Langmuir 2004, 20, 4885-4891 4885

In Situ Synchrotron Small-Angle X-ray Scattering/X-ray

Diffraction Study of the Formation of SBA-15 Mesoporous
Silica
Katarina Flodström,† Cilâine V. Teixeira,‡ Heinz Amenitsch,§
Viveka Alfredsson,† and Mika Lindén*,‡,|
Physical Chemistry 1, Lund University, P.O. Box 124, SE-221 00, Lund, Sweden,
Department of Physical Chemistry, Abo Akademi University, Porthansgatan 3-5,
FI-20 500, Turku, Finland, Institute of Biophysics and X-Ray Structure Research,
Austrian Academy of Sciences, Schmiedlstrasse 6, A-8043 Graz, Austria, and Department of
Heterogeneous Catalysis, Max-Planck-Institute for Coal Research, Kaiser-Wilhelm-Platz 1,
D-45470, Mülheim/Ruhr, Germany

Received February 11, 2004. In Final Form: April 5, 2004

Downloaded via COMISION NACIONAL ENERGIA ATOMICA on August 19, 2023 at 22:56:46 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

The initial stages of the formation of SBA-15 have been studied by in situ SAXS/XRD using synchrotron
radiation. Modeling of both the diffuse scattering and the X-ray diffraction patterns obtained at different
stages of the reaction results in a detailed description of the different reaction steps. The first step in the
formation is the liquid-liquid phase separation of spherical P123-silicate hybrid micelles after which
nucleation and growth of the 2D hexagonal phase occurs. Two-dimensional electron density maps calculated
on the basis of the intensities of the Bragg reflections suggest that changes in the degree of intermicellar
condensation are responsible for the time-dependent observations. The silica source (alkoxysilanes) may
be partially unhydrolyzed long after formation of the hexagonal structure, since the kinetics of the mesophase
evolution is notably slower when TEOS is used as the silica precursor as compared to TMOS. The results
obtained for SBA-15 are compared with other published data on the formation of SBA-15 and of the smaller
pore MCM-41 materials.

Introduction stable material. Furthermore, these materials usually

Ordered mesoporous silica materials have appealing
properties such as large internal surface areas and narrow
pore size distributions. Amphiphilic molecules are central
in the syntheses of these materials, where structures
similar to the mesostructures of surfactant-water systems
are formed. The amphiphiles function as scaffolds, or
structure directors, for the silica network, enabling the
formation of the mesoscopically ordered material. Syn-
theses with cationic amphiphiles, for example, cetyltri-
methylammonium chloride (CTACl), have been shown to
generate interesting structures such as the two-dimen-
sional hexagonal (p6mm) MCM-41 material and the
bicontinuous cubic (Ia3 h d) MCM-48 structure.1,2 Other
amphiphiles, for instance, nonionic surfactants and am-
phiphilic block copolymers, are today also widely used as
structure directors. The first synthesis of mesoporous silica
using nonionic surfactants was published in 1995.3
Subsequently, Pluronic-type triblock copolymers, (EO)x-
(PO)y-(EO)x (EO, ethylene oxide; PO, propylene oxide),
were shown to be excellent for creating highly ordered
mesostructures.4,5
Using block copolymers instead of ionic surfactants
results in wider pores, thicker silica walls, and a more
† (8) Flodström, K.; Alfredsson, V.; Källrot, N. J. Am. Chem. Soc. 2003,
Lund University.
‡ Abo Akademi University.
§ Austrian Academy of Sciences.
| Max-Planck-Institute for Coal Research.
(1) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck,
J. S. Nature 1992, 359, 710-712.
(2) Inagaki, S.; Fukushima, Y.; Kuroda, K. J. Chem. Soc., Chem.
Commun. 1993, 8, 680-682.
(3) Bagshaw, S. A.; Prouzet, E.; Pinnavaia, T. Science 1995, 269,
1242.
(4) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.;
Chmelka, B. F.; Stucky, G. D. Science 1998, 279, 548-552.
have interconnecting micropores between the ordered
mesopores,6 which seem to be absent in the MCM-type
materials. Different mesostructures of the inorganic
material can be obtained depending on, for example, the
Pluronic used, the temperature of the experiment, and
the addition of additives such as alcohol or salt.5,7-11 Zhao
et al.5 synthesized a two-dimensional hexagonal (p6mm)
silica now known as SBA-15 with the Pluronic polymer
P123 (EO20-PO70-EO20) and a micellar cubic (Im3 h m)
material called SBA-16 with Pluronic F127 (EO106-PO70-
EO106). Several other Pluronics can also be used for
formation of well ordered mesoporous silica.7,12 Recently,
the bicontinuous cubic Ia3 h d structure was formed with
Pluronic polymers.8,9,13
The determination of the mechanism of formation for
surfactant-assisted formation of mesoporous silica is a
challenging task. The initial concentration of the am-
phiphile is within the region of the micellar phase, but
the product achieves a highly ordered mesostructure that
(5) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D. J. Am.
Chem. Soc. 1998, 120, 6024-6036.
