Applied Catalysis A: General 312 (2006) 7785

www.elsevier.com/locate/apcata

Surfactant containing Si-MCM-41: An efficient basic catalyst

for the Knoevenagel condensation
Leandro Martins a, Tito J. Bonagamba b, Eduardo R. de Azevedo b,
Pascal Bargiela c, Dilson Cardoso a,*
a
Chemical Engineering Department, Federal University of Sao Carlos, Rodovia Washington Lus,
km 235, 13565-905 Sao Carlos, SP, Brazil
b
Instituto de Fsica de Sao Carlos, Universidade de Sao Paulo, P.O. Box 369, 13560-970 Sao Carlos, SP, Brazil
c
Instituto de Qumica, Universidade Federal da Bahia, R. Barao de Geremoabo s/n, 40170-290 Salvador, BA, Brazil
Received 3 March 2006; received in revised form 17 June 2006; accepted 19 June 2006
Available online 2 August 2006

Abstract
Si-MCM-41 molecular sieves, which are largely used in a number of acid and redox catalyzed reactions, as reported by the literature, were used
as basic catalyst for the Knoevenagel condensation. This reaction involves the condensation of carbonyl compounds with active methylene
compounds to generate unsaturated products. The catalyst, which pores were occluded by the organic template, provided high activity in mild
conditions, even under 10 8C. The high activity is due to the presence of siloxy anions, of high basicity, located in the pore cavity of the hexagonally
ordered mesopores. Ion exchange procedures evidenced the presence of these high basic sites as a consequence of the high pH increase of the
solution containing tetramethylammonium chloride, or the high H+ consumption, added to a the medium as HCl, to keep constant pH. O 1s XPS
and 29Si CP/MAS NMR measurements confirmed the existence of only
SiOCTA+ ionic pair, which is the basic site itself.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Si-MCM-41; Basic catalysis; Knoevenagel condensation; Molecular sieves

1. Introduction pore diameter ranging from 2 to 50 nm and a surface area in the

range of 1000 m2 g1. These new materials have provided the
Catalysis from solids with acid properties, when compared development of new catalysts with larger dispersion in the
to other solids containing basic sites have received, along the active area, a characteristic that is critical for the improvement
years, large attention as a result of the significance these of current industrial processes [6].
catalysts have in petroleum refinement. However, in many A way to develop basicity in mesoporous materials is by
reactions, basic catalysts may be used to replace acids with a dispersing alkali metal oxides in the channel. Kloetstra et al. [7]
selectivity gain, because they inhibit the development of other initially proposed to disperse cesium oxide particles in Si-
parallel reactions, such as cracking and coke formation [1]. MCM-41 pores. For the preparation, the molecular sieve is
The use of microporous solids with high aluminum content impregnated with a concentrated cesium acetate solution
such as zeolite 13 as basic catalysts is widely known [2,3]. followed by calcination so that the acetate ion can decompose.
Contrarily, materials with highly ordered pores in the The generated basic sites were evaluated by a programmed CO2
mesopores region containing silicon and oxygen in their desorption and showed a high basic strength. However, due to
chemical composition have been little studied. These materials the treatments used throughout the preparation, its structure was
are called MMS (mesoporous molecular sieves) and were first extensively damaged, as seen through X-ray diffraction and by
developed in the 1990s by Mobil Company [4,5]. They have a nitrogen adsorption isotherms [7]. It is widely known that SiO
bonds, present in these materials, hydrolyze in high pH
environments. Because mesoporous molecular sieves have high
* Corresponding author. Fax: +55 16 33518266. surface area, the hydrolysis in these bonds takes place more
E-mail address: dilsoncardoso@power.ufscar.br (D. Cardoso). easily, even in a moderately basic environment. Results show
0926-860X/$ see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2006.06.035
78 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785

