Use of immobilized organic base catalysts for continuous-flow fine chemical synthesis

Article abstract of DOI:10.1016/j.jcat.2008.07.005

A continuous-flow two-step process for the production of multifunctional intermediate compounds was performed with a combination of two column reactors packed with different organic immobilized catalysts. The first step of the reaction, nitroaldol condensation between aromatic aldehydes and nitromethane, is promoted by a silica gel-supported primary amine; the second step, Michael addition of β-dicarbonyl compounds to the nitrostyrenes obtained in the first step, is catalyzed by a silica gel-supported guanidine TBD. In both cases, the catalysts can be directly reused for several runs with no significant changes in activity.

Full text of DOI:10.1016/j.jcat.2008.07.005

Journal of Catalysis 258 (2008) 289–295
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Use of immobilized organic base catalysts for continuous-flow fine chemical
synthesis
L. Soldi , W. Ferstl , S. Loebbecke , R. Maggi , C. Malmassari , G. Sartori , S. Yada b
a
b
b
a
a
a,∗
a
Clean Synthetic Methodologies Group, Dipartimento di Chimica Organica ed Industriale dell’Università and Consorzio Interuniversitario “La Chimica per l’Ambiente” (INCA), UdR PR2,
Viale G.P. Usberti 17A, I-43100 Parma, Italy
b
Fraunhofer Institute for Chemical Technology ICT, Joseph von Fraunhofer Str. 7, 76327 Pfinztal, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
A continuous-flow two-step process for the production of multifunctional intermediate compounds was
performed with a combination of two column reactors packed with different organic immobilized
catalysts. The first step of the reaction, nitroaldol condensation between aromatic aldehydes and
nitromethane, is promoted by a silica gel-supported primary amine; the second step, Michael addition
of β-dicarbonyl compounds to the nitrostyrenes obtained in the first step, is catalyzed by a silica gel-
supported guanidine TBD. In both cases, the catalysts can be directly reused for several runs with no
significant changes in activity.
Received 21 May 2008
Revised 9 July 2008
Accepted 10 July 2008
Keywords:
Nitroaldol condensation
Michael addition
Heterogeneous catalysis
Organic–inorganic hybrid composites
Supported catalysts
© 2008 Elsevier Inc. All rights reserved.
Continuous-flow processes
1. Introduction
The use of catalysts immobilized onto solid supports for the
liquid-phase fine chemical synthesis allows their easy separation
Numerous products for fine, pharmaceutical, and material
chemistry with an ever-increasing level of complexity have been
developed over the past several decades through multistep syn-
thetic methods. In multistep synthesis, the preparative difficulties
increase significantly with the number of steps, due mainly to
extensive isolation and purification operations. These drawbacks
can be overcome by exploiting a multistep-sequential synthetic
methodology, namely by adding reagents and/or catalysts in a
sequential manner without isolating the previously formed inter-
mediates [1]. Moreover, in many cases, multicomponent reactions
can be used to assemble three or more starting compounds into
the final target molecule in a single reaction vessel.
The successful outcome of a given multistep sequential or mul-
ticomponent process depends on a fine balance of equilibria and
a convenient sequence of reversible and irreversible steps [2]. In
principle, complex organic compounds also can be prepared using
a combination of column reactors with supported catalysts and/or
reagents [3]. This approach uses columns packed with immobi-
lized organic or inorganic catalysts/reagents. Starting materials are
added to the top of the column and by means of gravity or pres-
sure passed through the packing, where they react, giving the
product recovered at the bottom of the setup [4].
from the reaction mixture and reuse for many runs, finally achiev-
ing great advantages, such as saving energy, minimizing waste
production, and delivering products of higher purity [5]. Moreover,
supported catalysts can be applied in packed-bed or monolithic re-
actors for continuous-flow processes [6,7].
These relatively new technologies offer many fundamental and
practical advantages of relevance to the pharmaceutical and fine
chemical industries, including facile automation, reproducibility,
safety, and process reliability [7–9]. The improved performance of
these catalytic reactors is attributed to faster heat transfer and
mixing as a result of the increased surface area-to-volume ra-
tio [10]. But despite the growing number of studies on this topic,
only a few reports to date have discussed the use of catalytic
packed-bed (micro) reactors applied to continuous-flow synthetic
chemistry [11,12].
As a result of a more general project aimed at achieving basic
information to design micro and monolithic reactors for organic
syntheses, we have recently patented the continuous-flow synthe-
sis of enatiomerically enriched (R,R)-chrysanthemic acid through a
tubular reactor packed with silica-supported chiral copper(II)-BOX
catalyst [13]. Here we present the results of our studies on the
setup of two catalytic reactors able to produce nitroaldol conden-
sation and the Michael reaction under continuous-flow conditions
(
Scheme 1) [14]. The combination of both processes allows the
Corresponding author. Fax: +39 0521 905472.
E-mail address: giovanni.sartori@unipr.it (G. Sartori).
*
production of multifunctional compounds 5, which are valuable
0
021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.07.005

290
L. Soldi et al. / Journal of Catalysis 258 (2008) 289–295
Scheme 1. Multistep synthesis of products 5.
◦
intermediates in fine and pharmaceutical chemistry, because ni-
trocompounds can be transformed into amines, ketones, carboxylic
acids, and other important components [15,16].
