Copper-free asymmetric allylic alkylation with a grignard reagent: Design of the ligand and mechanistic studies

Article abstract of DOI:10.1002/chem.201202318

The Cu-free asymmetric allylic alkylation, catalysed by NHC, with Grignard reagents is reported on allyl bromide derivatives with good results. The enantioselectivity was quite homogeneous (around 85 % ee) on large and various substrates, regardless of the nature of the Grignard reagent. The formation of stereogenic quaternary centres was highly regioselective for both aliphatic and aromatic derivatives with good enantiomeric excess (up to 92 % ee). The methodology developed was found to be complementary with the Cu-catalysed version. Several new NHCs were tested with improved efficiency. In addition, mechanistic studies, using NMR spectroscopy, led to the discovery of the catalytically active species. Copyright

Full text of DOI:10.1002/chem.201202318

DOI: 10.1002/chem.201202318
Copper-Free Asymmetric Allylic Alkylation with a Grignard Reagent:
Design of the Ligand and Mechanistic Studies
David Grassi, Chrysanthi Dolka, Olivier Jackowski, and Alexandre Alexakis*[a]
Abstract: The Cu-free asymmetric al-
lylic alkylation, catalysed by NHC,
with Grignard reagents is reported on
allyl bromide derivatives with good re-
sults. The enantioselectivity was quite
homogeneous (around 85% ee) on
large and various substrates, regardless
of the nature of the Grignard reagent.
The formation of stereogenic quaterna-
ry centres was highly regioselective for
both aliphatic and aromatic derivatives
with good enantiomeric excess (up to
92% ee). The methodology developed
was found to be complementary with
the Cu-catalysed version. Several new
NHCs were tested with improved effi-
ciency. In addition, mechanistic studies,
using NMR spectroscopy, led to the
discovery of the catalytically active
species.
Keywords: allylic compounds
·
asymmetric synthesis · Grignard re-
agents · heterocyclic carbenes · qua-
ternary stereocentres
Introduction
The asymmetric allylic alkylation (AAA), using copper and
a chiral ligand belonging to diverse families such as dipep-
tide, phosphite, phosphoramitides, ferrocene-based ligands
and N-heterocyclic carbenes (NHCs), has been reported in
the literature.[1] This powerful tool allowed the delivery of
enantioenriched compounds on a wide range of substrates
with a high level of regioselectivity.[2] The organometallics
that could be used in this transformation were diorganozinc,
triorganoaluminium and Grignard reagents.[2] Concerning
the use of NHC in the asymmetric allylic alkylation, gener-
ally the ligand was C2 symmetric.[3] In contrast, Hoveyda et
al.,[4] and others,[5] have developed a family of non-C2 sym-
metric bidentate ligands bearing a chelating group, which
could be an alcohol moiety to form a copper-alkoxy or sul-
fonate species (Scheme 1).[3]
By using such a family of ligands, Hoveyda and co-work-
ers also reported a copper-free AAA on trisubsituted allylic
chlorides, bearing an ester functionality, by using Grignard
reagents.[6] In general very good enantioselectivity and re-
gioselectivity could be achieved, but cyclopropanation of the
double bond occurred (as a side reaction), and the byprod-
uct was produced in the range of 7 to 26% yield during the
catalysis. In 2009, Hoveyda at al. reported the addition of di-
organozinc and triorganoaluminium to cinnamyl phosphate
derivatives, lacking the ester functionality, using a bidentate
Scheme 1. NHCs used in the copper-free AAA.
ligand with a sulfonate moiety.[6b] Importantly this methodol-
ogy could not be applied to Grignard reagents, regardless of
the nature of the leaving group employed. During the prep-
aration of this full report, Woodward and co-workers, using
a bidendate ligand bearing a sulfonate, in combination with
Grignard reagents, reported a copper-free methodology on
cinnamyl bromide derivatives (Scheme 1).[5b,c] Unfortunately
the scope of substrates, Grignard reagents and nucleophiles
were narrow and good results were accessed in one or two
examples, exemplifying the limit of this catalytic system. In
2010, our group reported a copper-free methodology using
Grignard reagents on a broad scope of substrates and
Grignard nucleophiles.[7] We decided to explore more
deeply the use of Grignard reagents as nucleophile for sev-
eral reasons. First of all, in contrast to diorganozinc re-
agents, a wide range of Grignard reagents could be readily
prepared in the laboratory or are commercially available.
Secondly Grignard reagents could be used to deprotonate
the imidazolium salt in situ and form the active catalyst.[8]
In our initial communication[7] we had discovered that the
NHC ligand used should have some intrinsic properties to
exhibit efficient catalytic activity (Scheme 2 and Scheme 3).
A three-carbon spacer on one arm of the NHC is required
[a] D. Grassi, Dr. C. Dolka, Dr. O. Jackowski, Prof. A. Alexakis
Dꢀpartement de Chimie Organique
Universitꢀ de Genꢁve
30 quai Ernest-Ansermet, CH-1211 Genꢁve-4 (Switzerland)
Fax: (+41)223796522
E-mail: alexandre.alexakis@unige.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202318.
1466
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1466 – 1475

FULL PAPER
gioselectivity was accessed (90% g regioselectivity), albeit
with lower enantiomeric values (33–84% ee). On quaternary
aromatic centres, perfect g regioselectivity and good enan-
tioselectivity (86–89% ee) was achieved in all cases. Un-
fortunately, on quaternary aliphatic derivatives moderate ee
values (64–73% ee) were often obtained, again with a per-
fect regioselectivity. We also proposed the following tenta-
tive mechanism (Scheme 4).
Scheme 2. NHC ligand used in our initial communication.
Scheme 4. Tentative mechanism.
We assumed that the NHC is doubly deprotonated by the
Grignard reagent to form intermediate A. The transmetalla-
tion with a third equivalent of Grignard reagent led to inter-
mediate B. The catalytic active species B forms a pseudo-
chair interaction with the substrate. The substrate is released
and the intermediate A is regenerated and ready for a new
catalytic cycle. It is crucial to mention that the carbene
probably acts as a Lewis base and somehow the intermedi-
ate B has an altered mode of reactivity compared with a
conventional Grignard reagent.[9]
Scheme 3. Ligand L1 in copper-free AAA.[7]
as well as a chelating hydroxyl group (L1 vs. L3, L4). The
mesityl part acts as a blocking group, which is supposed to
be necessary for both regioselectivity and enantioselectivity
(L1 vs. L2). The nature of the counteranion employed has a
negligible effect and the bromide as leaving group was the
best compromise between enantioselectivity and regioselec-
tivity. The solvent played an important role and ether, as
moderately coordinating solvent, was by far the best choice.
With these conditions in hand, the scope of substrates and
nucleophiles were then evaluated.
Results and Discussion
On cinnamyl bromide, good enantiomeric excess (ee)
values were obtained (average of 83% ee) but the regiose-
lectivity remained between 75–85% in favour of the desired
g adduct when primary Grignard reagents were used. The
secondary Grignard reagents, such as iPrMgBr and c-
HexMgBr afforded lower g regioselectivity values around
69% with 82–84% ee. Remarkably, tBuMgBr could be also
introduced with decent results (73% of g adduct, 68% ee), a
result that could not be reached using any copper-catalysed
system.[1] On cinnamyl bromide derivatives the ee values
and regioselectivities were comparable to cinnamyl bromide
yielding averages values of 75–81% g regioselectivity and
85% ee. On aliphatic derivatives, in most cases, a higher re-
Mechanistic studies: Before investigating the ligand design
(Scheme 5) it appeared crucial for us to try and understand
how our ligand could work without copper with such high
induction of enantioselectivity and regioselectivity. The first
Scheme 5. Structural modifications possible on the ligand.
