The Effect of Hydrogen Bonding on Allylic Alkylation and Isomerization Reactions in Ionic Liquids

Article abstract of DOI:10.1002/chem.200304895

Neutral allylic alkylation reactions, in which a base is generated in situ and which hence require no external bases, can significantly be retarded when carried out in the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate ([bmim][BF₄]). Evidence suggests that the base or base precursor enters into hydrogen bonding with the imidazolium cation and is thus made less readily available for deprotonation of pre-nucleophiles. However, the reaction proceeds well in the presence of stronger bases that are capable of deprotonation. Whilst the phenomenon of hydrogen bonding in ionic liquids can be detrimental to reactions such as allylic alkylation, it can be exploited to suppress unwanted allylic isomerization.

Full text of DOI:10.1002/chem.200304895

FULL PAPER
The Effect of Hydrogen Bonding on Allylic Alkylation and Isomerization
Reactions in Ionic Liquids
James Ross and Jianliang Xiao*[a]
Abstract: Neutral allylic alkylation re-
actions, in which a base is generated in
situ and which hence require no external
bases, can significantly be retarded when
carried out in the ionic liquid 1-butyl-3-
methylimidazolium tetrafluoroborate
([bmim][BF4]). Evidence suggests that
the base or base precursor enters into
hydrogen bonding with the imidazolium
cation and is thus made less readily
available for deprotonation of pre-nu-
cleophiles. However, the reaction pro-
ceeds well in the presence of stronger
bases that are capable of deprotonation.
Whilst the phenomenon of hydrogen
bonding in ionic liquids can be detri-
mental to reactions such as allylic alky-
lation, it can be exploited to suppress
unwanted allylic isomerization.
Keywords: allylation
¥ homogene-
ous catalysis ¥ hydrogen bonds ¥
ionic liquids ¥ palladium
past few years, little attention has been paid to the potential
effects of hydrogen bonding on catalyzed reactions in ionic
liquids.[7] Only recently have Diels Alder and nucleophilic
substitution reactions with no added catalysts been inves-
tigated in this context in imidazolium ionic liquids.[8] Follow-
ing on from our earlier studies into Pd-catalyzed reactions in
ionic liquids,[9] we have found that the capability of imidazo-
lium ionic liquids for hydrogen bonding can exert a remark-
able effect on neutral allylic alkylation reactions and,
interestingly, the effect can be harnessed to suppress Pd0-
catalyzed allylic isomerization,[10] a reaction that may diminish
the stereoselectivity of asymmetric allylic substitution.[11] The
details of our results are herein described.
Introduction
Hydrogen bonding embraces many important areas of
chemistry and biochemistry.[1] The effect of hydrogen bonding
on reaction chemistry in common molecular solvents is well
documented and understood.[1, 2] Hydrogen bonding in sol-
vents based on room-temperature ionic liquids is a relatively
new subject. In fact, the perception of hydrogen bonding in
imidazolium ionic liquids, the most extensively investigated
ionic liquids to date, was still controversial in the mid 1980s.[3]
Thanks to the pioneering studies of several research groups,[3]
it is now well established that imidazolium cations and their
associated anions form hydrogen bonds both in the solid state
and in solution.[4] The H2, H4 and H5 ring protons of the 1,3-
dialkylimidazolium cation can act as hydrogen bond donors
and interact with counteranions such as ClÀ, OTfÀ, and BF4 ,
À
Results and Discussion
which act as hydrogen-bond acceptors and can also enter into
hydrogen bonding with external hydrogen donors such as
H2O. Of the three imidazolium ring protons, the H2 proton
appears to form the strongest hydrogen bond. This feature has
been exploited for the assembly of a new class of anion
receptors based on imidazolium cations,[5] and has been
invoked to explain the selective transport/separation of
amines by membrane-supported imidazolium ionic liquids.[6]
Whilst the concept of hydrogen bonding in ionic liquids has
generally been accepted and explosive growth in research on
reaction chemistry in these solvents has been witnessed in the
Our previous investigation shows that Pd0-catalyzed allylic
alkylation or Tsuji Trost reactions with a variety of active
methylene compounds can be readily carried out in the ionic
liquid
1-butyl-3-methylimidazolium
tetrafluoroborate
([bmim][BF4]) under basic conditions.[9a,d] Similar results have
been reported from other laboratories.[12] In an effort to
determine the scope of reaction, the room-temperature allylic
alkylation of phenylallyl carbonate 1 with dimethyl malonate
2 catalyzed by Pd0 PAr3 was investigated in [bmim][BF4]
[Eq. (1)]. The catalytically active Pd0 species is expected to be
generated in situ from the starting Pd(OAc)2 upon reduction
with PAr3.[13] This is a neutral Tsuji Trost reaction; it requires
[a] Dr. J. Xiao, J. Ross
Leverhulme Centre for Innovative Catalysis
Department of Chemistry, University of Liverpool
Liverpool L69 7ZD (UK)
À
no external base, as decarboxylation of MeOCO2 , which
results from the oxidative addition of 1 to Pd0, generates CO2
Fax : (‡44)151 794-3589
E-mail: j.xiao@liv.ac.uk
and a strong base MeOÀ.[10, 11] Although methanol has a higher
4900
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304895
Chem. Eur. J. 2003, 9, 4900 4906

4900 4906
25 min (Table 1, entry 8). Even
more strikingly, the reaction in
THF was significantly inhibited
by the addition of a small
quantity of [bmim][BF4] and
the rate was progressively de-
creased by introducing more
[bmim][BF4] (Table 1, entries
9
11), indicating that some
key intermediate in the catalyt-
ic cycle is involved in a pre-
pKa than a dialkylimidazolium ion (29 versus ca. 24 in
DMSO),[14] the methoxide is expected to preferentially
deprotonate the malonate, which is about eight orders of
magnitude more acidic than the imidazolium cation.[15]
Furthermore, even if deprotonation of the solvent cations
took place, the so generated dialkylimidazol-2-ylidene would
readily deprotonate the malonate to give the required
nucleophile to attack the Pd allyl intermediate and complete
the catalytic cycle,[14b,c] unless it forms inactive palladium
complexes (vide infra).
