A highly regioselective salt-free iron-catalyzed allylic alkylation

Article abstract of DOI:10.1002/anie.200503274

(Chemical Equation Presented) Ironing out the kinks: A highly regioselective allylic alkylation can be performed in the presence of catalytic amounts of an iron(-II) complex and triphenylphosphane (see scheme; EWG = electron-withdrawing group). Allyl carbonates and pronucleophiles are coupled in high yields with a high regioselectivity comparable only with that obtained with Rh catalysts. The reaction is broadly applicable and does not require external base.

Full text of DOI:10.1002/anie.200503274

Angewandte
Chemie
Allylic Alkylation
DOI: 10.1002/anie.200503274
A Highly Regioselective Salt-Free Iron-Catalyzed
Allylic Alkylation**
Bernd Plietker*
Scheme 1. Allylic alkylation through s-allyl (k ,k @k ,k )or p-allyl metal mecha-
1
4
2
3
Transition-metal-catalyzed reactions belong to the most
important tools in synthetic organic chemistry. Over the
past 30 years the productive collaboration between organic
and inorganic chemists has led to a dynamic evolution in this
field and, as a consequence, to the development of a variety of
nism (k ,k @k ,k ). EWG=electron-withdrawing group.
2 3 4 5
leading to the observed regioselective CÀC bond formation in
[
13a]
the product V (s-allyl metal mechanism).
[1]
[15]
[16]
new transformations. Apart from the invention of new
reactions, the use of only small amounts of a catalyst is
advantageous with regard to environmental aspects.
Although the amount of catalyst employed is usually very
low the high price of these compounds often prevents their
application on a large scale. Hence, the search for new
catalysts based on inexpensive nontoxic metals such as iron
Roustan et al. and Xu and Zhou reported a regiose-
ÀII
lective, allylic alkylation in the presence of the formal Fe
complex [Bu N][Fe(CO) (NO)]. Although from an econom-
4
3
ical and ecological point of view this catalyst represents an
interesting alternative to the commonly used transition-metal
complexes, this catalytic process was not studied further
owing to the high catalyst loading (25 mol%) and the need for
[
2]
[17,18]
has gained increasing attention in the past years.
a toxic CO atmosphere.
Allylic alkylation is among the most successful catalytic
Within the course of the synthesis of a natural product, we
sought a regioselective allylic alkylation method in which the
new CÀC bond is formed at the carbon atom that bears the
[
3]
CÀC bond formations in organic chemistry. A variety of
transition metals can be used for the selective reaction
between a C nucleophile and allylic activated substrates.
Detailed investigations, especially in the field of palladium
catalysis, have made allylic alkylation a versatile synthetic
nucleofuge in the starting material. Although almost ignored,
ÀII
CÀC bond formation in the presence of a Fe complex
appeared remarkable to us with regard to both practicability
[
4]
[14,15]
tool in preparative organic chemistry. Although palladium is
the predominant metal in this field, various procedures
and stability of the catalyst.
Herein we summarize the
results of our investigations based on the pioneering work of
Roustan and Xu that have led to the development of an
efficient, regioselective, salt-free, iron-catalyzed allylic alky-
[
5]
[6]
[7]
[8]
involving ruthenium,
iridium,
nickel,
tungsten,
[
9]
[10]
copper, or molybdenum catalysts have been published.
In most cases, intermediate p-allyl metal complexes are
formed that allow asymmetric allylic alkylation through
addition of chiral ligands. However, the reaction of unsym-
metrically substituted substrates often leads to the formation
[
19]
lation. The scope of the reaction for different pronucleo-
[
20]
philes and allyl carbonates is discussed below.
The primary goal of our initial investigations in this field
was the replacement of the toxic CO atmosphere. The low
conversion in the absence of CO indicated a deactivation of
the catalyst within the first catalytic cycle (Table 1, entry 1).
Based on the assumption that the CO atmosphere should
suppress deactivating ligand-exchange processes at the metal
center, various ligands were tested in order to replace the CO
atmosphere (Table 1). A probable coordination at the metal
center could maintain the catalyst activity, thus taking over
the function of the CO atmosphere. Surprisingly, triphenyl-
phosphane, a monodentate basic ligand, showed optimum
[
11,12]
of regioisomers.