Chem. Eur. J. 2013, 19, 1466 – 1475
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1467

A. Alexakis et al.
hypothesis would be that the ratio of ligand compared to
the metal catalyst (Cu) is not sufficient to ensure complete
coordination to the copper alkoxide. The consequence of
that would be a background reaction of the cuprate, non-
complexed to the ligand, decreasing both the enantioselec-
tivity and regioselectivity. This hypothesis could explain why
the reaction worked with or without copper salts, giving the
same level of efficiency. The reaction was carried out with
twice more ligand than the copper species using Katsukiꢃs
ligand L5. We obtained the same results as with the copper
and ligand in a 1:1 ratio (Table 1, entries 2 and 3). Neverthe-
Scheme 6. Kinetic experiment.
rather puzzling result, as we found in many other examples
that without Cu, the enantioselectivity was much better.[11]
The next question we addressed was which method we
could use to show that a magnesium carbene species was re-
sponsible for the catalytic activity. In the literature, no mag-
nesium carbene species were observed and characterised
with such framework of the ligand. Nevertheless, in the last
decade Hoveyda et al. has characterised some aluminium-,
zinc- and silver-NHC species.[4,6] Firstly, in the 13C NMR
spectra, a chemical shift of the carbon of the imidazolium
salt, from d=160 to around 190–210 ppm, has been ob-
served when the NHC was complexed to a metal. Secondly
Table 1. Effect of ligand/copper ratio.
ꢀ
ꢀ
the Me Al or Et Zn group exhibited a downfield of the
chemical shift of the protons and of the carbons of the alkyl
moiety attached to the metal sulfonate (Al or Zn) when co-
ordinated to the carbene. This observation is consistent with
the s-donating effect of the NHC ligand, causing an increase
in the electron density of the carbon centre.[4]
Entry
Metal
[mol%]
L
Conv
[%]
g/a[a]
ee
G
[mol%]
[%][b]
Having in this information in hand, we carried out a
series of 13C and 1H NMR spectroscopic experiments of
ethyl magnesium bromide with L1 in different solvents.
Indeed, the solvent seemed to be also crucial not only on
the NMR composition of the observed species, but also on
the outcome of the reaction with cinnamyl bromide. There-
fore, we performed an experiment in pure deuterated ben-
zene (50:50 g/a, 67% ee), in diethylether/C6D6 10:1 (77:23 g/
a, 85% ee), in methyl tert-butyl ether (MTBE)/C6D6 10:1
(75:25 g/a, 72% ee) and in CH2Cl2/C6D6 10:1 (10:90 g/a,
40% ee). THF was not investigated as we already knew
from previous studies that it was efficient in this reaction.[7]
Besides these catalytic reactions, the intermediate organo-
metallic species were generated at ꢀ158C, in the required
solvent, and stirred for 10 min at the same temperature. The
mixture was transferred through a syringe into an NMR
tube containing C6D6 and placed under an inert atmosphere.
The 13C and 1H NMR spectra were then immediately record-
ed at room temperature (the formed catalyst is active in a
range of ꢀ308C to RT). By performing the reaction in C6D6
only (Table 2, entry 2 and Scheme 7) we could observe two
species, by 13C NMR, with a minor one at d=210 ppm and a
major one at d=207 ppm. It seems that in deuterated ben-
zene two different Mg carbene species were present in solu-
tion, probably A and B. In addition, all carbons of the NHC
ligand were present, which demonstrated that no decompo-
sition of the ligand had occurred (d=160 ppm for the qua-
ternary carbon next to alkoxy moiety, d=77 and 73 ppm for
the carbons of the diamine backbone, d=48 ppm for the
CH2 linker, d=19, 20, and 21 ppm for the mesityl part).
1
2
3
4
5
–
1 L5
1 L5
2 L5
0.5 L5Ag
1 L5Ag
70
99
99
99
99
70:30
69:31
69:31
67:33
67:33
57
55
55
57
57
1 CuTC[c]
1 CuTC[c]
1 CuCl
1 CuCl
[a] By 1H NMR spectroscopy. [b] By chiral GC. [c] CuTC=Copper thio-
phene carboxylate.
less to have a definitive answer to this point we decided to
form the silver complex using Katsukiꢃs ligand L5Ag to per-
form further transmetallation with copper. After a deep
screening of conditions, we found out that with freshly
formed silver oxide in a degassed mixture of THF/toluene
with powered 3ꢄ molecular sieves we could get the desired
silver carbene L5Ag with a decent yield of 75%. This silver
carbene exhibited a typical signal in the 13C NMR spectrum
at d=205 ppm, which is characteristic of metal–carbene
bond.[4–6]
These silver salts are known to display a dimeric form,
but unfortunately we could not manage to get suitable crys-
tals for an X-ray structure of this silver complex to confirm
this point. Consequently, we decided to transmetallate the
silver species to copper species using both 1:1 and 1:0.5
copper/ligand ratios. The results obtained were exactly the
same as those obtained with a copper-free or a copper/NHC
system (Table 1, entries 4 and 5). Additionally a kinetic
study showed a dramatic acceleration of the reaction when
copper is present (Scheme 6). Interestingly, on quaternary
centres, the copper-free system afforded a higher result than
the copper-catalysed system in terms of enantioselectivity,
but maintaining the same level of regioselectivity. This is a
1
Moreover, the H NMR spectra showed that the proton of
1468
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1466 – 1475

Copper-Free Asymmetric Allylic Alkylation
Table 2. Correlation of results with ratio of C/D.
FULL PAPER
a complex multiplet in the Grignard solution due to aggre-
gated monomers for clean triplet and quadruplet in the
magnesium-carbene species. Similarly,
a shift in the
13C NMR spectra was also observed (d=ꢀ0.88 to
ꢀ1.58 ppm for the CH2 linked to the magnesium, and a shift
from d=13.37 to 13 ppm for the CH3 group). This evidence
strongly suggested an insertion of the magnesium species
both to the alkoxy bond and into the carbene bond to ac-
count for such a downfield shift of the signal.
In MTBE a 1:1 ratio between A and B was observed
(Table 2, entry 3). In methylene chloride no magnesium-car-
bene species were detected, explaining the poor results ob-
tained (Table 2, entry 1).
This NMR study prompted us to draw the following con-
clusions: The copper-free catalysis is working through a
magnesium species inserted both into the alkoxy and into
the carbene bond. The choice of the solvent was crucial: in
diethyl ether, only the catalyti-
Entry
Solvent
Conv
[%][a]
g/a[a]
ee
Ratio
[%][b]
A/B[c]
1
2
3
4
CH2Cl2
Benzene
MTBE
Et2O
47
40
53
99
10:90
50:50
75:25
76:24
40
67
72
85
–
4.77:1
1:1
0:1
[a] By 1H NMR spectroscopy. [b] By chiral GC. [c] By integration of the
13C NMR spectra.
cally active species was present.
In contrast, in deuterated ben-
zene two species were present
and the catalytically active spe-
cies was present but only as
minor one. The difference of ef-
ficiency between the reactions
performed in benzene, MTBE,
methylene chloride, or diethyl
ether could be rationalised.
There is a clear correlation be-
tween the amount of B and the
conversion, the g/a ratio and
the ee value (Table 2). The two
magnesium carbene species
were in equilibrium and, de-
pending on the solvent used,
the equilibrium was shifted (or
not) toward the catalytically
active species. The Grignard re-
agent was also basic enough to
perform a double deprotona-
tion in a single pot, as the prod-
uct of mono-insertion into the
OH bond was not observed.