that pH 8.5 is the upper limit value to prevent the MCM-41 result, throughout titration, a constant pH value of 7.0 was kept
mesopore structure from undergoing significant destruction [8]. by continuously adding acid. The consumption of protons/g
Another way to get basic Si-MCM-41 sites is by solid was achieved by recording the volume of the acid source
functionalizing its surface with organic compounds, particu- consumed throughout the titration period.
larly with those containing terminal amines in their composi-
tion [9,10]. Before functionalizing, the ionic surfactant present 2.2. Characterization
in the pores is removed by calcination, generating silanol
groups, which are anchoring points. Concentration of silanol Small angle X-ray diffraction patterns were achieved with a
groups depends on how the ionic surfactant is removed. diffractometer D-5000 (Siemens), using powder method at
Calcined mesoporous silica has lower concentration of silanol 1.48 < 2u < 108 interval. Cu Ka (40 kV and 40 mA) radiation
groups as a result from their high condensation temperatures. If was used, nickel filter and 0.6 8/min speed.
the surfactant is removed by extraction with a solvent, their Nitrogen adsorption/desorption isotherms were achieved
pore surface is kept with plenty of silanol groups, with nearly with an equipment supplied by Quantachrome (Nova-1200).
100% coverage [9]. Before analysis, 50 mg of the sample, calcined or as-
All these procedures, which are used to generate basic sites synthesized, was vacuum treated at 110 8C for 2 h. Surface
on Si-MCM-41, are hard to process and not always lead to the areas were estimated through BET equation [14] and the pore
planned basic sites, stable enough for catalysis application. diameters were established via BJH method [15].
Recently, Kubota et al. [11,12] achieved excellent results Percent of organic material present in Si-MCM-41
during Knoevenagel condensation by using Si-MCM-41 molecular sieves, as-synthesized or acid titrated, was estab-
molecular sieve while keeping the surfactant inside the pores lished via C, H and N analyses in a Fisons equipment, model
called [CTA]Si-MCM-41. In this catalyst, active sites are high EA1108-CHNS-O.
basicity
SiO sites, which are in the channels. But, despite NMR analyses were performed at a VARIAN unit INOVA
good results, there is no further information available on the spectrometer operating at 79.5 and 400 MHz for 29Si and 1H,
acid-basic properties of Kubotas catalyst. respectively. Magic-angle spinning at 5 kHz was used in all
This work is designed to study the chemical behavior of experiments. In order to obtain more quantitative spectra a
catalyst [CTA]Si-MCM-41 by using it in the Knoevenagel cross-polarization excitation scheme using a radiofrequency
condensation model reaction in mild conditions. ramp (ramp CP/MAS) was used [16]. The ramp CP/MAS
spectra were obtained with 5 ms of contact time and recycle
2. Experimental delay of 2 s. 29Si direct polarization experiments (DP/MAS)
were performed using a 5-ms p/2 excitation pulse and 30 s of
2.1. Synthesis of [CTA]Si-MCM-41 molecular sieve recycle delay. In all experiments, during the 20-ms acquisition
time a 70 kHz 1H broadband two-pulse phase-modulation
[CTA]Si-MCM-41 catalyst was synthesized through Cheng (TPPM) [17] decoupling scheme was used.
et al. method [13]. Samples were prepared from aerosil silica, X-ray photoelectron spectroscopy measurements were done
cetyltrimethylammonium bromide (CTABr) and tetramethy- in equipment ESCALAB MKII de VG, source Al Ka, under
lammonium hydroxide (TMAOH). Composition of the reaction pressure of 108 mbar. C 1s was taken as reference at 285 eV.
medium was SiO2: 0.27, CTABr: 0.19, TMAOH: 40H2O.
First, 9.22 g TMAOH5H2O (Aldrich) and 26.35 g CTABr 2.3. Knoevenagel condensation
(Vetec) were diluted in 188.35 g deionized water. Then, the
mixture was kept stirring at 60 8C until a clear solution was This reaction was performed in a 20 mL using an initial
achieved. After cooling the solution to room temperature, reactants concentration of 0.5 mol L1 in the following way:
16.09 g of a silica source was added. The sample was aged for 1.07 g (10 mmol) benzaldehyde (Acros); 1.14 g (10 mmol)
24 h before going to thermal treatment, which was carried out ethyl cyanoacetate (Acros) and 15.42 g (17.9 mL) toluene
in autoclaves under autogenous pressure at 150 8C for 2 days. (Merck) as solvent were added to a borosicalate glass reactor
Finally, the sample was filtered, washed successively times with equipped with magnetic stirring and temperature control. In
deionized water up to pH < 10 and dried at 60 8C for 24 h. order to prevent from the presence of benzoic acid in the
Solid yield was approximately 10% compared to the gel mass. reaction media (oxidation of benzaldehyde) fresh reactant was
To compare with [CTA]Si-MCM-41, the surfactant occluded used. Additionally, chromatographic data revealed that, under
in the pores was removed through calcination, under an oxygen the mild reaction conditions used in this work, benzoic acid is
flow, through a heating ramp of 1 8C/min up to 520 8C, kept not present. The reaction was carried out at temperature
subsequently at this temperature for 6 h. intervals from 10 to 50 8C and atmospheric pressure. As soon as
Basic properties of the [CTA]Si-MCM-41 molecular sieve the reaction medium achieved planned temperature, 1 g of the
were established via titration with a hydrochloridric acid catalyst was added. Two different catalysts were evaluated in
solution (0.1 mol L1) from a suspension containing 1.5 g this study: (i) as-synthesized MCM-41, without any additional
[CTA]Si-MCM-41 in 200 mL deionized water or 0.5 mol L1 treatment and (ii) as-synthesized MCM-41 titrated with HCl.
of TMACl or NaCl solution. Preliminary assays showed that After 6 h reaction, the catalyst was retrieved for analysis and
when kept in an aqueous suspension, pH raised quickly. As a reuse by centrifuging and wash out with benzene to remove
 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785 79

