Enantioselective iridium-catalyzed allylic arylation

Article abstract of DOI:10.1002/chem.200801879

We describe herein the development of the first iridium-catalyzed allylic substitution using arylzinc nucleophiles. High enantioselectivities were obtained from the reactions, which used commercially available Grignard reagents as the starting materials. This methodology was also shown to be compatible with halogen/metal exchange reactions. Its synthetic potential is demonstrated by its application towards the formal synthesis of (+)-sertraline.

Full text of DOI:10.1002/chem.200801879

FULL PAPER
DOI: 10.1002/chem.200801879
Enantioselective Iridium-Catalyzed Allylic Arylation
Damien Polet, Xavier Rathgeb, Caroline A. Falciola, Jean-Baptiste Langlois,
Samir El Hajjaji, and Alexandre Alexakis*[a]
Abstract: We describe herein the development of the first iridium-catalyzed allylic
substitution using arylzinc nucleophiles. High enantioselectivities were obtained
from the reactions, which used commercially available Grignard reagents as the
starting materials. This methodology was also shown to be compatible with halo-
gen/metal exchange reactions. Its synthetic potential is demonstrated by its appli-
cation towards the formal synthesis of (+)-sertraline.
Keywords: allylic substitution
asymmetric catalysis · chiral ligands ·
cross-coupling · iridium
·
Introduction
enantiopure chiral allylic fluorinated carbonate as the sub-
strate, has been described by Evans.[15] Inspired by this
work, our group recently developed a general and efficient
iridium-catalyzed enantioselective allylic arylation reaction
with moderate regioselectivity and good to high enantiocon-
trol.[16]
We present herein a full account of the development of
this new methodology that has potential in the synthesis of a
number of biologically active compounds (Scheme 2), as
demonstrated by its utilization towards the formal synthesis
of (+)-sertraline.
The past several years have witnessed the development of a
large number of new asymmetric catalytic reactions. Our
group has contributed to this field and has studied, among
others, the copper-catalyzed allylic substitution reaction.[1] In
this reaction, palladium certainly stands as the most exten-
sively studied catalytic system.[2] Palladium, when associated
with chiral ligands, allows the allylation of the so-called soft
or stabilized nucleophiles (pKa <25) with high stereocontrol
and impressive turnover numbers. Despite notable efforts,
the use of hard nucleophiles (pKa >25) with palladium cata-
lysts has given rather limited results,[3] forcing research
groups to turn their attention to other metals, such as
copper. Since the pioneering work of Bꢀckvall and co-work-
ers in 1995,[4] this metal has been used to catalytically trans-
fer alkyl Grignard,[4,5] zinc,[6] and more recently aluminium[7]
reagents with high regio- and enantioselectivities
(Scheme 1). The copper system, however, could not be suc-
cessfully extended to aryl nucleophiles[2b,8] and despite the
work of many research groups using Co,[9] Ni,[10] Ti,[11] Pd,[12]
Rh,[13] or Cu[14] catalysts, there is still no general and effi-
cient system for such a transformation of prochiral allylic
compounds. A regio- and stereospecific version, with an
Results and Discussion
Thanks to the seminal achievements of the groups of Takeu-
chi,[20] Helmchen,[21] and Hartwig,[22] iridium is now being
employed with a large range of nucleophiles in the allylic
substitution reaction.[23] The g-selectivity of iridium allows
the transfer of stabilized nucleophiles, although more bor-
derline nucleophiles have been successfully used, namely,
ketone enolates,[24a] enamines,[24b] and silanolates.[25] Follow-
ing this trend, we started testing hard nucleophiles with the
iridium catalyst [{IrClACTHNUTRGNEUNG(cod)}2] (cod=cycloocta-l,5-diene) and
began our preliminary studies (Table 1) with ligand L1,
which generally affords high regio- and enantioselectivities
in both the Cu- and Ir-catalyzed allylic substitution.[1b,26]
At room temperature, the reaction carried out with four
equivalents of PhMgBr led, unsurprisingly, to the full depro-
tection of substrate 1a (Table 1, entry 1). The use of four
equivalents of PhZnBr, formed in situ, resulted in the de-
sired product with a promising enantiomeric excess (ee) of
2a, although the reaction was sluggish and the yield low
[a] D. Polet, X. Rathgeb, C. A. Falciola, J.-B. Langlois, S. E. Hajjaji,
Prof. A. Alexakis
Department of Organic Chemistry, University of Geneva
30 quai Ernest Ansermet, 1211 Genꢁve 4 (Switzerland)
Fax : (+41)22-379-32-15
E-mail: alexandre.alexakis@chiorg.unige.ch
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801879.
Chem. Eur. J. 2009, 15, 1205 – 1216
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1205

Having found the optimal
conditions for the allylic substi-
tution of 1a, we studied the be-
havior of different leaving
groups in the substitution reac-
tion (Table 2). The carbonate
derivative 1a (Table 2, entry 1),
although kinetically slower than
the chloride derivative 1ab
(Table 2, entry 2), gave a better
regioselectivity. The slightly less
reactive, allylic acetate 1ac
(Table 2, entry 3) required heat-
ing to 408C for the reaction to
proceed, but did afford the best
ee. In this case, however, the
ratio of products 2a/3a and the
conversion of the reaction were
lower than that for 1a. Un-
fortunately, neither a phospho-
nate (1ad, Table 2, entry 4), nor
Scheme 1. Examples of allylic substitution reactions using nonstabilized nucleophiles.
a carbamate (1ae, Table 2, entry 5)[28] were able to improve
the results obtained with 1a. We therefore selected the car-
bonate leaving group for further studies as it is the best
compromise between efficiency and reactivity.
Other solvents and organometallic species were tested
(Table 3) in order to optimize the reaction. Diethyl ether
(Table 3, entry 1) and dichloromethane (Table 3, entry 2)
were not as attractive as THF for this reaction. It is interest-
ing to observe that the chloride anion, when ZnCl2 was used
as the organometallic species (Table 3, entry 3) or when
LiCl was used as the additive (Table 3, entry 4), was signifi-
cantly detrimental to the reaction rate giving conversions
lower than 10% in each case. Copper (Table 3, entry 5) and
manganese (Table 3, entry 6) phenyl species were then tried
as the organometallic reagent, but without any success. An-
Scheme 2. Natural and non-natural biologically active compounds that
could arise from the asymmetric allylic transfer of an aryl group.
(Table 1, entry 2). The use of lithium halides as additives in
the reaction (Table 1, entries 3–5) led to the complete con-
version of compound 1a into mixtures of 2a and 3a. Lithi-
um bromide (Table 1, entry 4) was seen to give the best re-
sults (a point previously observed by the Evans group[15]);
the reaction being both regio- (2a/3a>2:1) and enantiose-
lective (74%). The substitution reaction was also found to
proceed under the more economical conditions in which
1.5 equivalents of the Grignard reagent, 0.75 equivalents of
ZnBr2, and 1.5 equivalents of LiBr, compared to that of the
substrate 1a, were used (Table 1, entry 6). The ratio of
ZnBr2 to the additive was found to be of importance; a ratio
of 2:1 in favor of the additive led to a lower conversion
(Table 1, entry 7) than that observed when the additive was
in equal or greater equivalents (Table 1, entries 4 and 6). Fi-
nally, the reactions in which commercially available PhLi
was used as the metal reagent gave decreased conversions
of 1a along with low regioselectivities and yields, both with
(Table 1, entry 8) and without an additive (entry 9).[27] From
the reactions summarized in Table 1, it was concluded that
the best conditions for further studies were those used in
entry 6.
other complex of iridium, [IrACHTNUGRTNEUNG(cod)2]BF4, with more cationic
character (Table 3, entry 7) also led to inferior performan-
ces. Co-solvents, such as N,N’-dimethyl-N,N’-propylene urea
(DMPU), N,N-dimethylacetamide (DMA), and 1,4-
diazabicycloACTHUNTGRNEUNG[2.2.2]octane (DABCO), or the slow addition of
the diphenylzinc species did not improve the results.[29]
We then screened different types of ligands (Scheme 3
and Table 4) to determine which was the most effective for
our system. Interestingly, the iridium complex devoid of any
ligand gave a satisfactory, selective conversion of 1a to the
linear isomer 3a (Table 4, entry 1). The screening of various
nonchiral phosphorus ligands (Table 4, entries 2–5) led to
the observation that unlike the initial studies by Takeuchi,[20]
the regioselectivity of the reaction is independent of the
electron-withdrawing properties of the ligand, a phenomen-
on recently noticed by Nomura.[30]
Of all the chiral ligands tested, none were found to give
better regioselectivity than L1, however, some, such as L8
(Table 4, entry 13), L9 (Table 4, entry 14) and L16 (Table 4,
entry 19), led to better enantioselectivities. Although L8 and
L9 were rather disappointing in terms of regioselectivity, the
1206
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1205 – 1216

