Supported ionic liquid-like phases as organocatalysts for the solvent-free cyanosilylation of carbonyl compounds: From batch to continuous flow process

Article abstract of DOI:10.1039/c3gc42238k

Supported ionic liquid-like phases (SILLPs) are able to efficiently catalyse the cyanosilylation of carbonyl compounds using trimethylsilyl cyanide under solvent-free conditions. These organic catalysts were efficient for various carbonyl compounds includ

Full text of DOI:10.1039/c3gc42238k

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Supported ionic liquid-like phases as
organocatalysts for the solvent-free
cyanosilylation of carbonyl compounds: from
Cite this: Green Chem., 2014, 16,
1639
batch to continuous flow process†
Sergio Martín, Raúl Porcar, Edgar Peris, María Isabel Burguete,
Eduardo García-Verdugo* and Santiago V. Luis*
Supported ionic liquid-like phases (SILLPs) are able to efficiently catalyse the cyanosilylation of carbonyl
compounds using trimethylsilyl cyanide under solvent-free conditions. These organic catalysts were
efficient for various carbonyl compounds including aromatic, aliphatic and α,β-unsaturated. This process
perfectly conforms to the features of green chemistry: no waste regarding side-products and uncon-
verted reactants, solvent-free, excellent catalytic activity, and no requirement for separation. Furthermore
the supported nature of the SILLPs allows for the development of continuous flow synthesis of cyano-
hydrin trimethylsilyl ethers. The effect of the morphology of the polymer (gel-type or macroporous bead
resins or, alternatively, a monolithic polymer) has also been studied, which has allowed selecting suitable
materials for the efficient continuous flow single pass synthesis of cyanohydrin trimethylsilyl ethers.
Received 30th October 2013,
Accepted 20th December 2013
DOI: 10.1039/c3gc42238k
www.rsc.org/greenchem
The use of ionic liquids (ILs) has been demonstrated to be
an efficient and environmentally friendly reaction media as
Introduction
Hydrocyanation and cyanosilylation of carbonyl compounds well as a promoter for the cyanosilylation of aldehydes under
are among the most important strategies for C–C bond- mild conditions without the need for a Lewis acid (or base) or
forming reactions in organic synthesis as, for instance, cyano- any other special activation.7 The presence of catalytic
hydrin trimethylsilyl ethers are industrially valuable and amounts of lanthanide triflates, particularly scandium triflate,
important intermediates for the synthesis of many valuable in the ILs increased largely their catalytic efficiency in terms of
molecules such as α-hydroxy acids, α-hydroxy aldehydes and the resulting turnover frequency and total turnover number.8
β-amino alcohols and other biologically active compounds.1,2
In the last few years, we have focused our efforts on the
Although there are different methods for the preparation of development of the so-called supported ionic liquid-like
trialkylsilyl cyanides, the addition to carbonyl compounds of phases (SILLPs) prepared by the immobilization of molecules
trimethylsilyl cyanide (TMSCN), which is a safe and easily with IL-like structures onto solid supports (Fig. 1). These
handled reagent compared to HCN, NaCN or KCN,3 allows advanced materials allow to transfer the IL properties to the
obtaining silyl protected products in high yields and readily solid phase leading to either monolithic or gel supported ionic
accessible for further manipulations.4 However, the use of silyl liquid-like phases (m- or g-SILLPs) sharing, therefore, the pro-
cyanides requires catalytic activation by an organometallic or perties of true ILs and the advantages of a solid support.9 Cata-
organocatalytic system. Thus, the traditional organometallic lytic processes based on the use of SILLPs allow (i) to
systems (Lewis acid catalysts) are based on the complexation of
a metal to both the carbonyl group and the cyano group of
TMSCN.5,6 However, the organocatalytic systems (Lewis basic
catalysts) often react with the silicon atom of TMSCN to
produce more reactive hypercoordinated silicon intermediates.
Departamento de Química Inorgánica y Orgánica, Universidad Jaume I, Av. Sos
Baynat s/n, 12071 Castellón, Spain. E-mail: luiss@uji.es, cepeda@uji.es
†Electronic supplementary information (ESI) available: GC methods and
1H NMR spectra of the products Table 2 and calculation Table 4. See DOI:
10.1039/c3gc42238k
Fig. 1 Design vectors used to fine-tune the organocatalytic properties
of supported ionic liquid-like phases (SILLPs).
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1639–1647 | 1639

View Article Online
Paper
Green Chemistry
minimize the amount of ILs used, (ii) easy separation and
recyclability and (iii) to develop mini-flow reactors for continu-
ous processes using either conventional solvents or supercriti-
cal fluids (scFs).10 Thus, a wide range of polymer-supported
ionic liquid-like phases (SILLPs) or “solid ionic solvents” with
tunable properties can be easily prepared in order to (1) gene-
rate novel (bio)catalytic species, (2) improve the stability of the
(bio)catalyst, (3) optimize immobilization and recyclability, (4)
assist in the activation of the (bio)catalyst, (5) facilitate product
isolation, and (6) influence the selectivity of the reaction,
which are some of the essential roles of conventional solvents.