Scheme 7. 13C NMR spectra of the reaction shown in Table 2, entry 4. D: Silver carbene. L1: Imidazolium salt.
A: Catalyst generated in mixture of Et2O/C6D6 1:10. B: Catalyst generated in mixture of Et2O/C6D6 10:1.
imidazolium and of the phenol were not anymore present
which demonstrated complete insertion of the magnesium
into the alkoxy bond and into the carbene bond.
The intrinsic properties of our bidendate ligand conferred a
dual mode of action with the same efficiency with or with-
out copper but through, probably, different mechanistic
pathways. Thanks to 13C NMR spectroscopy, all the different
intermediates of the postulated mechanism could be fully
characterised as depicted in Scheme 4.
When the reaction is performed in diethyl ether/C6D6, in
a ratio of 10/1 (Table 2, entry 4, and Scheme 7), we could
see only one magnesium carbene species with a chemical
shift of d=210 ppm. The peaks of all the other carbons of
the ligand were similarly observed on the 13C NMR spectra.
We also noticed a shift of the protons between EtMgBr,
alone in ether, and the magnesium carbene species (shift
from d=ꢀ0.1 to ꢀ0.52 ppm for the CH2 linked to the mag-
nesium in the 1H NMR spectra, with a shift from d=
1.70 ppm to 1.25 ppm for the CH3 group). Importantly, the
Ligand design: Next, it appeared reasonable to investigate
more deeply the ligand design to try to increase ideally both
ee value and regioselectivity on the four classes of substrates
(alkyl or aryl substitution, on a di- or trisubstituted double
bond). The solvent, the leaving group and the temperature
were kept unchanged. The analogues of L1 developed in
our laboratory were quite modular and modifications could
1
multiplicity of the signal in the H NMR spectra changed to
Chem. Eur. J. 2013, 19, 1466 – 1475
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1469

A. Alexakis et al.
hydes available, as well as the blocking part, which is de-
rived from aryl bromide derivatives. The ligands were syn-
thesised in a three-step procedure (Scheme 8). The monoal-
kylation step was carried out using Buchwald amination
conditions by using palladium diacetate and the S-Phos
ligand.[11] The coupling product was then reacted with the
corresponding aldehyde in a mixture of THF/MeOH in a
classical reductive amination conditions. The diamine was
then cyclised into the desired imidazolium salt using dry am-
monium chloride and trimethyl orthoformate. The following
ligand library could be synthesised (Scheme 9).
Every ligand was evaluated in our benchmark reaction,
the addition of EtMgBr to cinnamyl bromide. First, we
wanted to test if the cyclohexyl backbone, instead of the di-
phenyl one, would be valuable in terms of results. Thus,
ligand L7 (Table 3, entry 3) was tested, but unfortunately
provided almost a racemic mixture with a level of regiose-
lectivity equal to model NHC ligand L1. The effect of the
chelating group was then evaluated by trying different
P(Ph)2 and sulphonamine groups but keeping the cyclohexyl
backbone the same. The racemic ligands L8 or L9 (Table 3,
entries 4 and 5) gave very poor results in terms of regiose-
lectivity. Being aware of this failure, we returned to the di-
Scheme 8. Synthetic pathway of the ligand of the NHC. S-Phos=2-dicy-
clohexylphosphino-2’,6’-dimethoxybiphenyl.
be organised into four axis (see Scheme 5). The ligand back-
bone could be changed easily with a large panel of frame-
work available from cyclohexyl to an alkyl one. The left arm
could be highly tuneable due to wide number of benzalde-
Scheme 9. Synthesised library of NHCs.
1470
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1466 – 1475

Copper-Free Asymmetric Allylic Alkylation
Table 3. Evaluation of the new catalysts.
FULL PAPER
troduction of bulky group tBu L20 or a naphtyl group L21
(Table 3, entries 16 and 17) turned out to detrimental for
the efficiency of the reaction.
The blocking-aryl part (right part) was then examined. A
switch of the two methyl substituents in the meta position
L22 (Table 3, entry 18) afforded results comparable to
ligand L2, that is, very low g selectivity. We could conclude
from this experiment that at least one substitution in the
ortho position on the blocking part was required to reach
good levels of enantioselectivity and regioselectivity. Until
now the blocking part was always symmetric. We were eager
to know if introduction of dissymmetric substituents on the
blocking part could be valuable to the system.[4,5] With these
results in hand, we could hope that the analogue with a
methyl and ethyl in the ortho positions will surely provide
higher results in term of regioselectivity. Thus, ligand L23
(Table 3, entry 19) was tested and to our delight, the enan-
tioselectivity increased considerably (91%). We could sur-
pass the model ligand L1 in terms of enantioselectivity. The
ethyl/ethyl analogue L24 (Table 3, entry 20) was also tested
and we were pleased to see a noticeable increase of the re-
gioselectivity, while keeping the same range of ee value than
the one afforded by ligand L23. Further increase of the
steric bulk on the blocking part of the ligand by the replace-
ment of the two ethyl groups by two isopropyl groups L27
(Table 3, entry 23) led to a dramatic decrease in the ee
value. The biphenyl ligand L28 (Table 3, entry 24) also gave
disappointingly poor results in terms of regioselectivity,
whereas a modest ee value was obtained. The same trend
was observed for the naphtyl- and anthracyl-substituted li-
gands L30 and L31 (Table 3, entries 26 and 27). The intro-
duction of only one ethyl in ortho L29 (Table 3, entry 25)
gave an ee value and a regioselectivity comparable to L2.
We next attempted to combine the optimal blocking part
with the best aryloxy (left) part with a substitution in the
para position by using Ph and OBn moieties L25 and L26
(Table 3, entries 21 and 22) but this did not afford any val-
uable increase of the regio- or enantioselectivity. Finally, re-
placement of the hydroxyl group by a group CF3 or NO2
group L32, L33 (Table 3, entries 28 and 29) was detrimental
in both terms. In summary, the values of the enantioselectiv-
ity and of the regioselectivity were determined by subtle
changes on the ortho positions of the blocking arm and the
optimal chelating arm turned out to be the simplest one in
terms of substitution pattern and synthetic availability.
We decided to study further the scope of Grignard nu-
Entry
Ligand
Conv
[%][a]
g/a[a]
ee
[%][b]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
L1
L2
L7
L8
99
99
97
97
26
97
98
98
99
99
99
99
99
99
98
99
99
29
99
99
99
99
75
94
99
96
99
99
99
76:24
9:91
85
34
5
–
–
76:24[c]
60:40[c]
56:44[c]
69:31
72:28
78:22
62:38
74:26
100:0
84:16
23:77
11:89
22:78
82:18
66:34
14:86
73:27
77:23
84:16
83:17
67:33
17:83
11:89
19:81
17:83
62:38
73:27
L9
L10
L11
L12
L13
L14
L15
L16
L17
L18
L19
L20
L21
L22
L23
L24
L25
L26
L27
L28
L29
L30
L31
L32
L33
78
81
74
90
84
26
37
23
91
62
85
22
19
91
91
90
87
34
60
60
11
51
67
72
1
[a] By H NMR spectroscopy. [b] By chiral GC. [c] T=ꢀ308C.
phenylethane diamine backbone with the aim of tuning the
electronic properties of the sulphonamine group by the ad-
dition of electron poor/rich aryls or an alkyl group L10, L11,
and L12 (Table 3, entries 6–8). The selectivity obtained were
around 75:25 in favour of the g adduct and the ee value was
slightly lower than the one afforded by model NHC ligand
L1. Logically, we draw the conclusion that the hydroxyl
group was the best among chelating groups tested and
should remain untouched. To finish with the diamine back-
bone, ligand L17 (Table 3, entry 13) derived from the
amino-alcohol pool was also subjected to the model reac-
tion. The results obtained were lower than those obtained
using model carbene L1 showing the necessity of two phenyl
groups in the ligand backbone.