Allylic Arylation
FULL PAPER
Table 1. Preliminary studies with ligand L1 and substrate 1a.[a]
performance of our newly de-
scribed
Simplephos
L16[31]
(Table 4, entry 19) was very
close to that of L1.
Considering the results ob-
tained
for
L2
(Table 4,
entry 7),[32] the methoxy sub-
stituents of other ligands seem
to play an important role in ob-
taining higher regio- and enan-
tioselectivities.[33] Similarly, the
position and the nature of the
substituents on the amino part
of the ligand are crucial: L3
(Table 4, entry 8) with the me-
thoxy group in the para posi-
tion or L4 (Table 4, entry 9)
with methyl groups in the ortho
position gave rise to lower
enantio- and regioselectivities.
These results suggest that in L1,
the methoxy groups might play
a coordinating role; a situation
totally different to our observa-
tions with stabilized nucleo-
philes[33] and one in which
deeper investigation needs to
Entry
M (equiv)
ZnBr2 (equiv)
Additive (equiv)
2a/3a[b]
Conv. [%][b] (Yield [%])
100[d]
34
100
100
100
100 (72)
54
ee [%][c]
1
2
3
4
5
6
7
8
9
MgBr (4)
MgBr (4)
MgBr (4)
MgBr (4)
MgBr (4)
MgBr (1.5)
MgBr (1.5)
Li (1.2)
–
4
4
4
4
0.75
3
1.2
1.2
–
–
–
–
53:47
61:39
69:31
60:40
66:34
55:45
33:67
45:55
58
73
74
74
75
74
57
75
LiCl (4)
LiBr (4)
LiI (4)
LiBr (1.5)
LiBr (1.5)
LiBr (1.2)
–
43
52
Li (1.2)
[a] Each reaction was carried out on a 0.5 mmol scale of substrate in THF (2.5 mL) for 15–18 h before hydroly-
sis (see the Experimental Section). [b] Ratio of 2a/3a and conversion measured by GC–MS or H NMR spec-
troscopy. Yield combining 2a and 3a. Isolated yield in parentheses. [c] Measured by chiral GC. [d] Recovery
of the allylic alcohol.
1
Table 2. Evaluation of different leaving groups.[a]
be performed to comment any further.
Other ligands with either larger (Table 4, entry 10) or
smaller (Table 4, entries 11–12) amino moieties did not im-
prove the results. In the worst cases, ligands such as the fer-
rocenyl-based Josiphos (Josiphos=1-[2-(diphenylphospha-
nyl)ferrocenyl]ethyldicyclohexylphosphane) and Taniaphos
(Taniaphos=1-(S)-diphenylphosphanyl-2-(O-diphenylphos-
phanylphenylmethyl)ferrocene), as well as Binap (Binap=
2,2’-bis(diphenylphosphino)-1,1’-binaphthalene) gave less
than 10% conversion.
Entry X, Substrate
T [8C] t [h] 2a/3a[b] Conv. [%][b] ee [%][c]
1
2
3
4
5
OCO2Me, 1a
Cl, 1ab
OAc, 1ac
25
25
40
16
2
40
66:34
52:48
61:39
100
100
86
100
21
75
71
82
OPO
N
28[d] 44:56
68
OCONHPh, 1ae 25
20
25:75
n.d.[e]
An interesting outcome was obtained when ligands L8
and L9, diastereomers of L2 and L1, respectively, were used
(Table 4, entries 13–14). In L8 and L9 the binapthol moiety
has the aR configuration and the amine the S,S configura-
tion, whereas in L2 and L1 the
[a] Each reaction was carried out on a 0.5 mmol scale of substrate using
the optimal conditions (Table 1, entry 6) and kept at the indicated tem-
perature. [b] Measured by GC–MS or 1H NMR spectroscopy. [c] Mea-
sured by chiral GC. [d] Reaction time not optimized. [e] Not determined.
Table 3. Variation of solvents, anions, organometallic species, and iridium source.[a]
binapthol has the aS and the
amine the S,S configuration.
Both L8 and L9 generated a
significant increase in enantio-
selectivity (80 and 84%, respec-
tively) compared with L2
(66%) and L1 (75%), although
these ligands preferentially
gave the linear adduct (13:87
and 30:70 respectively). Biphe-
nol-type ligands, such as L10
and L11, also gave a larger pro-
portion of the linear isomer
(Table 4, entries 15–16).
Entry
M (equiv)
ZnX2 (equiv)
Additive (equiv)
Solvent
2a/3a[b]
Conv. [%][b]
ee [%][c]
1
2
3
4
MgBr (1.5)
MgBr (1.5)
MgBr (1.5)
MgCl (1.5)
MgBr (1.5)
MgBr (1.5)
MgBr (1.5)
ZnBr2 (0.75)
ZnBr2 (0.75)
ZnCl2 (0.75)
ZnCl2 0.75
CuBr (1.5)
MnCl2 (1.5)
ZnBr2 (0.75)
LiBr (1.5)
LiBr (1.5)
LiBr (1.5)
LiCl (1.5)
LiBr (1.5)
LiCl (1.5)
LiBr (1.5)
Et2O
CH2Cl2
THF
THF
THF
THF
THF
56:44
59:41
29:71
24:76
61:39
22:78
50:50
83
63
5
55
6
64
7
n.d.[d]
24
5
100
16
45
6
66
71
7[e]
[a] Each reaction was carried out on a 0.5 mmol scale of 1a in the indicated solvent using the optimal condi-
tions (Table 1, entry 6). [b] Measured by GC–MS or 1H NMR spectroscopy. [c] Measured by chiral GC.
[d] Not determined. [e] Source of iridium: [IrACTHNUGTRNEUNG(cod)2]BF4.
Chem. Eur. J. 2009, 15, 1205 – 1216
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1207

A. Alexakis et al.
sults and will be used as the
ligand of choice for further
studies.
Having the optimal ligand
and reaction conditions in
hand, we turned our attention
to different substrates. We were
particularly keen on testing 3-
arylprop-2-en-1-ol derivatives
because
the
corresponding
products, 3,3-diarylprop-1-enes,
represent useful entries to opti-
cally enriched a,a-diaryl prod-
ucts. For this study, we changed
both the steric and electronic
properties of the aryl groups
(Table 5).
At first glance of the various
substrates used, the reaction
scope seems to be quite gener-
al. High enantioselectivities
(>90%) are usually obtained,
with one example above 99%
(Table 5, entry 5). Also, the re-
gioselectivity usually stands
slightly above 50%, an unpre-
cedented level for this type of
reaction. The most striking ob-
servation could be made from
entries 1 and 9 in which a heter-
oatom is present in the ortho
position of the substrate
(Table 5, entry 1, ortho-OMe-
phenyl; entry 9, ortho-furanyl).
In both cases the position of
the substituent seemed to have
a detrimental impact on the
enantioselectivity of the reac-
tion (78 and 79% ee, respec-
tively). Low enantioselectivity
was also observed when using
other kinds of nucleophiles,
such as amines[22] or malo-
nates.[26] When the heteroatom
was in a different position of
the substrate, either in the meta
Scheme 3. Chiral ligands used in this study.
(Table 5, entry 2) or para
(Table 5, entry 3) positions, high
enantioselectivities were ob-
Low conversions of 1a were obtained when testing ligands
such as the ephedrine- or the a,a,a’,a’-tetraaryl-2,2-dimeth-
yl-1,3-dioxolane-4,5-dimethanol (TADDOL)-based families
(Table 4, entries 17–18). On the other hand, the newly de-
scribed “Simplephos” family of ligands[31] (Table 4, en-
tries 19–21) and particularly L16 (Table 4, entry 19) led to
good enantioselectivities together with acceptable regioiso-
meric ratios, although L1 was still seen to give the best re-
tained (90 and 91% respectively). We could not, however,
rationalize the unfavorable regioselectivity observed in the
case of the para-OMe substituent (Table 5, entry3).
The varying of the substituents of the substrates, with re-
gards to their electron withdrawing properties, was also un-
dertaken with no marked effect being observed (compare
entries 3–7). Entry 11 highlights the limitations of this new
methodology; when the substituent on the allylic carbonate
1208
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1205 – 1216

