Journal of Colloid and Interface Science 323 (2008) 359–364

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Synthesis of silicalite-1 from organo-silicic gels

Norma A. Sánchez-Flores a,∗ , Graciela Pacheco-Malagón a , Patricia Pérez-Romo c , Héctor Armendáriz c ,
Jaime S. Valente c , María de L. Guzmán-Castillo c , Jorge Alcaraz b , Leticia Baños d , J.M. Saniger Blesa a , J.J. Fripiat a,c
a
Centro de Ciencias Aplicadas y Desarrollo Tecnológico, Universidad Nacional Autónoma de México, Av. Universidad #3000, 70-186 CP 04510, México, Mexico
b
Facultad de Química, Edificio E, Universidad Nacional Autónoma de México, Av. Universidad #3000, 70-186 CP 04510, México, Mexico
c
Instituto Mexicano del Petróleo, Eje Central Lázaro Cárdenas No. 152, Col. San Bartola Atepehuacan, CP 07730, México, Mexico
d
Instituto de Materiales, Universidad Nacional Autónoma de México, Av. Universidad #3000, 70-186 CP 04510, México, Mexico

a r t i c l e i n f o a b s t r a c t

Article history: Well crystallized silicalite-1 has been obtained from three sources of amorphous silica, namely, rice hull
Received 27 February 2008 ashes, commercial Davisil, and a fume silica from Aldrich. The silicas were first dissolved in glycerol ac-
Accepted 16 April 2008 cording to a recently described reaction. This reaction transforms rapidly and efficiently large surface
Available online 23 April 2008
area silicates into poly-alkoxide gels. It can be schematized as an etherification of an alcohol func-
Keywords:
tion of glycerol by the weakly acid surface silanol groups. The facile hydrolysis of the alkoxide permits
Silicalite-1 synthesis the preparation of relatively pure and reactive silica, keeping the mesoporous character of the parent
Mesoporosity starting material. We insist on the mesoporous character of the solids obtained upon hydrolyzing the
organo-silicic gel because we believe the gel plays a role of template in the secondary synthesis of
mesoporous structures. The hydrolysis is carried out in presence of a structure directing agent, namely
tetra-propylammonium hydroxide, TPAOH. After aging, the residue is dried and calcined. The first advan-
tage of using the organo-silicic gel is probably related to the high degree of depolymerization of silica,
witness by the C/Si ratio. The second one, more subtle to define, is to provide an intermediate silica
with hydrophilic a hydrophobic regions, interfering differently with the surfactant. After calcination at
500 ◦ C, well crystallized silicalite-1 is obtained. The texture of the starting silica influences the textural
characteristics of the final silicalite-1.
© 2008 Elsevier Inc. All rights reserved.

1. Introduction should be obtained using a similar procedure, omitting the addi-

The first description of silicalite-1 was published by Flanigen
et al. in [1]. Silicalite-1 was grown from colloidal silica, autoclaved
at 100–200 ◦ C with tetra-propylammonium hydroxide (TPAOH) as
mineralizer. According to Barrer [2], “the versatility of hydrothermal
chemistry owes much to the mineralizing role of water.”
The use of a cheap, highly branched polysilanols, obtained by
reacting SiO2 with glycerol, will be investigated here in the pur-
pose to synthesize silicalite-1.
The interest of these highly branched polysilanols in the syn-
thesis of zeolite has been demonstrated recently by the successful
synthesis of ZSM-5 from an amorphous organo-alumino-silicic gel.
This gel was formed by reacting a mechanical mixture of RHA and
of natural clinoptilolite (working as aluminum donor), with glyc-
erol during 2 h at 200 ◦ C [3]. It was hydrolyzed by a solution of
tetra-propylammonium hydroxide (TPAOH) and then aged for two
days at 135 ◦ C in a stainless steel autoclave. Silicalite-1 having the
same MFI structure as ZSM-5, except for the absence of aluminum,
* Corresponding author.
E-mail address: norma.sanchez@ccadet.unam.mx (N.A. Sánchez-Flores).
tion the aluminum donor.
Tailoring the particle size of silicalite-1, by using specific struc-
ture directing agents (SDA) has been tried. Axnanda and Shantz [5]
outlining that microemulsions have been successful in the synthe-
sis of nanoparticles have focused their efforts on micro-emulsions
for manipulating zeolite particle morphology. A delicate balance
between competing factors has to be found. On the one hand,
micro-emulsion will function as a nanoreactor and will modu-
late nucleation and growth of the zeolite crystals. On the other
hand, the surfactant will affect the growth rates of certain crys-
tal faces by preferential adsorption. Micro-emulsions formed with
straight chain surfactant (CTAB) lead to amorphous silica, whereas
a branched surfactant leads to silicalite-1 [5].
These references to micro-emulsions are relevant here. In our
technique the polysilanol polymer has hydrophilic, Si–OH, and hy-
drophobic Si–O–CH3 regions, the extent of which depends upon
the extent of hydrolysis step.
About the depolymerization step, Qinghua [6] observed that
the nucleation period for the TPAOH–SiO2 –H2 O-ethanol system is
longer than that for the system TPAOH–TEOS–H2 O. Colloidal SiO2
must first depolymerize to reach a critical concentration before nu-
cleation can start. Hence, these authors pointed out that the use

