Silica-supported imines as mild, efficient base catalysts

Article abstract of DOI:10.1039/b002424o

Imine grafted silicas are found to be mild and effective base catalysts for Knoevenagel and Michael reactions.

Full text of DOI:10.1039/b002424o

View Online / Journal Homepage / Table of Contents for this issue
Silica-supported imines as mild, efficient base catalysts
Katherine A. Utting and Duncan J. Macquarrie*
Green Chemistry Group, Department of Chemistry, University of Y ork, Heslington, Y ork, UK
Y O10 5DD. E-mail: djm13=york.ac.uk; Fax: ]44 1904 434546
Received (in Montpellier, France) 29th March 2000, Accepted 31st May 2000
Published on the Web 5th July 2000
Imine grafted silicas are found to be mild and e†ective base catalysts for Knoevenagel and Michael reactions.
Heterogeneous base catalysis is becoming a major area of
chemical research due to the wide variety of organic trans-
formations that can be carried out under basic conditions, and
the continuing drive for cleaner, more efficient processes.1,2
Both amines and imines are well-known basic functions3,4
and amines have been used as the active basic site in many
supported base catalysts.5h7 The major drawback of these
catalysts is the irreversible formation of surface-bound amide
groups during the Knoevenagel reaction.5 In both amines and
imines the lone pair is able to act as a Bronsted base by
accepting a proton from a Bronsted acid. In imines the lone
pair can also act as a Lewis base by donating its electrons to
another compound or metal centre, aided by the electron
donating ability of the C2N bond.4,8 Basicity is diminished or
even absent when the unshared lone pair is involved in multi-
ple bonding in a delocalised system such as H-bonding.9h11
In this paper work has been carried out as part of a project
investigating the basic character of silica-supported phenol-
ates, which have been shown to be e†ective base catalysts for
the Michael reaction.12 The phenolate moiety is attached to a
silica surface via an imine bond, see Scheme 1. A series of
simple supported phenolate catalysts were prepared using this
original method12 (method a, Scheme 2). Improvements have
been made to increase the surface loading and catalytic activ-
ity of these Ðrst catalysts. Characterisation of the catalysts by
IR and UV methods revealed the susceptibility of the imine
bond to hydrolysis, leading to a reduced loading of surface
groups on these catalysts. New catalysts were prepared using
a modiÐed method of preparation and these catalysts showed
an increased surface loading of both organic groups and
sodium counterions. This increased loading, however, leads to
lower activity in the Knoevenagel reaction than found with
the original catalysts. However, after hydrolysis, it was found
that the catalytic activity increases despite the signiÐcant
reduction in catalytic sites. This implies that it is the few
remaining sites that are the most catalytically active. We also
present evidence that it is in fact the imine group that is the
basic site of the catalyst. In this paper we report the e†ec-
tiveness of the imine group to catalyse both Knoevenagel and
Michael reactions. We also show that while the groups
attached to the imine bond inÑuence the catalytic activity of
the catalysts, it is the imine group, and not the phenolate,
which is the basic site of the catalysts.
Experimental
All reagents were purchased from Aldrich Chemical
Company, Fluorochem and Lancaster Chemicals and used as
received. All solvents used were AR grade. Infrared spectros-
copy was carried out using a Bruker Equinox 55 IR. Thermal
Scheme 1 AMPS 1 and phenolate catalysts prepared via methods a
and b.
Scheme 2 Di†erences in methods of preparation of the phenolate catalysts.
DOI: 10.1039/b002424o
New J. Chem., 2000, 24, 591È595
591
This journal is ( The Royal Society of Chemistry and the Centre National de la Recherche ScientiÐque 2000

View Online
analysis was carried out using a Stanton STA 625 machine
and UV results were obtained using a CE 3055 3000 series
UV spectrometer. Product yields were calculated by GC,
using internal standards, with an HP 5890 series GC system
with an HP1 column.