To date, a catalytic regioselective allylic alkylation in
which the new CÀC bond is formed at the carbon atom that
bears the leaving group in the starting material was only
observed in the presence of rhodium catalysts by the groups
[
13]
[14]
of Evans and Martin. To favor such a reaction, a fast
alkylation of II (k ,k @ k ,k ; Scheme 1) might prevent the
1
4
2
3
formation of the undesired p-allyl metal intermediate III, thus
[
21]
results (Table 1, entry 3). Furthermore, the separate depro-
tonation of the pronucleophile proved to be unnecessary, thus
allowing a salt-free reaction procedure.
[
*] Priv.-Doz. Dr. B. Plietker
Organische Chemie II, FB Chemie
Universität Dortmund
Otto-Hahn-Strasse 6, 44221 Dortmund (Germany)
Fax: (+49)231-755-3884
E-mail: bernd.plietker@uni-dortmund.de
Subsequent investigations on the influence of the carbon-
ate structure indicated an isobutyl carbonate to be beneficial
for the reaction. The change in the carbonate structure
improved the yield significantly from 41% to 68%. A
subsequent optimization of the solvent indicated that strongly
coordinating solvents such as N,N-dimethylformamide
[
**] Financial support by the Fonds der Chemischen Industrie, the
Deutsche Forschungsgemeinschaft, the Otto-Röhm-Gedächtnis-
stiftung, and the Fonds zur Förderung des wissenschaftlichen
Nachwuchses im Fachbereich Chemie der Universität Dortmund is
gratefully acknowledged.
[
22]
increase the reactivity of the catalyst (Table 2, entry 8).
The moderate regioselectivity in the presence of acetonitrile
is surprising and suggests an alternative p-allyl metal mech-
anism (Table 2, entry 7). Investigations on the mechanism are
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 1469 –1473
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1469

Communications
[
a]
[a]
Table 1: Ligand effect in the allylic alkylation of 1.
Table 2: Solvent effect in the allylic alkylation.
[
b]
[c]
Entry
Solvent
2/3
Conversion [%]
1
2
3
4
5
6
7
8
9
THF
DME
96:4
92:8
n.d.
68
43
12
–
39
46
60
95
82
[
d]
acetone
CH Cl
n.d.
2
2
[
b]
[c]
toluene
DMSO
97:3
96:4
58:42
98:2
97:3
Entry
Ligand
2/3
Conversion [%]
[
d]
1
2
3
4
5
6
7
8
–
n.d.
n.d.
8
6
41
26
12
8
CH CN
3
[
d]
pyridine
PPh3
PBu3
DMF
NMP
[
d]
96:4
92:8
n.d.
n.d.
n.d.
n.d.
[a] All reaction were carried out in the presence of the Fe catalyst
P(OEt)
3
(10 mol%)and PPh 3 under otherwise identical conditions as described
[
e]
dppe
in Table 1. [b] Determined by GC. [c] Determined by GC relative to
undecane as internal standard.
[
e]
dppp
11
13
[
e]
dppf
[a] All reactions were performed on a 1-mmol scale under argon in the
presence of the Fe catalyst (10 mol%), ligand (10 mol%), and
pronucleophile (2 equiv)in absolute THF (5 mL)and stopped after
1
2 h. [b] Determined by GC. [c] Determined by GC relative to undecane
as internal standard. [d] Not determined. [e] dppe=Ethane-1,2-diylbis-
diphenylphosphane), dppp=propane-1,3-diylbis(diphenylphosphane),
(
dppf=1,1’-bis(diphenylphosphanyl)ferrocene.
Scheme 2. Iron-catalyzed allylic alkylation. Reagents and conditions:
substrate (1 mmol), pronucleophile (2 mmol), [Bu N][Fe(CO) (NO)]
4
3
currently being carried out in our laboratories. An increase in
the substrate concentration from 0.2 to 1molL allowed a
À1
(2.5 mol%), PPh
3
(3 mol%), DMF (1 mL), 808C, 24 h.
further decrease in the catalyst loading to as little as
2
.5 mol% while maintaining the regioselectivity. The opti-
proceeds in good yields (Table 3, entry 6). However, this CÀC
bond formation is not as regioselective as the formation of
only one quaternary center. Apparently, the steric repulsion
between the substituents at the allyl fragment and the
nucleophile leads to a lower selectivity. However, the
branched product is still the major product.
mized conditions are shown in Scheme 2.