After first confirming the reaction to be rapid in THF,
reaching complete conversion within 20 min,[16] we carried out
the same reaction in [bmim][BF4] under otherwise identical
conditions (Table 1, entries 1 5). Surprisingly, the reaction in
[bmim][BF4] was considerably slower, affording less than 1%
conversion after 1 h and complete conversion in a prolonged
time of about 30 h. PPh3 was found to be the best ligand out of
several phosphines tested in the ionic liquid. For example,
changing the ligand to the more ionic liquid-soluble and more
equilibrium with the imidazolium additive.
It was possible to accelerate the reaction by external bases.
This is clearly seen from the results obtained in the presence
of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) or OAcÀ (Ta-
ble 1, entries 12 and 13). However, the role of the two bases
must be different (vide infra); unlike the more basic DBU, the
acetate cannot deprotonate 2.[15, 17] A very revealing experi-
ment is the comparison of the reaction performed in
[bmim][BF4] with that in 1-butyl-2,3-dimethylimidazolium
tetrafluoroborate ([bdmim][BF4]), in which the H2 ring
proton is replaced with a methyl group (Table 1, entries 14
and 15). The reaction in the latter was considerably faster,
suggesting that the retarding effect of [bmim][BF4] relates to
its H2 proton.
The allylic alkylation of phenylallyl acetate 4 with methyl
nitroacetate 5 was similarly retarded when conducted in the
ionic liquid [Eq. (2)]. This is again a neutral reaction requiring
no external bases, because the OAcÀ ion generated in the
oxidative addition of 4 to Pd0 is basic enough to deprotonate 5
À
electron-rich P(4 MeOC6H5)3
under the same conditions gave
a negligible conversion after 1 h
and only 9% conversion after
5 h (Table 1, entries 6 and 7).
For comparison, the same li-
gand in THF brought about a
nearly complete reaction in
(pKa 8.0 in DMSO),[18] which is much more acidic than 2.
Indeed, the reaction in THF was complete within 0.5 h when
catalyzed by Pd0 PPh3 at 758C. Repeating this reaction in
[bmim][BF4], only 26% conversion was reached after 0.5 h
and complete conversion after an extended reaction time of
24 h.
According to the generally accepted mechanism for allylic
substitution, the oxidative addition of 1 or 4 to Pd0 PPh3
leads to two ionic species, MeOCO2À or OAcÀ and a Pd-allyl
cation [Eq. (1)], and as such should not be slowed down on
going from THF to an ionic medium. Therefore it is probably
the nucleophilic attack step that is affected in the ionic liquid.
Amatore, Jutand, and co-workers have recently shown that
the oxidative addition of allylic acetate to Pd0 is reversible,
and in the less polar THF, the resulting acetate anion and Pd
allyl cation form tight ion pairs rather than free ions as in
DMF.[19] The formation of such ion pairs could diminish the
positive charge on PdII and enhance the steric hindrance
around the metal atom, thus leading to a slower nucleophilic
attack. In the polar [bmim][BF4],[20] the formation of tight ion
Table 1. Effect of solvents on the neutral Tsuji Trost reaction between 1 and 2.[a]
Entry Solvent
Additive (mol%)[b] Ligand
Time [h] Conv[c] [%]
1
2
3
4
5
6
7
8
9
10
11
12
13
THF
-
-
-
-
-
-
-
-
PPh3
PPh3
PPh3
PPh3
0.33
1
5
16
30
1
100
< 1
38
[bmim][BF4]
[bmim][BF4]
[bmim][BF4]
[bmim][BF4]
[bmim][BF4]
[bmim][BF4]
THF
THF
THF
THF
75
PPh3
P(4-MeOC6H5)3
P(4-MeOC6H5)3
P(4-MeOC6H5)3 0.42
P(4-MeOC6H5)3 0.90
P(4-MeOC6H5)3 0.90
P(4-MeOC6H5)3 0.90
100[d]
< 1
9
5
99
63
46
33
100
71
89
[bmim][BF4] (4)
[bmim][BF4] (10)
[bmim][BF4] (20)
[bmim][BF4] DBU (200)
[bmim][BF4] [nBu4N][OAc] (100) P(4-MeOC6H5)3 0.83
P(4-MeOC6H5)3 0.50
P(4-MeOC6H5)3
PPh3
0.33
14[e] [bdmim][BF4] -
15[e] [bmim][BF4]
-
5
48
[a] All reactions were performed on a 1 mmol scale, 2 mol% Pd(OAc)2, and
8 mol% phosphine in [bmim][BF4] (2 mL) or THF (2 mL) at room temperature
unless otherwise indicated. [b] Relative to palladium catalyst. [c] Determined by
1H NMR spectroscopy of the crude reaction mixture. [d] Isolated yield for 3: 90%.
[e] At 508C.
Chem. Eur. J. 2003, 9, 4900 4906
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4901

J. Xiao and J. Ross
FULL PAPER
pairs between Pd allyl and acetate ions is less likely, as the
acetate would interact more favorably with the much more
abundant, less bulky solvent cations. Therefore, the sluggish
reaction in the ionic liquid could be due to the availability of
nucleophiles rather than a high barrier in the nucleophilic
attack step. This is of course consistent with the reaction being
fast when an external base is added, which could make the
deprotonated 2 more readily available (Table 1, entry 12).
In a very recent, detailed study, Welton and co-workers
reported that the nucleophilicity of halides in a noncatalyzed
SN2 reaction in [bmim][Tf2N] [Tf2N ˆ bis(trifluoromethylsul-
fonyl)amide] follows the order ClÀ < BrÀ < IÀ.[4h] This is
explained by the chloride forming the strongest hydrogen
bond to the imidazolium cations, in particular the H2 ring
protons, and so being made least available to undergo
nucleophilic attack. The observed inhibition of the neutral
Tsuji Trost reaction by [bmim][BF4] could be accounted for
in a similar manner by assuming that the MeOCO2À or OAcÀ
ions generated in the oxidative addition are strongly solvated
or ™trapped∫ by hydrogen bonding with the imidazolium
cations and are thus made unavailable to deprotonate the
hydrocarbon acids 2 or 5 to give the required nucleophile for
subsequent nucleophilic attack (Scheme 1). OAcÀ is known to
Figure 1. 1H NMR titration profile for addition of [bmim][BF4] (26
323 mm) to [nBu4N][OAc] (123 mm) in CDCl3 at 218C, the first point on
the chemical shift axis corresponding to 680 mm in [bmim][BF4].
clear, although it has bee suggested that chloride is surround-
ed by six imidazolium cations in related ionic liquids.[4h]
Titrating [bmim][BF4] with [MeOCO2][HDBU][21c] in
CDCl3 gave similar results (Figure 2). Although the change
in H2 chemical shift is less significant, indicating a weaker
interaction, it is clear that methyl carbonate forms hydrogen
Figure 2. 1H NMR titration profile for addition of [MeOCO2][HDBU]
(0 112 mm) to [bmim][BF4] (32.2 mm) in CDCl3 at 218C.
bonds to the imidazolium cations. Hence, one can reasonably
assume that it is this hydrogen bonding that prevents the
carbonate ions from decomposing readily to form the base
necessary for deprotonation of 2. A recent study by Amatore,
Jutand, and co-workers shows that the oxidative addition of
allylic carbonates to Pd0 is reversible and the resulting
carbonate anion does not decarboxylate as fast as previously
thought.[11] Hydrogen bonding should certainly enhance its
stability. Indeed, it is known that decarboxylation can
dramatically be decelerated by using dipolar protic sol-
vents.[24] In line with the hydrogen bonding assertion, the
imidazolium H2 ring proton moved immediately from d ˆ
Scheme 1. Proposed role of [bmim][BF4] as hydrogen-bond donor in
preventing efficient deprotonation of HNu, thus inhibiting nucleophilic
attack by NuÀ.
form a strong hydrogen bond to the H2 proton of imidazolium
ring in both solution and solid states.[21] To confirm this, a
1H NMR titration experiment was carried out, in which the
concentration of [nBu4N][OAc] was kept constant while that
of [bmim][BF4] was varied in CDCl3. The titration revealed
that the H2 proton chemical shift moved to lower field in a
sigmoid fashion with increasing acetate/bmim molar ratios
until a value of about 2, thereafter the H2 chemical shift was
little affected (Figure 1).[22] Only insignificant changes were
observed for the H4 and H5 ring protons. These observations
suggest that the H2 proton hydrogen bonds to OAcÀ and a 1:2
stoichiometric complex could be formed between the imida-
zolium and OAcÀ ions in CDCl3.[23] NMR spectroscopy has
1
8.6 ppm to d ˆ 10.4 ppm in the H NMR spectrum upon the
addition of Pd0 PPh3 to a NMR tube containing a 1/1 mixture
of 1 and [bmim][BF4] in CDCl3 [Pd(OAc)2, 10 mg; ratio of Pd/
PPh3/1 ˆ 1/4/1]. As aforementioned, methyl carbonate would
be generated from the oxidative addition of 1 to Pd0 under
these conditions. The major contribution to the downfield
shift comes probably from the acetate of Pd(OAc)2, however.
This experiment also indicates that the neutral allylic alkyla-
tion in [bmim][BF4] is not limited by the oxidative addition
step and supports the proposition that it is the nucleophilic
attack that is affected by the ionic liquids.
The notion that MeOCO2À could not function efficiently as
a base precursor due to hydrogen bonding explains the effects
of the added [bmim][BF4] on the reaction of 1 with 2 in THF.
The more imidazolium cations added, the easier the forma-
tion of the hydrogen bonding complex, and hence the lower
the concentration of base. Consequently, the concentration of
been most frequently used in probing hydrogen bonding
5, 21b]
involving imidazolium ionic liquids,[3
and in the case of
the imidazolium receptors mentioned earlier, similar obser-
vations have been made.[5] However, the nature of hydrogen
bonding complex in [bmim][BF4] shown in Scheme 1 is less
4902
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4900 4906

Allylic Alkylation in Ionic Liquids
4900 4906
nucleophile will be lowered, thus leading to slower nucleo-
philic addition (Scheme 1).
inactive dialkylimidazol-2-ylidene complexes of palladium by
deprotonation of the imidazolium cation by MeOÀ.[9b, 28] This
appears to be unlikely. First, the reaction of 7 and 2, which
involves an in situ generated alkoxide, proceeds equally well
in THF and the ionic liquid, suggesting that either N-
heterocyclic carbenes are not formed or they have no effect
on the palladium catalysis if formed. Second, the fact that 1
reacts with 2 much faster in the presence of DBU is
inconsistent with this hypothesis, because DBU would not
be expected to inhibit the formation of carbenes through
deprotonation by MeOÀ. Third, the effect of [bmim][BF4] on
the allylic alkylation in THF also casts doubts on this
possibility (Table 1), since the activity of the hypothesized
Pd carbene species would not be affected by additional
imidazolium cations. Indeed, no Pd-carbene complexes were
ever detected by NMR spectroscopy in the stoichiometric
reaction of Pd0 PPh3 with 1 in the presence of [bmim][BF4] at
room temperature. However,
It is also clear why DBU as an external base could bring
about an efficient reaction. DBU is a stronger base[17] than
OAcÀ and MeOCO2 (pKa 5.6 in water).[25] It does not form
À
stable hydrogen bonds with the acidic H2 proton of imidazo-
lium in the presence of the much more acidic 2; rather, as
expected, it deprotonates 2.[26] This is further supported by a
reaction involving an alkoxide base. Alkoxides are even
stronger bases; as such they are not expected to form a stable
hydrogen bond with [bmim][BF4] in the presence of 2. This is
clearly seen in the allylic alkylation of 2-methyl-2-vinyloxir-
ane 7 with 2, which generates in situ an alkoxide base by
oxidative addition to Pd0 and proceeds fast to give 8 as a
mixture of E,Z-isomers (E/Z ˆ 3/2) in room temperature
[bmim][BF4] or THF, reaching completion within 1 h in either
solvent under otherwise identical conditions [Eq. (3)].[27]
when the catalyst preparation
was performed with DBU
present at 808C rather than
being introduced upon cool-
ing the Pd0-PPh3-ionic liquid
mixture to room temperatu-
re,[9a,d] the resulting mixture
became colorless as opposed
to the more usual pale yellow.
This catalyst mixture did not
These results also show that, unlike Welton×s nucleophilic
substitution,[4h] it is the basicity of oxy anions, rather than the
uncleophilicity of hydrocarbon anions, that is primarily
affected by hydrogen bonding in the reactions involving 1 or
4 in [bmim][BF4]. The underlining mechanism for the
observed effects in the two different types of reactions is the
same, however, that is the nucleophiles or bases are made less
available for subsequent reactions due to hydrogen bonding
with the ionic liquid cations.
Acetate plays a different role. As indicated above, it cannot
deprotonate the acid 2 due to its low basicity. The accelerating
effect of acetate on the reaction of 1 and 2 most probably
results from the equilibrium shown in Equation (4), which
promote the reaction between 1 and 2 and it is probable that
the acidic H2 proton in [bmim][BF4] is being deprotonated at
elevated temperature, leading to inactive Pd carbene for-
mation.
Applying the hydrogen bonding proposition, it was possible
to suppress the Pd0-catalyzed isomerization of allylic acetates,
a reaction that can lead to the loss of regio- and stereo-
chemistry in stereospecific allylic substitution.[11] The isomer-
ization is thought to be due to the key oxidative addition step
being reversible, and unequivocal evidence for this has
recently been laid out.[11] As OAcÀ can be trapped by
hydrogen bonding with [bmim][BF4], allylic isomerization
resulting from reversible attack by the acetate would be
expected to be retarded when
the isomerization is carried
out in an imidazolium ionic
liquid. This is indeed the case.
Treating 9 with 5 mol% Pd0
PPh3 in CH2Cl2 afforded 35%
of product 10 after 1 h, com-
parable with the equilibrium
value reported in the litera-
increases the concentration of free carbonate and hence the
probability of its decomposition into the base MeOÀ. This is
reminiscent of some SN2 reactions carried out in protic
solvents, which can be accelerated upon introducing a basic
additive. The base competes with nucleophiles for hydrogen
bonding with the solvent and thus frees the former from
strong solvation by the latter.[2]
ture (Scheme 2).[29] However, the same reaction appears to be
completely suppressed in [bmim][BF4]. Thus, 10 was not
detected in the crude reaction mixture even after 20 h. The
isomerization of 9 in CH2Cl2 was even inhibited by a
substoichiometric quantity of [bmim][BF4]. Thus, in the
presence of only 0.15 equivalents of [bmim][BF4] relative to
9, only 12% of 10 was formed after 3 h. These observations
are again consistent with the acetate ion being involved in
equilibrium with imidazolium cations.
There exists a possibility that the sluggish reaction between
1 and 2 in [bmim][BF4] could stem from the formation of some
Chem. Eur. J. 2003, 9, 4900 4906
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4903

J. Xiao and J. Ross
FULL PAPER
À
MeOCO2 ions. Being strongly solvated by the ionic liquid
via hydrogen bonding, the anions could not function as
effective bases to deprotonate a HNu nucleophile, thus
rendering slow the nucleophlic attack at the PdII allyl
intermediate. However, this retarding effect can be alleviated
when a relatively strong base, either generated in situ or
added externally, is used, which favors deprotonation rather
than hydrogen bonding with the solvent. We further showed
that the phenomena of hydrogen bonding in the imidazolium
ionic liquids could be exploited to suppress unwanted, Pd0-
catalyzed isomerization of allylic acetates. Taken together,
these results highlight the potential effects of imidazolium
cations as hydrogen bond donors on catalytic reactions in
imidazolium ionic liquids and corroborate that ™The more a
solvent blocks up by hydrogen bond or otherwise the active
centers which take part in a chemical reaction, the less will be
the speed in it∫.[31]
Scheme 2. Isomerization of 9 and 10 by mechanistically distinct Pd0/PdII
catalysis.
Similar inhibition of isomerization was observed starting
from 10. Thus, while 58% of 9 was formed after 1 h in CH2Cl2,
10 remained intact in [bmim][BF4] even after 20 h. This is not
surprising, as the equilibrium constant K is close to 1 and so
the rate constant kÀ ꢀ k . Similar observations have been
‡
made with other allylic acetates in our laboratory. For
instance, the isomerization of 11, whilst not observed at room
temperature, proceeded smoothly at 508C in molecular
solvents [Eq. (5)]. Thus, exposing 11 to 5 mol% Pd0 PPh3
Experimental Section
General: All reactions were carried out in oven-dried glassware under
argon, using standard Schlenk and vacuum line techniques. [bmim][BF4]
and [bdmim][BF4] were prepared according to published procedures and
vacuum-dried and stored under argon.[32] THF and CH2Cl2 were freshly
distilled from sodium benzophenone and calcium hydride, respectively,
under nitrogen immediately prior to use. Phenylallyl carbonate 1 was
synthesized according to a literature method.[33] The synthesis of compound
10 was adapted from a literature method.[34] [MeOCO2][HDBU] was
prepared according to a recent procedure.[21c] Compounds 2, 4, 5, 7 and 9,
[nBu4N][OAc], Pd(OAc)2, and PPh3 were purchased from commercial
suppliers and used as received without further purification.
in CH2Cl2 lowered its initial cis/trans ratio from 87/13 to 43/57
after 1 h. An equilibrium value of 34/66 was measured after an
extended reaction time of 22 h. The same reaction was
suppressed in [bmim][BF4], though not entirely, probably
due to weakening of the hydrogen bonding interactions at
elevated temperature. Thus, the initial cis/trans ratio of 11 was
lowered from 87/13 to only 76/24 after 1 h and to a value of 54/
46 that is still far from equilibrium after 22 h.
Remarkably, the isomerization can be brought about when
a PdII catalyst is employed. Thus, treatment of 9 with
[PdCl2(MeCN)2] in [bmim][BF4] afforded 10 in 20% yield in
1 h reaction time; the same reaction in CH2Cl2 gave 10 in 33%
yield (Scheme 2). As is known in the literature, this reaction
proceeds intramolecularly via a cyclic intermediate and
involves no ionized acetate ions;[30] therefore it should not
be suppressed by hydrogen bonding. Differences between
CH2Cl2 and [bmim][BF4] in other solvent properties possibly
account for the different extent of isomerization in these
solvents. These results show that if the isomerization of an
allylic acetate is to be carried out in imidazolium ionic liquids,
PdII, rather than Pd0, should be the catalyst of choice.
Typical neutral Tsuji Trost reaction in [bmim][BF4] as exemplified for the
allylic alkylation between 1 and 2: Pd(OAc)2 (4.5 mg, 2 mol%) and PPh3
(21.0 mg, 8 mol%) were stirred in [bmim][BF4] (2 mL) at 808C for 20 min
under an atmosphere of argon and then allowed to cool to room
temperature. Compounds
1 (192.2 mg, 1.0 mmol) and 2 (198.2 mg,
1.5 mmol) were added and the mixture stirred vigorously under argon for
30 h. The reaction was not complete within a reaction time of 20 h, as
judged by 1H NMR monitoring. Upon completion, the reaction was
quenched with distilled water and the product extracted with EtOAc. The
organic layer was washed with water and brine and dried over MgSO4 to
give 3, which was isolated in 90% yield upon purification by flash column
chromatography (silica gel, eluent: n-hexane/EtOAc ˆ 10/1). The same
reaction was repeated in THF following a reported literature procedure.[16]
Starting with the same quantity of catalyst and reactants in 2 mL of THF,
the reaction was complete in 20 min. 1H NMR spectroscopy was used to
confirm the identity of the product by comparison with the literature.
The reaction of 4 and 5 in [bmim][BF4] on the same scale as above was
performed at 758C for an extended reaction time of 24 h to give a complete
conversion to 6. The reaction between 7 and 2 in [bmim][BF4] was complete
within 1 h reaction time furnishing 8, which was isolated in 94% yield
following purification by flash column chromatography (silica gel, eluent:
1
n-hexane/EtOAc ˆ 4/1). H NMR was used to confirm the identity of the
products by comparison with the literature, but in the case of 8, more data
are supplied, which do not appear to have been reported.
(E)-5-Phenyl-2-methoxycarbonyl-4-enoic acid methyl ester 3:[35] 1H NMR
(250 MHz, CDCl3, 258C, TMS): d ˆ 2.79 (dd, J ˆ 7.6, 7.2 Hz, 2H; CH2), 3.54
(d, J ˆ 7.6 Hz, 1H; CH), 3.71 (s, 6H; CH3), 6.12 (dt, J ˆ 15.8, 7.2 Hz, 1H;
Conclusion
ˆ
ˆ
CH CHCH2), 6.46 (d, J ˆ 15.8 Hz, 1H; CH CHCH2), 7.31 ppm (m, 5H;
C6H5).
In summary, neutral allylic alkylation reactions can be
considerably retarded in dialkylimidazolium ionic liquids
and our results suggest that this is due to hydrogen bonding
between the H2 proton of [bmim][BF4] and OAcÀ or
(E)-5-Phenyl-2-nitropent-4-enoic acid methyl ester 6:[18] 1H NMR
(250 MHz, CDCl3, 258C, TMS): d ˆ 3.12 (m, 2H; CH2), 3.85 (s, 3H;
CH3), 5.22 (dd, J ˆ 8.9, 5.8, Hz, 1H; CH), 6.08 (dt, J ˆ 15.8, 7.2 Hz, 1H;
ˆ
ˆ
CH CHCH2), 6.56 ppm (d, J ˆ 15.8 Hz, 1H; CH CHCH2).
4904
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4900 4906

Allylic Alkylation in Ionic Liquids
4900 4906
(E)/(Z')-2-(4-hydroxy-3-methyl-but-2-enyl)-malonic acid dimethyl ester 8:
1H NMR (200 MHz, CDCl3, 258C, TMS): d ˆ 1.67 (br s, 3H; CH3), 1.79 (br
s, 3 H; CH3'), 2.65 (br m, 4H; CH2CCH3, CH2CCH3'), 3.42 (m, 2H; CH',
CH), 3.74 (s, 12H; OCH3, OCH3'), 3.96 (br s, 2H; CH2), 4.10 (br s, 2H;
2002, 4, 107 111; g) S. N. Baker, G. A. Baker, F. V. Bright, Green
Chem. 2002, 4, 165 169; h) N. L. Lancaster, P. A. Salter, T. Welton,
G. B. Young, J. Org. Chem. 2002, 67, 8855 8861; i) C. Hardacre, J. D.
Holbrey, S. E. J. McMath, D. T. Bowron, A. K. Soper, J. Chem. Phys.
2003, 118, 273 278.
ˆ
CH2'), 5.19 (br t, J ˆ 7.7 Hz, 1H; C CH'), 5.35 ppm (br t, J ˆ 7.2 Hz, 1H;
C CH); 13C NMR (100 MHz, CDCl3): d ˆ 13.9, 21.8, 27.4, 27.4, 51.9, 52.0,
[5] a) K. Sato, S. Arai, T. Yamagishi, Tetrahedron Lett. 1999, 40, 5219
5222; b) J.-L. Thomas, J. Howarth, K. Hanlon, D. McGuirk, Tetrahe-
dron Lett. 2000, 41, 413 416; c) Y. Yuan, G. Gao, Z. L. Jiang, J. S.
You, Z. Y. Zhou, D. Q. Yuan, R. G. Xie, Tetrahedron 2002, 58, 8993
8999, and references therein.
ˆ
52.8, 52.9, 61.5, 68.3, 120.4, 122.9, 138.7, 138.8, 169.8, 169.9 ppm; MS (I.C./
‡
NH3): m/z (%): 234 [M ‡ NH4] (80); elemental analysis calcd (%) for
C10H16O5: C 55.55, H 7.47; found: C 55.55, H 7.54. The E/Z ratio for 8 was 3/
2
for the reaction in [bmim][BF4] and in THF based on 1H NMR
integration.
[6] L. C. Branco, J. G. Crespo, C. A. M. Afonso, Chem. Eur. J. 2002, 8,
3865 3871.
Typical Pd0-catalyzed isomerization reaction in [bmim][BF4] as exempli-
fied for 9: [Pd(dba)2] (28.7 mg, 5 mol%) and PPh3 (26.2 mg, 10 mol%)
were stirred in [bmim][BF4] (2 mL) at 808C for 20 min and then allowed to
cool to room temperature. 9 (114.2 mg, 1.0 mmol) was added and the
mixture stirred vigorously under an atmosphere of dry argon. 1H NMR
monitoring showed no formation of 10 after 20 h. The reaction was then
quenched with distilled water and the product extracted with Et2O. The
organic layer was washed with water and brine and dried over MgSO4 to
give 9 quantitatively. The same reaction in THF followed a reported
[7] For recent reviews, see: a) T. Welton, Chem. Rev. 1999, 99, 2071
2084; b) P. Wasserscheid, W. Keim, Angew. Chem. 2000, 112, 3926
3945; Angew. Chem. Int. Ed. 2000, 39, 3772 3789; c) M. J. Earle,
K. R. Seddon, Pure Appl. Chem. 2000, 72, 1391 1398; d) R. Sheldon,
Chem. Commun. 2001, 2399 2407; e) C. M. Gordon, Appl. Catal. A:
General 2001, 222, 101 117; f) H. Olivier-Bourbigou, L. Magna, J.
Mol. Catal. A: Chemical 2002, 182 183, 419 437; g) D. Zhao, M. Wu,
Y. Kou, E. Min, Catal. Today 2002, 74, 157 189; h) C. C. Tzschucke,
C. Markert, W. Bannwarth, S. Roller, A. Hebel, R. Haag, Angew.
Chem. 2002, 114, 4136 4173; Angew. Chem. Int. Ed. 2002, 41, 3964
4000.
[8] a) T. Fischer, A. Sethi, T. Welton, J. Woolf, Tetrahedron Lett. 1999, 40,
793 796; b) C. W. Lee, Tetrahedron Lett. 1999, 40, 2461 2464; c) A.
Aggarwal, N. L. Lancaster, A. R. Sethi, T. Welton, Green Chem. 2002,
4, 517 520; d) S. V. Dzyuba, R. A. Bartsch, Tetrahedron Lett. 2002,
43, 4657 4659; e) O. Acevedo, J. D. Evanseck, Abstr. Pap. Amer.
Chem. Soc. 2002, 224, 028-IEC.
[9] a) W. Chen, L. Xu, C. Chatterton, J. Xiao, Chem. Commun. 1999,
1247 1248; b) L. Xu, W. Chen, J. Xiao, Organometallics 2000, 19,
1123 1127; c) L. Xu, W. Chen, J. Ross, J. Xiao, Org. Lett. 2001, 3,
295 297; d) J. Ross, W. Chen, L. Xu, J. Xiao, Organometallics 2001,
20, 138 142; e) J. Ross, J. Xiao, Green. Chem. 2002, 4, 129 133.
[10] J. Tsuji, Palladium Reagents and Catalysts-Innovations in Organic
Synthesis, Wiley, Chichester, 1995.
[11] a) C. Amatore, S. Gamez, A. Jutand, G. Meyer, M. Moreno-Manas, L.
Morral, R. Pleixats, Chem. Eur. J. 2000, 6, 3372 3376, and references
therein; b) C. Amatore, S. Gamez, A. Jutand, Chem. Eur. J. 2001, 7,
1273 1280.
procedure,[29] with the ratio of
spectroscopy.
9
and 10 determined by 1H NMR
(E)-1-Acetoxy-but-2-ene (9):[36] 1H NMR (200 MHz, CDCl3, 258C, TMS):
d ˆ 1.73 (dd, J ˆ 6.1, 1.1 Hz, 3H; CH3), 2.05 (s, 3H; CH3CO2), 4.50 (d, J ˆ
6.6 Hz, 2H; CH2), 5.59 (td, J ˆ 15.4, 6.6, 1.1 Hz, 1H; CHCH2), 5.80 ppm (br
dq, J ˆ 15.4, 6.1 Hz, 1H; CH3CH).
3-Acetoxy-but-1-ene (10):[36] 1H NMR (200 MHz, CDCl3, 258C, TMS): d ˆ
1.31 (d, J ˆ 6.6 Hz, 3H; CH3), 2.06 (s, 3H; CH3CO2), 5.15 (dd, J ˆ 10.4,
ˆ
ˆ
1.1 Hz, 1H; CH CH2), 5.25 (dd, J ˆ 17.3, 1.1 Hz, 1H; CH CH2), 5.35 (br
ˆ
quintet, J ˆ 6.6 Hz, 1H; CH3CH), 5.85 ppm (ddd, CH CH2).
Acknowledgement
We thank the EPSRC for a studentship (J.R.) and the industrial partners of
the Leverhulme Centre for Innovative Catalysis for support. We are also
grateful to Professor Don Bethell and Dr. Richard Bonar-Law for helpful
discussions. Johnson Matthey is gratefully acknowledged for the loan of
palladium.
[12] a) C. de Bellefon, E. Pollet, P. Grenouillet, J. Mol. Catal. A: Chemical
¬
1999, 145, 121 126; b) S. Toma, B. Gotov, I. Kmentova, E.
¬
¬
¬
Solcaniova, Green. Chem. 2000, 2, 149 151; c) I. Kmentova, B .
œ
¬
¬
Gotov, E. Solcaniova, S. Toma, Green. Chem. 2002, 4, 103 106.
[1] a) G. A. Jeffrey, An Introduction to Hydrogen Bonding, Oxford
University Press, New York, 1997; b) S. Schreiner, Hydrogen Bonding,
Oxford University Press, New York, 1997. For some recent publica-
tions concerning hydrogen bonding and catalysis, see: c) S. P. de Viss-
er, F. Ogliaro, P. K. Sharma, S. Shaik, Angew. Chem. 2002, 114, 2027
2031; Angew. Chem. Int. Ed. 2002, 41, 1947 1951; d) Y. Huang, V. H.
Rawal, J. Am. Chem. Soc. 2002, 124, 9662 9663; e) R. D. Bach, J.
Phys. Chem. B 2002, 106, 4325 4335; f) C. E. Cannizzaro, K. N.
Houk, J. Am. Chem. Soc. 2002, 124, 7163 7169.
¬
[13] C. Amatore, E. Carre, A. Jutand, M. A. M'Barki, Organometallics
1995, 14, 1818 1826.
[14] a) W. N. Olmstead, Z. Margolin, F. G. Bordwell, J. Org. Chem. 1980,
45, 3295 3299; b) R. W. Alder, P. R. Allen, S. J. Williams, J. Chem.
Soc. Chem. Commun. 1995, 1267 1268; c) Y.-J. Kim, A. Streitwieser,
J. Am. Chem. Soc. 2002, 124, 5757 5761.
[15] The pKa of closely related diethyl malonate in DMSO is 16.4: F. G.
Bordwell, Acc. Chem. Res. 1988, 21, 456 463.
[16] a) J. Tsuji, I. Minami, Acc. Chem. Res. 1987, 20, 140 145; b) J. Tsuji,
Tetrahedron 1986, 42, 4361 4401.
[2] C. Reichardt, Solvents and Solvents Effects in Organic Chemistry, 2
ed., VCH, Weinheim, 1990.
[17] The pKa of DBU is 12.9 in H2O and 23.9 in CH3CN: W. Galezowski, A.
Jarczewski, M. Stanczyk, B. Brzezinski, F. Bartl, G. Zundel, J. Chem.
Soc. Faraday Trans. 1997, 93, 2515–2518, and references therein. The
pKa of HOAc is 4.8 in H2O and 12.3 in DMSO.[15]
[18] G. Giambastiani, G. Poli, J. Org. Chem. 1998, 63, 9608 9609.
[19] C. Amatore, A. Jutand, G. Meyer, L. Mottier, Chem. Eur. J. 1999, 5,
466 473.
[3] a) A. A. Fannin, L. A. King, J. A. Levisky, J. S. Wilkes, J. Phys. Chem.
1984, 88, 2609 2614; b) S. Tait, R. A. Osteryoung, Inorg. Chem. 1984,
23, 4352 4360; c) A. K. Abdul-Sada, A. M. Greenway, P. B. Hich-
cock, T. J. Mohammed, K. R. Seddon, J. A. Zora, J. Chem. Soc. Chem.
Commun. 1986, 1753 1754; d) K. M. Dieter, C. J. Dymek, Jr., N. E.
Heimer, J. W. Rovang, J. S. Wilkes, J. Am. Chem. Soc. 1988, 110,
2722 2726; e) A. G. Avent, P. A. Chaloner, M. P. Day, K. R. Seddon,
T. Welton, J. Chem. Soc. Dalton Trans. 1994, 3405 3413, and
references therein.
[20] J. L. Anderson, J. Ding, T. Welton, D. W. Armstrong, J. Am. Chem.
Soc. 2002, 124, 14247 14254, and references therein.
[21] a) J. S. Wilkes, M. J. Zaworotko, J. Chem. Soc. Chem. Commun. 1992,
[4] For recent examples, see: a) L. Cammarata, S. G. Kazarian, P. A.
Salter, T. Welton, Phys. Chem. Chem. Phys. 2001, 3, 5192 5200;
b) L. M. J. Muldoon, C. M. Gordon, I. R. Dunkin, J. Chem. Soc.
Perkin Trans. 2, 2001, 433 435; c) A. Elaiwi, P. B. Hitchcock, J.-F.
Huang, P.-Y. Chen, I.-W. Sun, S. P. Wang, Inorg. Chim. Acta 2001, 320,
√
965 967; b) P. Bonhote, A.-P. Dias, N. Papageorgiou, K. Kalyana-
sundaram, M. Gr‰tzel, Inorg. Chem. 1996, 35, 1168 1178. Methyl
carbonate forms a stable salt with protonated DBU, probably due to
hydrogen bonding: c) P. Munshi, A. D. Main, J. C. Linehan, C.-C. Tai,
P. G. Jessop, J. Am. Chem. Soc. 2002, 124, 7963 7971.
7
11; d) A. D. Headley, N. M. Jackson, J. Phys. Org. Chem. 2002, 15,
[22] The chemical shift of the imidazolium H2 proton moved upfield by
only 0.07 ppm when the concentration of [bmim][BF4] changed from
0.014 to 0.68m in CDCl3 in the absence of [nBu4N][OAc], showing that
52 55; e) A. E. Bradley, C. Hardacre, J. D. Holbrey, S. Johnston,
S. E. J. McMath, M. Nieuwenhuyzen, Chem. Mater. 2002, 14, 629
635; f) C. G. Hanke, N. A. Atamas, R. M. Lynden-Bell, Green Chem.
Chem. Eur. J. 2003, 9, 4900 4906
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4905

J. Xiao and J. Ross
FULL PAPER
the variation depicted in Figure 1 results primarily from interactions
with the acetate not from changes in the concentration of
[bmim][BF4].[3e]
[28] a) D. S. McGuinness, K. J. Cavell, B. F. Yates, Chem. Commun. 2001,
355 356; b) C. J. Mathews, P. J. Smith, T. Welton, A. J. P. White, D. J.
Williams, Organometallics 2001, 20, 3848; c) M. Hasan, I. V. Kozhev-
nikov, M. R. H. Siddiqui, C. Femoni, A. Stenier, N. Winterton, Inorg.
Chem. 2001, 40, 795 800; d) V. K. Aggarwal, I. Emme, A. Mereu,
Chem. Commun. 2002, 1612 1613.
[29] D. C. Braddock, A. J. Wildsmith, Tetrahedron Lett. 2001, 42, 3239
3242.
[30] S. Bouquillon, J. Muzart, Eur. J. Org. Chem. 2001, 3301 3305.
[31] S. R. Palit, J. Org. Chem. 1947, 12, 752 759.
[32] J. D. Holbrey, K. R. Seddon, J. Chem. Soc. Dalton Trans. 1999, 2133
2140.
[33] J. Lehmann, G. C. Lloyd-Jones, Tetrahedron 1995, 51, 8863 8874.
[34] T. Hayashi, A. Yamamoto, Y. Ito, E. Nishioka, H. Miura, K. Yanagi, J.
Am. Chem. Soc. 1989, 111, 6301 6311.
[35] F. Glorius, M. Neuburger, A. Pfaltz, Helv. Chim. Acta. 2001, 84, 3178
3196.
[23] a) S. N. Vinogradov, R. H. Linnell, Hydrogen Bonding, Van Nostrand
Reinhold, New York, 1971; b) D. H. Williams, I. Fleming, Spectro-
scopic Methods in Organic Chemistry, McGraw-Hill, London, 1996.
Assuming that the imidazolium and acetate ions are in equilibrium
with a 1:2 complex, the chemical shift of the imidazolium H2 proton
would vary in a sigmoid manner with the concentration of acetate ion.
[24] D. S. Kemp, D. D. Cox, K. G. Paul, J. Am. Chem. Soc. 1975, 97, 7312
7318.
[25] G. Gattow, W. Behrendt, Angew. Chem. 1972, 84, 549 550; Angew.
Chem. Int. Ed. Engl. 1972, 11, 534 535.
[26] In contrast to the behavior of acetate salts, addition of DBU to a
CD2Cl2 solution of [bmim][BF4] caused an upfield shift of the H2
resonance in the 1H NMR spectrum, suggesting that imidazolium ring
stacking rather than hydrogen bonding could be in operation.[3e]
However, weaker bases such as DABCO have been proposed to be
capable of deprotonating dialkylimidazolium salts in the presence of
an aldehyde.[28d]
[36] A. L. J. Beckwith, A. A. Zavitsas, J. Am. Chem. Soc. 1986, 108, 8230
8234.
[27] J. Tsuji, H. Kataoka, Y. Kobayashi, Tetrahedron Lett. 1981, 22, 2575
2578.
Received: February 28, 2003 [F4895]
Revised: June 6, 2003
4906
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4900 4906