Solution stability study of catalyst 4b
For the solution stability study, 0.25 g of catalyst 4b was
stirred overnight at 20 ¡C in 20 ml of each solution. The solu-
tions used were water, methanol, ethanol, dilute sodium
hydrogen carbonate and dilute hydrochloric acid. Each hour a
sample was taken (1 ml) and Ðltered. Each Ðltrate was then
standardised to pH 12 with 0.01 M NaOH solution. A set of
standards was produced to correlate and calculate the quan-
tity of phenolate lost from each of the experiments. UV
absorption at 254 nm was used to calculate the amount of
phenolate present in the Ðltrate.
Preparation of the catalysts
3-Aminopropyltrimethoxysilane (3.59 g, 20 mmol) was added
to Kieselgel 100 silica (20 g) in ethanol (250 ml) and stirred at
room temperature for 24 h. This was Ðltered and the solid
washed with ethanol and diethyl ether and dried at 100 ¡C, to
give AMPS 1.
Following a method previously published,12 a set of sup-
ported phenolate catalysts was prepared via method a: 4-
Hydroxybenzaldehyde (1.22 g, 10 mmol) was added to AMPS
(5 g) in ethanol (100 ml) and stirred at room temperature for
approximately 5 min. A distinct colour change was seen from
white to yellow. This mixture was Ðltered and the solid
washed with ethanol until the washings ran colourless. In 1 : 1
waterÈmethanol solvent mixture (50 ml : 50 ml), sodium
hydrogen carbonate was dissolved (1 g), and then the solid
from the previous stage was added and stirred for 1 h. This
mixture was then Ðltered and the solid washed with water (100
ml) and methanol (100 ml) and Ðnally dried at 100 ¡C to give
2a.
Method b was developed to prepare catalysts under more
anhydrous conditions: Sodium hydroxide (0.4 g, 10 mmol)
was dissolved in methanol (50 ml). 4-Hydroxybenzaldehyde
(1.22 g, 10 mmol) was then added to the sodium hydroxide
solution and stirred at room temperature for approximately
20 min. Then AMPS (5 g) was added to the solution along
with methanol (50 ml) and stirred for 24 h. This mixture was
then Ðltered and the solid washed with methanol, until the
washings ran colourless, and dried at 100 ¡C.
Catalytic procedures
Knoevenagel condensations between pentan-3-one (0.86 g, 10
mmol) and ethyl cyanoacetate (1.13 g, 10 mmol) were carried
out in reÑuxing cyclohexane (15 ml) with a Dean and Stark
trap to remove water. In all cases biphenyl was employed as
an internal standard (0.15 g, 1 mmol) and 0.25 g of each cata-
lyst was used.
Michael reactions were performed at reÑux temperature.
Nitromethane was employed as the solvent as well as a sub-
strate (15 ml). 2-Cyclohexen-1-one was the other reactant (0.96
g, 10 mmol), and nitrobenzene was used as an internal stan-
dard (0.1 ml, 1 mmol). In each case 0.25 g of catalyst was
employed. Samples were periodically withdrawn from both
reactions and analysed by GC.
Results and discussion
Catalysts prepared via method a
Elemental analysis of the catalysts prepared via method a
(Scheme 2) showed a low surface loading of the phenolate
groups with only traces of sodium; results are shown in Table
1.
Catalysts not containing an ÈONa unit were made using
the aldehyde of the group to be added to the surface. Each
aldehyde (10 mmol) was stirred in ethanol (100 ml) with
AMPS (5 g) overnight at room temperature. Each of the cata-
lysts was then Ðltered o† and washed with ethanol (approx.
100 ml), and dried at 100 ¡C.