ACHUTNGERNcNUG leoHCATUNGTRENNpUGN hiles on cinnamyl bromide to find out which of the two
We then turned our attention to phenoxy part (left part)
of the imidazolium. The use of a moderate electron-donat-
ing group such as OBn L14 in the para position (Table 3,
entry 10) did not improve things, being equal to L1. In con-
trast, a OMe group in ortho or meta position (L15 and L16)
had a dramatic effect on the enantioselectivity (27 and
37% ee, respectively, Table 3, entries 11 and 12). Surprising-
ly, L15 gave an excellent g regioselectivity. The use of an
electron-withdrawing group, CF3 L18 or NO2 L19 (Table 3,
entries 14 and 15), did not offer any improvement. The in-
selected ligands, L23 or L24, was optimal. Four different
and representative Grignard reagents were then selected.
EtMgBr and n-BuMgBr as primary ones, iPrMgBr as secon-
dary one and MeMgBr, which is known for its lack of selec-
tivity and low enantioselectivity obtained in a copper-cata-
lysed system. The ligand with two ethyl groups L24 dis-
played slight improvement of the ee values and clearly
higher g selectivities in all cases compared with ligands L1
and L23 (Table 4). We concluded that ligand L24 was the
most efficient one.
Chem. Eur. J. 2013, 19, 1466 – 1475
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1471

A. Alexakis et al.
Table 4. Further evaluation of catalysts L1, L23 and L24.
even achieved by using methyl Grignard (Table 5, entry 6).
Other Grignard reagents, including secondary, also displayed
good regioselectivity (9:1) in favour of the g adduct and
good ee values (88–91%) were obtained (Table 5, entries 5
and 7). Ligand L24 proved to be the best ligand, thus the
following studies were focussed on this ligand.
Entry
Ligand
RMgBr
Conv [%][a,c]
g/a[a]
ee
(R=)
[%][b]
1
2
3
4
5
6
7
8
L1
Et
Et
Et
99 (82)
99
99 (80)
99 (88)
99
99 (85)
99 (83)
99
99 (80)
99 (93)
99
76:24
73:27
77:23
83:17
88:12
90:10
77:23
76:24
81:19
69:31
73:27
84:16
85
91
91
70
77
80
85
89
91
82
87
87
We also turned our attention to trisubstituted substrates.
They are easily prepared according to the following way.
The addition of vinyl Grignard to a methyl aryl ketone gave
the tertiary alcohol. Treatment with an aqueous solution of
HBr gave the primary bromide through an SN’ reaction, af-
fording a 4:1 mixture of E and Z isomers. The E/Z mixture
proved difficult to separate, and further studies were per-
formed on this mixture. In Table 6, the number of equiva-
lents of Grignard reagent was screened to study whether or
not the Z isomer reacted and, if it did react, in what quanti-
ty. By keeping the reaction time constant to 45 min, it was
clearly shown that 2.40 equiv of Grignard reagent (Table 6,
entry 2) was detrimental to the enantiomeric excess as 50%
of the Z isomer reacted giving the opposite enantiomer.
Clearly, the E isomer reacted much faster. A smaller excess
of Grignard reagent from 1.4 to 1.8 equiv could be suitable
for the reaction (Table 6, entries 3 and 4). By lowering the
quantity of the Grignard to 1.20 equiv, we observed the total
recovery of the Z isomer (Table 6, entry 5). For other nu-
L23
L24
L1
L23
L24
L1
L23
L24
L1
L23
L24
Me
Me
Me
nBu
nBu
nBu
iPr
iPr
iPr
9
10
11
12
99 (90)
[a] By 1H NMR spectroscopy. [b] By chiral GC. [c] Yield of the adduct
mixture is in parenthesis.
We then investigated the electronic effects on analogues
of cinnamyl bromide by using EtMgBr. A slight improve-
ment of the ee value was noticed with L24, whereas clear
improvements in the regioselectivity were observed
(Table 5, entries 2 and 4). A perfect regioselectivity was
ACHUTNGERNcNUG leoAHCTUNGTRNENpUGN hiles and substrates 1.4 equiv of Grignard reagent will
be used. By using the same nucleophile, namely EtMgBr,
with other substrates we could see that the Z isomer did not
react anymore yielding a good ee value of 88% and a regio-
selectivity of 92:8 (Table 6, entry 9). The rest of Table 6 af-
forded compounds with ee values ranging from 78–85% and
regioselectivity from 78:22 to 100: 0. However, as shown on
the other substrates and Grignards, for quaternary centres
on aromatic substrates, it was difficult to control the amount
of reacted Z isomer because it was both nucleophile- and
substrate-dependent (Table 6, entry 6 vs. entry 8, entry 4 vs.
entry 9). Nevertheless the ee value (78–91%) and regiose-
lectivity (78:22 to 100:0 g/a) obtained were relatively higher
than those obtained using Cu/phosphoramidites system.[2c]
To conclude, it was evident that there was no major im-
provement compared with ligand L1 in this class
of substrate.
Table 5. Variation of the substrate and the Grignard nucleophile.
Entry
Ligand
R1
RMgBr
(R=)
Conv
g/a[a]
ee
[%][a,c]
[%][b]
1
2
3
4
5
6
7
L1
L24
L1
L24
L24
L24
L24
CF3
CF3
Me
Me
Me
Me
CF3
Et
Et
Et
Et
iPr
Me
nBu
99 (80)
99 (65)
99 (87)
99 (75)
99 (72)
99 (70)
99 (68)
77:23
88:12
77:23
85:15
86:14
100:0
88:12
91
92
88
91
90
80
88
[a] By 1H NMR spectroscopy. [b] By chiral GC. [c] Yield of the adduct
mixture is in parenthesis.
Table 6. Evaluation of E/Z mixture of trisubstituted substrates.
In contrast, for both tertiary and quaternary ali-
phatic derivatives, (Table 7), ligand L24 outper-
formed greatly ligand L1 in terms of ee values
(Table 7, entries 2, 4, 6, 8 and 10). The excellent re-
Entry
Ligand
R1
RMgBr
[equiv]
Conv E/Z
g/a[a,c]
ee
gioselectivity observed with L1[7] remained consis-
G
[%][a]
[%][b]
tent except for a couple of examples. Some impres-
sive improvements were noticed with crotyl bro-
mide (Table 7, entries 7 and 8) and hindered tri-
substituted substrates (Table 7, compare entries 5
and 9, with 6 and 10).
1
2
3
4
5
6
7
8
9
L1
H
H
H
H
H
H
H
Me
Me
1.8, R=Et
2.4, R=Et
1.8, R=Et
1.4, R=Et
1.2, R=Et
1.4, R=nBu
1.4, R=Ph
1.4, R=nBu
1.4, R=Et
99:0
99:50
99:0
98:0
92:0
98:2 (70)
97:3
97:3 (65)
97:3
89
76
L24
L24
L24
L24
L24
L24
L24
L24
91
91
91
81
78
85
88
97:3
61:0
100.0 (75)
78:22 (72)
89:11 (75)
92:8 (63)
(CH2)2
99:30
95:10
99:32
[a] By 1H NMR spectroscopy. [b] By chiral GC. [c] Yield of the adduct mixture is in
parenthesis.