adsorbed products. Analysis of reactants and products in liquid

samples was carried out with a Varian chromatograph equipped
with DB-1 capillary column and a flame ionization detector
(FID). Other analysis conditions included: injector tempera-
ture, 210 8C; column temperature, 140 8C for 12 min and
230 8C for 15 min, and He as carrier gas.

3. Results and discussion

Si-MCM-41 samples, calcined or not, showed a typical

diffractogram (Fig. 1a). As-synthesized sample showed a
specific area of 1.0 m2 g1 (Table 1) while calcined sample
showed a specific area of 990 m2 g1. Elemental C, H and N
analysis of as-synthesized Si-MCM-41 sample showed a C/N Fig. 1. X-ray diphratograma of mesoporous silica: (a) synthesized and (b)(d)
ratio close to 19, which corresponds to the CTA chemical exchanged in an aqueous medium while keeping pH at 7.
composition, indicating that tetramethylammonium hydroxide,
the mineralizing agent used in the synthesis, is not included in suspension. This exchange takes place very quickly: while 1.5 g
the solid phase. of the sample is added to 200 mL of 0.5 mol L1 TMACl
solution, pH rises to readings close to 10, with the need to
3.1. Ion exchange of [CTA]Si-MCM-41 molecular sieve frequently add HCl to keep pH at 7.0. Curve 1 in Fig. 3 shows
how H+/g [CTA]Si-MCM-41 consumption changed as a
Despite [CTA]Si-MCM-41 pores are completely filled with function of time in the presence of TMA cation, where it
the surfactant agent, Kubota et al. [11,12] reported that its can be seen the quick H+ consumption.
activity in Knoevenagel condensation is extremely high, even in Once the sample was washed with deionized water more
mild conditions. With the objective to develop even further the than 10 times and chemical analysis of the mesoporous material
activity of these materials, by increasing the contact of the indicated that it did not contain residual TMAOH, results from
reactants with active sites (larger surface area), the catalyst curve 1 in Fig. 3 suggest that surface
SiO sites, with high
[CTA]Si-MCM-41 was subjected to an ion exchange process basicity, are responsible for pH increase as a result of proton
with an organic cation with a lower volume than CTA+, i.e. consumption from water. In this process (Eqs. (1) and (2))
tetramethylammonium (TMA+). Removal of the surfactant formation
SiOH groups takes place and partial channel
occluded in Si-MCM-41 pores by ion exchange has been unblocking as a result from CTA removal:
performed by Hitz and Prins, by extraction with a hydro-
chloridric acid solution in ethanol [18]. (1)
In this study, during the ion exchange with a TMACl
aqueous solution, a sudden pH rise was evidenced. Essays with
different TMACl concentrations showed a relation between (2)
TMA cation concentration and the suspension pH, which was
measured 10 min after the mixture was completed (Fig. 2), that The presence of a considerable concentration of TMA cation
is, for higher TMA concentration, higher OH concentration. in the aqueous medium, forces CTA cation to exit the pores
To understand and better quantify this behavior, experiments (Eq. (1)), and the ion exchange takes place. However, because
were performed in a TMACl solution, by keeping pH at 7.0
SiOH groups are more stable than
SiOTMA groups, this
throughout the ionic exchange, by adding diluted HCl to the cation is quickly changed by a proton (Eq. (2)), with subsequent

Table 1
Modified MCM-41 molecular sieve composition
Sample CTA/Si (CTA+ + H+)a/Si BET area (m2 g1) Vp (mL/g) )
aH (A )
Dp (A
MCM-41, originalb 0.16 0.16  0.01 1 0 51.0 0
MCM-41, hydrolysis c 0.10 (37.5%)d 0.14  0.01 314 0.25 (25.2%)e 55.4 35.1
MCM-41, TMAClf 0.08 (50.0%) 0.16  0.01 374 0.28 (28.3%) 56.1 35.5
MCM-41, NaClf 0.07 (56.2%) 0.14  0.01 385 0.30 (30.3%) 57.0 35.6
MCM-41, calcinedg 0 (100.0%) 990 0.99 (100.0%) 46.4 33.8
a
CTA molar quantity present in the solid. Protons consumed in the titration.
b
C/N ratio = 19.5.
c
Sample taken by simple hydrolysis, according to Eq. (3).
d
In parenthesis: molar percent of removed CTA.
e
In parenthesis: volume of available pores.
f
Sample taken by ion exchange with TMACl or NaCl.
g
Sample taken through original MCM-41 calcination.
80 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785

approximately 5% of that expected, when taking into account

the quantity of CTA exchanged (Na/Si = 0.005). This result
confirms that despite most sodium is not in the solid, in