The iron-catalyzed allylic alkylation is applicable to a
variety of different C nucleophiles (Table 3). Nitrile-contain-
ing pronucleophiles proved to be most reactive. The presence
of only one CN group decreases the reaction time signifi-
cantly (Table 3, entries 3–6). The formation of two adjacent
quaternary carbon centers in the presence of the iron catalyst
The iron-catalyzed allylic alkylation is applicable to a
variety of differently substituted allyl carbonates (Table 4). In
[
a]
Table 3: Iron-catalyzed allylic alkylation of different pronucleophiles.
[
b]
[c]
Entry
1
Pronucleophile
Product A
t [h]
A/B
Yield [%]
5
6
2
7
24
98:2
96:4
81
2
3
24
12
71
74
8
9
92:8
4
5
10
12
11
13
12
12
94:6
94:6
79
92
6
14
15
12
87:13
81
[
a] All reactions were carried out on a 1-mmol scale in the presence of the Fe catalyst (2.5 mol%), PPh (3 mol%), and the pronucleophile (2 equiv) in
3
absolute DMF (1 mL)at 80 8C. [b] Determined by GC. [c] Yield of both regioisomers after isolation.
1
470
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1469 –1473

Angewandte
Chemie
[
a]
Table 4: Iron-catalyzed allylic alkylation of various carbonates.
[
b]
[c]
Entry
Substrate
Product A
t [h]
A:B
Yield [%]
1
2
16
18
17
19
12
48
93:7
–
84
53
3
20
21
48
–
69
4
5
6
22
24
26
23
25
27
24
36
24
98:2
96:4
98:2
78
76
78
7
28
29
48
92:8
56
8
9
30
30
3
48
24
93:7
61
65
31
84:16
[a] All reactions were performed as indicated in Table 3. [b] Determined by GC. [c] Combined yield of both regioisomers after isolation.
every case the substitution takes place at the carbon atom that
bears the original leaving group. The introduction of further
olefin substituents has a significant influence on the course of
the reaction. More substituents lead to longer reaction
[
23]
times. However, we were pleased to find that the regiose-
lectivity is not influenced (Table 4). Whereas aliphatic
substituents solely prolong the reaction times, a phenyl
substituent at C3 of the allyl fragment leads to a significantly
lower yield (Table 4, entries 2 and 7). The regioselectivity of
the reaction is defined exclusively by the substitution pattern
of the allyl carbonate (Table 4, entries 4/5 and 6/7). In
contrast, the nature of the pronucleophile does not influence
the regioselectivity (Table 3, entry 1and Table 4, entry 8;
Table 3, entry 6 and Table 4, entry 9).
Scheme 3. Postulated catalytic cycle for the iron-catalyzed allylic alkyla-
tion.
[24]
From the observed high regioselectivity in the alkylation it
can be concluded that, in contrast to most common transition-
metal-catalyzed allylic alkylations, no h -allyl intermediate is
[
23]
entries 1, 5, and 9). A preliminary mechanistic proposal is
shown in Scheme 3.
3
formed in this reaction. A selective double S 2’ addition both
of the metal VII (Scheme 3) in the first and of the nucleophile
X in the second step allows the observed high regioselectivity.
The proposed mechanism has stereochemical consequen-
ces. If indeed only a s-allyl metal species such as IX is formed
in situ, the allylic alkylation of enantiomerically enriched
carbonates should lead to the formation of enantiomerically
and regioisomerically enriched substitution products with
retention of the configuration. To test this hypothesis (S)-allyl
carbonate 33 was subjected to the alkylation conditions. The
N
An oxidative addition through an S 2 mechanism appears to
N
be unlikely as both reaction time and conversion rate in the
alkylation of primary allyl carbonates clearly depends on the
substitution pattern at C3 of the allyl fragment (Table 4,
Angew. Chem. Int. Ed. 2006, 45, 1469 –1473
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1471

Communications
enantiomeric ratios of the starting material and of the product
obtained which was purified by flash column chromatography
isohexane/diethyl ether, 5:1). Yield: 162 mg (0.81 mmol, 81%).