Allylic Arylation
FULL PAPER
Table 4. Comparison of different chiral and nonchiral ligands.[a]
entry 6) and it has been previ-
ously noted that copper is the
best catalyst for these re-
agents.[34]
The availability of functional-
ized Grignard reagents in the
laboratory has been extensively
demonstrated by Knochel.[35]
After testing commercially
available organomagnesium nu-
cleophiles, we became interest-
ed in implementing our meth-
odology by combining the halo-
gen/metal exchange reaction (I/
Mg or I/Li)[35] and our allylic
substitution reaction. The re-
sults obtained for selected ex-
Entry
L
2a/3a[b]
Conv. [%][b] (ee [%])[c]
Entry
L
2a/3a[b]
Conv. [%][b] (ee [%])[c]
1
2
3
4
5
6
7
8
–
5:95
89 (–)
42 (–)
38 (–)
89 (–)
89 (–)
12
13
14
15
16
17
18
19
20
21
L7
L8
L9
L10
L11
L12
L13
L14
L15
L16
40:60
13:87
30:70
18:82
16:84
48:52
40:60
59:41
40:60
58:42
79 (35)
96 (80)
100 (84)
85 (À74)
97 (71)
PPh3
20:80
31:69
5:95
P
A
P
A
6:94
L1
L2
L3
L4
L5
L6
66:34
47:53
50:50
39:61
34:66
9:91
72[d] (75)
98 (66)
96 (64)
100 (64)
99 (71)
71 (2)
28 (À15)
3 (n.d.[e]
88[d] (83)
71 (55)
)
9
10
11
88 (73)
[a] All experiments were run on a 0.5 mmol scale of 1a in THF (2.5 mL) according to the optimal conditions
(Table 1, entry 6). Reaction time: 16–20 h. [b] Measured by GC–MS or 1H NMR spectroscopy. [c] Measured
by chiral GC. [d] Yield obtained after column chromatography on SiO2. [e] Not determined.
amples
are
presented
in
Table 7. We were pleased to ob-
serve that the I/Mg exchange
procedure, followed by trans-
metallation with ZnBr2 and al-
lylic substitution using our stan-
dard aliphatic substrate 1a
(Table 7, entry 1), gave almost
identical results as with the
Table 5. Investigation into the scope of the substrates.[a]
Entry
1
R
L1
(aR,RR)
(aR,RR)
(aS,SS)
(aS,SS)
(aR,RR)
(aS,SS)
(aR,RR)
(aR,RR)
(aS,SS)
(aS,SS)
Product
2/3[b]
Conv. [%][b] (Yield [%])[c]
ee [%][d]
commercially
PhMgBr reagent
available
(Table 1,
1
2
3
4
5
6
7
9
1b
1c
1d
1e
1 f
1g
1h
1i
o-MeO-C6H4
m-MeO-C6H4
p-MeO-C6H4
p-F-C6H4
p-Cl-C6H4
3,4-Cl2-C6H3
p-CF3-C6H4
2-furyl
N
2b
2c
2d
2e
2 f
2g
2h
2i
42:58
49:51
33:67
50:50
55:45
57:43
56:44
73:27
53:47
54:46
100
78 (S)
90 (S)
100 (67)
100 (78)
100 (83)
100 (83)
100 (89)
100 (98)
100
entry 6). This was not the case,
however, for the aromatic sub-
strates used in this study. For
example, the p-MeO-phenyl
substituent gave a disappointing
91 (R)
93 (R)
99.2 (S)
95 (R)
97 (S)
79 (S)
outcome
with
significantly
10
1j
1k
2-naphthyl
nPr
N
2j
2k
100 (93)
100
92 (R)
11[e]
15 (n.d.[f])
lower selectivities (Table 7,
[a] All experiments were run on a 0.5 mmol scale of 1 in THF (2.5 mL) according to the optimal conditions entry 2). We then compared the
1
(Table 1, entry 6); Reaction time: 16–20 h. [b] Measured by GC–MS or H NMR spectroscopy. [c] Yield deter-
I/Mg and I/Li exchange proce-
mined by 1H NMR spectroscopy. [d] Measured by chiral GC. [e] Four equivalents of PhMgBr, ZnBr2, and
dures on the same aromatic
LiBr were used (reaction t=120 h, not optimized). [f] Not determined.
pronucleophile,
p-F-phenyl
(Table 7, entries 3 and 4). From
these two experiments, we were
was a linear aliphatic chain, such as an n-propyl group, the
enantioselectivity of the reaction dropped dramatically.
After investigating the scope and limitations of the sub-
strates, we tested several commercially available Grignard
reagents for their nucleophilic properties (Table 6). The se-
lected Grignard reagents were treated with zinc bromide
and lithium bromide according to the usual procedure (see
Table 1 and the Experimental Section). In general, the spe-
cies with aryl substituents gave results similar to those ob-
tained when R=Ph with enantioselectivities in the range
80–91% together with acceptable regioisomeric selectivity
of the desired adducts (Table 6, entries 1, 2, 4 and 5). The
only notable exception concerns the 3,4-dichlorophenyl de-
rivative (Table 6, entry 3), which gave reduced regio- and
enantioselectivities. Alkyl Grignard reagents, such as
MeMgBr, did not fit well into this type of system (Table 6,
able to conclude that the I/Li exchange reaction (Table 7,
entry 4) gave similar regioselectivities, but higher enantiose-
lectivities than the I/Mg procedure (Table 7, entry 3). The
experiments shown in entries 1–3 are among the first exam-
ples of the use of the Knochelꢃs exchange methodology in
asymmetric catalysis. The Br/Li exchange was also investi-
gated with the 2-bromonaphthalene-substituted compound
in which slightly lower enantio- and regioisomeric yields
than the commercially available 2-MgBr-naphthalene were
obtained (Table 7, entry 5).
The absolute configurations of adducts 2 were determined
by comparison of the rotation of 2a with known literature
data.[36] We then decided to apply our methodology to the
total synthesis of a known intermediate in the preparation
of (+)-sertraline, a serotonin uptake inhibitor used in the
treatment of depression (Scheme 4).[18,37] Compound 2g was
Chem. Eur. J. 2009, 15, 1205 – 1216
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1209