(6) Ryoo, R.; Ko, C. H.; Kruk, M.; Antochshuk, V.; Jaroniec, M. J.
Phys. Chem. B 2000, 104, 11465-11471.
(7) Kipkemboi, P.; Fogden, A.; Alfredsson, V.; Flodström, K. Langmuir
2001, 17, 5398-5402.
125, 4402-4403.
(9) Kleitz, F.; Choi, S. H.; Ryoo, R. Chem. Commun. 2003, 2136-
2137.
(10) Matos, J. R.; Kruk, M.; Mercuri, L. P.; Jaroniec, M.; Zhao, L.;
Kamiyama, T.; Terasaki, O.; Pinnavaia, T. J.; Liu, Y. J. Am. Chem. Soc.
2003, 125, 821.
(11) Kleitz, F.; Liu, D.; Anilkumar, G. M.; Park, I.-S.; Solovyov, L.
A.; Shmakov, A. N.; Ryoo, R. J. Phys. Chem. B 2003, 107, 14296.
(12) Flodström, K.; Alfredsson, V. Microporous Mesoporous Mater.
2003, 59, 167-176.
(13) Liu, X.; Tian, B.; Yu, C.; Gao, F.; Xie, S.; Tu, B.; Che, R.; Peng,
L.-M.; Zhao, D. Angew. Chem., Int. Ed. 2002, 41, 3876-3878.

10.1021/la049637c CCC: $27.50 © 2004 American Chemical Society

Published on Web 05/07/2004
4886 Langmuir, Vol. 20, No. 12, 2004
corresponds to higher concentrations in the phase diagram.
For the MCM-type materials, quite a large amount of work
has been performed to get insight into the mechanism of
formation, especially for the hexagonal MCM-41 material,
as recently reviewed by Patarin et al.14 Different experi-
mental techniques probe different length scales, and
therefore, a combination of techniques is often needed in
order to get a complete picture of the nucleation and growth
of the hybrid mesophase. Changes in the micelle environ-
ment during the synthesis can be detected by using
electron paramagnetic resonance (EPR), which was shown
by Zhang et al. on the MCM-41 synthesis.15,16 In situ
synchrotron small-angle X-ray diffraction probes the
appearance of long-range order in the system and has
allowed the kinetics of mesophase formation17,18 to be
followed in detail, also in connection to the kinetics of
solubilization of semipolar and hydrophobic additives.18-23
Fluorescence quenching methods using probes of different
polarities gives detailed information about the micellar
interface region and has been shown to be a very useful
tool for studying these systems.24 Cryogenic transmission
electron microscopy (cryo-TEM) is also an invaluable tool
for the study of the structures present in solution during
the early stages of the process, before long-range order
can be observed.25,26
However, to date, only a few publications focus on the
mechanism of formation of mesoporous silica formed with
nonionic amphiphiles. Boissière et al. used a fluoride-
assisted two-step synthesis for formation of hexagonal
mesoporous silica.27,28 They showed how small silicate
oligomers interact with the poly(ethylene oxide) PEO
chains of the micelles. Most of their work is concentrated
on material formed with Cx(EO)y surfactants. Experiments
were also made with Pluronic P123, but no details about
the kinetics in these syntheses were revealed. A detailed
discussion of interactions in these systems was presented
there. EPR was used by Ruthstein et al.29 to study Pluronic
P123 mediated synthesis of mesoporous silica. The most
significant changes appear within the first 100 min of the
synthesis, and a continuous depletion of water within the
corona-core interface was observed. In a previous work,
we presented results from 1H NMR and TEM that gave
(14) Patarin, J.; Lebeau, B.; Zana, R. Curr. Opin. Colloid Interface
Sci. 2002, 7, 107-115.
(15) Zhang, J.; Luz, Z.; Goldfarb, D. J. Phys. Chem. B 1997, 101,
7087-7094.
(16) Zhang, J.; Carl, P. J.; Zimmermann, H.; Goldfarb, D. J. Phys.
Chem. B 2002, 106, 5382-5389.
(17) Lindén, M.; Schunk, S. A.; Schüth, F. Angew. Chem., Int. Ed.
1998, 37, 821-823.
(18) A° gren, P.; Lindén, M.; Rosenholm, J. B.; Schwarzenbacher, R.;
Kriechbaum, M.; Amenitsch, H.; Laggner, P.; Blanchard, J.; Schüth, F.
J. Phys. Chem. B 1999, 103, 5943-5948.
(19) Lindén, M.; A° gren, P.; Karlsson, S.; Bussian, P.; Amenitsch, H.
Langmuir 2000, 16, 5831-5836.
(20) Kleitz, F.; Blanchard, J.; Zibrowius, B.; Schüth, F.; A° gren, P.;
Lindén, M. Langmuir 2002, 18, 4963-4971.
(21) Lind, A.; Andersson, J.; Karlsson, S.; A° gren, P.; Bussian, P.;
Amenitsch, H.; Lindén, M. Langmuir 2002, 18, 1380-1385.