0021-9797/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2008.04.039
360 N.A. Sánchez-Flores et al. / Journal of Colloid and Interface Science 323 (2008) 359–364

of a gel containing depolymerized silica should be beneficial by Table 1

shortening the aging time in secondary synthesis. In summary, this Textural characteristics of the starting silicas
paper will show that well crystallized silicalite-1 can be obtained Sample BET t-micro A Pore volume Average pore
easily from a cheap organic polysilanol prepared by depolymeriz- (m2 /g) (m2 /g) (cc/g) diameter (Å)
ing large surface area silicas in glycerol. RHA 173.8 0 0.268 62
Fume silica 202.1 7 .3 1.394 276
2. Experimental Davisil grade 480 na 0.75 na
634, Aldrich

2.1. Sources of silica

Table 2
Three sources have been chosen. In addition to biogenetic silica, Textural parameters obtained at 77.35 K from N2 adsorption–desorption isotherm
like that obtained by calcining the hull of the rice grain (RHA), two for the silicalite-1 samples, treated at 550 ◦ C. (a) SLF and (b) repetition
silicas, obtained by chemical processes, like Davisil or fume silica Parameters SLR SLD SLF
were used. Table 1 summarizes their textural characteristics.
t-Plot micropore A (m2 /g) 130.5 226.6 (a) 95.8
The properties of RHA and of the silica gels derived from it have (b) 118.4
been described separately [3]. The silica derived from the hydrol-
t-Plot external A (m2 /g) 101.7 151.2 (a) 242.4
ysis of the depolymerized RHA in water, has a BET surface area of (b) 208
237 m2 /g and a pore volume circa 0.46 cc/g.
t-Plot micropore volume V (cc/g) 0.065 0.114 0.055
Davisil grade 634, Aldrich catalogue 678, is a pure (99%) and
Total pore volume V (cc/g) 0.191 0.228 (a) 0.427
porous silica with a BET surface area of 480 m2 /g and a pore vol- (b) 0.394
ume of 0.75 cc/g.
% Micropore/pore volume V 43 50 12.8
The fume silica, obtained from Aldrich, CAS: 99.98% pure silica,
BET (m2 /g)
mp 1700 ◦ C, has a huge pore volume (1.394 cc/g) and very large
232.2 377.8 (a) 338.2
(b) 326.3
pores (diameter 276 Å).

(A) (B)

(C)

Fig. 1. X-ray diffraction patterns of (A) silicalite-1, SLR; (B) silicalite-1, SLD; (C) silicalite-1, SLF.
 N.A. Sánchez-Flores et al. / Journal of Colloid and Interface Science 323 (2008) 359–364 361

(A)

(B)

29 29
Fig. 2. (A) Comparison of the Si MAS NMR spectra of silicalite-1, SLR or SLD. (B) Si MAS NMR of silicalite-1, SLD, prepared from Davisil.