1472
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1466 – 1475

Copper-Free Asymmetric Allylic Alkylation
Table 7. Evaluation of alkyl-substituted substrates.
FULL PAPER
Experimental Section
General Procedures: All reactions were conducted under an
inert atmosphere. Unless otherwise stated, all reagents were
purchased from commercial suppliers and used without fur-
ther purification. All solvents employed in the reactions were
dried on alumina columns and degassed prior to use. Organic
solutions were concentrated under reduced pressure on a
Bꢅchi rotary evaporator. 1H (400 MHz) and 13C (100 MHz)
NMR spectra were recorded in CDCl3, and chemical shifts (d)
are given in ppm relative to residual CHCl3. Evolution of re-
action was followed by GC-MS Hewlett Packard (EI mode)
HP6890–5973. An asterisk on the GC-MS indicates the major
peak. Optical rotations were measured at 208C in a 1 cm cell
in the stated solvent; [a]D values are given in 10ꢀ1 degcm2 gꢀ1
Entry
Ligand
R1, R2
RMgBr
(R=)
Conv
[%][a]
g/a[a,c]
ee
[%][b]
1
2
L1
R2 =Cy
R1 =H
Et
99
99
99
99
80
78
99
99
85
86
78:22 (72)
89:11 (80)
95:5 (75)
100:0 (78)
98:2 (65)
98:2 (57)
86:14 (94)
84:16 (75)
98:2 (86)
100:0 (60)
78
83
80
90
73
82
33
63
64
79
L24
L1
R2 =Cy
R1 =H
Et
3
R2 =tBu
R1 =H
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
ACHTUNGTRENNUNG
4
L24
L1
R2 =tBu
R1 =H
ACHTUNGTRENNUNG
5
R2 =tBu
R1 =Me
R2 =tBu
R1 =Me
R2 =Me
R1 =H
ACHTUNGTRENNUNG
(concentration
c
given as g mLꢀ1). Enantiomeric excesses
6
L24
L1
ACHTUNGTRENNUNG
were determined by chiral-GC (capillary column, 10 psi H2).
Temperature programs are described as follows: initial tem-
perature (8C), initial time (min), temperature gradient
(8Cminꢀ1), final temperature (8C); retention time (Rt) are
given in min. All Grignard reagents (except ethyl and methyl
magnesium bromide (Aldrich)) were synthesised in Et2O by
addition of the corresponding bromide onto magnesium.
Flash chromatography were performed using silica gel 32–
63 mm, 60 ꢄ. The syntheses of imidazolium catalysts are de-
scribed in the Supporting Information.
7
ACHTUNGTRENNUNG
8
L24
L1
R2 =Me
R1 =H
ACHTUNGTRENNUNG
9
R2 =Cy
R1 =Me
R2 =Cy
R1 =Me
ACHTUNGTRENNUNG
10
L24
ACHTUNGTRENNUNG
[a] By 1H NMR spectroscopy. [b] By chiral GC. [c] Yield of the adduct mixture is in
parenthesis.
Representative procedure (Table 4, entry 1): In a flame-dried
Schlenk, under an N2 atmosphere, cinnamyl bromide
(0.5 mmol) and L1 (1 mol%) were suspended in dry Et2O
(2 mL) and cooled to ꢀ158C. A solution of EtMgBr (3m) in
Et2O (0.3 mL, 1.8 equiv) was added dropwise over 4 min.
Conclusion
After complete conversion, the mixture was quenched by addition of a
saturated solution of NH4Cl (2 mL) and stirred at RT for 15 min. The
aqueous layer was separated and extracted with Et2O (3ꢆ3 mL). The
combined organic fractions were dried over Na2SO4, filtered and concen-
trated in vacuo. The residue was purified by flash column chromatogra-
phy to offer a mixture of SN2’ and SN2 products.
½aꢁ2D0 =ꢀ30 (c=0.6, CHCl3) for the mixture. The ee value was measured
by chiral SFC (chiralcel OJ column, 2 mLminꢀ1, 200 bar, MeOH, 1%
during 2 min then 2%minꢀ1): Rt =5.56 (R), 6.11 (S) or chiral GC;
1H NMR (300 MHz, CDCl3): d=7.40–7.26 (m, 5H), 6.49 (d, SN2), 6.33
(m, SN2), 6.06 (m, SN2’), 5.13 (m, SN2’), 3.24 (q, J=7.4 Hz, SN2’), 2.29 (q,
J=7.0 Hz, SN2), 1.86 (m, SN2’), 1.59 (m, SN2), 1.07 (t, 3H, J=7.32 Hz,
SN2), 0.98 ppm (t, 3H, J=7.32 Hz, SN2’); 13C NMR (75 MHz, CDCl3): d=
144.4 (SN2’), 142.2 (SN2’), 130.9 (SN2), 129.9 (SN2), 128.4 (SN2’), 128.3
(SN2), 127.6 (SN2), 126.7 (SN2’), 126.1 (SN2’), 125.9 (SN2), 114.0 (SN2’),
51.7 (SN2’), 35.1 (SN2), 28.3 (SN2’), 22.5 (SN2), 13.7 (SN2), 12.1 ppm (SN2’).
In the first part of this report we tried to rationalise the
mechanistic pathway of the copper-free addition of Grignard
reagents in the AAA reaction. The species involved is a
magnesium–carbene species, characterised by NMR spectro-
scopy. Moreover, different magnesium–carbene species
could exist depending on the solvent. Indeed, the key step
in the efficiency of the reaction was the solvent control of
the equilibrium between two magnesium–carbene species,
that is, one active intermediate B and one (probably inac-
tive) intermediate A. Diethyl ether switched the equilibrium
towards the active species, whereas other solvents did not.
All the different intermediates of the mechanistic pathway
have been identified using simple NMR experiments. The
tentative mechanism in our initial communication seems,
therefore, to be viable. This fact proves that the copper-free
AAA works through a different mechanism compared with
the copper-catalysed version. In the second part, the design
modifications (>30 new ligands) of our initial best ligand
L1 allowed us to improve the existing results in three out of
four classes of substrates. The key change was the increase
of the steric bulk from methyl to ethyl groups in L24. The
copper-free allylic alkylation with Grignard reagents was
performed with good results, up to 92% ee. The enantiose-
lectivity was quite homogeneous (around 85% ee) on large
and various substrates, whatever the nature of Grignard re-
agents. The formation of stereogenic quaternary centres was
highly regioselective for both aliphatic and aromatic deriva-
tives with good enantiomeric excess (up to 92% ee). The
methodology developed turned out to be complementary
with the Cu-catalysed version.