SiONa form, its presence in the solution helps to develop the
hydroxide ion, similarly to that showed in Eq. (2). Such
behavior confirms the hypothesis based on Eqs. (1)(3),
indicating the presence of high basicity
SiO groups. Cation
molar balance at the end of the 24-h period showed that the ratio
(CTA+ + Hconsumed+)/Si is kept at approximately 0.15, close to
the original value of 0.16 (Table 1). This result confirms that for
each removed CTA+ a silanol group was created in the solid.
The ratio C/N equal to 18.5 achieved in the sample treated in a
TMACl containing solution suggests that, just as it happens
Fig. 2. Ion exchange without pH control 10 min after startup: suspension pH as with sodium, only a small amount of TMA cation remains in the
a function of tetramethylammonium chloride concentration.
solid as
SiOTMA. These results show that the initial objective
to develop even further the activity of these materials, by
release of OH. This quick proton consumption (Eq. (2)) can increasing the contact of the reactants with active sites
also be explained by the fact that the TMA cation, as opposed to (
SiOTMA) cannot be achieved.
CTA, provides high mobility in a reaction medium because of Fig. 1 shows X-ray diffractogram from as-synthesized as
its smaller volume and because it is not included among the well exchanged MCM-41 by titration procedures at pH 7.0,
micelles, as it happens with CTA. Therefore, Eqs. (1) and (2) indicating that the mesoporous material structure is kept
explain why pH increases as TMA cation concentration throughout the exchange.
increases (Fig. 2). As TMA concentration increases, the ionic CTA/Si results shown in Table 1 indicate that approximately
strength in the solution is higher, leading to larger TMA 50% of the CTA cation present in the original MCM-41 was
exchange (Eq. (1)). When no cation is added to the solution, exchanged by submitting it to a TMACl solution at pH 7.0.
only hydrolysis can be seen (Eq. (3)), and proton consumption However, the mesoporous volume built during the exchange
speed, to maintain pH at 7.0, is smaller, as shown in curve 3 of with TMACl is very small, compared to calcined MCM-41
Fig. 3: (approximately 28%) as shown in the nitrogen physisorption
isotherm (Fig. 4 and Table 1). The reason for such behavior
(3) could be explained, for example, by a preferential exchange of
the CTA present on the external surface of MCM-41 particles.
To determine the quantity of cation involved in the ion Exchanging CTA for a cation of lower volume would not
exchange, an exchange procedure was performed in a increase significantly the pore volume, because the external
0.5 mol L1 NaCl solution. Similarly to TMA cation, the surface of the particles does not significantly help this
presence of sodium in the liquid phase also increased the parameter. But this hypothesis is unlikely to be true because
consumption speed of H+ (curve 2 of Fig. 3), compared to CTA cations present in the external surface of the particles are
simple hydrolysis. Exchange speed was a little higher in the poorly connected to SiO groups and, therefore, are easily
presence of sodium cation, because it has a smaller volume and leached during washing process. A second hypothesis to
therefore diffuses more easily. explain the origin of the pore volume could be the removal of
After 24 h of ion exchange, the amount of sodium present in
the solid (determined by elemental chemical analysis) was

Fig. 4. Nitrogen physisorption isotherms of MCM-41 molecular sieves. Filled

points: adsorption and empty points: dessorption. (&) and (&) Calcined MCM-
Fig. 3. Proton consumption as a function of time throughout [CTA]Si-MCM-41 41, (~) and (~) MCM-41 exchanged for 24 h with TMACl at pH 7, (*) and
sample titration, with pH being controlled at 7.0. (*) MCM-41 as synthesized.
 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785 81

Fig. 5. Ion exchange process proposed between an M+ cation and CTA+ cation in sample [CTA]Si-MCM-41.

