One-pot formation of allylic chlorides from carbonyl derivatives

Article abstract of DOI:10.1021/ol802026u

(Chemical Equation Presented) An efficient, one-pot method for the conversion of carbonyl electrophiles to allylic chlorides has been developed, by activating magnesium alkoxides in situ using TiCl₄.

Full text of DOI:10.1021/ol802026u

ORGANIC
LETTERS
2008
Vol. 10, No. 21
4919-4922
One-Pot Formation of Allylic Chlorides
from Carbonyl Derivatives
Matthew J. Fuchter*,† and Jean-Noe¨l Levy
Department of Chemistry, Imperial College London, London SW7 2AZ
m.fuchter@imperial.ac.uk
Received August 29, 2008
ABSTRACT
An efficient, one-pot method for the conversion of carbonyl electrophiles to allylic chlorides has been developed, by activating magnesium
alkoxides in situ using TiCl4.
The general use of carbocations as synthetic intermediates
is limited by the strongly acidic conditions required for their
Scheme 1. C-O Bond Activation via Alkoxides
generation.1 Indeed, despite the relative stability of benzylic,
allylic, and propargylic carbocations,2 harsh conditions are
often required for their formation. During our recent studies
on the reactions of allylic alcohols, we wished to initiate
the formation of allylic cations under mild, nonacidic reaction
conditions. Since our desired allylic alcohol precursors could
readily be prepared from addition of a vinylic organometallic
reagent to a readily available benzaldehyde derivative 1, we
wondered whether it would be possible to affect direct
ꢀ-heterolytic cleavage of the C-O bond in situ, formally
eliminating a metal oxide as the leaving group (Scheme 1).
The fate of the subsequent carbocation 3 could be envisaged
to follow several pathways: (A) anion capture; (B) annulation
(to give an indene); (C) addition of an added nucleophile.
cations captured by allyl silanes.3 Related work by Murai
and co-workers has documented C-S bond cleavage.4 For
our cascade sequence however, we wished to focus on the
use of commercially available vinyl magnesium halides as
nucleophiles and hence the C-O cleavage of magnesium
alkoxides, a situation which was largely undocumented. In
one study regarding the synthesis of indene derivatives,
Tolbert had demonstrated that isolated, heavily substituted
magnesium alkoxides (generated upon addition of phenyl-
Recently, elegant work by Kabalka and co-workers has
demonstrated the ability of lithium alkoxides to undergo
C-O bond cleavage upon treatment with several Lewis acids
(most successfully BCl3 and FeCl3), with the subsequent
† Current address: Department of Pharmaceutical and Biological Chem-
istry, School of Pharmacy, University of London, 29-39 Brunswick Square,
London WC1N 1AX.
(1) Angel, S. R.; Louie, M. S. J. Org. Chem. 1991, 56, 2853–2866.
(2) (a) Kirmse, W.; Kund, K.; Ritzer, E.; Dorigo, A. E.; Houk, K. N.
J. Am. Chem. Soc. 1986, 108, 6045–6046. (b) Richey, H. G.; Philips, J. C.;
Rennick, L. E. J. Am. Chem. Soc. 1965, 87, 1381–1382. (c) Olah, G. A.;
Spear, R. J.; Westerman, P. W.; Denis, J. M. J. Am. Chem. Soc. 1974, 96,
5855–5959. (d) Pittman, C. U.; Wllmen, G.; Fojtasek, J. E.; Kispert, L. D.
J. Phys. Chem. 1975, 79, 2443–2446.
(3) (a) Kabalka, G. W.; Yao, M.-L.; Borella, S. J. Am. Chem. Soc. 2006,
128, 11320–11321. (b) Kabalka, G. W.; Yao, M.-L.; Borella, S.; Goins,
L. K. Organometallics 2007, 26, 4112–4114, and references therein.
(4) Murai, T.; Asai, F. J. Am. Chem. Soc. 2007, 129, 780.
10.1021/ol802026u CCC: $40.75
Published on Web 10/04/2008
2008 American Chemical Society