A. Alexakis et al.
Table 6. Investigations into the scope of the nucleophiles.[a]
submitted to cross-metathesis
conditions with methyl acrylate
and the Hoveyda–Grubbs cata-
lyst 5.[38] The double bond of
the resulting unsaturated ester
6 was then reduced under mild
conditions by using a copper-
catalyzed 1,4-reduction reac-
tion[39] leading to ester 7, a
known intermediate in Daviesꢃ
synthetic approach to sertrali-
ne.[37b]
Entry
L1
(aS,SS)
(aS,SS)
(aR,RR)
(aS,SS)
(aS,SS)
(aS,SS)
4
R
2
2/3[b]
Conv. [%][b] (Yield [%])[c]
ee [%][d]
1
2
3
4
5
6
N
4d
4e
4g
4h
4j
p-MeO-C6H4
p-F-C6H4
3,4-Cl2-C6H4
p-CF3-C6H4
2-naphthyl
Me
2d
2e
2g
2h
2j
57:43
45:55
16:84
40:60
63:37
18:82
100 (98)
100 (95)
100 (98)
100 (78)
100 (97)
100
91 (S)
82 (S)
63 (R)
80 (S)
91 (S)
6 (S)
4l
2l
Compared to stabilized nu-
cleophiles, such as malonates
and amines, the arylzinc nucleo-
philes studied in this work give
rise to the same facial selectivi-
ty.[27] This observation is quite
intriguing given the inherent
difference between the two nu-
cleophilic species. One could
possibly envisage the following
[a] All experiments were run on a 0.5 mmol scale of 1l in THF (2.5 mL) according to the optimal conditions
1
(Table 1, entry 6); Reaction time: 16–20 h. [b] Measured by GC–MS or H NMR spectroscopy. [c] Yield deter-
1
mined by H NMR spectroscopy. [d] Measured by chiral GC.
Table 7. Nucleophiles obtained by halogen/metal exchange.[a]
hypotheses,
among
others
(Figure 1): 1) The nature of the
iridium(I) catalyst is that of the
commonly accepted one;[23b]
the oxidative addition of the
iridium catalyst occurs on the
usual side (Re if R=cyclohex-
yl), which is followed by an anti
attack of the aryl nucleophile.
Entry Substrate
R
L1
Method[b] Ar
Product 2/3[c] Conv. [%][c] ee
Yield [%][d] [%][e]
1
2
3
4
5
1a
1d
1e
1e
1l
c-hex
A
A
A
A
B
B
Ph
Ph
Ph
Ph
2a
2d
2e
2e
54:46 100 (72)
18:82 100 (65)
46:54 100 (87)
44:56 100 (53)
40:60 100 (78)
77 (S)
44 (R)
78 (R)
p-MeO-C6H4
p-F-C6H4
p-F-C6H4
Ph
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
N
87 (R) 2) The selected conditions gen-
R
2-naphthyl[f] 2j
80 (S)
erate a different iridium(I) cat-
[a] All experiments were run on a 0.5 mmol scale of 1l in THF (2.5 mL) according to the optimal conditions
(Table 1, entry 6); Reaction time: 16–20 h. [b] Method A: I/Mg exchange; Method B: I/Li exchange. [c] Mea-
sured by GC–MS or 1H NMR spectroscopy. [d] Yield determined by 1H NMR spectroscopy. [e] Measured by
chiral GC. [f] Br/Li exchange.
alyst, already containing the
aryl group; the oxidative addi-
tion takes place on the opposite
side (Si if R=cyclohexyl),
which is followed by a reduc-
tive elimination process on the
same side. It can also be specu-
lated whether the branched and
linear regioisomers arise from
the same reaction intermediate
or from two different pathways:
the branched product following
pathway A, whereas the linear
product could arise from path-
way B.
Next, we tried our methodol-
ogy on acetate 8 under kinetic
resolution
0.55 equivalents
conditions
of
with
Ph2Zn
(Scheme 5). After the reaction
was stirred for two hours, at
51% conversion, the ratio of
adducts 2a to 3a (4:1) was
much greater than that found
Scheme 4. Towards the enantioselective formal synthesis of (+)-sertraline.
1210
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1205 – 1216

Allylic Arylation
FULL PAPER
Figure 2. Evolution of the regioisomeric proportion 2a/3a relative to 2a
*
under the usual conditions (Table 1, entry 6). : conversion, &: regiomer-
ic proportion.
Figure 1. Two possible pathways explaining the product stereochemistry;
A) anti attack pathway and B) reductive elimination pathway.
any further conclusions, we suppose that more than one cat-
alytic species are present in the reaction mixture.
when starting from 1a (2:1). The enantiomeric excess of 2a,
in the case of the acetate, was 68%. In addition, the reaction
was much faster than with 1a.[23b] This experiment shows
that the regiospecificity of the catalyst is of much greater
importance in the case of nonstabilized nucleophiles. On the
other hand, the selectivity factor (s=4)[40] remained close to
that obtained with stabilized nucleophiles, such as malo-
Since the mechanistic aspects of this reaction seemed
somewhat different from the classical malonate substitution,
we tested some previously unreactive allylic substrates.
Thus, we tried our methodology on cyclic substrates such as
rac-9. In Pd chemistry, such substrates are meso-p-allyl and
can be desymmetrized by the incoming nucleophile.[3] We
wondered if this would also be the case with Ir-catalyzed ar-
ylation because this would remove the linear/branched prob-
lem. For this study, three substrates were tested: the carbon-
ate 9a, the acetate 9b, and the bromide 9c (Table 8). The ar-
ylation of such substrates had already been examined with
various transition-metal complexes, excluding iridium, with
good to excellent results.[10e,12a,41]
nates[27]
.
Concerning the standard reaction (Table 1, entry 6), the
evolution of the regioisomeric ratio over time was followed
by gas chromatography (Figure 2). We were surprised to ob-
serve that the desired isomer represents a minor fraction of
the products up to around 35% conversion and gradually in-
creases over time to eventually become the major regioiso-
mer at the end of the reaction. This observation could make
one suppose that a chemical species is probably accumulat-
ing as the reaction progresses enabling the formation of the
desired isomer, however, all attempts to investigate the rea-
sons for the regioisomeric evolution were unfruitful. These
attempts involved running the reaction in the presence of a
stoichiometric or catalytic amount of adduct 2a, or to intro-
duce either MeOLi or AcOLi as additives, but none of these
perturbations led to an improvement of the regioisomeric
yield. Similarly, adding another equivalent of the starting
material 1a and the arylzinc species to the reaction mixture
after full conversion of the first equivalent was no more suc-
cessful. Although deeper investigation is needed to draw
Carbonate 9a reacted readily and cleanly under our stan-
dard conditions to afford compound 10. In contrast, we
were never able to obtain any adducts using stabilized nu-
cleophiles (such as malonates), even under forced conditions
(reflux temperature).[27] In this case, L1 is not an efficient
ligand (Table 8, entry 1). Among the usual phosphoramidite
ligands, L10 gave the best result with 76% ee (Table 8,
entry 4). Ligand L14, which has an extra nitrogen on the
amine part of the ligand, gave a similar result, particularly
when 1,5,7-triazabicycloACHTNUGRTNEUNG
[4.4.0]dec-5-ene (TDB)[42] was added
as cocatalyst (Table 8, entry 5). When PhZnBr/MgBr2 was
used, the ee could be improved to 82% (Table 8, entry 6).
Acetate 9b was clearly the least reactive of these sub-
strates. Often the conversions were not quantitative, howev-
Scheme 5. Allylic arylation under kinetic resolution conditions.
Chem. Eur. J. 2009, 15, 1205 – 1216
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1211