Table 3 The chemical analysis of the gel is not evident. Because the
29
Parameters used to simulate, by Gaussians functions, the Si MAS NMR spectrum primary reaction product is an organo-silicic gel dissolved in an
of silicalite-1, SLD, prepared from Davisil
excess glycerol, the chemical analysis of the gel must distinguish
Center of gravity FWHM Area fit the organic residues, attached to the inorganic framework, from
(ppm) (ppm) (%) the solvent, e.g., the excess glycerol. The removal of the water pro-
−116.8 1.406 8.7 duced by the etherification reaction, or of the physically adsorbed
−115.7 0.99 6.9 water, occurs between 100 and 200 ◦ C.
−114.6 1.04 16.7
−113.9 0.83 18.7
The removal of the excess ‘solvent,’ e.g., glycerol, is marked by
−113.3 0.83 15.9 a sharp endothermic peak between 200 and 300 ◦ C. The excess
−112.7 0.81 13.1 is taken into account in calculating the composition since it must
−111.9 1.09 13.9 be excluded in the calculation of the gel composition. Beyond 300
−110.2 1.41 5.9
and up to 1000 ◦ C, the organic fraction of the gel is removed by
cracking and combustion. The residue is dehydrated SiO2 . A broad
2.2. Synthesis procedure exothermic accident, observed near 350–550 ◦ C, is probably due to
the organic matter oxidation. The procedure to obtain the C/Si ratio
in the gel, excluding the carbon of the glycerol solvent has been
The procedure was quasi-independent of the nature of the described previously [7]. For instance, the C/Si ratio was ≈1 in the
starting silica: 2 g silica was depolymerized in 20 ml of glycerol at gel obtained from RHA before hydrolysis [3].
200 ◦ C for 2 h. The gels were subsequently hydrolyzed with 5 ml
of a 1 M TPAOH/water solution, separated from the excess glycerol 3. Characterization techniques
by decantation and aged under hydrothermal conditions at 135 ◦ C
for 3 days. The XRD of the crystalline silicalite-1 were recorded with a
The solid was separated from the liquid phase by centrifugation D-8-advance Bruker axs goniometer fitted with a graphite sec-
at 5000 rpm for 5 min. This separation was repeated 3 times. ondary beam monochromator filtering the CuK α radiation. The
Afterwards the solid was dried at 100 ◦ C for 24 h and calcined X-ray diffractogram was recorded from 2◦ and 70◦ with 2θ steps
at 500 ◦ C for 24 h also. of 0.02◦ and 1.2 s counting time.
362 N.A. Sánchez-Flores et al. / Journal of Colloid and Interface Science 323 (2008) 359–364
(A)
(C)
Fig. 3. (A) Large crystal (20/2 μm) in sample SLR. EDS indicates the absence of impurities in this crystal. (B) Smaller crystals and aggregate of nanoparticles of silica containing
the same inorganic impurities as the starting RHA, see text. (C) Single crystal electron diffraction pattern from SLD.
The 29 Si MAS NMR spectra were obtained at 79.492 MHz, while
spinning at 7.5 kHz Zirconium rotors, 4 mm i.d., 90◦ pulses with
20 s intervals were used. The number of transients was 250. The
NMR spectra deconvolution was achieved by using the Origin7
software.
Surface area and porosity were measured with a Quantachrome
Autosorb Automated Gas Sorption System. The pore size distribu-
tion was calculated from the desorption branch, according to the
BJH method [8].
Electron Microscopy was carried out with a Jeol 2200 FS
Transmission Electron Microscope with an acceleration voltage of
200 kV. Prior to the analysis, the sample was mounted onto a
carbon-coated formvar film, attached to a metal grid.
4. Results
Table 2 gives the nomenclature of the samples of silicalite-1
prepared from the silicas shown in Table 1. To the radical SL, indi-
cating that we deal with silicalite-1, added capital letters indicate
the origin of the starting silica: R, for rice hull ash, D, for Davisil,
and, F, for fume silica.
(B)
The SLR and SLD samples were identified as silicalite-1 without
ambiguity by comparing the XRD, Fig. 1, to the MFI XRD pub-
lished by van Koningsveld et al. [9], and reported by Treacy and
Higgins [10], as siliceous ZSM-5.
Several repetitions of the procedure gave similar results but
with some variations of the intensity of the XR reflections, either
due to uncontrolled variation in the extents of crystallization, or
to orientation effects, as suggested by the morphology of the flat
crystals.
The extent of microporous area gives a semi-quantitative ap-
proximate of the degree of crystallization. The orthorhombic phase
of silicalite-1 has twelve different (tetrahedral) sites [11]. Conse-
quently, twelve 29 Si MAS NMR lines should be observed in well
crystallized materials, this number being limited by instrumental
resolution [12,13].
The 29 Si MAS NMR spectra of SLR and of SLD are compared
in Fig. 2A. The 29 Si MAS NMR of SLD decomposed in eight lines,
in agreement with Fyfe [14], is represented in Fig. 2B. The pa-
rameters resulting from the simulation by Gaussian functions are
shown in Table 3. The shifts of the resonance lines in the 29 Si MAS
 N.A. Sánchez-Flores et al. / Journal of Colloid and Interface Science 323 (2008) 359–364 363