AMPS has a surface loading of approximately 1 mmol
g~1.6 A carbon : nitrogen ratio of 3 : 1 was thus expected for
AMPS, and 10 : 1 for the phenolate catalysts 2È4. The AMPS
catalyst showed the expected results. The “extraÏ carbon seen
in Table 1 is typical and can be attributed to non-hydrolysed
methoxy groups on the silicon of the grafted unit or on the
silicon dioxide surface.14 Catalyst 2a (OH form) is the 4-substi-
Characterisation of the catalysts
The infrared spectrum of AMPS was typical, showing a broad,
weak band at 1650È1560 cm~1, corresponding to OH
def
, as well as CÈH at
of
bound water and silanols and NH
tuted phenol, that is before treatment with NaHCO . Cata-
lysts 2a, 3a and 4a, however, show low C : N ratios and very
2 def
str
3
2800È2950 cm~1. The supported phenol and phenolate
spectra showed additional bands of moderate intensity, which
corresponded with imine stretches in the range of 1690È1640
small sodium loadings. This indicates that there is signiÐcantly
less than 100% conversion to imine and incomplete exchange
of OH to ONa. This may be due to an incomplete coupling
reaction, although such reactions are typically quantitative. A
more likely explanation is that the imine bonds, known to be
susceptible to hydrolysis,15,16 are partially cleaved from the
surface. During preparation, the solution of sodium hydrogen
carbonate in aqueous methanol facilitates the cleavage of the
imine bond. Some of the phenolate groups are thus removed
cm~1.13 Other bands (e.g., C ÈH ) from the phenolate
arom str
moiety were too weak to be assigned, due to the broad,
intense OÈH of the support. Elemental analysis showed a
str
low surface loading of phenol and phenolate groups for the
catalysts prepared via method a. Increased surface loadings
were seen for the catalysts prepared via method b. These
analyses are given in Table 1.
Table 1 Elemental analysis of the phenolate catalysts prepared via methods a and b
Loading of
N/mmol g~1
Loading of
C/mmol g~1
Loading of
Na/mmol g~1
Catalyst
C : N ratio
1
0.85
0.81
0.77
0.80
0.79
0.54
0.46
0.68
3.16
6.86
2.94
5.86
4.13
5.60
4.60
6.93
3.17
8.50
3.80
7.30
5.20
10.37
10.00
10.19
È
È
2a (OH form)
2a
3a
4a
2b
3b
4b
0.02
0.07
0.06
0.15
0.33
0.30
592
New J. Chem., 2000, 24, 591È595

View Online
bonds, since this is not consistent with the activity of the cata-
from the surface of the catalyst and then washed away in the
Ðltrate.
lysts before and after the removal of the “looseÏ imines (see
later). Rather, changes in the nature of the environment of the
remaining groups, such as altered hydrophobicity, hydrophi-
licity, organophilicity and reduced extent of lateral inter-
actions between the remaining groups are more likely to be
involved.
New catalysts prepared by method b
In light of these results, method b was developed for the prep-
aration of the catalysts under much more anhydrous condi-
tions, Scheme 2. An alternative source of sodium ions was
used, which would dissolve in a solvent other than water.
Also, the sodium ions were exchanged before the imine was
made, to minimise cleavage of the imine bond.
The elemental analysis results showed that higher surface
loadings were achieved by using the new method of prep-
aration. The above substituted phenolate catalysts were made
following the new method with increased surface loadings.
The results can be seen in Table 1.
The catalysts made via the new method, catalysts 2b, 3b and
4b, (Scheme 1), showed a high surface loading of phenol
groups with a higher conversion to phenolate groups than the
previous catalysts made via method a. The C : N ratio indi-
cates essentially 100% conversion to imine and the sodium
loadings are increased, which shows a conversion of ArOH to
ArONa of 28%, 36% and 44% for 2b, 3b and 4b, respectively.
IR analysis conÐrmed the presence of an imine band in all
three catalysts, as with the previous catalysts. Thermal
analysis showed a small loss in weight at about 100 ¡C, which
corresponds with the loss of either residual solvent and/or
water. There was also a small loss (ca. 1%) at approximately
250 ¡C, which could be the loss of physisorbed groups from
the surface. No other losses were seen up to a maximum tem-
perature of 625 ¡C.