CTA cation progressively to the interior of the channels, but in reactants (1 and 2) around unity, indicated that the subsequent
such case, both properties (CTA extraction and pore volume) Michael addition reaction (Eq. (5)) does not take place at the
should be proportional. A third and more likely hypothesis temperatures here used:
would be the uneven exchange of these cations along the
channels, as depicted in Fig. 5. In this case, the CTA extraction
not necessarily provides accessibility to the pore.
Results shown in Table 1 also shows that, as reported by (4)
other authors [13], throughout calcination, there is a strong
contraction of the mesoporous hexagonal arrangement para-
meter (aH), as a result from the condensation of neighboring
silanol groups. Results also show that after a partial ion
exchange of the CTA+ with H+, the mesoporous hexagonal
arrangement parameter and the diameter of the pores (Dp) in the
molecular sieves tend to increase. Such behavior may result
from lower CTA concentration, whose presence may contract
the pore diameter, as a result from its interaction to micelle (5)
formation.

3.2. Knoevenagel condensation in toluene

Results from Knoevenagel condensation (Eq. (4)) between The high reactant consumption speed, in aprotic medium,
benzaldehyde (1) and ethyl cyanoacetate (2) to form ethyl-2- confirms the presence of high basicity, as noticed during ion
cyano-3-phenylacrylate (3) are shown in Fig. 6. [CTA]Si- exchange studies in aqueous media (Fig. 3). When calcined Si-
MCM-41 catalyst had high activity, even in temperatures that MCM-41 was added to the reaction medium, a very low
were very below the ones reported by other authors [2]. conversion was seen (2%), suggesting that this material has
Throughout the reaction, the presence of a single product was very weak basic sites, possibility due to residual cations, which
always evidenced (3) and the ratio between the consumed are not completely removed. Even molecular sieve 13, which
has been extensively studied as a basic catalyst for the
Knoevenagel condensation [2], showed a low conversion at
50 8C (9% after 6 h, Fig. 6).
Fig. 7a shows benzaldehyde conversion as a function of time
for the catalyst reused in up to four successive batches. Its activity
is slightly reduced after every new use, probably as a result from
the leaching of part of the CTA to the reaction medium (step E in
Fig. 8). This result will be better discussed further.
Results from thermogravimetry shown in Fig. 9a and Table 2
indicate that CTA leaching to the reaction medium, in the first
use of a catalyst, is only ca. 0.3% m/m. The derivate curve
resulting from the loss of mass in as-synthesized [CTA]Si-
MCM-41 sample (Fig. 9b) provides a signal at approximately
160 8C (indicated by an arrow). This signal is related to CTA
decomposition taking place at pore-mouth [11,12,19]. For
Fig. 6. Benzaldehyde conversion as a function of time for different reaction samples used from the third catalytic test, no presence of this
temperatures. signal was seen, thus confirming that partial CTA leaching from
82 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785

Fig. 7. (a) Benzaldehyde conversion as a function of time and (b) after 6 h for [CTA]Si-MCM-41 catalyst reused in four successive cycles. Reaction
temperature = 50 8C.

pore-mouth to the reaction medium tends to reduce as the reaction medium. Finally, the third filtrate has practically no
catalyst is used. These cations are probably the ones most activity. These results differ from that of Kubota et al. [11,12],
weakly connected to the micelles and their leaching, therefore, which found no activity in the homogeneous phase, because they
could be one of the responsible for the slight reduction of the washed the catalyst exhaustively with hot water, before using it.
catalyst activity, as shown in Fig. 7b, curve 1. Curiously, despite the fact that in the third use of the catalyst,
To verify if the Knoevenagel reaction also occurred in no activity was found in the homogeneous phase, the activity of
homogeneous phase, the liquid phase resulting from a catalytic the solid remains practically the same (Fig. 7b, curves 2 and 1,
test was isolated from the as-synthesized solid catalyst and new respectively). This suggests that, after the third use of the
reactants were added to the filtrated. The new mixture was catalyst, no more leaching of CTA occurs. The explanation for
submitted to the same conditions as in the presence of the solid this stabilization can be achieved taking into account the
catalyst. The results are in curve 2 of Fig. 7b and shows that, hypothesis of Kubota et al. [11], that the losses of active sites
despite low quantity of leached CTA, by utilizing the filtrated occurs mainly at pore-mouth, similarly as shown in Fig. 5. The
from the catalyst used for the first time, a significant catalytic remaining CTA, at the interior of the pore, are not so easily
activity, but smaller than curve 1, can be seen. When the filtrated leached because they are more stable, as a consequence of a
from the second reaction is used, a much smaller activity is seen, higher interaction of the non-polar tails, as stated by Li et al.
thus confirming that smaller CTA+ quantity is leached to the [20] and Pool and Bolhuis [21].
The catalyst behavior in heterogeneous as well in
homogeneous phase, shown in Fig. 7b can be explained
through the scheme represented in Fig. 8. The steps AC are
similar to those proposed by Corma et al. [22], for Knoevenagel
reactions using zeolites. Depending on the stability of the CTA
cation in the micelle, the hydroxide ion eliminated from the
alkoxide ion (step D) can react in two different ways: forming
water and regenerating the siloxy active site (step A), or
forming CTAOH in the liquid phase.
Fig. 10 shows benzaldehyde conversion as a function of time
by using 100 mg of catalysts obtained at the end of the titration
with HCl (about 24 h, Fig. 3). It is significant to see that despite
the CTA content in exchanged samples being not very different
among them (Table 1), the catalytic activities are. The slightly
lower activity seen in the sample subjected to hydrolysis only
(curve 1 respect to 2) is possibly due to the need for the
reactants to diffuse to reach active sites
SiOCTA, which are
deeper inside the pores than in the original sample (curve 1).
For samples titrated in a TMACl or NaCl solutions (curves 3
and 4) the catalyst activity is much lower than previous ones
and, therefore, another unknown factor should prevail. The
activity calculated on a Si mol basis leads to the same
conclusions. It is necessary to mention that despite the samples
used in curves 24 (Fig. 10) have a much higher surface area
Fig. 8. Probable benzaldehyde condensation steps with ethyl cyanoacetate than the as-synthesized used in curve 1 (see Table 1), their
using [CTA]Si-MCM-41 catalyst. activities are lower.
 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785 83