A. Alexakis et al.
Table 8. Arylation of 9.[a]
gate addition to the parent al-
dehyde. Although the malo-
nate-type nucleophiles did not
react with under Ir catalysis
conditions, we did observe a re-
action
with
Ph2Zn/MgBr2
(Scheme 6). The reaction gave
the expected branched isomer
12, along with two other prod-
ucts, 13 and 13’, in a 78:6:16
ratio. Products 13 and 13’ arise
from the linear allylic acetate,
Entry
Substrate
L
Additive ([%])
Conv. [%][b]
(Yield [%])[c]
ee
G
[%][d]
1
2
3
4
5
6
7
8
9a
9a
9a
9a
9a
9a
9b
9b
9b
9b
9b
9b
9b
9b
9b
9c
9c
9c
9c
9c
L1 (aS,SS)
–
–
–
–
96
15 (S)
35 (S)
60 (S)
76 (S)
74 (R)
82 (R)
43 (S)
57 (S)
46 (S)
45 (S)
79 (S)
84 (R)
88 (S)
84 (S)
90 (R)
54 (S)
63 (S)
75 (S)
84 (S)
55 (S)
L11 (aS,SS)
L12 (aS,SS)
ent-L10 (SS)
L14
L14 PhZnBr/MgBr2
L1 (aS,SS)
L2 (aS,SS)
L8 (aR,SS)
L12
L12
100
100
100
97
100
98
76
15
91
82
77
83
96
85 (79)
100
99
98
97
followed by
a second allylic
substitution, both from a and g
attack. The ee value of 12 was
only 29%, and no further stud-
ies were done on this substrate.
We instead focused on 1,4-di-
functionalized allylic substrates.
For example, the Z dicarbonate
14 was unreactive with malo-
nates under Ir catalysis condi-
tions. In the arylation reaction
(Scheme 7), it gave mostly the
linear product (yield=90%,
TDB (8)
TDB (8)
–
–
–
–
TDB (8)
–
–
TDB (8)
TDB (8)
–
–
–
–
–
9
10
11
12
13
14
15
16
17
18[e]
19
20
L11 (SS)
ent-L10 (SS)
ent-L10 (SS)
L14
L1 (aS,SS)
L1 (aS,SS) PhZnBr/MgBr2
L1 (aS,SS) PhZnBr/MgBr2
ent-L10 (SS)
ent-L10 (SS) PhZnBr/MgBr2
branched/linear=7:93),
al-
100
though this was predictable be-
cause Helmchen and co-work-
ers demonstrated the different
[a] All experiments were run on a 0.5 mmol scale of 9 in THF (2.5 mL) according to the optimal conditions
(Table 1, entry 6); Reaction time: 16–20 h. [b] Measured by GC–MS or 1H NMR. [c] Yield determined by
1H NMR spectroscopy. [d] Measured by chiral GC. [e] Reaction run at 08C.
er, some of the best enantioselectivities were achieved with
this substrate. Again, L1 was not the best choice of ligands.
A match/mismatch situation could be seen with L2 and L8
(Table 8, entries 8 and 9), affecting, not the enantioselectivi-
ties, but the conversion. As was seen for carbonate 9a,
ligand L10 gave excellent results with 9b (Table 8, entry 14;
84% ee), although the addition of TDB showed no improve-
Scheme 6. Allylic arylation on a gem-difunctionalized substrate.
ment. Finally, the best selectivities were found with L14 and
TBD with 90% ee obtained (Table 8, entry 15).
The bromide 9c was the most reactive substrate. Even
without the catalyst present, the reaction proceeded at a sig-
nificant rate, with Ph2Zn/MgBr2 (the reaction was over in
10 min at room temperature). Using PhZnBr/MgBr2 as the
reagent allowed a slower background reaction. In the case
of 9c, L1 was more successful than for the other substrates.
The 54 % ee (Table 8, entry 16) could be improved to 63%
with PhZnBr/MgBr2 (Table 8, entry 17), and even to 75% at
08C (Table 8, entry 18). Here again, ligand L10 appeared to
be the best ligand with 84% ee (Table 8, entry 19).
Difunctional substrates are also of interest. For example,
the easily prepared diacetate 11 (Scheme 6), has previously
been used by Trost and Lee with stabilized nucleophiles.[43]
With a Pd-based catalyst, only the a (linear) product is ob-
tained, giving rise to an enantioenriched allylic acetate. In
behavior of E and Z isomers.[44] The E diacetate 15, pre-
pared from the dibromide 16,[45] gave only 25% of racemic
linear isomer. Finally, we turned our attention to the E di-
bromide 16, which is commercially available and has previ-
ously been exploited in Cu-catalyzed allylic substitution.[46]
Interestingly, in Cu-catalyzed reactions, compound E-16
gave 100% branched product, probably due to a favorable
interaction between the bromide and the copper reagent.
The first experiment with L1 showed, as expected, a moder-
ate preference for the branched product. Compared with
carbonates or acetates, however, the bromides are much
more reactive (Table 9) and part of the reaction outcome
was as a result of an uncatalyzed background reaction.
Indeed, the Grignard reagent itself was quite reactive, but
barely regioselective (Table 9, entry 1). Turning to the less
reactive PhZnBr/MgBr2 was more rewarding (Scheme 7 and
contrast, with Ir-based catalysts, we expected the
g
(branched) product, which formally corresponds to a conju-
1212
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1205 – 1216

Allylic Arylation
FULL PAPER
nucleophile to the reaction mix-
ture, followed by addition of
the substrate (the reverse of
above) led to our best result. In
this case we obtained almost
perfect regioselectivity (99:1)
and very good enantioselectivi-
ty (80% ee) with L8 (Table 9,
entry 8). This is, presently, the
best enantioselectivity achieved
Scheme 7. Allylic arylation on a difunctionalized allylic substrate.
for the arylation of this sub-
strate. Of particular interest
were our new SimplePhos li-
Table 9), both for the regioselectivity and the enantioselec-
tivity (compare Table 9, entries 2 and 3).
gands.[31] Thus, ligand L16 afforded 60% ee at room temper-
ature. Among the other SimplePhos ligands (L16–L19),
ligand L17 gave the best results (97:3 and 69% ee, entry 10).
It was particularly interesting to observe an inversion in the
configuration of the branched adduct obtained with phos-
phoramidite ligand compared with the SimplePhos ligands.
Table 9. Arylation of dibromide 16.[a]
Entry PhMgBr ZnBr2
(equiv) (equiv)
L
Ratio[b]17/ Conv.[%][c]
18
ee
ACHTUNGTRENNUNG
(Yield[%])[b] [%][d]
G
N
1
2
1.5
1.5
–
0.75
–
38:62
75
100
–
13
(R)
48
(R)
65
(R)
54
(R)
76
(R)
68
(R)
80
(R)
60
(S)
69
(S)
21
(S)
67
L1 (aR,RR) 65:35
L1 (aR,RR) 84:16
L1 (aR,RR) 97:3
Conclusion
3
1.5
1.5
2
1.5
1.5
1
100
100
100
100
84
We have developed an asymmetric allylic arylation reaction
using iridium. Associated with phosphoramidite or amino-
phosphine (SimplePhos) ligands, iridium forms a very effi-
cient catalyst for the transfer of arylzinc reagents. The de-
sired branched adducts were obtained in moderate regiose-
lectivities, but high enantioselectivities (up to 99% ee). The
use of diarylzinc reagents, obtained from the corresponding
Grignard reagents, were compatible with several halogen/
metal exchange reactions.
We applied our methodology to the synthesis of useful
diaryl alkenes and more particularly to an intermediate,
which was used in the synthesis of (+)-sertraline. Future
work aimed at improving the regioselectivities will be direct-
ed towards the understanding of the mechanism and the de-
velopment of new chiral ligands.
4[e]
5
L8
64:36
6
2
1
L8 + TBD 98:2
(8%)
L8 + TBD 94:6
(8%)
L8 + TBD 99:1
(8%)
L16 (SS)
L17 (SS)
L18 (SS)
L19 (SS)
7
1
1
8[f]
9
2
1
100
(85)
100
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
77:23
10
11
12
97:3
100
100
100
71:29
90:10
(S)
[a] All experiments were run on a 0.5 mmol scale of 14 in THF (2.5 mL)
according to the optimal conditions (Table 1, entry 6); Reaction time:
16–20 h, except for entries 6, 7 and 8 (1 h). [b] Measured by GC–MS or
1H NMR spectroscopy. [c] Yield determined by 1H NMR spectroscopy.
[d] Measured by chiral GC. [e] Reaction run at 08C. [f] Reverse addition
(substrate after nucleophile).
Experimental Section
General methods: All reactions were carried out under argon atmos-
phere with oven-dried glassware. Solvents were dried by filtration over
alumina (previously activated at 3508C over 12 h) under nitrogen before
use. Solvents were degassed by the bubbling through of nitrogen prior to
all experiments. H (400 MHz) and 13C (100 MHz) NMR spectra were re-
1
When using PhZnBr/MgBr2, lowering the temperature of
the reaction to 08C (Table 9, entry 4) yielded the best re-
sults, both for regio- and enantioselectivity (97:3 and
65% ee). In contrast to previous results (Table 4), ligand L8
appeared to be a good ligand, with respect to the obtained
ee value (Table 9, entry 5). In accordance with the work of
Helmchen and co-workers,[42] we observed that activation
with TBD improved both the regio- and enantioselectivities
(Table 9, compare entries 5 and 6). TBD activation of the
catalyst was tested with several ligands, including L1 and
L10, but L8 gave the best result. Finally, the addition of the
corded on a Bruker 400F NMR in CDCl3. Chemical shift values (d) are
given in ppm relative to residual CHCl3. Multiplicity is indicated as fol-
lows: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), dd
(doublet of doublet), dt (doublet of triplet), ddd (doublet of doublet of
doublet), qd (quartet of doublet), brs (broad singlet). Coupling constants
are reported in Hertz (Hz). The evolution of the reactions was followed
by GC–MS (EI mode) on an HP6890 instrument. Optical rotations were
recorded on a Perkin–Elmer 241 polarimeter at 208C in a 10 cm cell in
the stated solvent ; [a]D values are given in 10À1 degcm2 gÀ1 (concentra-
tion c given as g/100 mL). Enantiomeric excesses were determined by
two different methods. The first was by use of a chiral-SFC measurement
on a Berger SFC with the stated column. Gradient programs are de-
scribed as follows: initial methanol concentration (%)–initial time (min)–
Chem. Eur. J. 2009, 15, 1205 – 1216
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1213