Taking this into account, supported ionic liquid-like
materials (SILLPs) offer good potential as substitutes of bulk
ILs for the cyanosilylation reaction, and here we report our
efforts towards this goal. The process has been carried out by
the smooth addition of TMSCN to carbonyl compounds using
polymeric SILLPs as new bifunctional organocatalysts, under
different batch conditions and, preferentially, under continu-
ous flow solvent-free conditions.
Table 1 Effect of the solvent in the cyanosilylation reaction between
(14) and (15) catalysed by SILLP-7 at r.t. and 24 hours of reaction timea
Entry
Solvent
Yieldb (%)
TON
1
2
3
4
5
6
7c
CH2Cl2
Acetonitrile
2-MeTHF
H2O
MeOH
—
64
81
29
0
0
92
20
29
37
13
0
0
41
—
—
a 75 mg of SILLP 7, 3 mmol of acetophenone, 3.6 mmol of TMSCN and
5 mL of solvent, stirring at room temperature during 24 hours.
b Calculated by GC. c Reaction in the absence of catalyst.
like MeOH or water. The SILLP-7 (R = CH3, R′ = H, X = Cl, low
loading) was used as the initial catalyst. As shown in Table 1,
the cyanosilylation reaction was more efficient under solvent-
free conditions than using organic solvents or water. The reac-
tion proceeded smoothly at room temperature to afford the
corresponding cyanohydrin in 92% yield under solvent-free con-
ditions (Table 1, entry 6). It should be noted that the reaction
only led to low yields (20% yield) in the absence of the SILLP-7.
Regarding the use of solvents, both MeOH and water led to
non-conversion of the carbonyl compound. Low to modest
yields were obtained when either CH2Cl2 or 2-MeTHF was used
(Table 1, entries 1 and 3). The best solvent was clearly aceto-
nitrile, leading to 81% yield (Table 1, entry 2). In view of these
results, various aldehydes and ketones were tested under
solvent-free conditions at room temperature using SILLP-7 as
the catalyst to evaluate the general applicability of this protocol.
All the results are summarized in Table 2.
The SILLP-7 was able to catalyse the solvent-free cyanosilyla-
tion for substituted and unsubstituted benzaldehydes
(Table 2, entries 2–5) with similar results to those found for
acetophenone. Excellent yields of the corresponding products
were obtained, almost in pure form, without any additional
purification step except the catalyst filtering. In addition, a
cyclic ketone and an open chain aliphatic aldehyde (Table 2,
entries 1 and 7) were also converted into the corresponding
TMS-cyanohydrins in good yields. It should also be noted that
an aromatic α,β-unsaturated aldehyde also underwent the
cyanosilylation reaction efficiently (Table 2, entry 6). These pre-
liminary results suggest that these supported ionic liquid-like
phases are quite a promising system to develop simple,
efficient, room temperature and solvent free cyanosilylation
reactions for different classes of carbonyl compounds.
Results and discussion
Different PS-DVB polymers containing alkylimidazolium chlor-
ide subunits (7–12) were prepared from commercially available
Merrifield resins having different chloride loadings and cross-
linking degrees as previously reported by us.11 Initially, gel-
type microporous chloromethylated resins (1–3) in the form of
beads, bearing a low crosslinking degree (2% DVB) and with
chloride loadings ranging from 1.2 to 4.3 mmol Cl g−1 were
used. The synthetic protocol for the preparation of the SILLPs
has been previously optimized by our group and has been
described in detail, allowing a quantitative transformation of
the chloromethylated fragments into alkyl benzyl imidazolium
groups as is shown in Scheme 1.12
The cyanosilylation reaction between acetophenone (13)
and TMSCN (14) was selected as the benchmark reaction, as,
in general, most of the reported methods for the cyanosilyla-
tion of carbonyl compounds are largely limited to aldehydes as
well as to the more reactive aliphatic ketones.
First, the role of the solvent was studied by carrying out the
reaction either under solvent-free conditions or in a range of
solvents from a non-polar one, such as CH2Cl2, to polar ones
One of the potential advantages of the use of SILLPs is that,
like in the bulk ILs, it is easy to vary several of their structural
features (loading of IL-like units, substitution pattern, the
nature of the counterion, the type of polymeric backbone, etc.)
to fine-tune their physical and chemical properties. These
changes may have a big impact on the catalytic properties of
the SILLPs as organocatalysts for the cyanosilylation reaction
as we have already demonstrated to apply for other (bio)cata-
lytic systems.9,10,13 Hence, we studied the effects of some of
Scheme 1 Synthesis of SILLPs: (i) alkylimidazole, 90 °C; (ii) NaX or HX,
H2O. 1–3 resins are microporous beads with 2% crosslinking,
1 (1.2 mmol Cl g−1), 2 (2.1 mmol Cl g−1) and 3 (4.3 mmol Cl g−1); 4 macro-
porous resin (1.2 mmol Cl g−1).
1640 | Green Chem., 2014, 16, 1639–1647
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
moiety that can be achieved through the use of different alkyl
resides (R and R′). Thus, three different SILLPs (SILLP-8,
SILLP-9 and SILLP-10) were synthesized from the same starting
Merrifield resin. Therefore, three different resins having the
same loading (high loading) but a different substitution
pattern (R = CH3– and R′ = H– (SILLP-8); R = CH3–(CH2)3– and
R′ = H (SILLP-9) and R = CH3–(CH2)9– and R′ = CH3– (SILLP-
10)) were evaluated as catalysts for the benchmark cyanosilyla-
tion reaction (Scheme 2). In all the cases, the SILLPs were
active to catalyze the reaction. However, the time-course pro-
files of the product yield (see Fig. 2) present significant differ-
ences for each of the SILLPs. Thus, the SILLP-8 was the more
active one, reaching an 85% yield in ca. six hours, while the
presence of a longer alkyl chain at the imidazolium ring led to
a reduction of the catalytic efficiency.
Table 2 Cyanosilylation of various carbonyl compounds under solvent-
free conditions catalysed by SILLP-7 at r.t. and 24 hoursa
Entry
1
Reagent
Conv.b,c (%)
>99
2
>99
3
4
>99
99
It is noteworthy that the SILLP-8 containing the smaller
substituents showed some induction time, which is not notice-
able for the other SILLPs. Most likely, this observation must be
related to the highly hydrophilic character of SILLP-8 that
makes this resin rather hygroscopic. The observed induction
time should be related to the larger amount of residual water
present in this case in the polymer and that can interfere with
the reaction of interest, by reacting with TMSCN instead of the
chloride anion (the first step of the catalytic cycle). Indeed, the
water content for these SILLPs varied from ca. 11% for the
SILLP-8 to a maximum of 6–7% for the more hydrophobic
ones SILLP-9 and SILLP-10. As can be observed in Fig. 2, after
this induction period the SILLP-8 displayed the best catalytic
5
>99
6
7
99
>99
a 75 mg of SILLP 7, 3 mmol of carbonyl compound and 3.6 mmol of
TMSCN, stirring at room temperature during 24 hours. b Calculated by
GC. c A single product corresponding to the one expected for the
cyanosilylation reaction was detected by 1H NMR of the crude of the
reaction.
these structural variables on the catalytic efficiency for the
solvent free reaction between acetophenone and TMSCN.
First, the effect of the loading of the IL-like moieties in the
polymeric backbone was tested. At long reaction times
(24 hours) all the catalysts showed very good yields (ca. 90%)
independent of the SILLP-loading. However, the catalytic
efficiency of the SILLP measured in moles of products con-
verted per unit of time and mole of catalyst, before completion
of the reaction, was quite loading dependent. Indeed, when
the same amount of solid catalysts was used, the most efficient
SILLP was the one having a medium loading of imidazolium
groups (SILLP-11, 40% IL-like units by weight) showing a TOF
= 7.7 (mol 15 per mol cat h) calculated at two hours of reaction
time, while the SILLP with higher (SILLP-8, 63% IL-like units
by weight) and lower (SILLP-7, 18% IL-like units by weight)
amounts of functionality only presented a catalytic efficiency
(TOF) of 4.6 and 2.3 respectively. Most likely, this effect of the
loading can be associated with the accessibility to the catalytic
sites inside the polymeric beads. This can be difficult for low
loadings, because of the formation of polar/apolar domains in
the resin, but also in the resins containing a high degree of
IL-like functionalization, where the IL-fragments can be
strongly associated through the formation of anion–cation
supramolecular structures.
Scheme 2 Cyanosilylation between (14) and (15) organocatalyzed by
SILLPs.
Fig. 2 Effect of the substitution pattern of the SILLPs as catalysts for
the cyanosilylation reaction between (14) and (15) at r.t. and 24 hours of
Another factor that can be used to tune the catalytic activity
of the SILLPs is the substitution pattern of the imidazole
●
▲
R = CH3(CH2)3 and
reaction time.
R = CH3 and R’ = H (SILLP-8);
■
R’ = H (SILLP-9) and R = CH3(CH2)9 and R’ = CH3 (SILLP-10).
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1639–1647 | 1641

View Article Online
Paper
Green Chemistry
behavior. Thus, the exact microenvironment at the imidazo- the magnetic stirring bar. This often hampers an efficient fil-
lium moiety can have a key influence on the overall catalytic tration to separate the catalyst from the reagents and products,
cycle.
precluding the practical recycling of the supported catalyst.
Once established that the best results in terms of yield and Therefore, we designed a simple device allowing for a reactive
catalytic efficiency were obtained for the SILLP based on filtration process. The supported-catalyst was maintained in a
methyl-imidazolium, we decided to study the effect of the syringe body with an adequate filter at the top and the bottom
counterion. Our previous results have proved that very signifi- of the material, and then the solvent free mixture of the
cant variations in the physico-chemical properties of the reagents (ca. 3 mL) was eluted through the catalytic bed and
SILLPs can be obtained by varying the anion of the ionic- separated from the solid by simple gravimetric flow (Fig. 4).
liquid-like moieties grafted onto a polymeric matrix. Therefore, The contact time of the reagents with the SILLP-7 was long
SILLPs bearing five different anions and the same loading of enough to achieve an average product yield of 85% for three
the methyl-imidazolium cation were tested as catalysts for the consecutive batches of ca. 3 mL each. In summary, the SILLP-7
cyanosilylation reaction. The nature of the anion has a signifi- was stable enough to produce ca. 15 g of cyanosilylether per
cant effect on the catalytic efficiency as can be appreciated in gram of catalyst used during three consecutive recycles,
the time-course profiles of the product yield (Fig. 3). Hence, without showing any significant decrease in its activity.