Fig. 9. (a) Thermogravimetric analysis and (b) derivate curves resulting from the loss of mass for [CTA]Si-MCM-41 catalyst before and after its use.

In addition to higher stability of active sites in exchanged For as-synthesized, [CTA]Si-MCM-41 sample, almost no
samples, measurements for X-ray diffraction show that the difference between DP/ and CP/MAS NMR spectra is observed,
catalyst structure is kept intact after its use. Therefore, the as well in Q3/(Q2 + Q4) ratio (51.950.4%, Table 3). This is a
micelles present in the channels of the catalyst are significant strong indication that the siloxy groups, SiOM+, are not
for the mesopore structure to be kept as planned, both during hydrogen-bonded, but they are present major as SiOCTA+
the exchange (Fig. 3) and the reaction. ionic pairs, generating SiO basic sites. Because there is a
longer distance between 29Si in the lattice and 1H nuclei
3.3. Siloxy group characterization: 29
Si MAS NMR and O belonging to CTA+ chain and since the 1H nuclei in CTA+ are
1s/XPS spectroscopy rather mobile (as observed in the 1H spectra, not shown), the CP
transfer between these nuclei is not very effective. This makes
Solid-state 29Si nuclear magnetic resonance experiments the contribution of the SiOCTA+ ionic pair to the Q3 signal
29
( Si NMR) employing magic-angle spinning (MAS) with similar in both CP/MAS and DP/MAS spectra [26]. In contrast,
direct or cross polarization (DP or CP) has been shown to be a there is a very large difference between DP/ and CP/MAS
reliable mean to characterize various silicate materials [23]. spectra in the calcined sample, especially in the Q3 species (see
Particularly, CP process is interesting, thus it allows Fig. 11) and as well in Q3/(Q2 + Q4) ratio (36.365.5%,
discriminating against 29Si nuclei bonded close to hydrogen Table 3). Because silanol groups are expected to exhibit a very
atoms [24]. By 29Si NMR spectroscopy, the fraction of Si atoms effective 1H29Si CP transfer, the higher intensity of the peak at
bonded to nSi atoms, identified as Qn species (0  n  4), as 100 ppm in the CP/MAS as compared with the DP/MAS
well bonded to single and geminal SiOH groups, Q3 and Q2 spectra suggests the presence of a very high concentration of
silicon sites, respectively, can be discriminated sensitively [24 silanol groups in the calcined sample, not observed in the non-
27]. calcined one.
Fig. 11a and b, as well as Table 3 show 29Si DP and CP/MAS Additionally, in literature is well known that O 1s XPS is a
NMR spectra of as-synthesized and calcined MCM-41 way of directly determining Lewis basicity in zeolites and
samples. For both samples three peaks are observed, at related aluminosilicates materials [28,29] and that, small
91, 100 and 109 ppm. According to Sindorf and Maciel variations in oxygen binding energy means great difference in
[26], the low intensity peak at 91 ppm in calcined samples structural oxygen basicity [30,31].
corresponds to silicon atoms bonded to two siloxane bonds
SiOSi (Q2) and two geminal silanol groups. Similarly, the
resonance peak at 100 ppm is attributed to silicon atoms with
three siloxane bonds (Q3) and one silanol. Finally, the
resonance peak at 109 ppm is related to silicon atoms with
four siloxane bonds (Q4).

