Investigation on the drying induced phase transformation of mesoporous silica; A comprehensive understanding toward mesophase determination

Article abstract of DOI:10.1021/ja047785j

Phase transformation of mesoporous silica during the drying process is investigated. Assynthesized hexagonal p6mm obtained under the conditions used in this study is transformed to cubic Ia3d as drying proceeds, even at room temperature. Prolonged synthes

Full text of DOI:10.1021/ja047785j

Published on Web 08/17/2004
Investigation on the Drying Induced Phase Transformation of
Mesoporous Silica; A Comprehensive Understanding toward
Mesophase Determination
Masaru Ogura, Hayato Miyoshi, Sajo P. Naik, and Tatsuya Okubo*,†,‡
†
†
†,‡
Contribution from the Department of Chemical System Engineering, The UniVersity of Tokyo
and PRESTO, JST, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656 Japan
Received April 17, 2004; E-mail: okubo@chemsys.t.u-tokyo.ac.jp
Abstract: Phase transformation of mesoporous silica during the drying process is investigated. As-
synthesized hexagonal p6mm obtained under the conditions used in this study is transformed to cubic
Ia3d as drying proceeds, even at room temperature. Prolonged synthesis results in the formation of a
well-ordered hexagonal mesophase, with almost no phase transformation. Drying at a higher temperature
promotes the phase transformation of not only hexagonal to cubic, but also cubic to lamellar mesophases.
Release of water is detected during drying, which is followed by the phase transformation of the mesophases.
The phase transformations observed here proceed against the direction estimated on the basis of the
state-of-the-art understanding. Here, considering the degree of silicate condensation and the amount of
residual water in the as-synthesized mesoporous silica, a comprehensive explanation of mesophase
determination is proposed including thermodynamic and kinetic aspects to account for the results observed
here and those in the literature.
Introduction
separation membranes, as has already been demonstrated during
the past decade, but also for hosts or molds for functional
molecules and nanoparticles.
Mesoporous silica is a series of materials that contain well-
ordered and uniform-sized pores. It is generally synthesized in
the presence of cationic or block-copolymer surfactant micelles
in addition to a reactive source of silica. In 1990, Yanagisawa
Various earlier works have elucidated the formation mech-
anism of mesoporous silica and its structure can now be
controlled to some extent. As for the synthesis mechanism of
mesoporous silica, an LCT (liquid-crystal templating) mecha-
nism has been proposed, since its structure is similar to the
1
et al. discovered a mesoporous material with a hexagonal
mesophase; the material was synthesized using a long-chained,
trimethyl quaternary ammonium cation and kanemite, a layered
silicate compound. Two years later, Mobil researchers reported
the M41s family of silicate mesoporous molecular sieves with
3
lyotropic liquid-crystal phase of the surfactant. But the fact
that very low concentrations of surfactant also produce meso-
porous silica seems to support the mechanism of cooperative
organization of inorganic silica species and surfactant micelles.4
2
an exceptionally large and uniform pore-structure. Mesoporous
silica has various types of mesophases because of the different
interactions between silicate species and surfactant micelles as
a template, and because of the shapes of the micelles. By
changing the size and shape of the surfactant micelles, the
structure and pore size of mesoporous silica can be individually
controlled. Furthermore, the addition of a cosurfactant such as
trimethylbenzene (TMB) can expand the pore size of the
mesoporous silica. Due to the easily tunable pore size and
ordered structure, mesoporous silica has attracted considerable
interests toward many kinds of applications. The presence of
ordered pores in the mesoporous range makes mesoporous silica
a promising candidate not only for catalysts, adsorbents, and
Packing Parameters. As the parameter that determines the
structure of mesoporous silica, Huo et al. have proposed a
5
packing parameter. The packing parameter “g” is defined as g
) V/(a l), where V is the total volume of surfactant chains plus
0
any cosolvent organic molecule between chains, a0 is the
effective headgroup area at the micelle surface, and l is the
kinetic surfactant tail length or the curvature elastic energy. This
g factor originates from geometric constraints in the self-
6
assembly of hydrocarbon amphiphiles. The curvature of the
silicate wall is at last determined to optimize the charge balance
between the silicate and surfactant micelles. Accordingly, the
size, shape, and charge (pH of the solution) of a surfactant used
†
Department of Chemical System Engineering, The University of Tokyo.
PRESTO, JST.
‡
(
1) Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn.
(3) Raman, N. K.; Anderson, M. T.; Brinker, C. J. Chem. Mater. 1996, 8,
1682.
(4) Chen, C.-Y.; Burkett, S. L.; Li, H.-X.; Davis, M. E. Microporous Mater.
1993, 2, 27.
(5) Huo, Q.; Margolese, D. I.; Stucky, G. D. Chem. Mater. 1996, 8, 1147.
(6) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. J. Chem. Soc., Faraday
Trans. II 1976, 72, 1525.
1
990, 63, 988.
(
2) (a) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J.
S. Nature 1992, 359, 710, (b) Beck, J. S.; Vartuli, J. C.; Roth, W. J.;
Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T-W.; Olson,
D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L.
J. Am. