Chloride ion pairs as catalysts for the alkylation of aldehydes and ketones with C-H acidic compounds

Article abstract of DOI:10.1016/j.tetlet.2009.06.013

Chloride anions associated with various soft cations (like tetraalkyl ammoniums, alkyl imidazoliums or pyridiniums) were shown to be able to promote the alkylation of carbonyl derivatives with acidic compounds, as exemplified on Knoevenagel and aldol condensations under relatively mild conditions. This activity was attributed to an enhanced nucleophilicity of the chlorine anion, originating from a softness/hardness mismatch between the anion and the cation.

Full text of DOI:10.1016/j.tetlet.2009.06.013

Tetrahedron Letters 50 (2009) 4833–4837
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Chloride ion pairs as catalysts for the alkylation of aldehydes and ketones
with C–H acidic compounds
a
b
b
a,*
Camille Carrignon , Philippe Makowski , Markus Antonietti , Frédéric Goettmann
a
Institut de Chimie Separative de Marcoule, UMR 5257, ICSM site de Marcoule, BP 17171, 30207 BAGNOLS SUR CEZE, France
Colloid Chemistry Department, Max-Planck Institute of Colloids and Interfaces, Scientific Campus Golm, 14476 POTSDAM, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Chloride anions associated with various soft cations (like tetraalkyl ammoniums, alkyl imidazoliums or
pyridiniums) were shown to be able to promote the alkylation of carbonyl derivatives with acidic com-
pounds, as exemplified on Knoevenagel and aldol condensations under relatively mild conditions. This
activity was attributed to an enhanced nucleophilicity of the chlorine anion, originating from a soft-
ness/hardness mismatch between the anion and the cation.
Received 4 March 2009
Revised 2 June 2009
Accepted 3 June 2009
Available online 7 June 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Organocatalysis
C–C bond forming reactions
Ionic liquids
Lewis bases
While Lewis acidic catalysts have been used for a long time
especially in heterogeneous catalysis, Lewis basic catalysts only
to the fact that ion pairs with mismatching softness and hardness
were forming weakly associated ion pairs even in relatively non
polar solvents. The resulting unshielded and, thus, nucleophilic an-
1
recently started unfolding their potentialities.2 Nevertheless,
numerous applications of Lewis basic catalysts can already be
found in the literature. This can be exemplified for the case of
two important archetypical C–C bond forming reactions, namely
the Knoevenagel and the aldol condensations. Knoevenagel con-
densations, the reaction of aldehydes or ketones with C–H acidic
malonates (like malonodinitrile or diethylmalonate), are usually
catalysed by Brönsted bases, typically amines (Scheme 1a). Charles
however reported as early as 19633 an alternative mechanism
involving the attack of a primary amine on an aldehyde to form
an imine, which then reacts with the methylene moiety (Scheme
ions were used to promote the conversion of epoxides and CO
2
into
1
2
organic carbonates and the aromatisation of partially unsatu-
1
3
rated 6-rings. In an attempt to test the versatility of this property
of halides, we investigated their activity in C–C bond forming reac-
tions. Quite interestingly, there already have been some reports on
the use of halides to promote Knoevenagel condensations. LiBr was
employed as a heterogeneous catalyst for this reaction under sol-
,4
14
vent free conditions. The authors of this report did not make
any statement on the mechanism of the described reactions. How-
ever it can be assumed that the strong oxophilicity of the lithium
was promoting the reaction and that the bromide anion only
1
b). The possibility to promote Knoevenagel condensations with
1
5
16
nucleophilic compounds was also evidenced by the use of tri-
played a minor role. Iodine and tetrabutylammoniumbromide
5
phenyl phosphine as a catalyst for this reaction. Aldol condensa-
were also reported to promote Knoevenagel reactions in water.
However, in both cases the reactions were carried out in the pres-
ence of K CO , making it difficult to assess exactly what was the
2 3
role of the halogen. This contribution, thus, aims at demonstrating
the possibility to catalyse the alkylation of carbonyl groups with C–
H acidic compounds with chlorine ions, avoiding any other co-
reactant.
tions, the reaction of two aldehydes or ketones to yield an enone,
can be catalysed by Brönsted acids or strong bases, but also by ba-
6
,7
sic solids.
Lewis basic catalysts also proved very effective for enantioselec-
tive aldol condensation; this includes the use of amino acids, espe-
8–11
cially proline, phosphoramides and phosphine oxides.
It was recently shown that halides associated with very soft cat-
ions (like tetraalkyl ammoniums or alkyl imidazoliums) were able
to act as catalysts in various reactions. This property was attributed
In the first part of this work, it was assessed whether halide ion
pairs were able to promote one of the easiest Knoevenagel conden-
sations, namely the reaction of benzaldehyde with malonodinitrile
to yield benzylidenemalononitrile (Scheme 2) at 80 °C in 4 h. Fig-
ure 1 displays the obtained conversions for various ion pairs (cal-
culated as the amount of formed product determined by GC–MS
*
Corresponding author. Tel.: +33 466 905 758; fax: +33 466 797 611.
E-mail address: Frederic.goettmann@cea.fr (F. Goettmann).
0
040-4039/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2009.06.013

4
834
C. Carrignon et al. / Tetrahedron Letters 50 (2009) 4833–4837
EWG1
EWG2
EWG1
EWG2
O
R
O
a-
+
R
BH
B-
EWG1
EWG2
EWG1
EWG2
EWG1
EWG2
HO
2
-H O
O
+
R
R
R
Nu-
- Nu-
EWG1
EWG1
EWG2
O
OH
Nu
+
H
+
b-
R
R
EWG2
Nu
Scheme 1. Accepted mechanisms for the (a) basic and (b) nucleophilic pathways for the Knoevenagel condensation.
relative to a toluene reference). These investigations were facili-
tated by the fact that many ion pairs with organic cations are com-
mercially available because of their use as ionic liquids.
In our previous study on the aromatisation of Diels–Alder reac-
1
3
tion products we noticed that: (i) the anions have to be halides to
promote the reaction but the nature of the halide did not strongly
affect the reaction; (ii) the cation did not strongly impact the reac-
tion rates as soon as solubility was provided. In the case of Knoe-
venagel condensations, a given solubility of the ion pair was also
required. Indeed NaCl, which is not soluble in acetonitrile at this
temperature, failed to provide any rate acceleration as compared
with the test without catalyst. But, contrary to our previous work,
even with good solubility, the nature of the cation and the nature
of the halide affected the outcome of our test reaction. Indeed,
while hexadecyltrimethyl ammonium chloride (CTAC), hexade-
cylmethyl imidazolium chloride (C16mimCl) and ethylmethyl imi-
dazolium chloride (EmimCl) featured almost the same activity,
the pyridinium-based salts (butylmethylpyridinium chloride and
hexadecylpyridinium chloride) featured a stronger catalytic activ-
ity, reaching 54% and 37% of conversion. These observations are
in agreement with our hypothesis that the activity of the tested
ion pairs correlates with the softness/hardness mismatch of our
ions. Indeed, pyridinium species can be regarded as softer cations
as alkylammoniums or even imidazoliums. Unfortunately we were
not able to find any comprehensive study on the dissociation con-
stants of these ion pairs in organic solvents in the literature and
further experimental work is required to measure it.
Figure 1. Results of the reaction of benzaldehyde with malononitrile in the
presence of various catalysts. (i) General reaction conditions: 0.2 mmol of catalyst,
2
mmol of aldehyde, 2 mmol nitrile, in 5 mL of acetonitrile, 80 °C for 4 h. (ii) CTAC
stands for cetyltrimetylammonium chloride, CTAB for cetyltrimetylammonium
bromide, C16mimCl for cetylmethylimidazolium chloride, C16pyrCl, for cetylpyri-
dinium chloride EmimCl for ethylmethylimidazolium chloride, EmimBF
4
for
ethylmethylimidazolium tetrafluoroborate and BumpyrCl for butylmethylpyridini-
um chloride. (iii) The yields were determined by GC–MS as the amount of formed
benzylidene with toluene as an internal standard. Reaction of benzaldehyde with
malononitrile.
activity of our chlorine-based ion pairs originates from the fact that
they are weakly associated, then increasing the dielectric constant
of the used solvent should help increasing the conversion rates. For
this investigation, C16pyrCl was selected as a catalyst, because it
features a good solubility in a wide range of solvents. We tested
the Knoevenagel condensation of benzaldehyde and malononitrile
at 80 °C for 12 h with C16pyrCl in heptane, toluene, diisopropyl
ether, THF, isopropanol and acetonitrile. Figure 2 displays the ob-
Hexadecyltrimethylammonium bromide (CTAB) did not yield
any detectable rate enhancement nor did ethylmethylimidazolium
tetrafluoroborate (EmimBF ), showing that only chlorine was the
4
active anion for this reaction. It is also worth being noticed that
the obtained conversions are relatively low (a primary amine
would perform much better under such conditions) and are only
interesting from an academic point of view as hints on the actual
mechanism.
Although butylmethyl pyridinium chloride showed the best
activity in this catalysts screening, its low miscibility with many
solvents prompted us to pursue our study with ion pairs featuring
long alkyl chains. We thus selected C16pyrCl and CTAC as they were
the second and third best choices (with respect to the measured
activity).
In a second step, we decided to investigate the impact of the
solvent on the observed conversion enhancements. Indeed, if the
Figure 2. Influence of the dielectric constant of the solvent on conversion of rate of
Scheme 2. Reaction of benzaldehyde with malononitrile.
the Knoevenagel condensation.

C. Carrignon et al. / Tetrahedron Letters 50 (2009) 4833–4837
4835
Table 1
Results of various Knoevenagel condensations catalysed by CTAC
a
Entry
Aldehyde
Nucleophile
Product
Conv.b (%)
CTAC
C
16PyrCl
3
CH CN
THF
91
CH
3
CN
THF
O
CN
CN
1
64
98
71
H
H
CN
CN
CN
O
O
COOEt
COOEt
2
3
4
5
6
7
54
0
46
0
69
0
50
0
CN
COOEt
/
H
O
COOEt
CN
CN
CN
81
56
50
71
100
100
50
97
82
73
74
82
87
15
54
H
CN
O
COOEt
COOEt
H
O
CN
CN
CN
CN
H
CN
CN
CN
MeO
MeO
O
NC
O
CN
O
47
H
CN
CN
H
CN
H
H
8
9
61
6
44
0
41
40
38
32
26
34
<1
13
<1
0
CN
O
O
CN
Also traces of aldol condensation product
COOEt
COOEt
CN
CN
CN
CN
H
1
1
1
0
1
2
41
65
13
43
21
0
O
CN
CN
CN
CN
H
O
CN
CN
CN
NC
O
CN
CN
NC
NC
CN
O
O
CN
CN
1
3
4
15
31
10
32
51
43
28
30
COOEt
COOEt
CN
1
a
General reaction conditions: 0.2 mmol of CTAC, 2 mmol of aldehyde, 2 mmol of donor, in 5 mL of acetonitrile, at 80 °C for 12 h.
Conversions were determined as the consumption of donor as measured per GC–MS using toluene as an internal standard.
b

4
836
C. Carrignon et al. / Tetrahedron Letters 50 (2009) 4833–4837
Table 2
Results of various Knoevenagel condensations catalysed by CTACa
Entry
Electrophile
Nucleophile
Products
Conv.b (%)
CTAC
C
16pyrCl
Bu
Bu
O
Bu
1
2
3
4
5c
6
/
/
74
0
100
O
(
Traces of the E-isomer)
O
/
0
O
O
Bu
85
0
58
28
70
60
40
O
Bu
(Traces of the hexanal dimmer)
O
O
O
Ph
Ph
Ph
O
O
O
96
95
91
Ph
O
O
O
O
Ph
O
O
7
Ph
General reaction conditions: 0.2 mmol of CTAC, 4 mmol of pure aldehyde or ketone or a mixture of 2 mmol of aldehyde and 2 mmol of ketone, without solvent, at 120 °C
a
for 20 h.
b
c
The conversion rates were determined as the amount of consumed reactant by GC–MS using biphenyl as an internal standard.
In this case only 1 mmol of ketone was used.
tained conversion rates as a function of the dielectric constant of
the solvents. As can be seen an initial increase in the conversion
Knoevenagel substrates. A closer look at the reactions however
shows surprising features. Indeed, while in the case of benzalde-
hyde the usual decrease of conversion when passing from malon-
onitrile, to ethyl cyanoacetate to diethyl malonate is observed
(Table 1, entries 1–3) the yields with p-tolualdehyde are much bet-
ter (Table 1, entries 4 and 5,) than with benzaldehyde. This is unu-
sual as p-tolualdehyde, being more electron rich, is a worse
electrophile. In the case of p-methoxy benzaldehyde, however,
the conversion decreases, as expected (Table 1, entry 6). A moder-
ate conversion is also observed in the case of furaldehyde (Table 1,
entry 7, up to 74%). The great advantage of using chlorine-based
ion pairs as catalysts for Knoevenagel reactions, however, seems
to be the fact that relatively unreactive electrophiles can be used.
Indeed non aromatic aldehydes such as hexanal, 2-methylbutanal
or pivaldehyde (Table 2, entries 8, 10 and 11) and even ketones
(cyclohexanone, Table 2, entries 13 and 14) could yield the conden-
sation product in low to moderate yields. Here again, strong devi-
ations from the expected activities are observed. Acetophenone, for
instance is a relatively active ketone and shows very poor yields
(Table 2, entry 12, 26% at best). Cyclohexanone, on the contrary
is deactivated through both sigma donation and sterical hindrance,
but reacts with malononitrile and even better with ethyl cyanoac-
rates with increasing
e was observed reaching a maximum with
THF and isopropanol (with 86% and 93% of conversion). These data
fit well with the hypothesis that ion pair dissociation is the reason
for the activity of our chlorides. Nevertheless acetonitrile, which is
the most polar solvent we tested, gave a lower conversion rate of
7
1%. It is relatively difficult to decide at this stage, whether this lit-
tle loss of activity is due to a second order effect (e.g., a shielding of
the chlorine by very polar solvents), or just has to be considered as
non relevant (because of experimental error). Although isopropa-
nol gave the best result, it proved difficult to deal with in the rest
of the study due to the formation of small amounts of acetals. The
next reactions were thus tested in both acetonitrile and THF.
We then investigated the activity of CTAC and C16pyrCl for a
broad range of Knoevenagel substrates. We thus heated various
combinations of electrophile and nucleophile in the presence of
catalytic amounts of CTAC and C16pyrCl both in acetonitrile and
THF at 80 °C for 12 h. Table 1 displays some of the results we
obtained.
A quick look at the results reported in Table 1 shows that both
CTAC and C16pyrCl can actually act as catalysts for a large range of

C. Carrignon et al. / Tetrahedron Letters 50 (2009) 4833–4837
4837
etate (Table 1, entries 13 and 14 up to 51% of conversion). These
unusual reactivities confirm that the Knoevenagel reactions do
not proceed via the usual basic mechanism (Scheme 1a). They also
indicate that the reaction is probably concerted. Regarding the
activity of both ion pairs, the pyridinium proves more active than
the ammonium in most cases. This is consistent with our hypoth-
esis that a softness/hardness mismatch is governing the catalytic
activity. The influence of the solvent remains moderate, but THF
tends to be the better solvent. Theoretical investigations are cur-
quired to shed some light on the reasons for these observations.
We also currently focus on immobilised chlorine ion pairs in order
to yield stable heterogenous aldol condensation catalysts, which
would be of interest for industrial applications.
Acknowledgements
The Max-Planck Society and the CEA are gratefully acknowl-
edged for financially supporting this work.
rently running to find
observations.
a
suitable explanation for these
Supplementary data
As aldol condensations are, in principle, closely related to Knoe-
venagel reactions, we also attempted to promote various conden-
sations between aldehydes and/or ketones with CTAC and
Supplementary data (experimental procedures) associated with
this article can be found, in the online version, at doi:10.1016/
j.tetlet.2009.06.013.
C
16pyrCl at 120 °C for 20 h. Table 2 summarises the results we
obtained.
As can be seen, our two ion pairs also proved to act as effective
References and notes
promoters for aldol condensations. Interestingly, while the nature
of the aldehyde did not strongly impact the conversion rate (hex-
anal, benzaldehyde and cinnamaldehyde were converted from
1. Corma, A. Chem. Rev. 1997, 97, 2373–2419.
2
3
4
.
.
.
Denmark, S. E.; Beutner, G. L. Angew. Chem., Int. Ed. 2008, 47, 1560–1638.
Charles, G. Bull. Soc. Chim. Fr. 1963, 1559.
Charles, G. Bull. Soc. Chim. Fr. 1963, 1576.
7
4% to 96% with CTAC), the nature of the ketones proved to be
determining. Acetophenone, which is a typical aldol substrate,
did only poorly react (Table 2, entries 2 and 4). On the contrary,
5. Yadav, J. S.; Reddy, B. V. S.; Basak, A. K.; Visali, B.; Narsaiah, A. V.; Nagaiah, K.
Eur. J. Org. Chem. 2004, 546–551.
6.
Figueras, F.; Kantam, M. L.; Choudary, B. M. Curr. Org. Chem. 2006, 10, 1627–
637.
1
-indanone, which is more sterically hindered, or cyclohexanone,
1
which is very electron rich and thus usually a relatively bad nucle-
ophile, both did react (Table 2, entries 5–7) even yielding the high-
est conversion rates in this study (from 85% to 96%). In the case of
aldol reactions, CTAC performed better than C16pyrCl. These obser-
vations again have to be rationalised.
In this work, it has been evidenced that chlorine, when associ-
ated with soft organic cations, could act as an effective catalyst
to promote Knoevenagel and aldol condensations in, sometimes,
high yields under mild conditions. Interestingly, the use of CTAC
and C16PyrCl as a catalyst for these reactions enabled to use rela-
tively unreactive aldehydes such as hexanal, 2-methylbutanal
and pivaldehyde. Detailed experimental and theoretical work is re-
7. Fischer, A.; Makowski, P.; Mueller, J. O.; Antonietti, M.; Thomas, A.; Goettmann,
F. ChemSusChem 2008, 1, 444–449.
8.
Denmark, S. E.; Su, X.; Nishigaichi, Y.; Coe, D. M.; Wong, K. T.; Winter, S. B.;
Choi, J. Y. J. Organomet. Chem. 1999, 64, 1958–1967.
9. Kotani, S.; Hashimoto, S.; Nakajima, M. Tetrahedron 2007, 63, 3122–3132.
0. Sakthivel, K.; Notz, W.; Bui, T.; Barbas, C. F. J. Am. Chem. Soc. 2001, 123, 5260–
267.
1. Seayad, J. B. L. Org. Biomol. Chem. 2005, 3, 719–724.
1
5
1
12. Xie, Y.; Zhang, Z. F.; Jiang, T.; He, J. L.; Han, B. X.; Wu, T. B.; Ding, K. L. Angew.
Chem., Int. Ed. 2007, 46, 7255–7258.
1
1
3. Kaper, H.; Antonietti, M.; Goettmann, F. Tetrahedron Lett. 2008, 49, 4546–4549.
4. Prajapati, D.; Lekhok, K. C.; Sandhu, J. S.; Ghosh, A. C. J. Chem. Soc., Perkin Trans.
1 1996, 959–960.
15. Ren, Y. M.; Cai, C. Catal. Lett. 2007, 118, 134–138.
6. Gupta, M.; Wakhloo, B. P. Arkivoc 2007, 94–98.
1