According to Scheme 1, the multifunctional compounds 5 can
be produced by nitroaldol condensation directly between aromatic
aldehydes 1 and nitromethane 2, giving nitrostyrenes 3 (step 1),
followed by Michael addition of β-dicarbonyl compounds 4 to the
electron-poor carbon–carbon double bond of compounds 3 (step 2)
of the KG-60 silica, previously dried at 350 C overnight, using
trialkoxyorganosilanes. In particular, 3-aminopropyltriethoxysilane,
N-methyl-3-aminopropyltrimethoxysilane, N,N-diethyl-3-amino-
propyltrimethoxysilane, and 3-(2-aminoethylamino)propyltrimeth-
oxysilane gave solid basic catalysts, referred to as KG-60-NH2 [21],
KG-60-NHMe [22], KG-60-NEt2 [23], and KG-60-NH(CH ) NH [24],
2
2
2
respectively.
A typical preparation was as follows. Dried silica (10 g) was
added to toluene (150 ml), followed by the selected trialkoxysilane
[
17,18].
(
30 mmol). The mixture was refluxed for 18 h under mechani-
2
. Experimental
cal stirring. The resulting slurry was cooled to room temperature
and the modified silica was filtered, washed with toluene (50 ml),
acetone (50 ml), water (50 ml), and finally acetone (2 ×50 ml). Af-
ter being allowed to dry at room temperature under vacuum, the
2
.1. Materials
All chemicals and solvents were used as received except ben-
◦
modified silica was dried in an oven at 100 C for 10 h.
zaldehyde, which was distilled before use. Thin-layer chromatogra-
phy (TLC) was performed with Merck 60 F254 silica gel plates, and
compounds were visualized with UV light.
2
1
.2.2. Preparation of silica-supported
,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD)
Starting materials for catalyst preparation were as follows: silica
The supported catalyst was prepared by modifying the method
of Jacobs et al. [25] in two steps. First, (3-glycidyloxypropyl)-
trimethoxysilane (10 ml, 46.5 mmol) was added dropwise under
nitrogen to a solution of TBD (6.5 g, 46.9 mmol) in dry DMF
2
gel KG-60 (size 0.040–0.080 mm; surface area 480–540 m /g; pore
volume 0.74–0.84 cm /g), (3-aminopropyl)triethoxysilane (99%,
3
Aldrich), (3-diethylaminopropyl)trimethoxysilane (ꢀ90%, Fluka),
(
3-methylaminopropyl)trimethoxysilane (ꢀ97%, Fluka), 3-(2-ami-
(65 ml). The resulting mixture was stirred for 24 h at room
nomethylaminopropyl)trimethoxysilane (>98%, Fluka), (3-glyci-
dyloxypropyl)trimethoxysilane (>97%, Fluka), 1,5,7-triazabicyclo-
temperature. Second, a suspension of KG-60 silica (20.1 g pre-
viously sieved and dried) in dry toluene (150 ml) was refluxed
for 1 h under nitrogen, after which the solution obtained in the
first step was added and refluxed for additional 16 h under ni-
trogen. The modified silica thus obtained was filtered and washed
with toluene (120 ml), dichloromethane (200 ml), and methanol
[4.4.0]dec-5-ene (98%, Aldrich), and 1,1,1,3,3,3-hexamethyldisila-
zane (99.9%, Aldrich).
Starting materials for synthetic studies were as follows: benz-
aldehyde (ꢀ99%, Fluka), 4-methylbenzaldehyde (ꢀ96%, Fluka),
4
-chlorobenzaldehyde (ꢀ97%, Aldrich), 4-methoxybenzaldehyde
(100 ml), then extracted with a diethyl ether dichloromethane
(
98%, Aldrich), ethyl acetoacetate (ꢀ99%, Riedel–de-Haën), 2,4-pen-
mixture (150/150 ml) in a Soxhlet apparatus overnight. Finally, the
modified silica was dried.
tanedione (ꢀ99%, Aldrich), diethylmalonate (99%, Aldrich), and ni-
tromethane (96%, Aldrich).
2
.2.3. Preparation of the capped KG-60-TBD-SiMe3 [26]
Dried silica gel KG-60 (8 g), previously heated at 350 C for 3 h
2.2. Catalyst preparation
◦
and dried by vacuum at room temperature for 2 h, was introduced
into a flask containing dry toluene (120 ml). Then (3-glycidyloxy-
propyl)trimethoxysilane (4.0 ml, 17 mmol) was added, and the
mixture was heated at reflux under stirring for 18 h. The modi-
fied silica was filtered and washed with dichloromethane (20 ml).
The powder was then mixed with HMDS (19.4 g, 120 mmol),
and the mixture was heated at 60 C under stirring for 2 h. The
powder was filtered and washed with dichloromethane (20 ml),
then introduced into a flask containing a solution of TBD (2.45 g,
Silica-supported organic catalysts can be prepared by a sol–gel
or postmodification method. We used the postmodification ap-
proach for the preparation of all silica-supported bases, because
catalysts with lower base loads gave better yields in shorter re-
action times [19]. The higher activity of the catalysts prepared by
the postmodification method compared with those prepared by the
2
◦
sol–gel method likely is due to the larger surface area (∼250 m /g
2
for the postmodification method vs. ∼135 m /g for the sol–gel
method [20]).
17.58 mmol) in dry toluene (120 ml), with the resulting mixture
2
.2.1. Preparation of KG-60-supported amines
Catalysts were prepared according to the methods reported in
stirred at room temperature for 10 h. The catalyst was filtered,
washed with dichloromethane (20 ml), and extracted with a Soxh-
let apparatus for 10 h using a 1:1 dichloromethane/diethyl ether
the literature by grafting the organic groups to the surface silanols

L. Soldi et al. / Journal of Catalysis 258 (2008) 289–295
291
solution (200 ml). Finally, the catalyst was filtered and dried under
vacuum at room temperature.
supported amine; 4.5% mol) mixed with KG-60 silica sieved at
◦
80 mesh and previously heated at 350 C for 3 h (0.65 g), with ni-
◦
tromethane (10 ml). The column was heated to 110 C, after which
2.3. Catalyst characterisation
a solution of distilled benzaldehyde (0.87 g, 8.20 mmol) in dry ni-
tromethane (4.2 ml) was added. The solution was passed through
a syringe pump at 0.5 bar pressure with a 0.25 ml/min flow
rate. The crude was recovered after the column was rinsed with
All materials were characterised with respect to their composi-
tional, textural, and surface properties. The loadings of the organic
groups were calculated from the nitrogen content by elemental
analysis performed with a Carlo Erba CHNS-0 EA1108 elemental
analyzer.
◦
dichloromethane (30 ml) at 35 C. Nitrostyrene 3a was produced
in a 97% yield (GC analysis). All solvents were distilled off, and the
resulting yellow oil was diluted with a solution of ethyl acetoac-
etate (3.1 g, 24.0 mmol) in THF (15 ml). This solution was passed
through a second column, similar to that used previously, packed
with KG-60-TBD-SiMe3 (1.08 g, 0.79 mmol) in THF (15 ml) for 1 h
at room temperature under 0.2 bar pressure. The crude was re-
covered, and the column was rinsed with ethyl acetate (20 ml).
Product 5b was obtained in an 85% overall yield. Products 3a–d
and 5a–c gave melting points and spectral data consistent with
values reported previously [22,27,28].
Thermogravimetric analyses, performed with a Setaram TG-DTA
92 thermobalance in He/O2 flow, revealed that all of these ma-
terials were thermally stable and thus did not release organic
◦
fragments until 300 C. The N2 adsorption–desorption isotherms,
◦
carried out at −196 C on a Micromeritics ASAP 2010, were used to
determine the specific surface areas, SABET. Before each measure-
◦
−4
Pa for
ment, the samples were outgassed at 110 C and 1.33 ×10
12 h.
2
2
.4. Reaction procedures
3
. Results and discussion
.4.1. Nitroaldol condensation under batch conditions
To a solution of the selected benzaldehyde (2.46 mmol) in dry
Amines supported on silica and organic polymers have been
nitromethane (3 ml), the catalyst (3% mmol of supported amine
used occasionally to promote nitroaldol and Michael reactions un-
der batch conditions [19,23,29,30]. However, the problem of per-
forming the reactions under continuous-flow conditions remains
to be addressed. To effectively compare the catalytic activity of
supported basic organic catalysts in both reactions with that ob-
served in previous studies, we carried out a comparison using an
uniform series of solid catalysts prepared by anchoring different
organic bases to the same support, namely amorphous silica KG-
with respect to aldehyde) was added. The mixture was heated at
◦
7
5 C and stirred for 1 h. The crude was then filtered and the cat-
alyst recovered after washing with ethyl acetate (20 ml). Finally,
yield, conversion and selectivity were determined by gas chro-
matography (GC) through internal standard analysis.
2
.4.2. Nitroaldol condensation under continuous-flow conditions
KG-60 silica (0.77 or 0.65 g depending on KG-60-NH2 amount),
60, through the same postsynthetic method. In contrast, in the
◦
previously sieved to 80 mesh and dried at 350 C for 3 h, was
mixed with the KG-60-NH2 catalyst (0.23 g, 0.25 mmol or 0.35 g,
previous studies on this topic, reactions were carried out with
catalysts supported on different materials (e.g., amorphous silica,
micelle-templated silica, mesoporous MCM-41 silica, polystyrene)
and/or prepared with different methods (e.g., postsynthesis, sol–
gel, copolymerization), yielding results that are not directly com-
parable.
A nonswelling material such as amorphous silica KG-60, sieved
at 80 mesh, was used as the support, because a major issue en-
countered in packing channels with supported catalysts is the
pressure drop across the channel caused by either swelling or the
size of the packing material [31,32]. In the nitroaldol condensation
between benzaldehyde and nitromethane [20] (Scheme 1, step 1),
primary, secondary, and tertiary amines, as well as TBD, were used
as catalysts supported on amorphous silica KG-60 through a con-
venient spacer.
0.37 mmol; 3.0% and 4.5% mol with respect to benzaldehyde).
The powder was packed in the reactor, an oil-jacketed steel HPLC
column (6.0 cm length × 0.8 cm inner diameter), along with
nitromethane (10 ml). The jacketed column was heated to the
selected temperature, and a solution of the aromatic aldehyde
(8.20 mmol) in dry nitromethane (4.2 ml) was passed under pres-
sure at a selected flow rate using a syringe pump. The crude was
collected at the bottom of the column after rinsing with ethyl ac-
etate (20 ml).
2
.4.3. Michael addition of β-dicarbonyl compounds to nitrostyrene
under batch conditions
To a solution of nitrostyrene (0.30 g, 2.0 mmol) in the selected
solvent (1.5 ml), β-dicarbonyl compound (2.0 mmol) and the cat-
alyst (10% mmol) were added. The mixture was stirred at room
temperature under nitrogen for 4 h. The catalyst was filtered and
washed with ethyl acetate (15 ml), after which the crude was col-
lected.
Batch reactions were carried out by stirring freshly distilled
benzaldehyde (2.46 mmol) in nitromethane (3 ml) used as a
◦
solvent-reagent for 1 h at 75 C in the presence of 3% mol catalyst
(with respect to benzaldehyde) under nitrogen, to avoid oxidation
of the benzaldehyde to benzoic acid, which could react with the
supported amine giving the amide and definitively poison the cat-
alyst [33]. The reaction directly gave (E)-nitrostyrene 3a as the sole
isomer, accompanied by polynitroderivative byproducts due to the
polymerization of 3a and/or addition of a second molecule of ni-
tromethane to 3a.
The results reported in Table 1 confirm that the yield and se-
lectivity of product 3a were affected mainly by the nature of the
supported amine rather than by the physicochemical parameters
of the catalysts, such as surface area or basic strength. Indeed,
different supported amine catalysts with comparable surface areas
and similar amine loadings gave product 3a with completely dif-
ferent yield and selectivity (Table 1, entries 2 and 4). Moreover, the
supported propylamine gave product 3a with yield and selectivity
greater than that of the more basic supported TBD (propylamine
2
.4.4. Michael addition of β-dicarbonyl compounds to nitrostyrene
under continuous-flow conditions
KG-60-TBD-SiMe3 catalyst sieved at 80 mesh (1.00 g, 0.73 mmol)
was packed in an oil-jacketed steel HPLC column (similar to that
used for the nitroaldol condensation) with THF (10 ml). Then a so-
lution of the β-dicarbonyl compound (22.0 mmol) and nitrostyrene
(1.08 g, 7.3 mmol) in THF (14 ml) was passed through under low
pressure by a syringe pump (∼0.2 bar) at room temperature for
1
h. The crude was collected at the bottom of the column.
2
3
.4.5. Continuous-flow two-step catalytic synthesis of
-carbethoxy-4-phenyl-5-nitro-2-pentanone (5b)
An oil-jacketed steel HPLC column (6.0 cm long × 0.8 cm i.d.)
was packed with KG-60-NH2 catalyst (0.35 g, 0.37 mmol of the

292
L. Soldi et al. / Journal of Catalysis 258 (2008) 289–295
Table 1
Catalytic efficiency of organic bases supported on amorphous silica KG-60 in the batch nitroaldol condensation between benzaldehyde and nitromethane as solvent-reagent
4
Entry
Catalyst
Loading
mmol/g)
r0 · 10
Surface area
(m /g)
1a conv.
(%)
3a yield
(Sel. ) (%)
a
2
b
(
(mol/(min g))
1
2
3
4
5
6
KG-60-NH2
KG-60-NH2
KG-60-NH(CH2)2NH2
KG-60-NHMe
KG-60-NEt2
0.65
1.06
1.24
0.88
1.12
0.67
4.24
6.55
254
254
225
245
210
260
76
100
52
75
28
75 (99)
98 (98)
36 (70)
28 (37)
4 (15)
KG-60-TBD
67
26 (38)
a
Initial rate of formation of 3a per gram of catalyst.
b
(
Yield/Conversion) × 100.
Table 2
efficiency (namely yield and selectivity in product 5a) depends on
the pKb value.
Catalytic efficiency of organic bases supported on amorphous silica KG-60 in the
batch Michael addition of acetylacetone to nitrostyrene
Concerning aliphatic amines, the secondary amine (pK ≈ 3.2)
b
Entry Catalyst
Solvent (EN)a
Loading
3a conv.b 5a yield
afforded product 5a in a 76% yield, whereas the less basic tertiary
T
c
(mmol/g) (%)
(Sel. ) (%)
amine (pK ≈ 3.4) gave product 5a in a lower yield (43%) (Table 2,
b
1
2
3
4
5
6
7
8
KG-60-NHMe
THF (0.207)
THF (0.207)
THF (0.207)
CH2Cl2 (0.309)
EtOH (0.654)
Toluene (0.099)
0.88
1.12
0.67
0.67
0.67
0.67
92
81
80
86
92
63
100
89
76 (82)
43 (53)
69 (86)
78 (90)
67 (73)
40 (63)
99 (99)
entries 1 and 2).
KG-60-NEt2
KG-60-TBD
KG-60-TBD
KG-60-TBD
KG-60-TBD
KG-60-TBD
The mostly higher yields and selectivities observed with TBD-
supported catalysts can be attributed mainly to their higher ba-
sicity, resulting in a greater concentration of the active acetylace-
tone carbanion [38]. This increases the Michael addition rate and
shortens the permanence time of the very reactive nitrostyrene in
contact with the siliceous material, thereby minimising byproduct
formation (Table 2, entries 2 and 3).
However, the activity of KG-60-TBD catalyst was strongly af-
fected by the polarity of the solvent, because the reaction in-
volves charged species in the transitions state (Table 2, entries 3,
Acetylacetone (0.571) 0.67
KG-60-TBD-SiMe3d THF (0.207)
0.73
87 (98)
e
a
b
c
Normalised ET(30) parameter (Reichardt) [37].
Relative to the conversion of nitrostyrene.
(
Yield/Conversion) × 100.
d
e
2
Surface area: 265 m /g.
Acetylacetone/nitrostyrene 3:1 molar ratio.
4, and 6). In particular, it was found that the higher the polarity
of the solvent, the greater the nitrostyrene conversion and product
yield.
The maximum catalytic efficiency was observed when acety-
lacetone was utilised as solvent-reagent, because under these con-
ditions, the concentration of the acetylacetone anion likely reached
the maximum value (Table 2, entry 7).
A seemingly anomalous behaviour was observed when the re-
action was carried out in ethanol (Table 2, entry 5): High ni-
trostyrene conversion (92%) was accompanied by a surprisingly
low 5a yield value (67%). This result can be rationalised by taking
into account that ethanol is a protic solvent able to catalyse the
nitrostyrene polymerisation, which also can be promoted by silica
surface free silanols [39]. Interesting results were achieved by pre-
venting the negative effect of the surface silanols through silaniza-
tion with hexamethyldisilazane (HMDS) using KG-60-TBD-SiMe3 as
catalyst with a molar ratio acetylacetone/nitrostyrene = 3:1 (Ta-
ble 2, entry 8) [40].
For both reactions (nitroaldol condensation and Michael addi-
tion), we performed the Sheldon test [41] to confirm the inactivity
of the filtered reaction mixture and the reusability of the catalyst
in both cases.
In light of the foregoing results, we planned to perform both re-
actions under continuous-flow conditions using the best catalysts
identified in the batch experiments. First, nitroaldol condensation
pK = 3.4, 3a yield = 98%, 3a selectivity = 98%; TBD pK = −11.0,
b
b
3
a yield = 26%, 3a selectivity = 38%) (Table 1, entries 2 and 6).
This behaviour has been attributed to the fast initial production
of a supported reactive imine intermediate through reaction of the
supported propylamine with benzaldehyde, as can be easily veri-
fied by FT-IR spectra [19]. Product 3a arose from the nitro-Mannich
addition of nitromethane (activated as nitronate anion) to the sup-
ported imine, followed by the β-scission [34]. As expected, when
the reaction was carried out in the presence of KG-60-NH2 samples
containing different amounts of anchored propylamine, the initial
reaction rate [35] was proportional to the amount of the supported
amine, indicating that the reaction should be of first order with re-
spect to the catalyst concentration (Table 1, entries 1 and 2).
We also analyzed the progress of benzaldehyde 1a conversion
versus time in the model reaction performed in the presence of
KG-60-NH2 catalysts. The conversion reached 100% after 60 min;
the data are better described by a first-order irreversible kinetic
−2
−3
−1
.
with k = 1.2 × 10
(1.9 × 10 ) min
Concerning the effect of the supported base, it is well recog-
nised that different basic catalysts, such as KG-60-NHMe, KG-60-
NEt2 and KG-60-TBD, that exhibit basic sites with medium to high
basic strength can promote the reaction through a typical nitroal-
dol mechanism involving addition of the nitronate anion to the
aldehyde [17]. The significant difference between the two chemi-
cal pathways (i.e., the nitro Mannich and nitroaldol mechanisms)
is the main factor accounting for the different catalytic activities.
The optimization of step 2 under batch conditions was suc-
cessively studied. Reactions were performed at room temperature
using nitrostyrene (2 mmol) and acetylacetone (2 mmol) as model
reagents in the presence of the supported catalysts used in the
previous step (10% molar ratio) in different solvents [36] (Table 2).
Table 2 reports only the results obtained with secondary and ter-
tiary amines as well as TBD, because the primary amine gave ex-
tremely poor yield and selectivity. In this case, the basic strength
of the supported catalyst plays an important role, and the catalytic
(
Scheme 1, step 1) was studied using a tubular reactor. The reac-
tor was prepared with an oil-jacketed steel HPLC column (6.0 cm
long × 0.8 cm i.d.) packed with 0.23 g of KG-60-NH2 “diluted” into
◦
0
.77 g of KG-60 silica heated previously to 350 C for 3 h. Both
siliceous materials were sieved at 80 mesh before use. The column
reactor was washed with nitromethane and used immediately, tak-
ing into account the optimum conditions found in the previous
batch experiments. A solution of benzaldehyde (0.87 g, 8.2 mmol)
in nitromethane (4.2 ml) was passed through the column at differ-
ent flow rates and temperatures by a syringe pump. Low pressure
(∼0.2 bar) was applied at the top of the column to keep the flow
constant during the entire residence time. The results are reported
in Table 3.

L. Soldi et al. / Journal of Catalysis 258 (2008) 289–295
293
Table 3
Table 4
Continuous-flow nitroaldol condensation of benzaldehyde and nitromethane
Preparation of substituted nitrostyrenes via continuous-flow nitroaldol condensation
over KG-60-NH2 catalyst (1.06 mmol/g propylamine loading)
a
(
solvent-reagent) over KG-60-NH2 catalyst (1.06 mmol/g propylamine loading)
b
Entry
Catalyst/1
molar ratio
Residence
time (min)
Flow rate
(ml/min)
Temp.