(R)-(3-Methyl-pent-4-enyl)benzene: The ee value was measured by chiral
GC (Hydrodex B3P column: 70-0-1-105-0-20-170-5 45 cmsꢀ1) Rt =26.12
(S), 26.87 (R); 1H NMR (300 MHz, CDCl3): d=7.46–7.30 (m, 5H), 5.90
(m, 1H, SN2’), 5.59 (m, 2H, SN2), 5.15 (m, 2H, SN2’), 2.74 (m, 3H), 2.28
(m, 1H), 1.78 (m, 2H), 1.19 ppm (d, J=6.8 Hz, 3H); 13C NMR (75 MHz,
CDCl3): d=144.4 (SN2’), 142.7 (SN2’), 128.4 (SN2’), 128.2 (SN2’), 125.6
(SN2’), 113.0 (SN2’), 38.4 (SN2’), 37.4 (SN2’), 33.6 (SN2’), 20.2 ppm (SN2’);
(R)-(3-tert-Butyl-pent-4-enyl)benzene: ½aꢁ2D0 =0 (c=1.2, CHCl3) for the
mixture. The ee value was measured by chiral GC (Hydrodex B3P
column: 75-0-20-90-80-20-170-5 30 cmsꢀ1
) Rt: 54.51(R), 55.40 (S);
1H NMR (300 MHz, CDCl3): d=7.39–7.25 (m, 5H), 5.71 (m, 1H, SN2’),
5.20 (dd, J=2.2, 10.2 Hz, 1H, SN2’), 5.07 (dd, J=2.2 and 17.0 Hz, 1H,
SN2’), 4.95 (m, 1H, SN2’), 2.76 (m, 1H, SN2’), 2.47 (m, 2H, SN2’), 2.10 (m,
2H, SN2), 1.92 (m, 1H, SN2’), 1.76 (m, 1H, SN2’), 2.76 (m, 1H, SN2’), 1.49
(m, 1H), 0.94 ppm (s, 9H); 13C NMR (75 MHz, CDCl3): d=143.0, 140.1,
128.4, 128.2, 125.5, 116.4, 54.8, 34.5, 32.6, 30.8, 27.7 ppm; HRMS (EI
mode): m/z calculated for C15H22: 202.1721 [M+1]; found: 202.1722.
(R)-(3,3-tert-Butylmethyl-pent-4-enyl)benzene:
½aꢁ2D0 =+48
(c=0.95,
CHCl3). The ee value was measured by chiral GC (Lipodex E column:
1
F60–1 method) Rt: 76.40 (R), 76.78 (S); H NMR (300 MHz, CDCl3): d=
Chem. Eur. J. 2013, 19, 1466 – 1475
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1473

A. Alexakis et al.
7.35–7.21 (m, 5H), 5.93 (dd, J=10.9, 17.5 Hz, 1H), 5.23(dd, J=1.7,
10.9 Hz, 1H), 5.04 (dd, J=1.7 and 17.5 Hz, 1H), 2.49 (m, 2H), 1.74 (m,
2H), 1.09 (s,3H), 0.94 ppm (s, 9H). 13C NMR (75 MHz, CDCl3): 144.1,
143.8, 128.4, 128.3, 125.5, 113.6, 44.6, 37.2, 35.8, 31.5, 25.8, 16.3 ppm;
HRMS (ESI+ mode): m/z calculated for C16H26: 216.1878 [M+1]; found:
216.1880.
2H, SN2), 7.45 (d, J=8.1 Hz, 2H, SN2), 7.32 (d, J=8.1 Hz, 2H, SN2’), 6.40
(m, 2H, SN2), 5.95 (m,1H, SN2’), 5.07 (m, 2H, SN2’), 3.24 (q, J=7.4 Hz,
1H, SN2’), 2.25 (q, J=7.1 and 13.7 Hz, 2H, SN2), 1.78(m, 2H, SN2’), 1.55
(m, 2H, SN2), 0.99 (t, J=7.4 Hz, 3H, SN2), 0.90 ppm (t, J=7.4 Hz, 3H,
SN2); 13C NMR (75 MHz, CDCl3): d=148.6 (SN2’), 141.2 (SN2’), 133.9
(SN2), 128.8 (SN2), 128.0 (SN2’), 125.7 (q, J=3.4 Hz, SN2’), 125.3 (SN2),
(ꢀ)-(R)-3-cyclohexyl-pentene: 1H NMR (300 MHz, CDCl3): d=5.55 (m,
1H, SN2’), 5.36 (m, 2H, SN2), 4.95 (m, 2H, SN2’), 1.70–0.80 ppm (m,
16H); 13C NMR (75 MHz, CDCl3): d=141.4, 114.9, 52.0, 41.5, 31.2, 29.7,
26.7, 24.3, 12.1 ppm; ½aꢁ2D0 =ꢀ7 (c=0.8, CHCl3) for the mixture. The ee
value was measured by chiral GC (Chirasil DEX-CB column: 75–50–20–
170–5 30 cmsꢀ1) Rt: 33.84 (R), 35.11 (S).
125.2 (SN2), 114.9 (SN2’), 51.5 (SN2’), 35.1 (SN2), 28.2 (SN2’), 22.3
13.7 (SN2), 12.0 ppm (SN2’).
ACHTUNGTRENNUNG(SN2),
(ꢀ)-(R)-(3-methyl-3-phenyl)- pentene: ½aꢁ2D0 =ꢀ11 (c=1, CHCl3). The ee
value was measured by chiral GC (Chirasil DEX-CB column: 80-0-1-110-
0-20-170-5 30 cmsꢀ1) Rt: 19.63 (R), 20.27 (S).1H NMR (300 MHz, CDCl3):
d=7.38–7.24 (m, 5H), 6.11 (m, 1H), 5.17 (m, 2H), 1.89 (m, 2H),1.44 (s,
3H), 0.86 ppm (t, J=5.7 Hz, 3H). 13C NMR (100 MHz, CDCl3): d=
147.1, 146.9, 128.0, 126.7, 125.7, 111.7, 44.5, 33.7, 24.3, 8.9 ppm.
(+)-(R)-(3-cyclohexyl-3-methylpent-4-enyl)benzene: ½aꢁ2D0 =+17 (c=1.1,
CHCl3). The ee value was measured on the two diastereoisomer couples
after epoxidation by chiral SFC (chiralcelOJ column, 2 mLminꢀ1
,
(R)-3-(4-methylphenyl)-pentene: The ee value was measured by chiral
200 bar, MeOH, 2% during 2 min then 2%minꢀ1): 7.19 (R,R), 7.99 (S,S),
9.08 (R,S), 9.70 (S,R); 1H NMR (300 MHz, CDCl3): d=7.37–7.22 (m,
5H), 5.85 (dd, J=10.9, 17.6 Hz, 1H), 5.17(dd, J=1.4, 10.9 Hz, 1H), 5.02
(dd, J=1.4, 17.6 Hz, 1H), 2.73 (m, 1H), 2.57 (m,2H), 1.87–1.63 (m, 7H),
1.39–1.13 (m, 5H), 1.07 (s, 3H); 13C NMR (75 MHz, CDCl3): d=146.3,
143.6, 128.3, 127.8, 125.5, 46.4, 42.3, 41.2, 30.7, 27.7, 27.2, 26.8,18.7 ppm;
HRMS (EI mode): m/z calculated for C18H26: 242.2032 [M+1]; found:
242.2035 ppm.
GC (Chirasil-Dex CB column: 70-0-1-115-0-20-170-5 40 cmsꢀ1
) Rt:
27.10(R), 27.64 (S); 1H NMR (300 MHz, CDCl3): d=7.28 (m„ 4H, SN2),
7.18 and 7.14 (d, J=8.3 Hz, 2H, SN2),6.42 (d, J=15.8 Hz, 1H, SN2), 6.23
(m, 1H, SN2), 6.00 (m,1H, SN2’), 5.08 (m, 2H, SN2’), 3.17 (q, J=7.4 Hz,
1H, SN2’), 3.38 (s, 3H), 2.24 (q, J=7.5 Hz, 2H, SN2), 1.78 (m, 2H, SN2’),
1.55 (m, 2H, SN2),1.02 (t, J=7.4 Hz, 3H, SN2), 0.93 ppm (t, J=7.4 Hz,
3H, SN2); 13C NMR (75 MHz, CDCl3): d=142.5 (SN2’), 141.4 (SN2’),
135.5 (SN2’), 129.9 (SN2), 129.7 (SN2),129.1 (SN2’ and SN2), 127.5 (SN2’),
125.8 (SN2), 113.8 (SN2’), 51.3 (SN2’), 35.1 (SN2), 28.3 (SN2’), 22.6 (SN2),
21.0 (SN2’), 13.7 (SN2), 12.2 ppm (SN2’).