Table 2
Loss of mass seen in the thermogravimmetric analysis
Sample Loss of mass (%)a
[CTA]SiMCM-41 54.2
First use 53.9
Second use 53.6
Third use 53.5
Fourth use 53.5
Fig. 10. Benzaldehyde conversion as a function of time at 50 8C for catalysts
a
At temperatures ranging from 120 to 800 8C. modified by ion exchange (using 100 mg mass).
84 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785

29
Fig. 11. Si DP/ and CP/MAS NMR spectra for (a) as-synthesized and (b) calcined MCM-41.

Table 3 Additionally to these results, measurements of 29Si DP/MAS

Relative peak areas (%) in the 29Si DP/ and CP/MAS NMR spectra of MCM-41
NMR, performed by Kubota et al. [11] in as-synthesized
as-synthesized and calcined
[CTA]Si-MCM-41, showed two peaks Q3 and Q4, more than
Sample Signal Q2 (%) Q3 (%) Q4 (%) Q3/(Q2 + Q4) 50% in the form of Q3, attributed to the presence of Si(3-OSi, 1-
[CTA]Si-MCM-41 DP 17.7 51.9 30.4 1.1  0.1 O) moieties. Therefore, these authors indirectly suggested that
CP 11.6 50.4 38.0 1.0  0.1
SiO sites are the ones responsible for the high basic activity
Si-MCM-41 calcined DP 15.7 36.3 48.0 0.6  0.1 of this catalyst. This proposal was also presented by Brunel
CP 18.0 65.5 16.5 1.9  0.1 et al. [33], that using XPS suggested the existence of
SiOH3N+R ionic pairs in MCM-41 containing grafted amine
O 1s XPS measurements of as-synthesized and calcined groups. In both works, the authors suggest the existence of
MCM-41, shown in Fig. 12a and b, respectively, indicate that
SiOM+ groups, where M+ (instead of H+) is CTA+ or an
binding energy of O 1s is much lower in presence of the organic cation, by indirect means and differently of the methods
voluminous CTA cation (532.5 eV), which means that the employed in this work.
Lewis basicity of the SiO oxygen atom, belonging to [CTA]Si-
MCM-41 sample, is much higher. Similarly, Li et al. [32] 4. Conclusions
studied the influence of Ba content in MCM-41 basic catalysts
and found that the binding energy of O 1s decreases from 533.0 Mesoporous silica Si-MCM-41, with its pores still filled
to 532.5 eV as the Ba content increases. Consequently, as CTA+ with the surfactant cation provided strong
SiO basic sites.
is a voluminous cation, its interaction with
SiO sites is weak, The high catalyst activity in toluene and low temperatures was
moving the chemical balance SiOCTA $ SiO + CTA+ to the evidenced by the conversion of 99% of the reactants in a short
right and increasing
SiO basicity. Because the [CTA]Si- reaction period. Despite this molecular sieve not showing shape
MCM-41 catalyst channels are filled with the surfactant, selectivity property, as the reaction takes place at the pore-
according to Kubota et al. [11,12], the high activity seen in mouth only, it showed to be a very promising basic catalyst for
Knoevenagel condensation is mostly due to the basic sites fine chemistry synthesis.
present in the pore-mouth. Ion exchange in aqueous media at a neutral and constant pH
evidenced the presence of very strong
SiO basic sites. To
keep pH constant throughout the exchange, there was a need to
continuously add hydrochloridric acid. Although the contribu-
tion to the basicity due to CTA+OH ion pair, occluded insided
of the MCM-41 channels, cannot be excluded, the existence of
siloxy basic sites was supported by O 1s XPS and 29Si CP/MAS
NMR measurements.
Due to the presence of part of the CTA remaining inside the
channels, becoming part of the micelles, no change was seen in
the X-ray diffraction pattern of the samples, both in the ion
exchange and in the reaction.