( C)
1a conv.
(%)
3a yield
(Sel. ) (%)
Entry
1
Product
3
Yield (Sel. ) (%)
95 (97)
◦
a
3a
1
2
3
4
5
3/100
3/100
3/100
3/100
4.5/100
25
35
50
35
12
0.12
0.08
0.06
0.08
0.25
56
56
56
75
110
40
60
65
76
98
33 (82)
45 (75)
50 (77)
61 (80)
95 (97)
2
3
3b
3c
3d
97 (98)
95 (98)
65 (93)
a
(
Yield/Conversion) × 100.
Benzaldehyde 1a conversion and nitrostyrene 3a yield were
evaluated from the entire reaction mixture recovered at the end
of the flow-through process. By using 3% mol of KG-60-NH2 with
respect to benzaldehyde passed, the conversion of 1a and the yield
of 3a increased with increasing residence time from 25 to 50 min,
reaching modest values in all cases (40%–65% 1a conversion; 75%–
4
82% 3a selectivity) (Table 3, entries 1–3).
Ideally, the flow rate should be slow enough to guarantee full
a
conversion of the starting materials and to give only the desired
product at the end of the reaction. However, by directly applying
the best reaction parameters achieved with batch experiments to
the column reactor, the selectivity toward nitrostyrene 3a dropped
down mainly due to the production of 1,3-dinitro-2-phenylpropane
by the further addition of nitromethane to nitrostyrene. To over-
come this drawback, the residence time was shortened to 12 min,
Experimental conditions: all reactions were performed using 8.2 mmol of the
selected benzaldehyde in 4.2 ml of nitromethane in the presence of 0.35 g of KG-
0-NH2 and 0.65 g of KG-60; residence time: 12 min, temperature: 110 C.
b
◦
6
(
Yield/Conversion) × 100.
Table 5
Preparation of some Michael adducts of nitrostyrene with β-dicarbonyl compounds
via continuous-flow procedure
a
◦
the reaction temperature was increased to 110 C, and the amount
b
Entry
1
Product
5
Yield (Sel. ) (%)
90 (98)
of catalyst was increased to 4.5%. Under these conditions, a low
pressure of ∼0.5 bar was applied to ensure a regular flow inside
the column. The efficiency of the column reactor was enhanced
significantly, and product 3a was obtained with 95% yield and 97%
selectivity without the need to isolate the intermediate nitroalco-
hol (Table 3, entry 5) [42].
5a
The catalytic reactor could be reused efficiently after simple
rinsing with nitromethane. A second reaction mixture was then
passed through the reactor under the same conditions, giving
product 3a in a 93% yield and 96% selectivity.
2
5bc
92 (98)
The continuous-flow condensation process was extended to
some aromatic aldehydes under the optimized conditions of
◦
1
2 min residence time and 110 C, giving the small series of substi-
tuted nitrostyrenes reported in Table 4 with interesting yields and
selectivities.
3
5c
60 (72)
Table 4 shows that the introduction of an electron-withdrawing
substituent at the para position of the aromatic aldehyde resulted
in a lower yield of the corresponding nitrostyrene (entry 4). This
phenomenon is in good agreement with the presence of supported
imine intermediacy. Indeed, in an early comparative study, San-
terre et al. [43] reported that the rate of imine formation from
substituted aromatic aldehydes and aliphatic amines decreased sig-
nificantly as the electron-withdrawing power of the substituent on
the aromatic ring increased.
We next investigated the continuous-flow Michael addition of
acetylacetone to nitrostyrene under the best conditions found in
the batch experiments (Scheme 1, step 2). We performed reac-
tions over the KG-60-TBD-SiMe3 catalyst using THF as the solvent.
Preparation of the packed column-reactor was similar to that de-
scribed above, using a tube with the same dimensions packed with
a
Experimental conditions: all reactions were performed using 7.3 mmol of ni-
trostyrene, 22.0 mmol of the β-dicarbonyl compound in 14.0 ml of THF in the
presence of 1.00 g of KG-60-TBD-SiMe3; residence time: 35 min, temperature: 25 C.
b
◦
(
4
Yield/Conversion) × 100.
c
8:52 diastereomeric mixture.
nitrostyrene and β-dicarbonyl compounds. Illustrative synthetic re-
sults are reported in Table 5. The lower reactivity observed with
diethyl malonate (Table 5, entry 3) likely can be ascribed to the
lower pKa value of the active methylene function [44].
Finally, the preparation of the multifunctional compound 5b
was studied by sequential two-step column catalytic methodol-
ogy (Scheme 2). Benzaldehyde (0.9 g, 8.6 mmol) diluted in ni-
tromethane (5 ml) was reacted through the KG-60-NH2 column-
1
.0 g of KG-60-TBD-SiMe3 (loading value, 0.73 mmol/g) sieved at
0 mesh. After the column-reactor was washed with THF, a solu-
tion of acetylacetone (2.2 g, 22.0 mmol) and nitrostyrene (1.08 g,
.3 mmol) in THF (14 ml) was passed through with a syringe pump
8
7
under low pressure (∼0.2 bar) at room temperature (residence
time, 35 min). Product 5a was obtained in 90% yield (98% selectiv-
ity). The same column-reactor could be used, after accurate rinsing
with ethyl acetate, to prepare some other Michael adducts between
◦
reactor at 110 C for 12 min of residence time under nitrogen,
affording nitrostyrene 3a in a 97% yield. The reaction mixture was
then subjected to distillation to remove the excess of nitromethane

294
L. Soldi et al. / Journal of Catalysis 258 (2008) 289–295
catalitici”). The authors thank the Centro Interdipartimentale Mis-
ure (CIM), “G. Casnati,” for the NMR measurements.