(
were analyzed by means of chiral HPLC (Scheme 4). The
enantiomeric excess decreased to a certain extent; however,
the S-configured malonic ester derivative 34 was formed as
Received: September 15, 2005
Revised: October 26, 2005
Published online: January 27, 2006
Keywords: alkylation · carboxylic acids · iron · nucleophiles ·
.
regioselectivity
[
1] a) Transition Metals for Organic Synthesis, Vol. 1 and 2 (Eds.: M.
Beller, C. Bolm), Wiley-VCH, Weinheim, 1998; b) Comprehen-
sive Asymmetric Catalysis I–III (Eds.: E. N. Jacobsen, A. Pfaltz,
H. Yamamoto), Springer, Berlin, 1999.
Scheme 4. Iron-catalyzed allylic alkylation of enantiopure carbonate
(S)-33.
[
2] For a review on iron-catalyzed cross-coupling, see: A. Fürstner,
R. Martin, Chem. Lett. 2005, 34, 624.
[
[
3] B. M. Trost, M. L. Crawley, Chem. Rev. 2003, 103, 2921.
4] J. Tsuji in Palladium Reagents and Catalysts (Ed.: J. Tsuji), Wiley,
New York, 1995, p. 290.
the main product. The lower ee values are probably the result
of epimerization of the starting material under the reaction
conditions. This hypothesis was proven by a control experi-
ment in which (S)-33 was subjected to the reaction conditions
in the absence of the iron catalyst. After 48 h a decrease in the
enantiopurity to 83% ee was observed. Hence, at the current
state of research such an epimerization event might be
responsible for the observed loss in enantioselectivity. How-
ever, this test reaction serves as a first indication that a s-allyl
metal mechanism takes place in the iron-catalyzed allylic
alkylation.
[5] a) Y. Morisaki, T. Kondo, T.-A. Misudo, Organometallics 1999,
18, 4742; b) B. M. Trost, P. Fraisse, Z. Ball Angew. Chem. 2002,
114, 1101; Angew. Chem. Int. Ed. 2002, 41, 1059.
[
6] a) R. Takeuchi, N. Shiga, Org. Lett. 1999, 1, 265; b) B. Bartels, G.
Helmchen, Chem. Commun. 1999, 741.
7] a) K.-G. Chung, Y. Miyake, S. Uemura, J. Chem. Soc. Perkin
Trans. 1 2000, 15; b) M. T. Didiuk, J. P. Morken, A. H. Hoveyda,
J. Am. Chem. Soc. 1995, 117, 7273.
[
[8] a) G. C. Lloyd-Jones, A. Pfaltz, Angew. Chem. 1995, 107, 534;
Angew. Chem. Int. Ed. Engl. 1995, 34, 462; b) B. M. Trost, M. -
Hung, J. Am. Chem. Soc. 1987, 109, 2176; c) B. M. Trost, M.-H.
Hung, J. Am. Chem. Soc. 1983, 105, 7757.
In summary, we have described an efficient, salt-free,
regioselective iron-catalyzed allylic alkylation of allyl carbo-
nates. The addition of triphenylphosphane led to a significant
[
9] E. S. M. Persson, M. van Kaveren, D. M. Grove, J.-E. Bꢀckvall,
G. van Koten, Chem. Eur. J. 1995, 1, 351.
increase in reactivity of the [Fe(CO) ]-derived stable iron(Àii)
5
[10] B. M. Trost, K. Dogra, I. Hachiya, T. Emura, D. L. Hughes, S. W.
Krska, R. A. Reamer, M. Palucki, N. Yasuda, P. J. Reider,
Angew. Chem. 2002, 114, 2009; Angew. Chem. Int. Ed. 2002, 41,
complex [Bu N][Fe(CO) (NO)]. In the presence of the
4
3
catalyst (2.5 mol%) and PPh (3 mol%) a variety of different
3
1929.
allyl carbonates and pronucleophiles were coupled in good to
excellent yields. The use of a carbonate as leaving group
avoids a separate deprotonation of the pronucleophile by
external bases. Apart from the possible salt-free reaction
course, a fact that is especially attractive for applications on a
preparative scale, the high regioselectivity of the alkylation is
remarkable. The new CÀC bond is formed selectively at the
[
11] The use of ligands sometimes leads to a shift in the regioisomer
ratio towards the branched products; for example, see: J. W.