Knoevenagel reaction (Scheme 3)
The catalysts were compared in the Knoevenagel reaction of
ethyl cyanoacetate and pentan-3-one, a reaction in which they
have not previously been evaluated. As can be seen in Table 2,
the activity of the phenolate imine catalysts is excellent, being
signiÐcantly more active than AMPS (one of the most active
base catalysts for this reaction). This demonstrates that the
catalysis observed is not due to the residual amino groups. It
can also be seen that the catalysts produced by method b are
less active than those made using method a.
In order to investigate whether it is the few remaining sites
that are the most active, catalyst 4b was washed with water
for 36 h, to take o† the labile surface groups. In doing this the
activity of catalyst 4b was increased. The results can be seen in
Graph 2. This was repeated for all of the catalysts produced
via method b and the same increase in activity was seen.
E†ect of changing or removing the phenolate cation. The
cation was changed (by reaction of phenols with KOH, LiOH
and NMe OH) to test the e†ect it would have on the activity
4
of the catalyst in the Knoevenagel reaction. These catalysts
(catalysts 5, 6 and 7, Scheme 4), showed activities similar to
those of the sodium containing catalysts, showing that the
cation has little e†ect on activity. Equivalent catalysts were
also made that either contained the OH group, rather than
the phenolate, or had no phenol groups and these were also
used in the Knoevenagel reaction (catalysts 8 and 9, Scheme
Solution stability study
To prove that the imine bond within the catalysts is suscep-
tible to cleavage in aqueous solutions, and to deÐne reaction
conditions under which the catalyst would be stable, a study
into catalyst stability was undertaken. Catalyst 4b was used in
all cases. Graph 1 shows the results of this study.
The results show a large loss in phenolate groups in
aqueous basic and acidic solutions, but negligible losses in
alcohol. The main loss in surface groups occurs very rapidly,
with negligible losses afterwards. This shows that rapid cleav-
age of the imine bond occurs. This is therefore likely to be the
reason for the low surface loading associated with method a
preparation. When only 2 molar equiv. of water are employed,
under the conditions used to produce the imine bond initially,
only 5% of surface groups are lost. E†ective deprotonation of
the phenol without cleavage of the imine is thus likely to be
very sensitive to the exact reaction conditions, thus use of an
alcoholic base may be more successful. This also indicates that
while some of the imine bonds are cleaved easily there are
others that are held di†erently or bound more strongly to the
surface and are not cleaved so readily. These groups could
also be located in varying chemical environments, making
cleavage easier or harder. Lack of accessibility is unlikely to
be a possible reason for the stability of the remaining imine
Scheme 3 Knoevenagel reaction of ethyl cyanoacetate and pentan-3-
one.
Table 2 Comparison of various catalysts in the Knoevenagel reac-
tion of ethyl cyanoacetate and pentan-3-one
%Yield after time/h
0.5
1
2
3
4
5
6
Phenolate catalysts prepared via methods a and b
1
10.6
38.0
15.3
43.1
31.8
5.6
16.5
53.0
22.6
57.5
43.8
7.7
23.9
71.8
31.0
72.6
55.2
10.4
45.3
30.1
75.8
36.7
76.5
60.8
11.7
49.6
36.8
78.8
38.5
82.8
64.2
13.0
51.5
43.7
87.1
39.7
82.0
66.7
14.0
52.1
49.1
88.3
44.6
86.8
67.1
14.6
52.5
2a
3a
4a
2b
3b
4b
28.6
37.4
Phenolate catalysts with no or di†erent active cations
5
6
7
8
9
28.9
31.2
24.4
39.2
33.0
39.2
41.5
32.5
54.4
41.0
49.3
52.4
40.0
62.0
49.1
54.7
58.6
45.2
69.3
54.5
58.4
62.6
47.7
69.6
60.0
60.0
63.7
48.9
71.9
60.0
60.7
64.8
49.7
81.1
60.0
Imine catalysts
10
11
12
13
14
15
16
6.0
27.5
44.0
16.5
38.0
67.2
37.6
9.2
50.6
58.1
23.4
53.7
69.3
49.4
13.6
69.1
69.8
29.8
67.1
81.5
70.5
16.3
69.2
72.7
31.6
72.7
88.2
74.9
17.9
76.0
76.7
32.5
74.4
92.9
80.3
19.4
75.7
82.8
33.7
79.0
98.6
86.8
23.3
È
È
34.8
80.1
99.5
86.2
Graph 1 Catalyst stability when stirred in di†erent solutions; (…)
water, (=) methanol, (>) dilute NaHCO , (+) dilute HCl.