References
[
[
[
[
1] L.F. Tietze, V. Beifuss, Angew. Chem. Int. Ed. Engl. 32 (1993) 131.
2] A. Domling, I. Ugi, Angew. Chem. Int. Ed. 39 (2000) 3168.
3] A.M. Hafez, A.E. Taggi, T. Lectka, Chem. Eur. J. 8 (2002) 4114.
4] A.M. Hafez, A.E. Taggi, T. Dudding, T. Lectka, J. Am. Chem. Soc. 123 (2001)
10853.
[5] S.V. Ley, I.R. Baxendale, R.N. Bream, P.S. Jackson, A.G. Leach, D.A. Longbottom,
M. Nesi, J.S. Scott, R.I. Storer, S.J. Taylor, J. Chem. Soc. Perkin Trans. 1 (2000)
3815.
[6] G. Jas, A. Kirschning, Chem. Eur. J. 9 (2003) 5708.
[7] B.P. Mason, K.E. Price, J.L. Steinbacher, A.R. Bogdan, D.T. McQuade, Chem. Rev.
107 (2007) 2300.
[
[
8] S.J. Haswell, P. Watts, Green Chem. 5 (2003) 240.
9] A. Kirschning, H. Monenschein, R. Wittenberg, Angew. Chem. Int. Ed. 40 (2001)
6
50.
10] V. Hessel, H. Loewe, Chem. Eng. Technol. 28 (2005) 267.
11] A.R. Bogdan, B.P. Mason, K.T. Sylvester, D.T. McQuade, Angew. Chem. Int. Ed. 46
2007) 1698.
[
[
(
[12] C. Wiles, P. Watts, S.J. Haswell, Tetrahedron 60 (2004) 8421.
[13] S. Carloni, V. Borzatta, L. Moroni, G. Tanzi, G. Sartori, R. Maggi, WO patent 123
254 (2005).
[
14] When we were writing this paper, a quite attractive study on a similar topic
based on sophisticated multicatalyst was published: S.L. Poe, M. Kobašlija, D.T.
McQuade, J. Am. Chem. Soc. 129 (2007) 9216.
Scheme 2. Preparation of product 5b via sequential column catalytic methodology.
[
15] O.M. Berner, L. Tedeschi, D. Enders, Eur. J. Org. Chem. (2002) 1877.
16] A.G.M. Barrett, G.G. Graboski, Chem. Rev. 86 (1986) 751.
[
(
to avoid formation of byproducts derived from further addition
of nitromethane to nitrostyrene). The resulting oil was diluted in
THF (16 ml) containing ethyl acetoacetate (3.3 g, 25.2 mmol) and
passed at room temperature through the second column-reactor
filled with KG-60-TBD-SiMe3, washed previously with THF, with
a residence time of 35 min. The final reaction mixture was col-
lected, giving compound 5b in an 85% overall yield with respect
to benzaldehyde. Compounds like 5 represent useful precursors to
various complex organic molecules of industrial interest [45] with
functionalities that can be derived from both the nitro group and
the β-dicarbonyl system [46].
[
17] G. Rosini, Comprehensive Organic Synthesis, vol. 2, Pergamon Press, Oxford,
1991, p. 321.
[
18] R. Ballini, L. Barboni, F. Fringuelli, A. Palmieri, F. Pizzo, L. Vaccaro, Green Chem.
9 (2007) 823.
[19] G. Sartori, F. Bigi, R. Maggi, R. Sartorio, D.J. Macquarrie, M. Lenarda, L. Storaro,
S. Coluccia, G. Martra, J. Catal. 222 (2004) 410.
[
20] D.J. Macquarrie, R. Maggi, A. Mazzacani, G. Sartori, R. Sartorio, Appl. Catal. A
General 246 (2003) 183.
[
21] E. Angeletti, C. Canepa, G. Martinetti, P. Venturello, J. Chem. Soc. Perkin Trans. 1
1989) 105.
(
[
[
[
22] G. Demicheli, R. Maggi, A. Mazzacani, P. Righi, G. Sartori, F. Bigi, Tetrahedron
Lett. 42 (2001) 2401.
23] R. Ballini, G. Bosica, D. Livi, A. Palmieri, R. Maggi, G. Sartori, Tetrahedron Lett.
44 (2003) 2271.
4
. Conclusion
24] M.L. Kantam, P. Sreekanth, P. Lakshmi Santhi, Green Chem. 2 (2000) 47.
In this study, we have investigated the preparation of two
[25] Y.V. Subba Rao, D.E. De Vos, P.A. Jacobs, Angew. Chem. Int. Ed. Engl. 36 (1997)
661.