A. Alexakis et al.
percent gradient of methanol (%minÀ1)–final methanol concentration
(%); retention times (tR) are given in min. The second method was by
chiral GC measurements either on a HP6890 (H2 as vector gas) or
HP6850 (H2 as vector gas) instrument with the stated column. Tempera-
ture programs are described as follows: initial temperature (8C)–initial
time (min)–temperature gradient (8CminÀ1)–final temperature (8C); re-
tention times (tR) are given in min. Flash chromatography was performed
using silica gel 32–63 mm, 60 ꢄ.
2H), 2.27 ppm (m, 2H); 13C NMR (100 MHz, CDCl3, 258C): d=173.4,
144.5, 142.6, 132.4, 130.3, 130.2, 129.6, 128.7, 127.6, 127.1, 126.8, 51.5,
49.5, 32.1, 30.1 ppm.
Typical procedure for allylic arylation of carbonate 9b
(S)-Cyclohex-2-enylbenzene (10): (According to Table 8, entry 15) In a
flame-dried Schlenk tube, the chiral ligand L14 (0.02 mmol, 0.04 equiv),
[{IrClACTHUNTRGNENG(U cod)Cl}2] (6.7 mg, 0.01 mmol, 0.02 equiv), and freshly sublimed
TBD (5.6 mg, 0.04 mmol, 0.08 equiv) were dissolved in dry THF
(0.5 mL). The resulting orange solution was stirred for 2 h at room tem-
perature under a nitrogen atmosphere. Acetate 9b (70.1 mg, 0.5 mmol,
1 equiv) was added and the mixture was stirred for 15 min. During this
time another solution was prepared in a second flame-dried flask con-
taining freshly dried ZnBr2 (112.6 mg, 0.5 mmol, 1 equiv), LiBr (86.8 mg,
1 mmol, 2 equiv), and dry THF (2 mL). The phenyl Grignard reagent
(1 mL, 1 mmol, 2 equiv) was added to the latter flask and the resulting
solution was stirred for 2 h before being introduced into the first Schlenk
flask. The combined mixture was stirred overnight at 408C and quenched
with a saturated aqueous solution of NH4Cl (10 mL). The product was
extracted with Et2O (3ꢅ10 mL) and the organic layer was dried and con-
centrated in vacuo. The crude product was purified on a silica gel chro-
matography column (cyclohexane) and 10[10e] was isolated as a colorless
All chiral ligands were prepared according to the literature procedures
(for ligands L1–L11, see references [26], [33], [47], and [48]; for ligands
L12–L15, see references [49] and [50]; for ligands L16–L19, see refer-
ence [31]). Lithium bromide and lithium chloride were dried at 808C
over 24 h prior to use. [{IrClACHTUNTRGNEUNG(cod)}2] was purchased from Strem and used
as received. Allylic carbonates and acetates were synthesized by known
experimental procedures.[51]
Typical procedure for allylic arylation of carbonates 1a–1k
Preparation of the nucleophilic solution: In a flame-dried flask containing
dry THF (2 mL for each mmol of starting material), ZnBr2 (0.75 equiv)
and LiBr (1.5 equiv) were added (equivalents are relative to substrate).
The solution was stirred for 10–15 min at room temperature. The corre-
sponding solution of Grignard reagent (1.5 equiv) was then added to the
flask. The resulting solution was ready to be used.
1
liquid (62 mg). H NMR (400 MHz, CDCl3, 258C): d=7.36–7.23 (m, 5H),
5.93 (m, 1H), 5.75 (m, 1H), 3.44 (m, 1H), 2.14–1.55 ppm (m, 6H).
13C NMR (100 MHz, CDCl3, 258C): d=146.7, 130.2, 128.4, 128.3, 127.8,
126.0, 41.9, 32.7, 25.1, 21.2 ppm; IR (CHCl3): n˜ =3064, 3009, 2932, 2860,
1601, 1491, 1450, 737, 702 cmÀ1; MS (EI): m/z (%): 158 (100) [M]+, 143
(57), 129 (95), 115 (57), 104 (18), 91 (28).
Procedure for the Ir-catalyzed arylation of 1g: A flame-dried Schlenk
tube was charged with ligand L1 (S,SS) (4.4%), [{IrClACHTUNGRTNEUNG(cod)Cl}2] (2%),
and dry THF (1 mL for each mmol of starting material). The resulting
orange solution was stirred for 10–15 min at room temperature. The sub-
strate 1g (4 mmol) was then added and the reaction mixture was treated
with the nucleophilic solution. The resulting mixture was stirred at room
temperature overnight. The mixture was then hydrolyzed with water, a
few drops of HCl (10%) and extracted with Et2O. The combined organic
layers were dried over anhydrous MgSO4, filtered, and concentrated in
vacuo. The product was purified by flash column chromatography (silica
gel, pentane), to afford compound 2g (551.2 mg, 47%) as a colorless
liquid and a mixture of 2g/3g (510 mg, 42%). An ee value of 95% was
measured by chiral SFC with an OJ-H column (program: 2% MeOH-1’-
Typical procedure for allylic arylation of dibromide 16
(S)-(1-Bromobut-3-en-2-yl)benzene (17): (According to Table 9, entry 8)
In a flame-dried Schlenk tube, chiral L8 (aR,SS) (21.6 mg, 0.04 mmol,
0.08 equiv), [{IrClACTHNUTRGNENUG(cod)Cl}2] (13.4 mg, 0.02 mmol, 0.04 equiv), and freshly
sublimed TBD (5.6 mg, 0.04 mmol, 0.08 equiv), were dissolved in dry
THF (0.5 mL). The resulting orange solution was stirred for 2 h at RT
until intense red coloration appeared. Immediately following the colora-
tion, the substrate (106 mg, 0.5 mmol, 1 equiv) was added to the Schlenk
tube and the mixture stirred for 15 min. During this time another solu-
tion was prepared in a second flame-dried flask containing freshly dried
ZnBr2 (112.6 mg, 0.5 mmol, 1 equiv) and LiBr (86.8 mg, 1 mmol, 1 equiv)
in dry THF (2 mL). The phenyl Grignard reagent (1 mL, 1 mmol,
2 equiv) was then added to the flask and the resulting mixture was stirred
for 2 h before being introduced into the first Schlenk. The combined re-
action mixture was stirred at RT for 2 min before Et2O (10 mL) and a sa-
turated aqueous solution of NH4Cl (10 mL) were added to the reaction
mixture. The organic layer was washed with brine (10 mL) and water
(10 mL), then dried over Na2SO4 and concentrated in vacuo. The crude
product was purified on silica gel chromatography column (SiO2, pen-
tane, Rf =0.79) to recover products 17 (89 mg). Enantiomeric excesses
was measured by chiral GC with a Chirasil-Dex CB, Helium flow instru-
ment (program: 70–0–1–170) tR = 47.55 (À), 48.33 (+); [a]2D2 =À22.18
(c=0.28 in CHCl3) for 65% ee; 1H NMR (400 MHz, CDCl3, 258C): d=
7.38–7.34 (m, 2H), 7.30–7.29 (m, 1H), 7.28–7.22 (m, 2H), 6.03 (ddd, J=
10.4, 7.3, 17.4 Hz, 1H), 5.22 (dd, J=1.0, 10.4 Hz, 1H), 5.17 (dd, 1H, J=
1.3, 17.2 Hz), 3.71 (m, 1H), 3.65 (s, 1H), 3.63 ppm (d, J=1.8 Hz, 1H);
13C NMR (100 MHz, CDCl3, 258C): d=141.32, 138.6, 128.9, 128.9, 127.8,
127.8, 127.4, 117.2, 51.9, 36.5 ppm; IR (neat): n˜ =3063, 2959, 2923, 1873,
1640, 1601, 1493, 1453, 1416, 1258, 1220, 1074, 989, 920, 762, 748, 698,
650 cmÀ1; MS (EI) m/z (%): 212 (21), 210 (21), 154 (17), 131 (22), 118
(20), 117 (100), 116 (14), 115 (42), 91 (30), 77 (16), 51 (19); HRMS (EI):
m/z calcd for C10H10Br: 210.0044; found=210.0044.
2–15%, 175 bar, 2 mLminÀ1, 308C). tR = 6.25 (S) and 6.53 (R); [a]D20
=
+3.0 (c=1.00 in CHCl3); 1H NMR (400 MHz, CDCl3, 258C): d=6.80–
7.80 (m, 8H), 6.27 (m, 1H), 5.30 (dd, J=10.4, 1.0 Hz, 1H), 5.06 (dd, J=
10.4 Hz, 1H), 4.71 ppm (d, J=7.1 Hz, 1H); 13C NMR (100 MHz, CDCl3,
258C): d=143.5, 142.0, 139.4, 132.3, 130.5, 130.3, 130.2, 128.6 (2C), 128.4
(2C), 128.0, 126.8, 117.3, 54.0 ppm.
ACHTUNGTRENNUNG(R,E)-Methyl 4-(3,4-dichlorophenyl)-4-phenylbut-2-enoate (6): Com-
pound 2g (551.2 mg, 2.1 mmol, 1 equiv) was added, under argon, to a
flask containing CH2Cl2 (3 mL). Methylacrylate (0.43 mL, 4.75 mmol,
2.3 equiv) and catalyst 5 (29.8 mg, 2.3%) were then added and the mix-
ture was stirred at reflux overnight before being diluted with Et2O and
hydrolyzed with HCl (10%). The organic layer was separated and the
aqueous layer was extracted with Et2O (3ꢅ). The combined organic
layers were washed with brine and dried over anhydrous Na2SO4, filtered,
and concentrated in vacuo. The product was purified by flash column
chromatography (silica, pentane/Et2O 9:1), affording the desired com-
pound 6 as a colorless oil (568.1 mg, 84%). [a]2D0 =+1.0 (c=1.12 in
CHCl3); 1H NMR (400 MHz, CDCl3, 258C): d=7.40–7.20 (m, 6H), 7.15
(d, 2H, J=7.8 Hz), 7.01 (d, J=8.1 Hz, 1H), 5.76 (d, 1H, J=15.6 Hz),
4.84 ppm (d, J=7.3 Hz, 1H); 13C NMR (100 MHz, CDCl3, 258C): d=
165.2, 148.5, 141.7, 140.2, 133.3, 131.0, 130.5, 130.3, 128.8 (2C), 128.4
(2C), 127.9, 127.3, 123.2, 52.2, 51.6 ppm.
(R)-Methyl 4-(3,4-dichlorophenyl)-4-phenylbutanoate (7): CuACHTUNGRTNEUNG(OAc)2·H2O
(3.8 mg, 5%) and (Æ)-BINAP (12.6 mg, 5%) were added to a flask con-
taining THF (4 mL). After stirring for 5 min, poly(methylhydrosiloxane)
(100 mL, 1.62 mmol, 4 equiv) was added into the flask, followed by 6
(130.1 mg, 0.405 mmol, 1 equiv) and tBuOH (124.3 mg, 1.62 mmol,
4 equiv). The resulting mixture was stirred at room temperature for 5 h
and concentrated in vacuo. The crude mixture was purified by flash
column chromatography (silica gel, pentane/Et2O 9:1) affording com-
pound 7 as a colorless oil (130.9 mg, 100%). [a]2D0 =À6.1 (c=1.12 in
CHCl3); 1H NMR (400 MHz, CDCl3, 258C): d=7.36–7.20 (m, 7H), 7.08
(dd, J=8.4, 1.8 Hz, 1H), 3.91 (t, J=7.3 Hz, 1H), 3.65 (s, 3H), 2.35 (m,
All other products are described in the Supporting Information, including
chromatographic and spectral data.
Acknowledgements
The authors thank the Swiss National Research Foundation (grant no.
200020-113332) for financial support, and Laetitia Palais for providing
1214
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1205 – 1216