(22) A° gren, P.; Lindén, M.; Rosenholm, J. B.; Blanchard, J.; Schüth,
F.; Amenitsch, H. Langmuir 2000, 16, 8809-8813.
(23) Tiemann, M.; Goletto, V.; Blum, R.; Babonneau, F.; Amenitsch,
H.; Lindén, M. Langmuir 2002, 18, 10053-10057.
(24) Frasch, J.; Lebeau, B.; Soulard, M.; Patarin, J.; Zana, R.
Langmuir 2000, 16, 9049-9057.
(25) Regev, O. Langmuir 1996, 12, 4940-4944.
(26) Pevzner, S.; Regev, O.; Lind, A.; Lindén, M. J. Am. Chem. Soc.
2003, 125, 652.
(27) Boissière, C.; Larbot, A.; Van der Lee, A.; Kooyman, P. J.; Prouzet,
E. Chem. Mater. 2000, 12, 2902-2913.
(28) Boissière, C.; Larbot, A.; Bourgaux, C.; Prouzet, E.; Bunton, C.
A. Chem. Mater. 2001, 13, 3580-3586.
(29) Ruthstein, S.; Frydman, V.; Kababya, S.; Landau, M.; Goldfarb,
D. J. Phys. Chem. B 2003, 107, 1739-1748.
Flodström et al.
a good overview of the formation process.30 The process
involves silicate adsorption to globular micelles, associa-
tion of the globular micelles into flocs, precipitation of the
flocs, and finally micelle-micelle coalescence that forms
the two-dimensional packing. However, although the
individual steps involved in the formation of the SBA-15
material could be identified, several structural aspects
remained unaddressed, especially with respect to the time
dependency of the packing of the silica-surfactant micelles
in the 2D hexagonal phase, but also at times shorter than
that at which long-range order is observed. We therefore
decided to perform a combined in situ small-angle X-ray
scattering/X-ray diffraction (SAXS/XRD) study on the
same system. Synchrotron SAXS/XRD is a very useful
technique to obtain time-resolved information about
structural changes during the mesophase formation. Using
synchrotron X-ray results as a base for calculations, it is
possible, for example, to reconstruct electron density
distribution maps revealing details of the silica network,
as previously shown by Solovyov et al. for MCM-41.31 In
a very recent article by Zholobenko et al.,32 ex situ
synchrotron X-ray results in connection to the formation
of SBA-15 are presented. However, in contrast to our study,
Zholobenko et al. mainly focus on the stages after the
formation of the hexagonal phase. We are focusing on the
earlier stages of the synthesis, where details about the
aggregation process, the silica polymerization, and the
change in structure can be obtained. Besides, our mea-
surements are performed in a continuous fashion, whereas
Zholobenko et al. performed their measurements on
samples taken out from the reaction beaker at specific
times. In this paper, we present synchrotron SAXS results
from two different SBA-15 syntheses, using two different
silicon alkoxides as the precursor for silica.
Experimental Section
Synthesis. Syntheses of the SBA-15 materials were performed
according to well-known literature procedures.4,5,12 A 0.48 g
portion of Pluronic 123 (EO20PO70EO20, BASF) was dissolved in
18.75 g of 1.6 M HCl, resulting in a 2.5 wt % solution. The
temperature was adjusted to 35 °C. A 1 g portion of tetraethyl
orthosilicate (TEOS) or alternatively 0.73 g of tetramethyl
orthosilicate (TMOS) (the same amount on a molar basis) was
thereafter added during vigorous stirring. The polymers (Plu-
ronics) were obtained as a gift from BASF and used as received
without further purification. TEOS and TMOS (>98%) were
obtained from Aldrich, and HCl was obtained from Merck.
The reactions were carried out in a batch reactor. The liquid
was pumped through a 1 mm diameter quartz capillary, through
which the X-ray beam passed, and then back to the reactor at
a pumping rate of 70 mL/min. This setup has previously been
described in the literature.16 The measurements started directly
after the addition of TEOS or TMOS. Both the reactor and the
capillary were thermostated to a desired temperature of 35 °C.
The synchrotron X-ray measurements were performed at the
Austrian SAXS beamline at the ELETTRA synchrotron (2 GeV
electron storage ring), Trieste, Italy.33A linear position-sensitive
Gabriel detector was used, and the photon energy was 8 keV.
Data Treatment. The measured curves were normalized for
variations of the primary intensity, and the background was
subtracted using 1.6 M HCl in water as the reference.
(30) Flodström, K.; Wennerström, H.; Alfredsson, V. Langmuir 2004,
20, 680.
(31) Solovyov, L. A.; Kirik, S. D.; Shmakov, A. N.; Romannikov, V.
N. Microporous Mesoporous Mater. 2001, 44-45, 17-23.
(32) Zholobenko, V. L.; Khodakov, A. Y.; Durand, D. Microporous
Mesoporous Mater. 2003, 66, 297-302.
(33) Amenitsch, H.; Rappolt, M.; Kriechbaum, M.; Mio, H.; Laggner,
P.; Bernstorff, S. J. Synchrotron Radiat. 1998, 5, 506.