2
packed-bed column reactors containing propylamine (KG-60-NH2)
and guanidine TBD (KG-60-TBD-SiMe3) immobilized onto amor-
phous silica KG-60. The first catalyst, KG-60-NH2, was used to effi-
ciently perform continuous-flow nitroaldol condensation between
nitromethane and aromatic aldehydes, giving a small series of
[26] J.S. Swan, M. Azadpur, A.J. Bharucha, M.A. Krafczyk, J. Liq. Chromatogr. Related
Technol. 11 (1988) 3385.
[27] C.P. Fei, T.H. Chan, Synthesis (1982) 467.
[
[
[
28] D.A. Evans, S. Mito, D. Seidel, J. Am. Chem. Soc. 129 (2007) 11583.
29] A.P. Wight, M.E. Davis, Chem. Rev. 102 (2002) 3589.
(
9
E)-nitrostyrenes variously substituted on the aromatic ring (65%–
7% yield, 93%–98% selectivity). The second catalyst, KG-60-TBD-
30] M.R. Buchmeiser (Ed.), Polymeric Materials in Organic Synthesis and Catalysis,
Wiley, Weinheim, 2003.
SiMe3, gave a Michael addition of β-dicarbonyl compounds to ni-
trostyrenes affording polyfunctional compounds under continuous-
flow conditions (60%–92% yield, 72%–98% selectivity).
The two-step reaction can be performed sequentially in a flow-
trough mode, ensuring that any excess nitromethane is distilled
off before β-dicarbonyl compound is added to the first reaction
mixture. The catalytic reactors were reused for several runs, with
no decrease in activity.
[31] N. Nikbin, P. Watts, Org. Process Res. Dev. 8 (2004) 942.
[32] N.T.S. Phan, D.H. Brown, P. Styring, Green Chem. 6 (2004) 526.
[33] D.J. Macquarrie, J.H. Clark, A. Lambert, J.E.G. Mdoe, A. Priest, React. Funct.
Polym. 35 (1997) 153.
[
34] H.H. Baer, L. Urbas, in: S. Patai (Ed.), The Chemistry of Nitro and Nitroso
Groups, Part 2, Interscience, New York, 1970, p. 117.
4
[
35] r0 ·10 values were calculated after 15 min reaction time in the process carried
out with 0.07 g of catalyst (KG-60-NH2 with 0.65 mmol/g loading: 18% yield;
KG-60-NH2 with 1.06 mmol/g loading: 28% yield).
These products represent valuable intermediates being con-
verted by classic methods into variously substituted β-aminoke-
tones and β-aminoacids, useful building blocks in pharmaceutical
chemistry.
[
36] The (stereo)selective Michael addition of 1,3-dicarbonyl compounds to ni-
troalkenes under homogeneous catalysis has been extensively studied:
(
(
a) [14];
b) M. Watanabe, A. Ikagawa, H. Wang, K. Murata, T. Ikariya, J. Am. Chem. Soc.
126 (2004) 11148.
[37] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, VCH, Wein-
heim, 1988.
Acknowledgments
[
38] The high catalytic activity of the silica-supported strongly basic TBD catalyst
was shown in carbamate and carbonate synthesis: S. Carloni, D.E. De Vos, P.A.
Jacobs, R. Maggi, G. Sartori, R. Sartorio, J. Catal. 205 (2002) 199.
This research was partly supported by the MUR (National
Project “Sintesi organiche ecosostenibili mediate da nuovi sistemi

L. Soldi et al. / Journal of Catalysis 258 (2008) 289–295
295
[
39] Blank experiment confirmed that nitrostyrene was telomerized in ∼15% yield
upon prolonged stirring with KG-60 silica at room temperature.
40] It is indeed reported that the surface of pure silica gel is covered with silanol
groups, at a maximum concentration of 8 Brönsted acid OH groups per nm2
with a pKa value ≈6: G. Fink, B. Steinmez, J. Zechlin, C. Przybyla, B. Tesche,
Chem. Rev. 100 (2000) 1377.
[45] Michael adducts 5 can be converted (through conventional methods) into
γ -aminocarbonyl compounds which constitute valuable intermediates for
[
the preparation of
a variety of 1,4-difunctionalized derivatives. As an ex-
ample, compound 5b was hydrolyzed and decarboxylated by treatment
with aqueous sodium hydroxide at room temperature affording 4-phenyl-
5-nitro-2-pentanone in 90% yield (from 5b) that was then reduced to
4-phenyl-5-amino-2-pentanone hydrochloride in 88% yield by treatment
with hydrogen over Pt–S–C in EtOH/HCl: R. Tamura, D. Oda, H. Kurokawa,
Tetrahedron Lett. 27 (1986) 5759.
[
41] H.E.B. Lempers, R.A. Sheldon, J. Catal. 175 (1998) 62.
[42] M. Caldarelli, J. Habermann, S.V. Ley, J. Chem. Soc. Perkin Trans. 1 (1999) 107.
[43] G.M. Santerre, C.J. Hansrote, T.I. Crowell, J. Am. Chem. Soc. 80 (1958) 1254.
[44] P. Tundo, P. Venturello, E. Angeletti, J. Chem. Soc. Perkin Trans. 1 (1987)
[46] N. Ono, The Nitro Group in Organic Synthesis, Wiley–VCH, New York, 2001,
chap. 4.
2159.