Acknowledgements

The authors would like to thank Heloise O. Pastore (IQ/

Fig. 12. O 1s XPS spectra of (a) as-synthesized and (b) calcined MCM-41. Unicamp) for taking DRX measurements and Sonia Maria
 L. Martins et al. / Applied Catalysis A: General 312 (2006) 7785 85

Cabral de Menezes (Cenpes/Petrobras) for suggestions about [14] S. Brunauer, P.H. Emmet, E. Teller, J. Am. Chem. Soc. 60 (1938) 309.
29 [15] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373.
Si NMR experiments. The Brazilian research agencies
[16] O.B. Peersen, X.L. Wu, I. Kustanovich, S.O. Smith, J. Magn. Res., Ser. A
FAPESP and CNPq supported this project. 104 (1993) 334.
[17] A.E. Bennett, C.M. Rienstra, M. Auger, K.V. Lakshmi, R.G. Griffin, J.
References Chem. Phys. 103 (1995) 6951.
[18] S. Hitz, R. Prins, J. Catal. 168 (1997) 194.
[1] K. Tanabe, W.F. Holderich, Appl. Catal. A: Gen. 181 (1999) 399. [19] Y. Kubota, Y. Nishizaki, H. Ikeya, J. Nagaya, Y. Sugi, Stud. Surf. Sci.
[2] U.D. Joshi, P.N. Joshi, S.S. Tamhankar, V.V. Joshi, C.V. Rode, V.P. Catal. 141 (2002) 553.
Shiralkar, Appl. Catal. A: Gen. 239 (2003) 209. [20] Z.H. Li, S.J. Roy, Y.Q. Zou, R.S. Bowman, Environ. Sci. Technol. 32
[3] D. Barthomeuf, Micropor. Mesopor. Mater. 66 (2003) 1. (1998) 2628.
[4] J. Weitkamp, M. Hunger, U. Rymsa, Micropor. Mesopor. Mater. 48 (2001) [21] R. Pool, P.G. Bolhuis, J. Phys. Chem. B 109 (2005) 6650.
255. [22] A. Corma, V. Fornes, R.M. Martn-Aranda, H. Garca, J. Primo, Appl.
[5] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Catal. 59 (1990) 237.
Schmitt, C.T.-W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. [23] S. Haukka, E.L. Lakomma, A. Root, J. Phys. Chem. 97 (1993) 5085.
Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834. [24] X.S. Zhao, G.Q. Lu, A.K. Whittaker, G.J. Millar, H.Y. Zhu, J. Phys. Chem.
[6] H. Van Bekkum, E.M. Flanigen, P.A. Jacobs, J.C. Jansen, Introduction to B 101 (1997) 6525.
Zeolite Science and Practice, Elsevier, 2001. [25] C. Cannas, M. Casu, A. Musinu, G. Piccaluga, J. Non-Cryst. Solids 351
[7] K.R. Kloetstra, M. van Laren, H. van Bekkum, J. Chem. Soc., Faraday (2005) 3476.
Trans. 93 (1997) 1211. [26] X.S. Zhao, G.Q. Lu, J. Phys. Chem. B 102 (1998) 1556.
[8] C.N. Perez, E. Moreno, C.A. Henriques, S. Valange, Z. Gabelica, J.L.F. [27] D.W. Sindorf, G.E. Maciel, J. Am. Chem. Soc. 103 (1981) 4263.
Monteiro, Micropor. Mesopor. Mater. 41 (2000) 137. [28] M. Stocker, Micropor. Mater. 6 (1996) 235.
[9] C.-M. Yang, K.-J. Chao, J. Chin. Chem. Soc. 49 (2002) 883. [29] J. Xie, M. Huang, S. Kaliaguine, Appl. Surf. Sci. 115 (1997) 157.
[10] F. Fajula, D. Brunel, Micropor. Mesopor. Mater. 48 (2001) 119. [30] J. Tsou, P. Magnoux, M. Guisnet, J.J.M. O rfao, J.L. Figueiredo, Appl.
[11] Y. Kubota, Y. Nishizaki, H. Ikeya, M. Saeki, T. Hida, S. Kawazu, M. Catal. B: Environ. 51 (2004) 129.
Yoshida, H. Fujii, Y. Sugi, Micropor. Mesopor. Mater. 70 (2004) 135. [31] M. Huang, A. Adnot, S. Kaliaguine, J. Catal. 137 (1992) 322.
[12] Y. Kubota, H. Ikeya, Y. Sugi, T. Yamada, T. Tatsumi, J. Mol. Catal. A: [32] Q. Li, S.E. Brown, L.J. Broadbelt, J.-G. Zheng, N.Q. Wu, Micropor.
Chem. 249 (2006) 181. Mesopor. Mater. 59 (2003) 105.
[13] C.F. Cheng, D.H. Park, J. Klinowski, J. Chem. Soc., Faraday Trans. 93 [33] D. Brunel, A.C. Blanc, E. Garrone, B. Onida, M. Rocchina, J.B. Nagy, D.J.
(1997) 193. Macquarrie, Stud. Surf. Sci. Catal. 142 (2002) 1395.