Formation of SBA-15 Mesoporous Silica
Figure 1. Time-resolved SAXS pattern of the (a) P123-TMOS
and (b) P123-TEOS systems. See text for details. The sudden
jump in intensity of the micellar scattering reflection in part
a originates from a 3 min break in the data collection. A data
accumulation time of 1 min per frame was used.
The micellar domain was analyzed by fitting a form factor of
spheres34 (see later):
P(q) ) B
4π
3 [
(Fpol - Fpar)R3par3
sin(qRpar) - qRpar cos(qRpar)
(qRpar)3
+
]
sin(qRT) - qRT cos(qRT) 2
(Fsol - Fpol)R3T3 (1)
(qRT) 3 initially coexists with the (10) reflection, but its intensity
where B is a scaling factor to adjust the relative scale of the
measured intensity, Fpol, the electron density of the corona, Fpar,
the electron density of the core region, Fsol, the electron density
of the solvent, Rpar, the radius of the core, and RT, the total radius
of the micelle. Fsol was calculated for the water with 1.6 M HCl
and kept fixed, and Fpar was calculated for the PO segments and
kept constant during the fitting process. Fpol was calculated for
an EO segment and used as a starting value for the fittings, but
was left free, as well as Rpol and Rpar. To check the exponential
behavior of the scattering curve at very low q values, a correction
was made to eliminate the attenuation of the absorber used to
avoid damage of the detector at very small angles.
Results and Discussion
In Situ SAXS/XRD. The in situ synchrotron SAXS/
XRD pattern measured during the formation of a 2D
hexagonal silica-P123 mesophase is shown in Figure 1.
Here, TMOS (Figure 1a) or TEOS (Figure 1b) was used
as the silica precursor. The diffractograms are very well
resolved and reveal many of the steps involved in the
formation of the final material. We will first briefly discuss
(34) Porod, G. In Small Angle X-ray Scattering; Glatter, O., Kratky,
O., Eds.; Academic Press: New York and London, 1982.
Langmuir, Vol. 20, No. 12, 2004 4887
Figure 2. Time evolution of the d spacing and integrated
intensities of the respective reflections of the 2D hexagonal
phase, as well as the micellar contribution for the P123-TMOS
system.
the overall characteristics of the time-resolved diffrac-
tograms before discussing the key features in more detail.
At short reaction times (up to ∼40 min for TMOS-P123
and ∼50 min for TEOS-P123), only a broad hump is
observed, centered around s ) 0.1 nm-1 (s ) 1/d, d ) 10
nm). This is the scattering originating from the P123
micelles aggregated in flocs by the polymerizing silica.
The intensity of the micellar scattering increases with
time, in both cases, which has been attributed to an
increased electron density contrast due to the adsorption
of silicate species to the PEO groups of the amphiphile,28
and possibly also due to the size increment of the flocs.
However, the solution turned turbid already after a
reaction time of ∼25 min, which is important for the
mechanistic implications discussed later. A more detailed
analysis of this micellar scattering data is given below. A
sudden appearance of a sharp reflection appears at s )
0.080 nm-1 (d ) 12.5 nm) together with two other less
intense reflections at s ) 0.138 nm-1 (d ) 7.25 nm) and
s ) 0.160 nm-1 (d ) 6.25 nm) which can be indexed as the
(10), (11), and (21) reflections, respectively, of a 2D
hexagonal mesophase. Note that the (20) reflection does
not appear at this stage. The micellar scattering reflection
decreases rapidly, and it is no longer present at a reaction
time of 45 min. The full width at half-maximum (fwhm)
of the (10) reflection was 6.5 × 10-3 nm-1 at a reaction
time of 45 min for the P123-TMOS system and 50 min
for the P123-TEOS system, but it reaches a constant
value of 6 × 10-3 nm-1 after ∼55 min in both cases. The
continuous increase in the intensity of the (10) reflection
reflects an increase in the amount of ordered matter. The
intensity of the low-angle scattering also decreases
dramatically at this stage, indicating that the transition
from the micellar arrangement to the 2D hexagonal
structure is fairly rapid, as can be observed in Figure 2.
The (20) reflection appears 65 min into the reaction for
the P123-TMOS system and 85 min into the reaction for
the P123-TEOS system, after which the intensity of this
reflection increases continuously with time. These results
indicate that the silica sources are not fully hydrolyzed
even at this stage of the reaction, since it is well-known
that the rate of hydrolysis of TEOS is slower than that of
TMOS. The intensities of the (11) and (21) reflections, on
the other hand, actually decrease with time, as shown in
Figure 3. This behavior is in strong contrast to what has
been previously observed for a range of MCM-41-type
materials, which are synthesized in the presence of simple
4888 Langmuir, Vol. 20, No. 12, 2004
Table 1. Parameters Obtained by Fitting the Form Factor of the Spherical Micelles to the P123-TMOS System for
Curves Measured at Different Times after the Addition of TMOSa
parameter 15 20
Rpar (Å) 44 44
RT (Å) 68 68
scaling factor 4.20 × 10-5 4.20 × 10-5
Fpol (el/Å3) 0.360 0.361
a Rpar is the radius of the micelle core, RT, the total radius of the micelle, and Fpol, the electron density of the corona region.