reactions performed with catalysts based on SILLPs containing
The former results also suggest that these catalytic SILLPs
the harder Lewis bases as the anions (Cl− and TfO−) showed a can be stable enough as to develop a continuous flow
better activity than those with catalysts bearing softer Lewis system.16–18 For this purpose, a simple set-up allowing a forced
base anions such as NTf2−, BF4 or SbF6−. Indeed, the calcu- continuous flow-through process was evaluated.19 The system
−
lated reaction rates change followed the sequence Cl− (r = was based on a HPLC pump, a fix-bed catalytic reactor and a
0.5848 h−1) ≫ TfO− (r = 0.0224 h−1) > NTf2− (r = 0.0417 h−1) ≫ continuous flow ATR-FTIR cell to facilitate the monitoring and
BF4 (r = 0.0065 h−1) > SbF6 (r = 0.0050 h−1) being the optimisation of the process (see Fig. 5).20
−
−
enhancement in the reaction rate up to ca. 117 folds going
A careful analysis of the nature and shape of the catalyst
from SbF6− to Cl− (see Fig. 3). According to known mechanistic under consideration is always needed for a proper design of
data, the anion must coordinate to the Si atom to form a more the corresponding mini-flow reactors (mfr). Indeed, the mor-
reactive pentacoordinated silicate for the reaction to occur. phology and shape of the catalyst can lead, in some cases, to
This binding leads to a polarization of the adjacent bonds, significant differences in catalytic performance.16,21–23 There-
thereby increasing the electron density at the peripheral fore, for the preparation of mini-flow fix-bed catalytic reactors,
atoms, activating the CN− to attack the carbonyl group.14,15 we decided to study polymers with different morphologies.
Thus, the presence of a hard Lewis base, for instance the First, the use of gel-type lightly crosslinked resins was evalu-
chloride in the SILLP, should facilitate this mechanism of ated for the preparation of the reactor (mfr-SILLP-7-gel).24 This
action. On the other hand, in the catalytic cycle, the imidazo- was the type of polymer successfully used for batch experi-
lium cation may contribute to stabilize the more reactive ments. Indeed, both reagents and catalysts based on gel-type
pentacoordinated silicate through additional weak interactions resin have been successfully used for different C–C bond for-
such as hydrogen bonding, hydrophobic and van der Waals mation reactions.25,26 However, it should be taken into
attraction together with ionic interactions with the anion.
account that these polymers only swell in certain solvents. In
The possible recycling of the catalyst was also evaluated. the absence of a good swelling, intraparticle diffusion of the
The nature of the support, being a gel type resin, makes it very
sensitive to mechanical damage by the abrasion produced by
Fig. 3 Effect of the SILLP counterion on the cyanosilylation reaction
between (13) and (14) catalysed by SILLP-8 and SILLP-13a–d at r.t. and
Fig. 4 Recycling of the SILLP-7 for the cyanosilylation reaction
between (14) and (15) at r.t. during 24 hours.
●
▼
■
◆
▲
24 hours. (X = Cl), (X = TfO), (X = NTf2), (X = SbF6), (X = BF4).
1642 | Green Chem., 2014, 16, 1639–1647
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Table 3 Continuous flow solventless cyanosilylation reaction between
14 and 15 at r.t. using different polymeric materialsa
Flow mL Yieldb Productivity
Entry Reactor
min−1
(%)
(g 16 per g cat) h−1 c
1
2
3
4
5
6
7
8
9
mfr-SILLP-7-gel
mfr-SILLP-7-gel
mfr-SILLP-7-gel
mfr-SILLP-12-macro 0.1
mfr-SILLP-12-macro 0.3
mfr-SILLP-12-macro 0.5
mfr-SILLP-17-molth 0.01
mfr-SILLP-17-molth 0.03
mfr-SILLP-17-molth 0.05
0.1
0.3
0.5
99
99
99
99
98
96
95
85
75
9.0
27.1
45.1
9.0
26.8
43.7
0.2
0.3
0.6
a mfr-SILLP-7-gel: 600 mg of SILLP-7, 1.01 meq g−1; mfr-SILLP-12-
macro: 600 mg of SILLP-12 1.09 meq g−1, mfr-SILLP-17-molth: 3.38 g,
3.03 meq g−1; 3.67 M of 14 in a mixture of acetophenone–TMSCN
1 : 1 molar. b Calculated by GC. c Calculated as [flow rate (mL min−1) ×
concentration (mmol mL−1) × (yield%/100) × Fw(16) (mg mmol−1) ×
(1/1000)]/[cat. (g) × loading cat. (mmol g−1)].
point the flow can be significantly increased (from 0.1 to
0.5 mL min−1) maintaining the same outcome for the reaction.
The results obtained show a productivity for the reactor with
this particular configuration of 45.1 g of 16 per g SILLP-7 h−1
for a flow of 0.5 mL min−1 (Table 3, entry 3).
As mentioned above, proper swelling of resin is essential to
provide accessibility to the catalytic sites that are embedded in
the interior of the polymeric beads. The swelling of this gel-
type polymeric catalyst (SILLP-7, 600 mg used) is significant,
with ca. 100% swelling in the presence of an equimolecular
mixture of 14 and 15 for 24 hours (vi = 1 mL, vf = 2 mL).28
Fig. 5 Continuous flow reactor, including continuous monitoring, for
the cyanosilylation reaction between (14) and (15) catalysed by SILLP-7
at r.t. (a) Photo and scheme of the set-up. (b) Evolution with time of the
characteristic peaks of both reactant (14) and product (16) for the model
cyanosilylation in the mini-flow reactor mfr-SILLP-7-g at r.t. and under
Alternatively,
a mini-flow reactor (mfr-SILLP-12-macro)
reactor was packed with an analogous supported catalyst pre-
pared from a macroporous resin. In contrast to gel-type or
microporous resins, the beads from macroporous polymers
display a permanent porosity regardless of the solvents or
reagents, which usually makes them more suitable for flow
different flow conditions.
(CvO at 1768 cm−1),
●
■
(C–OTMS at
1115 cm−1), (Si–O at 995 cm−1).
▲
reagents and substrates can be the limiting rate factor. There- processes. The reactor showed a similar behaviour as the first
fore, suitable swelling conditions are required to allow the flow one, leading to yields of 99–95% of the desired product at flow
through the generated internal microchannels of the resin rates ranging from 0.1 to 0.5 mL min−1 (Table 3, entries 4–6).
(usually beads).21,26 Furthermore, the swelling is accompanied The slight difference observed for both reactors can be
by a change in volume that can be very important, thus assigned to the different void volume observed for each
affecting the packing.21 In this regard, a regulating device was reactor. Although the total volume occupied by the wet and
used to properly adjust the volume of the reactor to that of the swollen polymer is not very different (2 mL for mfr-SILLP-7-gel
swollen resin. In this way, it is feasible to achieve a reproduci- and 1.8 mL for mfr-SILLP-12-macro), there is a significant
ble flow through systems based on gel-type polymers.27
difference in the void volume due to the different swelling
The initial results for the fix-bed reactor (mfr-SILLP-7-gel) observed. Thus, in the case of the gel-type resin (mfr-SILLP-7-
prepared with the gel-type resin SILLP-7 are depicted in Fig. 5. gel), the void volume can be estimated to be ca. 0.4 mL, while
The continuous monitoring of the effluent by FT-IR allowed a for the reactor packed with the macroporous catalyst (mfr-
direct analysis of the results for this continuous flow process. SILLP-12-macro) it is ca. 1.2 mL. These differences provide
The graph in Fig. 5 displays the evolution with time of the slightly different residence times for each reactor. Thus, the
characteristic peaks for both the reactant (14) and the product residence time for a flow rate of 0.1 mL min−1 is 4 min for mfr-
(16). The data show the presence of an important induction SILLP-7-gel and 12 min for mfr-SILLP-12-macro.
time before the stationary regime is reached. Once this is
Fig. 6 depicts the yield vs. time course for the cyanosilyla-
achieved, the flow system can be maintained with the same tion reaction between 14 and 15 using either mfr-SILLP-7-g or
productivity and high yields without observing any decrease in mfr-SILLP-12-macro at 0.1 mL min−1. In both cases the induc-
the activity of the catalyst with time. As a matter of fact, at this tion time observed was ca. 3–2.5 fold the residence time
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1639–1647 | 1643

View Article Online
Paper
Green Chemistry
Fig. 7 Stability study: evolution with time of the yield of the product for
the cyanosilylation between benzaldehyde and (15) obtained by the
min-flow reactor (a) mfr-SILLP-7-gel at 0.01 mL min−1
.
Fig. 6 Evolution with time of the yield of the product (16) for the model
cyanosilylation between (14) and (15) obtained by the min-flow reactors
(a) mfr-SILLP-7-gel and (b) mfr-SILLP-12-macro at 0.1 mL min−1
.
divinylbenzene in the presence of the right porogenic
agent.12,29 Further modification of the chloromethyl groups of
the monolith with methylimidazole led to a monolithic reactor
with an analogous chemical structure as SILLP-7 and SILLP-
required to fill the reactor void volume. It must be noted that
the induction time observed is much higher than that detected 12.29 This mini-flow reactor (mfr-SILLP-17-molth) was then eval-
uated using the same experimental continuous flow set-up
and different flow rates ranging from 0.5 to 0.01 mL min−1
in the previous batch experiments. In this regard, several
factors need to be taken into consideration to understand the
.
process of catalyst conditioning before reaching the optimal In this case, only moderate yields (30%) were detected for
0.5 mL min−1 flow rate and smaller flow rates (longer resi-
operational behaviour. The first factor to be taken into con-
dence time) were needed to achieve satisfactory yields (Table 3,
(tubing and interparticle void-spaces in the reactor) and this entries 7–9).
sideration is the process of filling the void spaces of the system
Table 3 summarises the results obtained for the three mini-
flow reactors prepared with SILLPs presenting similar func-
can be associated with the initial period for which an increase
in the amount of starting material in the effluent is observed.
Volumetric calculations based on the volume spent during tional loadings but different polymer morphologies. Although
previous studies from different groups, including our own
this period of time afforded a total volume for the system
results, have revealed that monolithic flow-through reactors
which is in good agreement with the one estimated according
to the length and diameter of the tubing and to the volume of can provide, in some instances, significantly better perform-
ances;13,21,25 in this case, the fix-bed reactors prepared from
the reactor and the density (in the dry state) of the polymeric
beads (ca. 0.4 mL and 1.2 mL overall void volume). On the polymeric beads delivered the most efficient system. No sig-
other hand, as has been mentioned above, reaction of the nificant differences were observed between the gel-type and
macroporous catalysts with
a maximum productivity of
chloride with TMSCN represents the first step of the catalytic
cycle, but water can efficiently compete with chloride for the 43–45 g of 16 per g cat h−1. This must be associated with the
reaction with TMSCN when present at the polymeric beads, very low flow rates required to achieve reasonable conversions,
which suggest a reduced accessibility of the catalytic sites in
significantly reducing the efficiency of the process. Thus, the
induction period can be associated with two main factors: (a) the monolith. The presence of the functional groups on the
the wetting and swelling of the polymeric beads by the fluid surface of the porous structure of the monolith can provide a
containing the substrate and the reagent;21 (b) the reaction of
high local concentration of IL-like moieties strongly associated
the TMSCN with any residual water. Indeed, if the columns are and ordered, substantially hampering the participation of the
preheated in an oven at 100 °C overnight, the induction time chloride anions in the first step of the catalytic cycle. Thus,
this effect should be reminiscent of the one found for high
observed is significantly reduced to 1–1.5 fold the residence
time required to fill the reactor void volume.
loading resins. The use of a porogenic solvent like dodecanol
Furthermore, the systems showed a good stability after is known to provide monoliths with large macropores, which
is accompanied by a reduction in the total surface. Besides,
deactivation, allowing the synthesis during this time of 23.8 g polar fragments tend to be located at the interface with the
48 hours of continuous use without any significant catalytic
of the corresponding cyanosilylated compound (see Fig. 7).
Finally, the use of a mini-flow reactor prepared from a
monolithic polymeric rod was also studied. The monolithic
porogenic agent becoming the accessible surface of the final
polymer.
Table 4 shows a comparison of some of the more efficient
reactor was prepared as reported by us and others by poly- methodologies reported up to now for the cyanosilylation reac-
tion here considered. All these previous methods combine at
merisation of
a
mixture of 4-chlorovinylbezene and
1644 | Green Chem., 2014, 16, 1639–1647
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Table 4 Comparison between the different catalytic processes developed for the cyanosilylation reaction of carbonyl compounds
Ref
Solvent
Cat.
Conditions
Prod.a
E factor
7b
8b
[OMIM][PF6]
[BMIM][SbF6]
[OMIM][PF6]
Sc(OTf)3
r.t., batch, 24 h
r.t., batch, 30 min
0.55
20 000
156
96
30c
SolFC
60 °C, flow, 0.04 mL min−1
25
42
0.16
c
—
SolFC
r.t. flow, 0.5 mL min−1
0.05
a Productivity as (mol 16 per mol cat) h−1 b Made with benzaldehyde. c Made with acetophenone.
.
least two of the elements applied in our system, namely: ionic implementing the productivity of the system, the product iso-
liquids, (organo)catalysis, solventless conditions (SolFC) and lation and a simple catalyst recycling. The use of a flow system
flow chemistry. Two main parameters have been selected for has also been shown to minimize the negative effects associ-
comparison: productivity and E factor. The first one is related ated with the low mechanical stability of the polymeric
with the chemical efficiency of the process, while the E factor materials employed in this work, which precludes proper
highlights its greenness. Entries 1 and 2 summarise the recovery of the catalyst under pure batch processes. The con-
results obtained using ILs as the solvent. Even when the ILs nection of the effluent tubing to a flow-cell in a FT-IR has
could be recycled, the use of an additional organic solvent was allowed continuous monitoring of the reaction, assisting in its
required to isolate the product, which is reflected in the sig- optimization. This process perfectly conforms to the features
nificant increase of the E factor, in spite of the really high pro- of green chemistry: 100% atom economy, no waste regarding
ductivity obtained when Sc(OTf)3 is present. The use of solid/ side-products and unconverted reactants, solvent-free, excel-
supported catalysts (entries 3 and 4, Table 4) facilitates the iso- lent catalytic activity, and no requirement for separation, along
lation of the product and the reuse of the catalyst by simple fil- with continuous monitoring.
tration, which is reflected by a large reduction in the E factor
(>100 fold reduction). In the case of the SILLP-7, besides, the
productivity observed is twofold that of the recently reported
Experimental
Synthesis and characterization of SILLPs
supported phosphine.30 No additional solvent is required
neither for the reaction nor for the isolation step leading to a
further E factor reduction.
Polymers modified with ionic-liquid-like moieties were syn-
thesized by chemical modification of polymer-bound chloro-
methyl groups with the corresponding substituted imidazole
(4, 5 or 6) following the experimental procedures previously
described.12 Commercially available Merrifield-gel polymers
with different loadings (1.2, 2.1, 4.3 mmol Cl− g−1, 1–3) were
The two polymer-supported systems do not provide the
impressive productivity of the reaction catalysed by Sc(OTf)3 in
ILs; however both are organocatalytic solvent-free and metal
free reactions with very low E factors. Although a comparison
of this type can be excessively simplistic, at the level con-
sidered, it allows illustrating how the combination of different
tools from the Green Chemistry “tool box” can lead to syner-
gies helping us to move forward towards more sustainable
processes.
used to afford the SILLPs 7–12 (Scheme 1). The metathesis of
chloride by different anions (BF4−, TfO−, NTf2−, and SbF6
)
−
yielded a new series of SILLPs with an exchanged anion
(13a–d).
Synthesis and characterization of SILLP-17-monolith
A solution of divinylbenzene (2.4 mL; 80% technical grade), 4-
vinylbenzene chloride (1.6 mL; 90% purity grade), 1-dodecanol
Conclusions
The results here presented highlight the advantages of the use (4.0 mL) and AIBN (42 mg) was prepared. A 15 mm × 100 mm
of supported ionic liquid-like phases (SILLPs) to develop Omnifit® column was filled with this stock solution, both ends
greener and more efficient catalytic processes. Indeed such of the column were sealed and the system was introduced into a
SILLPs allowed us to design organocatalytic species for the bath at 70 °C for 24 h. A white polymeric monolith was
solvent free cyanosilylation reaction of carbonyl compounds. obtained. This monolithic column was connected to a HPLC
The modular structural features of the SILLPs allow to fine- pump and washed with THF at 1 mL min−1 for 2 h at r.t. Then,
tune the catalytic efficiency by varying parameters such as the methylimidazole (25 mL) was pumped at a flow rate of 0.2 mL
SILLP loading, the substitution pattern of the imidazolium min−1 for 24 h, the column being maintained at 60 °C. The
units and the nature of the anion. The optimization of these column was then washed with THF at 1 mL min−1 for 2 h at r.t.
parameters led to stable organocatalytic supported systems
All of the polymeric materials were characterized by FT-IR
that facilitate the development of a continuous flow system spectroscopy, Raman microspectroscopy, and elemental
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1639–1647 | 1645

View Article Online
Paper
Green Chemistry
analysis, which confirmed the expected structures. The details
of the synthesis and characterization of these compounds can
be found in our previous publications.9,11,12,29
6 (a) C. Wiles and P. Watts, ChemSusChem, 2012, 5, 332–338;
(b) C. Wiles and P. Watts, Org. Process Res. Dev., 2008, 12,
1001–1100.
7 Z. L. Shen, S. J. Jib and T. P. Loh, Tetrahedron Lett., 2005,
46, 3137–3139.
Batch cyanosilylation reaction
8 Y. P. Boyoung, Y. R. Ka, H. P. Jung and L. Sang-gi, Green
Chem., 2009, 11, 946–948.
9 V. Sans, N. Karbass, M. I. Burguete, V. Compañ, E. García-
Verdugo, S. V. Luis and M. Pawlak, Chem.–Eur. J., 2011, 17,
1894–1906.
10 M. I. Burguete, E. García-Verdugo, I. Garcia-Villar, F. Gelat,
P. Licence, S. V. Luis and V. Sans, J. Catal., 2010, 269, 150–
160.
3 mmol of the corresponding carbonyl compound were added
over 75 mg of the catalyst followed by the addition of 3.1 mmol
of TMSCN. The mixture was gently stirred during 24 hours at
25 °C. To determine the kinetic profiles an aliquot of 20 μL
was regularly taken (30 min, 1 hour, 2 hours, 4 hours, 6 hours,
8 hours and 24 hours) and diluted in 1 mL of CH2Cl2. Samples
were then analysed by G.C. and 1H-NMR (see ESI†).
11 M. I. Burguete, E. García-Verdugo, S. V. Luis and
J. A. Restrepo, Phys. Chem. Chem. Phys., 2011, 13, 14831–
14838.
12 M. I. Burguete, H. Erythropel, E. García-Verdugo, S. V. Luis
and V. Sans, Green Chem., 2008, 10, 401–407.
13 P. Lozano, E. García-Verdugo, N. Karbass, K. Montague,
T. De Diego, M. I. Burguete and S. V. Luis, Green Chem.,
2010, 12, 1803–1810.
Reactive filtration cyanosilylation reaction
A regular 5 mL syringe equipped with a nylon filter was filled
with 0.4764 mg of the catalyst SILLP-7. The mixture of the
reagents (1 : 1 mol acetophenone–TMSCN) was added to the
top of the syringe and was eluted through the catalytic bed
and separated from the solid by simple gravimetric flow.
Aliquots were taken at constant time intervals (15 min) and
analysed by G.C.
14 M. Kira, K. Sato and H. Sakurai, J. Am. Chem. Soc., 1988,
110, 4599–4602.
15 S. Rendler and M. Oestreich, Synthesis, 2005, 1727–1747.
Continuous flow cyanosilylation reaction
The reactor was set up by introducing the SILLP-7 (ca. 600 mg) 16 S. V. Luis and E. García-Verdugo, Chemical Reactions
in a glass Omnifit® column (5 mm × 10 cm), which was con-
nected at the head to a Jasco HPLC pump (PU2080 plus) and
at the bottom to a continuous flow cell for a Pike-ATR
and Processes under Flow Conditions, RSC Green Chem-
istry Series, The Royal Society of Chemistry, Cambridge,
2010.