magnesium bromide to R,ꢀ-unsaturated ketones) undergo
ꢀ-heterolytic cleavage when pyrolized at 160 °C under
vacuum, to give indenes in moderate yield.5 Several problems
exist with this strategy as a synthetic method, however,
including to the need to isolate the moisture- and air-sensitive
crystalline magnesium alkoxides prior to reaction, the high
temperatures required, and the need for significant substitu-
tion to avoid side reactions such as elimination and hydride
abstraction. We reasoned that an alternative metal alkoxide
may mediate the correct reactivity profile to allow mild, low-
temperature ꢀ-heterolytic cleavage to form allylic cations
directly from benzaldehydes. With this in mind, we selected
commercially inexpensive titanium tetrachloride as a reaction
partner, in light of its strong dehydrating capabilities6 and
precent for initating C-O bond cleavage, for example in
Reetz’s pinoeering work on the geminal dimethylation of
ketones.7 To our delight, exposure of the magnesium
alkoxide 10 to titanium tetrachloride at low temperature,
followed by warming to room temperature, resulted in the
formation of (E)-allylic chloride 11 (see Scheme 2). The
form the desired allylic chloride 4. One complication of this
approach, however, could be the propensity of the Grignard
reagent to attack the electrophilic titanium center, leading
to titanium ate complexes,11 and mediating homocoupling
of the Grignard reagent.12 Alternatively, addition of vinyl-
magnesium chloride to benzaldehyde 1 should give the
corresponding magnesium alkoxide, which can undergo
metathesis upon addition of titanium tetrachloride to give
the titanium alkoxide. Subsequent elimination should form
allylic chloride 4. To find the optimum conditions, both
experimental methods were investigated in parallel.
Addition of titanium tetrachloride (1-2 equiv) to benzal-
dehyde (7) at -78 °C, followed by addition of vinyl
magnesium chloride (1-2 equiv) and warming to room
temperature, gave moderate yields of the product 11. In
almost all cases however, residual benzaldehyde (7) was
observed. We attributed this to the attack of the Grignard
reagent onto the titanium center, initiating subsequent side
reactions such as homocoupling12 and reducing the equiva-
lents of the active vinyl nucleophile. To overcome this hurdle,
a two-step, one-pot procedure was developed whereby the
Grignard reagent was added slowly to benzaldehyde (7) at
ambient temperature. The mixture was then cooled to -78 °C
for the addition of titanium tetrachloride. The low temper-
ature was simply to control the exotherm of the addition.
Warming to ambient temperature and then stirring overnight
resulted in efficient production of (E)-allylic chloride 11.
Alternatively, by heating to reflux, the reaction was complete
within 20 min. Under optimized conditions, the stoichiometry
of titanium tetrachloride was lowered to 0.5, and the product
was isolated in 77% yield (Scheme 2).13
Scheme 2
.
One-Pot Formation of (E)-Allylic Chloride C-O
Bond Activation via Alkoxides
Using the developed conditions, the scope of the reaction
was surveyed varying the substitution pattern on the aromatic
ring (Table 1). The reaction readily tolerated heavier
substitution (entries 1 and 2), the use of electron-rich
(11) (a) Kulinkovich, O. G.; Kananovich, D. G. Eur. J. Org. Chem.
2007, 212, 1–2132. (b) Mahrwald, R. Tetrahedron 1995, 51, 9015–9022.
(12) Inoue, A.; Kitagawa, K.; Shinokubo, H.; Oshima, K. Tetrahedron
2000, 56, 9601–9605.
allylic chloride products are important substrates in a variety
of reactions including metal-catalyzed allylation reactions.8
For example, they have been recently been reported as
substrates in nickel-catalyzed Negishi reactions9 and in
stereospecific zirconium-mediated SN2′ substutions.10 We
therefore decided to examine the scope of this reaction
further.
From the outset, the ideal order of addition of the reagents
was unclear. Addition of titanium tetrachloride to benzal-
dehyde 1 should increase the electrophilicity of the carbonyl
carbon following Lewis acid coordination. This should
facilitate the low-temperature nucleophilic addition of vi-
nylmagnesium chloride to give the corresponding titanium
alkoxide (see 2, MXn ) ClTiXn), which should eliminate to
(13) Standard procedure for the one-pot conversion of benzaldehyde
derivatives to (E)-allylic chlorides. (E)-(3-Chloroprop-1-enyl)benzene (11).
Vinylmagnesium chloride (1.2 mL, 1.6 M in THF, 1.90 mmol) was added
slowly to benzaldehyde (0.2 mL, 1.88 mmol) in dry THF (5 mL) at ambient
temperature. The mixture was stirred and monitored by TLC. After full
consumption of benzaldehyde, the mixture was cooled to -78 °C, and TiCl4
(0.1 mL, 0.95 mmol) was added. The reaction was allowed to rise to ambient
temperature, before warming to 80 °C. The reaction was stirred for 20 min
at 80 °C and then quenched with water (10 mL). EtOAc (10 mL) was added
and the resulting phases separated. The organic layer was washed with water
and the aqueous layer extracted EtOAc (3 × 10 mL). The combined organic
phase was dried over anhydrous MgSO4, and the solvent was evaporated
under reduced pressure. Flash column chromatography (SiO2, EtOAc:
Hexane, 1:4) afforded allylic chloride 11 (218 mg, 77%) as a yellow oil:
Rf ) 0.4 (SiO2, EtOAc:Hexane, 1:4); 1H NMR (400 MHz, CDCl3) d
7.38-7.41 (m, 2H), 7.31-7.35 (m, 2H), 7.26-7.29 (m, 1H), 6.66 (d, J )
15.6 Hz, 1H), 6.32 (dt, J ) 15.6, 7.2 Hz, 1H), 4.25 (dd, J ) 7.2, 1.2 Hz,
2H); 13C NMR (100 MHz, CDCl3) d 135.5, 134.2, 128.6, 128.3, 126.7,
124.9, 45.4. (See Supporting Information for full experimental and
characterization data.)
(5) Tolbert, L. M. J. Org. Chem. 1979, 44, 4584–4588.
(6) Karunakar, G. V.; Periasamy, M. J. Org. Chem. 2006, 71, 7463–
7466, and references therein.
(7) Reetz, M. T. Top. Curr. Chem. 1982, 106, 1.
(14) Bartoli, G.; Palmieri, G.; Bosco, M.; Dalpozzo, R. Tetrahedron
Lett. 1989, 30, 2129–2132.
(8) For reviews, see: (a) Pfaltz, A.; Lautens, M. in ComprehensiVe
Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A.; Yamamoto, H., Eds.;
Springer: New York, 1999; Vol. 2. Chapter 24. (b) Trost, B. M.; Van
Vranken, D. L. Chem. ReV. 1996, 96, 395–422.
(15) Periasamy, M.; Jayakumar, K. N.; Bharathi, P. J. Org. Chem. 2000,
65, 3548–3550.
(16) Kharasch, M. S.; Weinhouse, S. J. Org. Chem. 1936, 1, 209–230.
(17) Georgoulis, C.; Gross, B.; Ziegler, J.-C. C. R. Acad. Sci. 1971,
273, 378–381.
(9) Sunghee, S.; Fu, G. C. J. Am. Chem. Soc. 2008, 130, 2756–2757.
(10) Fox, R. J.; Lalic, G.; Bergman, R. G. J. Am. Chem. Soc. 2007,
129, 14144–14145.
(18) Flynn, A. B.; Ogilvie, W. W. Chem. ReV. 2007, 107, 4698–4745.
4920
Org. Lett., Vol. 10, No. 21, 2008