Figure 3. Standard fitting of the form factor of a spherical
micelle. The curves correspond to 2.5 wt % P123 in 1.6 M HCl
in water (a) 35 min and (b) 40 min after the addition of TMOS.
Crosses represent the experimental curves, and solid lines are
the calculated form factors. Obtained parameters are presented
in Table 1.
cationic surfactants,18,19,35,36 and recently also for SBA-
15-type silica with a smaller d spacing,32 where the
intensity of the (11) reflection is initially lower than that
of the (20) reflection, but the reverse is observed at longer
reaction times. Furthermore, the structure contracts
continuously with time and has reached an s value of 0.086
nm-1 (11.6 nm) 90 min into the reaction, as shown in
Figure 3. The rate at which the intensity of the (11)
reflection decreases follows the kinetics of the decrease in
the d spacing of the hexagonal phase and that of the
increase in the intensity of the (10) reflection. However,
as it will be shown later, the material is not fully developed
at this stage, and this prevents us from carrying out kinetic
modeling of the reaction according to, for example, the
Avrami equation, which has recently been used to describe
the nucleation of a 2D hexagonal zirconia-CTAB (CTAB,
hexadecyltrimethylammonium bromide) mesophase.36
However, we note that, also in that case, the reaction may
not have reached completeness.
Analysis of the Evolution of the SAXS Micellar
Scattering Region. The first step in the fitting process
(35) Lindén, M.; Blanchard, J.; Schacht, S.; Schunk, S. A.; Schüth,
F. Chem. Mater. 1999, 11, 3002.
(36) Né, F.; Testard, F.; Zemb, Th.; Grillo, I. Langmuir 2003, 19,
8503.
Flodström et al.
time (min)
30 35 40
46 46 46
70 70 70
4.20 × 10-5 4.20 × 10-5 2.81 × 10-5
0.366 0.368 0.368
of the scattering data is to determine the shape, and thus
the form factor, of the micelles. It has been shown that
P123 micelles can be cylindrical in pure water solution.37
Thus, the first attempt to reduce the scattering data was
to fit a form factor of finite cylinders34 to the curves.
However, the data could not be fitted assuming the
presence of finite cylindrical micelles. Furthermore, the
absence of elongated cylindrical micelles was also verified
by the fact that a q-1 behavior was not present at the
beginning of the curves. The absence of elongated micelles
at short reaction times is also corroborated by TEM
analysis of the growing particles.30 Therefore, the system
is certainly composed of spherical micelles, and the form
factor of spheres was successfully fitted according to eq
1. It is worth mentioning that not even a q-4 dependence
was observed at the beginning of the curves, so no Porod
analysis could be performed either.
A standard fit is shown in Figure 3, and the obtained
parameters are summarized in Tables 1 and 2 for P123
with TMOS and TEOS, respectively. Fifteen minutes after
the addition of TEOS, a total micellar radius (RT) of 68 Å
for P123-TMOS and 70 Å for P123-TEOS is reached,
with a micellar core radius (Rpar) close to 45 Å. As already
explained in the Experimental Section, the fitting was
performed having only the micellar radii and Fpol as free
parameters. At first, the scaling factor was also let free,
and once the best results were obtained, as judged from
the minimum of the χ2- value, this parameter was also
fixed. The obtained values are close to the dimensions of
the P123 micelle presented by Nagarajan,37 who found a
radius of 42.1 Å for the core region and 66.5 Å for the
whole micelle. The difference of 2 Å obtained for the
samples containing TMOS, at the initial stages, is inside
an error of 2 Å by considering an average value between
our results and the ones obtained by Nagarajan,37 so we
can consider that the micelle size remained constant.
However, no reasonable results were obtained for the
scattering curve measured after a reaction time of 30 min
if all three parameters (Fpol, Rpar, and RT) were left free.
Therefore, the micellar radii were kept fixed in order to
calculate the electron density in the palisade layer. Fixing
higher values for the micelle dimension was also tried,
but the best results were obtained with the constant size.
Furthermore, it is sensible to state that the size of the
micelles remains constant, since, as shown below, the
micellar radius at 15 min was very close to that observed
initially for the precipitating hexagonal phase. The
increase in the electron density is evidenced by the increase
in the peak intensity and its shift toward higher q values.
At 40 min for TMOS and 45 min for TEOS, the micelles,
with a diameter of 140 Å, precipitate in the hexagonal
phase with a d spacing of 124 Å, equivalent to a cell
parameter of a ) 143 Å. The number of micelles present
in solution decreases concomitantly (the factor for scaling
the measured intensity decreases, as a consequence of
the decrease in the number of scatterers, as demonstrated
in Figure 2) until the micellar peak vanishes completely.
(37) Nagarajan, R. Colloids Surf., B 1999, 16, 55.