Allylic Arylation
FULL PAPER
the SimplePhos ligands, as well as BASF for the generous gift of chiral
amines.
[17] A. D. Patil, A. J. Freyer, L. Killmer, P. Offen, B. Carte, A. J. Jure-
wicz, R. K. Johnson, Tetrahedron 1997, 53, 5047–5060.
[18] W. M. Welch, A. R. Kraska, R. Sarges, K. B. Coe, J. Med. Chem.
1984, 27, 1508–1515.
[19] S. P. Shrestha, Y. Amano, Y. Narukawa, T. Takeda, J. Nat. Prod.
2008, 71, 98–101.
[20] R. Takeuchi, M. Kashio, Angew. Chem. 1997, 109, 268–270; Angew.
Chem. Int. Ed. Engl. 1997, 36, 263–265.
[21] J. P. Janssen, G. Helmchen, Tetrahedron Lett. 1997, 38, 8025–8026.
[22] T. Ohmura, J. F. Hartwig, J. Am. Chem. Soc. 2002, 124, 15164–
15165.
[23] For seminal reviews, see: a) R. Takeuchi, S. Kezuka, Synthesis 2006,
3349–3366; b) G. Helmchen, A. Dahnz, P. Dꢆbon, M. Schelwies, R.
Weihofen, Chem. Commun. 2007, 675–691.
[24] a) T. Graening, J. F. Hartwig, J. Am. Chem. Soc. 2005, 127, 17192–
17193; b) D. J. Weix, J. F. Hartwig, J. Am. Chem. Soc. 2007, 129,
7720–7721.
[25] I. Lyothier, C. Defieber, E. M. Carreira, Angew. Chem. 2006, 118,
6350–6353; Angew. Chem. Int. Ed. 2006, 45, 6204–6207.
[26] K. Tissot-Croset, D. Polet, S. Gille, C. Hawner, A. Alexakis, Synthe-
sis 2004, 2586–2590.
[27] D. Polet, A. Alexakis, K. Tissot-Croset, C. Corminboeuf, K. Ditrich,
Chem. Eur. J. 2006, 12, 3596–3609.
[28] For the use of carbamates in allylic substitution and their beneficial
effect towards regioselectivity, see: H. L. Goering, S. S. Kantner,
C. C. Tseng, J. Org. Chem. 1983, 48, 715–721.
[29] X. Rathgeb, PhD Dissertation, University of Geneva (Switzerland),
2007.
[30] N. Nomura, S. Komiyama, H. Kasugai, M. Saba, J. Am. Chem. Soc.
2008, 130, 812–814.
[31] L. Palais, I. S. Mikhel, C. Bournaud, L. Micouin, C. A. Falciola, M.
Vuagnoux-dꢃAugustin, S. Rosset, G. Bernardinelli, A. Alexakis,
Angew. Chem. 2007, 119, 7606–7609; Angew. Chem. Int. Ed. 2007,
46, 7462–7465.
[32] B. L. Feringa, Acc. Chem. Res. 2000, 33, 346–353.
[33] D. Polet, A. Alexakis, Org. Lett. 2005, 7, 1621–1624.
[34] K. Tissot-Croset, A. Alexakis, Tetrahedron Lett. 2004, 45, 7375–
7378.
[35] A. Boudier, L. A. Bromm, M. Lotz, P. Knochel, Angew. Chem. 2000,
112, 4584–4606; Angew. Chem. Int. Ed. 2000, 39, 4414–4435.
[36] a) B. Breit, D. Breuninger, Synthesis 2005, 147–157; b) T. Hayashi,
H. Iwamura, M. Naito, Y. Matsumoto, Y. Uozumi, M. Miki, K.
Yanaki, J. Am. Chem. Soc. 1994, 116, 775–776.
[1] a) A. Alexakis, K. Croset, Org. Lett. 2002, 4, 4147–4149; b) K.
Tissot-Croset, D. Polet, A. Alexakis, Angew. Chem. 2004, 116, 2480–
2482; Angew. Chem. Int. Ed. 2004, 43, 2426–2428; c) K. Tissot-
Croset, A. Alexakis, Tetrahedron Lett. 2004, 45, 7375–7378; d) C. A.
Falciola, K. Tissot-Croset, A. Alexakis, Angew. Chem. 2006, 118,
6141–6144; Angew. Chem. Int. Ed. 2006, 45, 5995–5998; e) C. A.
Falciola, A. Alexakis, Angew. Chem. 2007, 119, 2673–2676; Angew.
Chem. Int. Ed. 2007, 46, 2619–2622; f) C. A. Falciola, K. Tissot-
Croset, H. Reyneri, A. Alexakis, Adv. Synth. Catal. 2008, 350, 1090–
1100.
[2] a) B. M. Trost in Allylic substitution reactions, Vol. 2 (Eds.: A. Pfaltz,
M. Lautens), 1999, pp. 833–884; b) Z. Lu, S. Ma, Angew. Chem.
2008, 120, 264–303; Angew. Chem. Int. Ed. 2008, 47, 258–297.
[3] B. M. Trost, D. L. Van Vranken, Chem. Rev. 1996, 96, 395–422.
[4] M. van Klaveren, E. S. M. Persson, A. del Villar, D. M. Grove, J.-E.
Bꢀckvall, G. van Koten, Tetrahedron Lett. 1995, 36, 3059–3062; for
reviews on Cu, see: a) A. Alexakis, C. Malan, L. Lea, K. Tissot-
Croset, D. Polet, C. Falciola, Chimia 2006, 60, 124–130; b) C. Falcio-
la, A. Alexakis, Eur. J. Org. Chem. 2008, 3765–3780.
[5] a) A. Alexakis, C. Malan, L. Lea, C. Benhaim, X. Fournioux, Synlett
2001, 0927–0930; b) S. Tominaga, Y. Oi, T. Kato, D. K. An, S. Oka-
moto, Tetrahedron Lett. 2004, 45, 5585–5588; c) F. Lopez, A. W. Van
Zijl, A. J. Minnaard, B. L. Feringa, Chem. Commun. 2006, 409–411.
[6] a) F. Dꢆbner, P. Knochel, Angew. Chem. 1999, 111, 391–393;
Angew. Chem. Int. Ed. 1999, 38, 379–381; b) H. Malda, A. W. van
Zijl, L. A. Arnold, B. L. Feringa, Org. Lett. 2001, 3, 1169–1172;
c) C. A. Luchaco-Cullis, H. Mizutani, K. E. Murphy, A. H. Hoveyda
Angew. Chem. 2001, 113, 1504–1508; Angew. Chem. Int. Ed. 2001,
40, 1456–1459; Angew. Chem. Int. Ed. 2001, 40, 1456–1459.