3
New J. Chem., 2000, 24, 591È595
593

View Online
Graph 2 Comparison of phenolate catalysts prepared via methods a
and b and washed catalysts prepared via method b: (+) 2a, (=) 2b,
(+) 2b washed.
4). These catalysts displayed similar behaviour to the phenol-
ate catalysts. The results can be seen in Table 2 and Graph 3.
Thus it can be seen that changing the cation or removing it
completely has little e†ect on the activity of the catalyst. It is
the imine group, therefore, which is providing the basicity that
catalyses the reaction. This has also been found in other sup-
ported systems.17 This explains why the 2,4-substituted phe-
nolate catalysts are lower in activity in this case. The nitrogen
lone pair is delocalised within the H-bonding system and
therefore, the catalyst is far less basic.4,5,11 This can be seen in
Scheme 5. Similar results have been obtained for the 2-
substituted phenolate catalyst. Also, it is the lower loading
catalysts that are more active so the few imine bonded sites
that are not cleaved by hydrolysis are actually responsible for
the catalytic action of the catalyst. Intermolecular H-bonding
Scheme 6 Imine catalysts with di†erent substituent groups, com-
pared in the Knoevenagel reaction of ethyl cyanoacetate and pentan-
3-one.
through lateral interactions could explain the reduced activity
of the catalysts with a higher surface loading.
Non-phenolic catalysts. To investigate this theory, a series of
imine-containing catalysts, shown in Scheme 6, were prepared
from di†erent aldehydes. The results in Table 2 show that the
di†erent groups have an inÑuence on the activity of the cata-
lysts. They also show the e†ectiveness of the imine group as a
base catalyst. While the reasons for the di†erent activities seen
in this group are not fully understood, it appears that more
bulky substituents such as the tert-butyl group favour higher
activity, a factor that may reduce SiOHÈimine H-bonding
interactions, making the imine more available for base cataly-
sis.
Michael reaction
A selection of catalysts were tried in the Michael reaction of
nitromethane and 2-cyclohexen-1-one, Scheme 7, and the
results can be seen in Table 3. Initially, the imine catalysts
were more active than the phenolate catalysts, though after
6 h there was little di†erence in conversion. Heavier products
were seen in all the reactions though it was only the phenolate
catalysts that showed a decrease in yield. This indicates that
some subtle di†erences do exist between the phenolate and
imine catalysts in some cases. Again, the imine group appears
to be the catalytic centre of the catalysts.
Scheme 4 Phenolate catalysts with di†erent active cations and with
the active cations removed, compared in the Knoevenagel reaction of
ethyl cyanoacetate and pentan-3-one.
Scheme 7 Michael reaction of nitromethane and 2-cyclohexen-1-one.
Table 3 Comparison of phenolate catalysts prepared via methods a
and b and other imine catalysts in the Michael reaction of nitro-
methane and 2-cyclohexen-1-one
Graph 3 Comparison of phenolate catalyst, phenol catalyst and
catalyst with no active cation: (+) 2b, (=) 8, (+) 9.
%Yield after time/h
Catalyst
0.5
1
2
3
4
5
6
2a
2b
12
14
16
27.5
24.5
23.6
13.6
22.0
36.6
34.1
32.7
21.3
31.5
48.0
44.1
42.2
31.8
37.8
52.4
47.8
47.7
39.1
42.5
57.6
48.0
53.0
45.8
43.2
52.8
47.2
54.0
48.2
44.0
50.8
45.8
57.6
53.0
45.1
Scheme 5 Intramolecular hydrogen bonding found in the 2,4-substi-
tuted phenolate catalyst.