Formation of SBA-15 Mesoporous Silica
Table 2. Parameters Obtained by Fitting the Form Factor of the Spherical Micelles to the P123-TEOS System for
Curves Measured at Different Times after the Addition of TEOSa
parameter 15 20
Rpar (Å) 43 44
RT (Å) 70 70
scaling factor 4.59 × 10-5 4.59 × 10-5
Fpol (el/Å3) 0.359 0.361
a Rpar is the radius of the micelle core, RT, the total radius of the micelle, and Fpol, the electron density of the corona region.
Analysis of the Time Dependency of the XRD
Pattern. One means to derive electron density distribu-
tions from a limited number of reflections obtained by
powder X-ray diffraction has been described by Harper et
al.38 We have used this model for calculating the electron
density reconstructions (2D electron density maps) of the
2D hexagonal P123-silica phase to determine time-
dependent parameters such as the radius of the hydro-
phobic core, the thickness of the silica layer, and the degree
of intermicellar condensation. We will shortly review the
most important parts of this model before proceeding to
the analysis.
The 2D hexagonal phase has a centrosymmetric unit
cell, which means that Fe(r) ) Fe(-r), where Fe(r) is the
electron density and r is the vector to the position in the
unit cell, and the electron density can then be expressed
as a Fourier series of cosines by
Fe(r) ) Fav + ∑q Aq cos(q‚r)
where Aq is the Fourier coefficient, q is the Fourier vector
of the hexagonal lattice, and Fav is the average electron
density. X-ray powder diffraction generates a pattern of
concentric rings from an ordered sample. The rings are
reduced to single reflections if a one-dimensional detector
is used, as in our case. Each integrated intensity (I(|q|))
of a reflection corresponds to the Fourier component (Aq)
of the electron density
I(|q|) ) KM(h, k)|Aq|2|q|-2
with
|q| ) 2πs ) 2/x3a(h2 + k2 + hk)
where K is an arbitrary scaling factor giving the difference
from the diffraction intensity in theory (absolute units) to
the relative intensity (counts) in an experiment, M(h, k)
is the multiplicity of the respective reflection, s is the s
spacing, and a is the unit cell parameter. The factor |q|-2
stands for the Lorentz factor for the applied scattering
geometry.39 For the 2D hexagonal phase, the multiplicity
(M(h, k)) is 6 for h, k ) 0 or h ) k, and the other
multiplicities are equal to 12; in particular, the (10), (11),
and (20) reflections all have a multiplicity of 6, while the
(21) reflection has a multiplicity of 12. This approach has
also been used by Impéror-Clerc et al.40 to study the
presence of micropores in calcined SBA-15 materials.
However, under our analysis, the value of the arbitrary
constant K is kept the same throughout the process, and
no further normalization as in ref 40 is performed. The
remaining step is to assign the correct sign, + or -, for
(38) Harper, P. E.; Mannock, D. A.; Lewis, R. N. A. H.; McElhaney,
R. N.; Gruner, S. M. Biophys. J. 2001, 81, 2693.
(39) Warren, B. E. X-ray diffraction; Addison-Wesley: Reading, MA,
1969.
(40) Impéror-Clerc, M.; Davidson, P.; Davidson, A. J. Am. Chem.
Soc. 2000, 122, 11925.
Langmuir, Vol. 20, No. 12, 2004 4889
time (min)
30 35 40 45
44 44 44 44
70 70 70 70
4.59 × 10-5 4.59 × 10-5 4.59 × 10-5 4.41 × 10-5
0.364 0.366 0.368 0.368
(2)
Figure 4. Two-dimensional electron density maps (4 × 4 unit
cells) created using the intensities and positions of the (10),
(11), (20), and (21) reflections at the following specific reaction
times as input data for the P123-TMOS system: (a) 40 min,
(b) 71 min, and (c) 92 min. See text for details. White areas
(3) correspond to areas of high electron density and dark areas to
areas of low electron density. The straight line in part a indicates
the range for the one-dimensional electron density plot shown
in Figure 5.
(4) the different Fourier coefficients, that is, the correct phase.
We are limited to four terms in the series of eq 2, since
we observe four reflections in the diffractogram. In our
case, the electron density in the PEO-silica layers is
definitely higher than the electron density in the poly-
(propylene oxide) (PPO) core, which is why it is evident
that the sign of the first Aq term is negative, provided that
the origin is set to the center of the cross section of the
cylinder. By iteration, - - + + was found to be the correct
phasing for the initial stages of the developing 2D
hexagonal silica-P123 phase. Further support for the
correctness of the phase assignment is given by the fact
that this phasing gives a physically reasonable description
of the electron density variations for all reaction times.
This is mandatory, since no reflections (temporarily)
disappear during the time window studied, which is a
prerequisite for a phase change.
The calculated electron density distribution plots for 4
× 4 unit cells are shown in Figure 4. Note that the images
are not scaled to real space distance. Only electron density
maps for the P123-TMOS system are shown, but similar
electron density maps were observed for P123-TEOS,
although here the kinetics of the reaction was slower, as
already stated. In Figure 4a, corresponding to the time
4890 Langmuir, Vol. 20, No. 12, 2004
Figure 5. One-dimensional real space electron density maps
of the x-direction of the density maps shown in Figure 4. The
zero position corresponds to the middle of the hydrophobic core.