[7] a) C. Bçrner, P. J. Goldsmith, S. Woodward, J. Gimeno, S. Gladiali,
D. Ramazzotti, Chem. Commun. 2000, 2433–2434; b) D. G. Gilling-
ham, A. H. Hoveyda, Angew. Chem. 2007, 119, 3934–3938; Angew.
Chem. Int. Ed. 2007, 46, 3860–3864; c) Y. Lee, K. Akiyama, D. G.
Gillingham, M. K. Brown, Amir H. Hoveyda, J. Am. Chem. Soc.
2008, 130, 446–447.
[8] R. M. Magid, Tetrahedron 1980, 36, 1901–1930.
[9] For Co, see: H. Yasui, K. Mizutani, H. Yorimitsu, K. Oshima, Tetra-
hedron 2006, 62, 1410–1415.
[10] For Ni, see: a) G. Consiglio, O. Piccolo, L. Roncetti, F. Morandini,
Tetrahedron 1986, 42, 2043–2053; b) T. Hiyama, N. Wakasa, Tetrahe-
dron Lett. 1985, 26, 3259–3262; c) E. Gomez-Bengoa, N. M. Heron,
M. T. Didiuk, C. A. Luchaco, A. H. Hoveyda, J. Am. Chem. Soc.
1998, 120, 7649–7650; d) U. Nagel, H. G. Nedden, Inorg. Chim.
Acta 1998, 269, 34–42; e) K.-G. Chung, Y. Miyake, S. Uemura, J.
Chem. Soc. Perkin Trans. 1 2000, 15–18 and K.-G. Chung, Y.
Miyake, S. Uemura, J. Chem. Soc. Perkin Trans. 1 2000, 2725–2729;
f) S. Son, G. C. Fu, J. Am. Chem. Soc. 2008, 130, 2756–2757.
[11] For Ti, see: T. Mukaiyama, H. Ishikawa, Chem. Lett. 1974, 1077–
1078.
[12] For Pd, see: a) J. C. Fiaud, L. Aribi-Zouioueche, J. Organomet.
Chem. 1985, 295, 383–387; b) M. E. Hoke, M.-R. Brescia, S. Bogac-
zyk, P. DeShong, B. W. King, M. T. Crimmins, J. Org. Chem. 2002,
67, 328–335.
[13] For Rh, see: M. Lautens, M. C. Dockendorff, K. Fagnou, M. Malicki,
Org. Lett. 2002, 4, 1311–1314.
[14] For Cu, see: a) S. Tominaga, Y. Oi, T. Kato, D. K. An, S. Okamoto,
Tetrahedron Lett. 2004, 45, 5585–5588; b) U. Piarulli, P. Daubos, C.
Claverie, C. Monti, C. Gennari, Eur. J. Org. Chem. 2005, 895–906;
c) M. A. Kacprzynski, T. L. May, S. A. Kazane, A. H. Hoveyda,
Angew. Chem. 2007, 119, 4638–4642; Angew. Chem. Int. Ed. 2007,
46, 4554–4558.
[15] P. A. Evans, D. Uraguchi, J. Am. Chem. Soc. 2003, 125, 7158–7159.
[16] A. Alexakis, S. E. Hajjaji, D. Polet, X. Rathgeb, Org. Lett. 2007, 9,
3393–3395.
[37] For other enantioselective syntheses of sertraline, see: a) E. J. Corey,
T. J. Gant, Tetrahedron Lett. 1994, 35, 5373–5376; b) H. M. L.
Davies, D. G. Stafford, T. Hansen, Org. Lett. 1999, 1, 233–236 and
references therein.
[38] S. B. Garber, J. S. Kingsbury, B. L. Gray, A. H. Hoveyda, J. Am.
Chem. Soc. 2000, 122, 8168–8179.
[39] M. P. Rainka, Y. Aye, S. L. Buchwald, Proc. Natl. Acad. Sci. USA
2004, 101, 5821–5823; to prevent the occurrence of racemization,
we avoided using a palladium-catalyzed hydrogenation reaction, for
details, see: C. Fehr, J. Galindo, I. Farris, A. Cuenca, Helv. Chim.
Acta 2004, 87, 1737–1747.
[40] H. B. Kagan, J. C. Fiaud in Topics in Stereochemistry, Vol. 18, Wiley,
New York, 1988, p. 249.
[41] a) G. Consiglio, F. Morandini, O. Piccolo, J. Chem. Soc. Chem.
Commun.. 1983, 112–114; b) K. H. Shukla, D. J. Boehmler, S. Boga-
cyzk, B. R. Duvall, W. A. Peterson, W. T. McElroy, P. DeShong, Org.
Lett. 2006, 8, 4183–4186; c) F. Menard, T. M. Chapman, C. Docken-
dorff, M. Lautens, Org. Lett. 2006, 8, 4569–4572.
[42] C. A. Welter, O. Koch, G. Lipowsky, G. Helmchen, Chem. Commun.
2004, 896–897.
[43] a) B. M. Trost, C. B. Lee, J. Am. Chem. Soc. 2001, 123, 3671–3686;
b) B. M. Trost, C. B. Lee, J. Am. Chem. Soc. 2001, 123, 3687–3696.
[44] B. Bartels, C. Garcia-Yebra, F. Rominger, G. Helmchen, Eur. J.
Inorg. Chem. 2002, 2569–2586.
[45] S. Bouquillon, J. Muzart, Eur. J. Org. Chem. 2001, 3301–3305.
Chem. Eur. J. 2009, 15, 1205 – 1216
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1215

A. Alexakis et al.
[46] C. A. Falciola, A. Alexakis, Angew. Chem. 2007, 119, 2673–2676;
Angew. Chem. Int. Ed. 2007, 46, 2619–2622.
[49] A. Alexakis, J. Vastra, J. Burton, C. Benhaim, P. Mangeney, Tetrahe-
dron Lett. 1998, 39, 7869–7872.
[50] K. Tissot-Croset, PhD Dissertation, University of Geneva (Switzer-
land), 2005.
[51] J. Lehmann, G. C. Lloyd-Jones, Tetrahedron 1995, 51, 8863–8874.
[47] B. L. Feringa, M. Pineschi, L. A. Arnold, R. Imbos, A. H. M. De V-
ries, Angew. Chem. 1997, 109, 2733–2736; Angew. Chem. Int. Ed.
Engl. 1997, 36, 2620–2623.
[48] A. Alexakis, D. Polet, S. Rosset, S. March, J. Org. Chem. 2004, 69,
5660–5667.
Received: September 12, 2008
Published online: December 15, 2008
1216
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1205 – 1216