 10, 2010
Published online: January 5, 2011
Chem. Eur. J. 2011, 17, 1586 – 1595
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1595