The curves correspond to a reaction time of (a) 40 min (thin
line), (b) 60 min (double dash), (c) 71 min (dotted line), and (d)
92 min (thick line).
immediately after the formation of the hexagonal me-
sophase, the cylinders are weakly, if at all, cross-linked;
that is, the degree of intermicellar silicate condensation
is very low. The degree of intermicellar condensation
increases gradually with time, leading to a continuous
increase in the electron density along the 3-fold axis and
to an overall much more homogeneous electron density
distribution in the silica walls, as can be expected (see
Figure 4). The increase in the electron density of the
innermost part of the hydrophobic core with time is an
artifact related to the limited number of terms in the
Fourier series. The silica is, at this stage, mainly located
in the palisade layer of the P123 cylinders. Furthermore,
the micellar cross section has a hexagonal shape. This
has previously been observed by TEM for calcined high-
quality samples of MCM-41.41 The occurrence of such a
geometry may, however, be an artifact due to the limited
number of terms in the Fourier series.
A cut in the x-direction through the 2D electron density
plots is shown in Figure 5. The electron density is plotted
as a function of real space distances with the origin in the
center of the micellar cross section. A slight increase in
the diameter of the hydrophobic PPO core with time is
observed. However, the thickness of the high electron
density region, that is, the silica wall, is dramatically
decreasing. This indicates that the main reason for the
observed decrease in the d spacing of the hexagonal phase
with time is connected to ongoing silica condensation
reactions. The artifact previously seen in the electron
density maps (Figure 4) is also obvious in Figure 5. With
time, the electron density of the artifact increases, but it
is clear that the contrast caused by this erroneous density
is small compared to the overall electron density contrast
between the inner hydrophobic core and the silica-PEO
corona. Thus, this density will not interfere in any
significant way with the interpretation of the electron
density maps.
The observed phasing, - - + +, is different from the
phasing observed for the 2D hexagonal structure of
siliceous MCM-41 synthesized in the presence of cetylt-
rimethylammonium bromide, which was found to be
- + + -.42 However, as already discussed, the intensity
(41) Alfredsson, V.; Keung, M.; Monnier, A.; Stucky, G. D.; Unger,
K. K.; Schüth, F. J. Chem. Soc., Chem. Commun. 1994, 921.
(42) Beurroies, I.; A° gren, P.; Rosenholm, J. B.; Amenitsch, H.; Denoyel,
R.; Lindén, M. Manuscript in preparation.
Flodström et al.
of the (11) and (21) reflections continuously decrease with
time after the initial stages of the formation of the P123-
silica structure, which can seem counterintuitive, as the
degree of silica condensation increases with time. These
two reflections correspond to the second and fourth term
in the Fourier series, and in order for a phase change to
occur, the intensities of the corresponding reflections have
to go through zero. Therefore, an obvious reason for the
observed time dependence of the reflections is that the
phase of the P123-silica structure will eventually shift
from - - + + to - + + -, that is, the same phasing as
was observed for MCM-41.42 Physically, the reason for
this phase shift is that the ratio (p) between the corona
thickness and the core radius will decrease with time,
due to the increasing degree of silica condensation. It is
well established that the ratio between the thickness of
the silica layer and the inner core radius is lower for MCM-
41 than for SBA-15,5 and our observation that the phasing
of SBA-15 is initially different from that of MCM-41 is
therefore logical. Phase changes have also been demon-
strated both theoretically and experimentally for calcined
samples of unmodified and modified MCM-41 and SBA-
15 samples as a function of p.43,44
However, the phasing, associated with the wall thick-
ness to hydrophobic core radius ratio, may vary depending
on the synthesis conditions. The formation of SBA-15
under different synthesis conditions was recently studied
by synchrotron XRD.32 The intensity of the (11) reflection
in this case was found to increase continuously with time.
This can be explained by the fact that the synthesis
conditions are slightly different, causing the phase of the
(11) reflection to be +, in contrast to our findings. Another
explanation is that the phase shift occurs between the
first (18 min after TEOS addition) and second measure-
ments (72 min after TEOS addition) and, hence, cannot
be detected, as the data were not collected in situ.
The intensity ratio between the (11) and (20) reflections
increases with time during the formation of silica,18,19
titania,35 and zirconia,35,36 2D hexagonal mesophases
formed with cationic surfactants. This effect has been
attributed to the increasing degree of condensation in the
inorganic corona, leading to increased contrast in the
electron density between the core and the corona. As shown
here, we observe a decrease in the intensity ratio between
the (11) and (20) reflections for SBA-15. We ascribe this
to the same physical effect. This highlights the importance
of the phasing of the observed intensities of the reflections.
The reflections of the 2D hexagonal mesophases formed
with cationic surfactants, mentioned above, would nor-
mally be assigned to the - + + - phasing. The increase
in the (11)/(20) intensity ratio, as previously suggested,
would then reflect a densification of the inorganic portion
due to condensation reactions, while the opposite is true
for the - - + + phase. Therefore, this ratio cannot be
taken as a general measure for the time-dependent
densification of the inorganic portion unless the correct
phasing is known.