(A)

Fig. 4. Large and thin silicalite-1 crystal in SLD.

NMR spectra, Fig. 2, scale as the charges brought by the silicon

atoms [15].
The transmission electron microscopy image of silicalite-1
(SLR), shows two types of morphology. Beside well formed and
elongated crystals, having typically a length and width of a few mi-
crometers and a thickness of a fraction of micrometer (Fig. 3A), ag-
glomerates of amorphous nanoparticles are also detected (Fig. 3B).
The electron diffraction pattern by a single crystal is shown in
Fig. 3C. EDS chemical analyses show that large crystals are com-
posed of pure silica, while the amorphous nanoparticles contain
the same impurities as those found in the hydrolysis product of
the RHA gels [4], namely Na, Ca, Mg, and K, coming from the orig-
inal plant tissues. The sum of their concentrations is less than 2%.
The sum of their concentrations is 2.8% and the total contribution
of the amorphous nanoparticles, in SLR, is estimated to less than
10% by weight.
In silicalite-1, SLD, the crystals are, in average, larger, and thin-
ner. They are made of pure silica with evanescent traces of alu- (B)
minum, Fig. 4.
The textural analysis, performed by measuring the N2 physical Fig. 5. N2 adsorption–desorption isotherm on (A) silicalite-1, SLR, from RHA; (B) sili-
adsorption and by applying the BJH algorithm [8], yields the results calite-1, SLD, from Davisil.
shown in Tables 1 and 2. The t-plot micropore volume of SLR is
smaller that of SLD, but both values are smaller than that reported transformation of the RHA gel is apparently not as advanced as
for porous carbon template silicalite-1 [13]. The micropore surface that of the Davisil gel since, on the basis of XRD, 29 Si MAS NMR,
area is much larger in SLD than in SLR. Thus the material having and electron microscopy, SLD, is better crystallized than SLR ob-
zeolitic features, namely a large porous volume with opening of tained from RHA. A more advanced crystallization degree in Davisil
≈2 nm, is less abundant when starting from SLR. than in RHA, as witness also by their BET areas (Table 2), may be
The t-plot micropore surface area is also smaller than in the assigned to a more divided starting material.
ZSM-5. For the latter, the degree of transformation of the mixed According to earlier investigations [7], the reaction of silicates
gel of RHA and clinoptilolite [3] into ZSM-5 is high. The exter- with glycerol results in the formation of polysilanols or alkoxides
nal surface area is of the same order of magnitude in silicalite-1 and in the depolymerization of the original silica with the partial
as in ZSM-5. The secondary maximum observed near 3.5 nm in conservation of the original ordering. The similarity between the 29 Si
the pore size distribution shown in Fig. 5A, is reminiscent of the (and 27 Al) MAS NMR spectra of the gels and those of the corre-
singularity observed at the same radius in the RHA and in ZSM-5 sponding starting material had been interpreted as if a local order-
synthesized from RHA. This singularity is not observed in silicalite- ing persisted in the gels. In particular, in a gel, obtained from a
1, SLD, Fig. 5B. HY zeolite [13], the observation of tetrahedral aluminum has sug-
gested that the smallest organized structural unit has a volume
5. Discussion about 1 nm3 .
An interesting observation is the apparent existence of a corre-
Since the alumino-silicic gel, derived from the simultaneous lation between the porous volumes of the starting silica and those