594
New J. Chem., 2000, 24, 591È595

View Online
Conclusions
References
1
2
3
H. Hattori, Chem. Rev., 1995, 95, 537.
J. H. Clark and D. J. Macquarrie, Chem. Soc. Rev., 1996, 303.
J. W. Smith, in T he Chemistry of the Amino Group ed. S. Patai,
Interscience, London, 1966, pp. 235È253.
J. W. Smith, in T he Chemistry of the C2N Bond, ed. S. Patai,
Interscience, London, 1966, pp. 161È204.
D. J. Macquarrie, J. H. Clark, A. Lambert, J. E. G. Mdoe and A.
Priest, React. Polym., 1997, 35, 153 and references therein. D. J.
Macquarrie, J. H. Clark, A. Lambert, J. E. G. Mdoe and A.
Priest, Supported Reagents and Catalysis in Chemistry, ed. B. K.
Hodnett, A. P. Kybett, J. H. Clark and J. K. Smith, Royal Society
of Chemistry, Cambridge, 1998, 174.
D. J. Macquarrie and D. B. Jackson, Chem. Commun., 1997, 1781.
M. Lasperas, T. Llorett, L. Chaves, I. Rodriguez, A. Cauvel and
D. Brunel, Stud. Surf. Sci. Catal., 1995, 108, 75.
A series of supported phenolate catalysts were prepared by
attaching formyl phenols to aminopropyl silica. These imine-
bonded catalysts outperform AMPS in catalytic activity in the
Knoevenagel reaction. The stability of the imine bond was
investigated. A large initial loss of surface groups is seen
through hydrolysis, whereas only small losses are seen after.
This abrupt change indicates that a small amount of imine
groups are held in a di†erent environment or bound more
strongly to the surface. It is these strongly held groups that
appear to be the most catalytically active centres of the cata-
lyst. Changing the active cation, and even removing it, has
little or no e†ect on the activity of the catalyst. Changing the
groups attached to the imine bond has a signiÐcant e†ect on
the reactivity of the catalyst, increasing or decreasing the
basicity due to the mesomeric and inductive e†ects of the “end
groupsÏ. The 2,4-substituted phenolate demonstrates that H-
bonding e†ectively “ties upÏ the nitrogen lone pair, which in
turn decreases activity. Imines that are bound in a certain way
and that are not susceptible to hydrolysis are the source of the
catalytic activity of the catalyst.
4
5
6
7
8
9
J. Weinstein and E. McIninch, J. Am. Chem. Soc., 1960, 82, 6064.
N. V. Sidgwick, Organic Chemistry of Nitrogen, Clarendon Press,
Oxford, 3rd edn., 1966, pp. 19È36.
10 P. A. Kollman and L. C. Allen, Chem. Rev., 1972, 72, 283.
11 A. W. Baker and A. T. Shulgin, J. Am. Chem. Soc., 1959, 81, 1523.
12 D. J. Macquarrie, T etrahedron L ett., 1998, 39, 4125.
13 D. H. Williams and I. Fleming, Spectroscopic Methods in Organic
Chemistry, McGrawÈHill, London, 5th edn., 1995, pp. 28È62.
14 D. Brunel, A. Cauvel, F. Fajula and F. DiRenzo, Stud. Surf. Sci.
Catal., 1995, 97, 173.
15 R. W. Layer, Chem. Rev., 1963, 63, 489.
Acknowledgements
D.J.M. thanks the Royal Society for a University Research
Fellowship, K.A.U. thanks EPSRC for a project studentship.
16 A. Bruylants and E. FeytmantsÈDe Medicis, in T he Chemistry of
the CarbonÈNitrogen Double Bond, ed. S. Patai, Interscience,
London, 1966, pp. 465È504.
17 D. Brunel and A. Blanc, in preparation.
New J. Chem., 2000, 24, 591È595
595