Mechanistic Implications. Recently, there has been
much development in the understanding of the mechanism
of formation of mesoscopically ordered, hybrid surfactant-
inorganic materials due to an increasing number of in
situ studies. Our results can be explained by considering
a series of chemical events leading to the formation of the
2D hexagonal phase: (a) hydrolysis of the silicon alkoxide,
(b) adsorption of hydrolyzed silicate species to the PEO
(43) Feuston, B. P.; Higgins, J. B. J. Phys. Chem. 1994, 98, 4459.
(44) Sauer, J.; Marlow, F.; Schüth, F. Phys. Chem. Chem. Phys. 2001,
3, 5579.
Formation of SBA-15 Mesoporous Silica
units of spherical P123 surfactant micelles, (c) clustering
of poorly condensed silica-surfactant hybrid micelles into
flocs, (d) elongation of the surfactant micelles and parallel
formation of domains exhibiting hexagonal order, and (e)
growth of the domains. This interpretation is well in
agreement with the outcome of recent studies25,36,45-48 and
especially with the recent 1H NMR and TEM study30 of
the same P123-TMOS system. Since the formation
mechanism of this system is discussed in detail in ref 30,
we will focus on the new findings of the present study.
Two possibilities were given in ref 30 in order to explain
the formation of the hexagonal phase, a direct precipitation
of a hexagonal phase in equilibrium with a dilute solution
or an initial liquid-liquid phase separation into a
concentrated and a dilute phase of (nearly) spherical
aggregates, which, upon a further increase in the attrac-
tion, changes to an equilibrium dilute solution-hexagonal
phase. Our results are better explained by the second
suggestion, in agreement with the findings in ref 30, since
(a) the solution turns turbid well before any sign of the
hexagonal structure appears in the diffractograms and
(b) the spherical silica-P123 micelles coexist initially with
the 2D hexagonal mesophase. Furthermore, we observe
virtually the same value for the diameter of spherical
P123-silica micelles and the distance between the
cylinders in the hexagonal phase at early stages, which
indicates that the cylinders are fairly close-packed already
at this stage. In addition, the fact that the fwhm of all the
reflections reaches a virtually constant value soon after
the formation of the hexagonal phase, while there is a
continuous increase in the intensity of the (10) reflection,
indicates that fairly large, ordered domains are formed.
This conclusion is also supported by TEM images taken
(45) Chan, H. B. S.; Budd, P. M.; Naylor, T. deV. J. Mater. Chem.
2001, 11, 951.
(46) Yao, N.; Ku, A. Y.; Nakagawa, N.; Lee, T.; Saville, D. A.; Aksay,
I. A.; Chem. Mater. 2000, 12, 1536.
(47) Sadasivan, S.; Fowler, C. E.; Khushalani, D.; Mann, S. Angew.
Chem., Int. Ed. 2002, 41, 2151.
(48) Brennan, T.; Hughes, A. V.; Roser, S. J.; Mann, S.; Edler, K. J.
Langmuir 2002, 18, 9838.
Langmuir, Vol. 20, No. 12, 2004 4891
for the final material,30 that exhibits highly aligned 2D
hexagonal domains several 100 nm in size, that is, infinite
on the X-ray diffraction length scale. Our data further
suggest that the silica source may be partially unhydro-
lyzed long after formation of the hexagonal structure, since
the kinetics of the mesophase evolution is notably slower
when TEOS is used as the silica precursor as compared
to TMOS. However, it is well-known that, while the
hydrolysis of the first alkoxide group is fast under acidic
conditions, the rate of hydrolysis decreases successively
for the remaining alkoxide groups.49 The continuous
decrease in the d spacing of the hexagonal mesophase
after its formation can be directly ascribed to the increase
in the density of the silica walls due to increasing intra-
and intermicellar condensation reactions. The d spacing
of the hexagonal mesophase decreases continuously, which
is obvious in Figure 5. This is the result of two opposing
effects: the increase in the hydrophobic core radius and
the decrease in the silica wall thickness. Both effects can
be explained by the silica condensation. For the silica wall,
this naturally leads to a decreased thickness, due to the
densification of the silica network. On the other hand, the
silica condensation together with the subsequent dehy-
dration of parts of the EO palisade layer lead to an
enlargement of the core. However, the first effect domi-
nates over the second, leading to the observed decrease
in d spacing.
Acknowledgment. The financial support from the EU
projects (Elettra) HPRI-CT-1999-00033 and HPRN-CT-
1999-00025 (Nucleus) (C.V.T.) is gratefully acknowledged,
as well as the financial support from the SSF program
Colloid and Interface Technology (K.F.), the Swedish
Research Council (V.A.), and the Humboldt Foundation
(M.L.).
LA049637C
(49) Brinker, C. J.; Scherer, G. W. Sol-Gel Science; Academic Press:
San Diego, CA, 1990.