(ꢀ)-(R)-3-phenyl-butene: ½aꢁ2D0 =ꢀ3 (c=0.5, CHCl3) for the mixture.ee
was measured by chiral GC (Hydrodex B3P column, 60-10-1-90-20-170-5
45 cmsꢀ1): Rt: 19.49 (R), 20.14 (S); H NMR (300 MHz, CDCl3): d=7.33–
(ꢀ)-(R)-3-(4-trifluromethylphenyl)-pentene: ½aꢁ2D0 =ꢀ32 (c=0.9, CHCl3)
for the mixture. The ee value was measured by chiral GC (Chirasil-Dex
CB column: 110-20-20-170-5 40 cmsꢀ1) Rt: 5.98 (R), 6.36 (S); 1H NMR
(300 MHz, CDCl3): d=7.58 (d, J=8.1 Hz, 2H, SN2’), 7.55 (d, J=8.1 Hz,
2H, SN2), 7.45 (d, J=8.1 Hz, 2H, SN2), 7.32 (d, J=8.1 Hz, 2H, SN2’), 6.40
(m, 2H, SN2), 5.95 (m,1H, SN2’), 5.07 (m, 2H, SN2’), 3.24 (q, J=7.4 Hz,
1H, SN2’), 2.25 (q, J=7.1 and 13.7 Hz, 2H, SN2), 1.78(m, 2H, SN2’), 1.55
(m, 2H, SN2), 0.99 (t, J=7.4 Hz, 3H, SN2), 0.90 ppm (t, J=7.4 Hz, 3H,
SN2); 13C NMR (75 MHz, CDCl3): d=148.6 (SN2’), 141.2 (SN2’), 133.9
(SN2), 128.8 (SN2), 128.0 (SN2’), 125.7 (q, J=3.4 Hz, SN2’), 125.3 (SN2),
1
7.26 (m, 5H), 6.53–6.32 (m, 2H, SN2), 6.10 (m, 1H, SN2’), 5.16 (m, 2H,
SN2’), 3.58 (m, 1H, SN2’), 2.34 (m, 2H, SN2), 1.47 (d, J=7.0 Hz, 3H,
SN2’), 1.21 ppm (t, J=7.5 Hz, 3H, SN2); 13C NMR (75 MHz, CDCl3): d=
145.5 (SN2’), 143.2 (SN2’), 132.6 (SN2), 128.8 (SN2), 128.5 (SN2), 128.4-
A
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
(R)-3-phenyl-4-methyl-pentene: The ee value was measured by chiral GC
(Hydrodex B-3P column, 50-0-1-170-5 45 cmsꢀ1) Rt: 31.95 (R), 33.08 (S);
1H NMR (300 MHz, CDCl3): d=7.48–7.23 (m, 5H), 6.47 (d, J=15.7 Hz,
1H, SN2), 6.33 (m,1H, SN2), 6.09 (m, 1H, SN2’), 5.11 (m, 2H, SN2’), 2.98
(t, J=8.9 Hz, 1H, SN2’), 2.20 (t, J=6.8 Hz, 2H, SN2), 2.04 (m,1H, SN2’),
1.82 (m, 1H, SN2), 1.06 (d, J=5.7 Hz, 3H, SN2), 1.04 and 0.86 ppm (d, J=
6.7 Hz,3H, SN2’); 13C NMR (75 MHz, CDCl3): d=144.3 (SN2’), 141.2
125.2 (SN2), 114.9 (SN2’), 51.5 (SN2’), 35.1 (SN2), 28.2 (SN2’), 22.3
13.7 (SN2), 12.0 ppm (SN2’).
ACHTUNGTRENNUNG(SN2),
(SN2’), 130.8 (SN2), 129.8 (SN2), 128.4 (SN2), 128.3
126.8 (SN2), 125.9 (SN2’ and SN2), 114.9 (SN2’), 58.5 (SN2’), 42.4 (SN2),
32.6(SN2’), 28.6 (SN2), 22.4 (SN2), 21.0 (SN2’), 20.7 ppm (SN2’).
ACHTUGNTRENN(NUG SN2’), 127.9 (SN2’),
Acknowledgements
ACHTUNGTRENNUNG
(ꢀ)-(R)-3-phenyl-heptene: ½aꢁ2D0 =ꢀ32 (c=0.8, CHCl3) for the mixture.
The ee value was measured by chiral GC (chirasil DEX-CB7 column, 90-
45-20-170-5 35 cmsꢀ1): Rt: 23.83 (R), 24.49 (S); 1H NMR (300 MHz,
CDCl3): d=7.46–7.22 (m, 5H), 6.46 (d, J=15.8 Hz, 1H, SN2), 6.31
(m,1H, SN2), 6.02 (m, 1H, SN2’), 5.10 (m, 2H, SN2’), 3.31 (q, J=7.5 Hz,
1H, SN2’), 2.28 (q, J=7.0 Hz, 2H, SN2), 1.57–1.19 (m, 6H, SN2’ and SN2),
0.98 ppm (m, 3H, SN2’ and SN2); 13C NMR (75 MHz, CDCl3): d=144.7
(SN2’), 142.5 (SN2’), 131.2 (SN2), 129.7 (SN2), 128.4 (SN2’ and SN2),127.6
(SN2’), 126.7 (SN2), 126.0 (SN2’), 125.9 (SN2), 113.8 (SN2’), 49.9 (SN2’), 35.1
(SN2’), 29.7 (SN2’),29.1 (SN2), 22.6 (SN2’ and SN2), 14.0 ppm (SN2’ and
SN2).
(ꢀ)-(R)-3-phenyl-pentene: ½aꢁ2D0 =ꢀ30 (c=0.6, CHCl3) for the mixture.
The ee value was measured by chiral SFC (chiralcel OJ column,
2 mLminꢀ1, 200 bar, MeOH, 1% during 2 min, then 2%minꢀ1): Rt: 5.56
(R), 6.11 (S); 1H NMR (300 MHz, CDCl3): d=7.40–7.26 (m, 5H), 6.49
(d, 1H, SN2), 6.33 (m, 1H, SN2), 6.06(m, 1H, SN2’), 5.13 (m, 2H, SN2’),
3.24 (q, J=7.4 Hz, 1H, SN2’), 2.29 (q, J=7.0 Hz, 2H, SN2), 1.86 (m, 2H,
SN2’), 1.59 (m, 2H, SN2), 1.07 (t, J=7.3 Hz, 3H, SN2), 0.98 ppm (t, J=
7.3 Hz, 3H, SN2’); 13C NMR (75 MHz, CDCl3): d=144.4 (SN2’), 142.2
The authors thank the Swiss National Research Foundation (grant
number 200020-126663) and COST action D40 (SER contract number
C07.0097) for financial support, as well as Dr Andreas Zumbuhl for loan
of the Young valve Schlenk glassware and Andrꢀ Pinto for the NMR ex-
periments.
[1] For reviews see: a) A. Alexakis, C. Malan, L. Lea, K. Tissot-Croset,
D. Polet, C. Falciola, Chimia 2006, 60, 124–130; b) C. A. Falciola, A.
Alexakis, Eur. J. Org. Chem. 2008, 3765–3780; c) A. Alexakis, J. E.