Table 1. Synthetic Scope of the Methodology
Table 2. Further Synthetic Scope of the Methodology
entry
R1
R2
R3
substrate product yield (%)
1
2
3
4
5
6
7
8
Me
Me
OMe
H
H
H
H
H
H
OMe
H
H
H
H
H
Me
H
H
CN
CO2Me
H
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
67
73
95a
59
78b
76
0
NO2
H
NMe2
0
a Product isolated crude (ca. 95% pure) as unstable to SiO2
chromatography. b Reaction performed in two steps. Yield refers to second
step.
aromatics (entries 3 and 4), and the use of electron-poor
aromatics (entries 5 and 6), with the products isolated in
moderate to excellent yield. Nitro-aromatic substrates (entry
7) proved problematic, presumably due to the known
reactivity of nitro groups with vinyl magnesium chloride,14
whereas amino substrates (entry 8) also gave extensive
decomposition due to the known oxidative reactions of
aniline derivatives with TiCl4.15
a Reagents and conditions: TiCl4 (0.5 equiv), THF, -78 to 80 °C.
b Product isolated crude (>95% pure) as elimination occurred upon
purification. c Reaction time, 36 h.
To extend this methodology further, we investigated the
use of more heavily substituted Grignard reagents as nu-
cleophiles as well as the use of both aromatic aldehydes and
ketones, and aliphatic aldehydes as substrates. Unfortunately,
using only 1 equiv of more heavily substituted Grignard
reagents, such as commercially available 1-methyl-1-prope-
nylmagnesium bromide, we consistently saw competing
reduction of our carbonyl substrate.16 The use of additives
such as HMPA17 failed to combat this shortcoming. Increas-
ing the number of equivalents of the Grignard reagent
resulted in the minimization of this reduction side-reaction
but was not compatible for our one-pot system. For this
reason, we synthesized a variety of allylic alcohols from the
corresponding bulky Grignard reagents and aldehydes or
ketones (see Supporting Information) and then used these
in our developed reaction, regenerating the magnesium
alkoxides via deprotonation using iso-propyl magnesium
chloride (Table 2).
In general, substitution patterns were well tolerated.
Substrates bearing tertiary alkoxide centers (entries 1 and
2) readily underwent the reaction, and impressively, substrate
30 (entry 2) resulted in formation of a tetrasubstituted double
bond, which are known to be challenging synthetic targets.18
Monosubstitution of the terminal alkene position posed no
problem (entry 3), and the product 38 was isolated in
excellent yield. However, in general, for the more heavily
substituted products, elimination of HCl to give an olefin
was an issue and precluded further purification of 38 (see
Supporting Information). Disubstitution of the terminal
alkene resulted in unavoidable elimination (entries 4 and 5),
although the corresponding dienes were isolated in good yield
(with product 40 isolated as a mixture of regioisomers). To
our delight, aliphatic systems also proved applicable to the
rearrangement (entries 6 and 7), albeit with longer reaction
times, and the products were isolated in moderate yield.
Although the development of this methodology demon-
strated the ability of TiCl4 to mediate the C-O bond cleavage
of magnesium alkoxides, it was still unclear whether het-
erolytic C-O bond cleavage to give a free allylic cation 43
was occurring, with subsequent capture of the most stable
conformer of this species,5 to produce an (E)-allylic chloride
selectively or whether the reaction was proceeding through
a closed, six-membered transition state 44, either somewhat
concerted19 or SNi′ (Scheme 3). We therefore initiated a
number of experiments to determine the mechanistic path-
way.
Kabalka and others have detailed the ability of allyl
trimethylsilane to capture cationic intermediates;3 however,
its presence under our reaction protocol resulted in no such
capture, and the expected allylic chloride was the only
reaction product. Interestingly, we were also unable to
scavenge and free chloride anions from solution using a
variety of silver salts. Conversely with analogous substrates
but using either lithium alkoxides or FeCl3 or BCl3 instead
of TiCl4 resulted in the formation of the same products albeit
with a lower degree of reactivity. These conditions have been
(19) For example: Kolundzic, F.; Micalizio, G. C. J. Am. Chem. Soc.
2007, 129, 15112–15113.
(20) (a) Pocker, Y.; Hill, M. J. J. Am. Chem. Soc. 1969, 91, 3243–
3248. (b) Kenyon, J.; Partridge, S. M.; Phillips, H. J. Chem. Soc. 1937,
207–218.
Org. Lett., Vol. 10, No. 21, 2008
4921