Fuller, J. C. Wilt, J. Parr, Org. Lett. 2004, 6, 1301.
12] The character of the metal influences the regiochemical course
of allylic alkylations. Ruthenium, molybdenum, rhodium, tung-
sten, and iridium catalysts often lead to the formation of the
branched product, regardless of the substitution pattern of the
starting material.
13] a) P. A. Evans, J. D. Nelson, J. Am. Chem. Soc. 1998, 120, 5581;
b) P. A. Evans, J. D. Nelson, Tetrahedron Lett. 1998, 39, 1725.
14] B. L. Ashfeld, K. A. Miller, S. F. Martin, Org. Lett. 2004, 6, 1321.
15] J. L. Roustan, J. Y. Merour, F. Houlihan, Tetrahedron Lett. 1979,
3721.
[
carbon atom in the allyl fragment that bears the leaving group
in the starting material. Comparable high regioselectivities
were previously only observed in the presence of rhodium
catalysts. Further investigations on the mechanism, stereo-
and regioselectivity aspects, and scope and limitations are
currently underway in our laboratories.
[
[
[
[16] a) Y. Xu, B. Zhou, J. Org. Chem. 1987, 52, 974; b) B. Zhou, Y.
Xu, J. Org. Chem. 1988, 53, 4421.
[
17] For reviews on the chemistry of tricarbonyl iron complexes, see:
a) M. Periasamy, C. Rameshkumar, U. Radhahrishnan, A.
Devasagayaraj, Curr. Sci. 2002, 71, 1307; b) H.-J. Knꢁlker,
Chem. Rev. 2000, 100, 2941.
Experimental Section
2
0
: A 10-mL Schlenk tube was charged with the iron catalyst (10.6 mg,
.025 mmol, 2.5 mol%) and PPh3 (7.3 mg, 0.03 mmol, 3 mol%),
[18] For a comprehensive study on iron-mediated allylic alkylations,
see: D. Enders, B. Jandeleit, S. von Berg, G. Raabe, J. Runsink,
Organometallics 2001, 20, 4312.
[19] For an interesting Fe-catalyzed allylic alkylation of allyl halides,
see: R. Martin, A. Fürstner, Angew. Chem. 2004, 116, 4045;
Angew. Chem. Int. Ed. 2004, 43, 3955.
which were then dissolved under argon in dry DMF (1mL) and
heated to 808C for 30 min. The mixture was cooled to room
temperature, and the allyl carbonate 4 (186 mg, 1 mmol) and dimethyl
malonate (232 mg, 2 mmol) were then added. The reaction mixture
was heated in the sealed Schlenk tube for 24 h at 808C. The mixture
was cooled to room temperature, and dichloromethane (20 mL) was
then added. The organic layer was extracted with water (2 ” 10 mL)
and dried over a mixture of Na SO and charcoal (1:1; ꢀ 5 g). After
[20] B. Plietker, patent pending (102005040752.8)
[21] A ligand exchange in the presence of P(OPh) was previously
3
reported: K. Itoh, S. Nakanishi, Y. Otsuji, Bull. Chem. Soc. Jpn.
1991, 64, 2965.
2
4
filtration and evaporation of the solvent a pale yellow liquid was
1
472
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1469 –1473

Angewandte
Chemie
[
[
[
22] The positive effect of N,N-dimethylformamide on the ligand
[
21]
exchange has been described.
23] A similar effect was observed by Evans and Nelson in rhodium-
[
13a]
catalyzed allylic alkylations.
24] The presented mechanism, including the molecular structures,
represents a model. Detailed spectroscopic investigations on the
structure of the active catalyst as well as on reactive intermedi-
ates are currently underway in our laboratories.
Angew. Chem. Int. Ed. 2006, 45, 1469 –1473
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1473