Bꢇckvall, N. Krause, O. Pꢈmies, M. Diꢀguez, Chem. Rev. 2008, 108,
2796–2823; d) S. R. Harutyunyan, T. den Hartog, K. Geurts, A. J.
Minnaard, B. L. Feringa, Chem. Rev. 2008, 108, 2824–2852;
e) “Transition Metal Catalyzed Enantioselective Allylic Substitution
in Organic Synthesis”: A. Alexakis in Topics in Organometallic
Chemistry, Vol. 38 (Ed.: U. Kazmaier), Springer, Berlin, 2012.
[2] For recent examples, see: a) C. Falciola, A. Alexakis, Angew. Chem.
2007, 119, 2673–2676; Angew. Chem. Int. Ed. 2007, 46, 2619–2676;
b) A. W. van Zijl, W. Szymanski, F. Lopez, A. J. Minnaard, B. L. Fer-
inga, J. Org. Chem. 2008, 73, 6994–7002; c) C. Falciola, A. Alexakis,
Chem. Eur. J. 2008, 14, 10615–10627; d) N. Yoshikai, K. Miura, E.
Nakamura, Adv. Synth. Catal. 2009, 351, 1014–1018; e) T. den Har-
tog, B. Macia, A. Minnaard, B. L. Feringa, Adv. Synth. Catal. 2010,
352, 999–1013; f) W. Lçlsberg, S. Ye, H.-G. Schmalz, Adv. Synth.
Catal. 2010, 352, 2023–2031; g) F. Giacomina, D. Riat, A. Alexakis,
Org. Lett. 2010, 12, 1156–1159; h) F. Fang, H. Zhang, F. Xie, G.
(SN2’), 130.9 (SN2), 129.9 (SN2), 128.4 (SN2’), 128.3
126.7 (SN2’), 126.1 (SN2’), 125.9 (SN2), 114.0 (SN2’), 51.7 (SN2’), 35.1 (SN2),
28.3(SN2’), 22.5 (SN2), 13.7 (SN2), 12.1 ppm (SN2’).
ACHTUGNTRENN(UGN SN2), 127.6 (SN2),
ACHTUNGTRENNUNG
(ꢀ)-R)-3-(4-trifluromethylphenyl)-pentene: ½aꢁ2D0 =ꢀ32 (c=0.9, CHCl3)
for the mixture. The ee value was measured by chiral GC (Chirasil-Dex
CB column: 110-20-20-170-5 40 cmsꢀ1): Rt: 5.98 (R), 6.36(S); 1H NMR
(300 MHz, CDCl3): d=7.58 (d, J=8.1 Hz, 2H, SN2’), 7.55 (d, J=8.1 Hz,
1474
www.chemeurj.org
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1466 – 1475

Copper-Free Asymmetric Allylic Alkylation
FULL PAPER
Yang, W. Zhang, Tetrahedron 2010, 66, 3593–3598; i) J. F. Teichert,
S. Zhang, A. W. van Zijl, J. W. Slaa, A. J. Minnaard, B. L. Feringa,
Org. Lett. 2010, 12, 4658–4660; j) B. Mao, K. Geurts, M. FaÇanꢉs-
Mastral, A. W. van Zijl, S. P. Fletcher, A. J. Minnaard, B. L. Feringa,
Org. Lett. 2011, 13, 948–951; k) H. Li, A. Alexakis, Angew. Chem.
2012, 124, 1079–1082; Angew. Chem. Int. Ed. 2012, 51, 1055–1058.
[3] a) S. Tominaga, Y. Oi, T. Kato, D. K. An, S. Okamoto, Tetrahedron
Lett. 2004, 45, 5585–5588; b) S. Okamoto, S. Tominaga, N. Saino, K.
Kase, K. Shimoda, J. Organomet. Chem. 2005, 690, 6001–6007; c) H.
Seo, D. Hirsch-Weil, K. A. Abboud, S. Hong, J. Org. Chem. 2008,
73, 1983–1986; d) K. B. Selim, K. Yamada, K. Tomioka, Chem.
Commun. 2008, 5140–5142; e) D. Hirsch-Weil, D. R. Snead, S. Ina-
gaki, H. Seo, K. A. Abboud, S. Hong, Chem. Commun. 2009, 2475–
2477; f) K. B. Selim, Y. Matsumoto, K. Yamada, K. Tomioka,
Angew. Chem. 2009, 121, 8889–8891; Angew. Chem. Int. Ed. 2009,
48, 8733–8735; g) K. B. Selim, H. Nakanishi, Y. Matsumoto, Y. Ya-
mamoto, K. Yamada, K. Tomioka, J. Org. Chem. 2011, 76, 1398–
1408.
[4] a) A. O. Larsen, W. Leu, C. N. Oberhuber, J. E. Campbell, A. H.
Hoveyda, J. Am. Chem. Soc. 2004, 126, 11130–11131; b) J. J. Van
Veldhuizen, J. E. Campbell, R. E. Giudici, A. H. Hoveyda, J. Am.
Chem. Soc. 2005, 127, 6877–6882; c) D. G. Gillingham, A. H. Hov-
eyda, Angew. Chem. 2007, 119, 3934–3938; Angew. Chem. Int. Ed.
2007, 46, 3860–3864; d) Y. Lee, K. Akiyama, D. G. Gillingham,
M. K. Brown, A. H. Hoveyda, J. Am. Chem. Soc. 2008, 130, 446–
447.
[5] a) T. Jennequin, J. Wencel-Delord, D. Rix, J. Daubignard, C. Crꢀvi-
sy, M. Mauduit, Synlett 2010, 1661–1665; b) J. N. Levy, C. M.
Latham, L. Roisin, N. Kandziora, P. Di Fruscia, A. J. P. White, S.
Woodward, M. J. Fuchter, Org. Biomol. Chem. 2012, 10, 512–515;
c) C. M. Latham, A. J. Blake, W. Lewis, M. Lawrence, S. Woodward,
Eur. J. Org. Chem. 2012, 699–707.
[6] a) Y. Lee, A. H. Hoveyda, J. Am. Chem. Soc. 2006, 128, 15604–
15605; b) Y. Lee, B. Li, A. H. Hoveyda, J. Am. Chem. Soc. 2009,
131, 11625–11633.
[7] O. Jackowski, A. Alexaki, Angew. Chem. 2010, 122, 3418–3422;
Angew. Chem. Int. Ed. 2010, 49, 3346–3350.
[8] a) D. Martin, S. Kehrli, M. d’Augustin, H. Clavier, M. Mauduit, A.
Alexakis, J. Am. Chem. Soc. 2006, 128, 8416–8417; b) H. Hꢀnon, M.
Mauduit, A. Alexakis, Angew. Chem. 2008, 120, 9262–9264; Angew.
Chem. Int. Ed. 2008, 47, 9122–9124.
[9] S. E. Denmark, G. L. Beutner, Angew. Chem. 2008, 120, 1584–1663;
Angew. Chem. Int. Ed. 2008, 47, 1560–1638.
[10] D. Grassi, A. Alexakis, Org. Lett. 2012, 14, 1568–1571.
[11] a) B. P. Fors, P. Krattinger, E. Strieter, S. L. Buchwald, Org. Lett.
2008, 10, 3505–3508; b) B. P. Fors, D. A. Watson, M. R. Biscoe, S. L.
Buchwald, J. Am. Chem. Soc. 2008, 130, 13552–13554; c) B. P. Fors,
S. L. Buchwald, J. Am. Chem. Soc. 2010, 132, 15914–15917.
Received: June 29, 2012
Revised: October 29, 2012
Published online: December 6, 2012
Chem. Eur. J. 2013, 19, 1466 – 1475
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1475