dissolution of RHA and clinoptilolite was quite active in the syn- of the crystalline silicalite, Fig. 6. We tried to offer a simple ex-
thesis of ZSM-5, it was not surprising that the silicic gels obtained planation to this curious observation but, obviously more work is
from RHA or from DAVISIL were good precursors of silicalite-1. The necessary before accepting it. The walls of the pores existing in
364 N.A. Sánchez-Flores et al. / Journal of Colloid and Interface Science 323 (2008) 359–364
Fig. 6. Correlation between the porous volumes (cc/g) of the starting silicas and of
the silicalite-1, SLD, SLR, SLF.
the starting silica get decorated by the hydrophilic fragments of
the gels, for instance containing SiOH groups. Further calcination
would regenerate a network of secondary cavities, replacing the
original ones, inside the crystalline silicalite-1.
Thus the choice of the silica starting material rules to some
extent the textural characteristics of the final silicalite.
6. Conclusions
(1) The simultaneous use of an organo-silicic gel and of a struc-
ture directing agent (TPAOH) has resulted in a new synthesis
procedure for obtaining rapidly a micron-sized well crystal-
lized and pure silicalite-1.
(2) The porous volumes of the starting silicas, measured before
depolymerization in glycerol, are correlated to the porous vol-
umes of the resulting silicalite-1. Tailoring the particle size of
silicalite-1, could be thus achieved more simply by the choice
of the porous characteristics of the starting silica than by using
specific structure directing agents.
Acknowledgments
We thank deeply for their help: Ing. Carlos López (Instituto
Mexicano del Petróleo) for NMR spectra and Hugo A. Sánchez
(CCADET, Universidad Nacional Autónoma de México) for techni-
cal support.
References
[1] E.M. Flanigen, J.M. Bennett, R.W. Growse, J.P. Cohen, R.I. Patton, R.M. Kirchner,
J.V. Smith, Nature 271 (1978) 512.
[2] R.M. Barrer, Hydrothermal Chemistry of Zeolites, Academic Press, London, 1982.
[3] G. Pacheco-Malagón, N.A. Sánchez-Flores, J.M. Saniger Blesa, L. Baños, P. Pérez-
Romo, J.S. Valente, M.L. Guzmán-Castillo, F. Hernández-Beltrán, J.J. Fripiat, Mi-
cropor. Mesopor. Mater. 100 (2007) 70.
[4] N.A. Sánchez-Flores, G. Pacheco-Malagón, P. Pérez-Romo, H. Armendariz, M.L.
Guzmán-Castillo, J.M. Saniger, J.J. Fripiat, J. Chem. Technol. Biotechnol. 82
(2007) 614.
[5] S. Axnanda, D.F. Shantz, Micropor. Mesopor. Mater. 84 (2005) 236.
[6] L. Qinghua, Crystallization of colloidal TPA-silicalite-1 by a two-stage varying-
temperature synthesis, Licentiate thesis, ISNN 1402-1757/ISRN LTU-LIC-00/04-
SE/NR 2000:4.
[7] G. Pacheco-Malagón, P. Pérez-Romo, N.A. Sánchez-Flores, J.M. Saniger Blesa,
M.L. Guzmán-Castillo, F. Hernández-Beltrán, J.J. Fripiat, Inorg. Chem. 44 (2005)
8486.
[8] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373.
[9] H. van Koningsveld, J.C. Jansen, H. van Bekkum, Zeolite 10 (1990) 235.
[10] M.M.J. Treacy, J.B. Higgins, Collection of Simulated XRD Powder Patterns for
Zeolites, fourth ed., Elsevier, Amsterdam, 2001.
[11] C.A. Fyfe, G.C. Gobi, J. Klinowsky, J.M. Thomas, S. Ramadas, Nature 296 (1982)
530.
[12] E. Martinez, J.J. Fripiat, L.J. Alvarez, Chem. Phys. Lett. 349 (2001) 286.
[13] G. Pacheco-Malagón, P. Pérez-Romo, N.A. Sánchez-Flores, M.L. Guzmán-Castillo,
C. López-Franco, J.M. Saniger Blesa, F. Hernández-Beltrán, J.J. Fripiat, Inorg.
Chem. 45 (2006) 3408.
[14] C.A. Fyfe, J.H. O’Brien, H. Strobi, Nature 326 (1987) 281.
[15] V.L. Zholobenko, S.M. Holmes, C. Cundy, J. Dwyer, Micropor. Mesopor. Mater. 11
(1997) 83.