MIRacle™ single reflection ATR (diamond/ZnSe) adapted to a 17 For examples of continuous flow systems see: Beilstein
Jasco-6300 FT-IR spectrometer by standard tubing connectors.
J. Org. Chem., Chemistry in flow systems I and II, ed.
A solution of the reagents (1 : 1.1 mol acetophenone–TMSCN)
A. Kirschning, 2011.
was pumped through the catalytic bed at different flow rates. 18 (a) B. P. Mason, K. E. Price, J. L. Steinbacher, A. R. Bogdan
The IR spectra were recorded at continuous intervals of time
(2–4 min) during the run. In all the cases, aliquots were
collected at the outlet of the reactor and were analysed by GC.
and D. T. McQuade, Chem. Rev., 2007, 107, 2300–2318;
(b) B. Ahmed-Omer, J. C. Brandtand and T. Wirth, Org.
Biomol. Chem., 2007, 5, 733–740; (c) P. Watts and C. Wiles,
Chem. Commun., 2007, 443–467; (d) T. Fukuyama,
M. T. Rahman, M. Sato and I. Ryu, Synlett, 2008, 151–163;
(e) K. Geyer, T. Gustafsson and P. H. Seeberger, Synlett,
2009, 2382–2391; (f) S. Marrea and K. F. Jensen, Chem. Soc.
Rev., 2010, 39, 1183–1202; (g) D. Webb and T. F. Jamison,
Chem. Sci., 2010, 1, 675–680; (h) J. Yoshida, H. Kim and
A. Nagaki, ChemSusChem, 2011, 4, 331–340.
Acknowledgements
This work was partially supported by MINECO, Spain (Ref:
CTQ2011-28903) and Generalitat Valenciana (PROMETEO
2012/020 and ACOMP GV-2013/207). Cooperation of the SCIC
of the UJI for instrumental analyses is also acknowledged.
19 J. Wegner, S. Ceylan and A. Kirschning, Chem. Commun.,
2011, 47, 4583–4592.
20 (a) S. Suga, M. Okajima, K. Fujiwara and J. Yoshida, J. Am.
Chem. Soc., 2001, 123, 7941–7942; (b) M. Rueping,
T. Bootwicha and E. Sugiono, Beilstein J. Org. Chem., 2012,
8, 300–307.
21 V. Sans, N. Karbass, M. I. Burguete, E. García-Verdugo and
S. V. Luis, RSC Adv., 2012, 2, 8721–8728.
Notes and references
1 J. M. Brunel and I. P. Holmes, Angew. Chem., Int. Ed., 2004,
43, 2752–2778.
2 R. J. H. Gregory, Chem. Rev., 1999, 99, 3649–3682.
3 R. I. Kureshy, S. H. R. Abdi, S. Agrawal and R. V. Jasra, 22 A. R. Bogdan, B. P. Mason, K. T. Sylvester and
Chem. Rev., 2008, 252, 593–623.
4 J. Gawronski, N. Wascinska and J. Gajewy, Chem. Rev.,
2008, 108, 5227–5252.
5 A. D. Dilman and S. L. Loffe, Chem. Rev., 2003, 103, 733–
772.
D. T. McQuade, Angew. Chem., Int. Ed., 2007, 46, 1698–
1701.
23 A. El Kadib, R. Chimenton, A. Sachse, F. Fajula,
A. Galarneau and B. Coq, Angew. Chem., Int. Ed., 2009, 48,
4969–4972.
1646 | Green Chem., 2014, 16, 1639–1647
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
24 The reactors are named as mfr-SILLP-N-type, where mfr
Escrich, S. Sayalero, A. Bastero and M. A. Pericàs, Chem.–
Eur. J., 2009, 15, 10167–10172.
indicates a mini-flow reactor, SILLP-N refers to the number
assigned to the catalyst (N); and type defines the 27 Commercially available Omnifit(R) glass chromatography
morphology of the polymer: gel for microporous beads (gel
type), macro for macroporous beads, and molth for the
macroporous monolith.
columns with adjustable height end pieces (lunger) were
used.
28 Swelling (vf − vi/vi × 100).
25 S. V. Luis and E. García-Verdugo, Continuous Flow Systems 29 (a) N. Karbass, V. Sans, E. García-Verdugo, M. I. Burguete
Using Polymer-Supported Chiral Catalysts, in Polymeric
Chiral Catalyst Design and Chiral Polymer Synthesis, ed.
and S. V. Luis, Chem. Commun., 2006, 3095–3097;
(b) P. Lozano, E. García-Verdugo, R. Piamtongkam,
N. Karbass, T. de Diego, M. I. Burguete, S. V. Luis and
J. L. Iborra, Adv. Synth. Catal., 2007, 7, 1077–1084.
S. Itsuno, John Wiley
9781118063965.ch5, 2012.
& Sons, Inc., DOI: 10.1002/
26 (a) P. Kasaplar, C. Rodríguez-Escrich and M. A. Pericàs, 30 G. Strappaveccia, D. Lanari, D. Gelman, F. Pizzo, O. Rosati,
Org. Lett., 2013, 15, 3498–3501; (b) E. Alza, C. Rodríguez-
M. Curini and L. Vaccaro, Green Chem., 2013, 15, 199–204.
This journal is © The Royal Society of Chemistry 2014
Green Chem., 2014, 16, 1639–1647 | 1647