43 (Scheme 3, path A) is occurring in our developed reaction.
However, it is germane to state that since the chromatogram
from the HPLC indicated the presence of contaminants (i.e.,
diene) arising from HCl elimination we cannot rule out the
possibility that product 38 racemizes on the HPLC column.
Unfortunately, attempts at alternative chiral analyses, such
as the use of chiral shift reagents, or chiral GC failed to
resolve the enantiomers of 38. Flash chromatography resulted
in both elimination and decomposition, and the average crude
optical rotation of the product was 4 degrees.
Scheme 3. Potential Reaction Mechanisms
In conclusion, we have developed a novel method of
converting either allylic alcohols or benzaldehyde derivatives
into (E)-allylic chlorides via TiCl4-mediated C-O bond
activation of magnesium alkoxides. Initial mechanistic studies
have indicated that, as in the related systems of Kabalka and
co-workers,3 this reaction appears to proceed via cationic
intermediates and thus contributes to the notion that Lewis
acids can be used to generate cations from alkoxides under
mild, nonacidic conditions.
previously demonstrated by Kabalka and co-workers,3 to
react via a cationic pathway.
As a more conclusive probe, we prepared a chiral alkoxide
derived from alcohol 4520 and measured the degree of
chirality transfer upon reaction (Scheme 4). No transfer of
Acknowledgment. We thank the School of Pharmacy,
University of London and Imperial College London, for
financial support and Professor A. G. M. Barrett, FRS
FMedSci, and Mr S. Kroll (Imperial College London) for
their assistance in obtaining MS data.
Scheme 4. Mechanistic Studies
Supporting Information Available: Experimental pro-
cedures and tabulated spectral data and scanned images of
1H and 13C NMR spectra for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
chiral information (75% ee to 0% ee) was observed upon
reaction, as judged by chiral HPLC. This suggets that
heterolytic C-O bond cleavage to give a free allylic cation
OL802026U
4922
Org. Lett., Vol. 10, No. 21, 2008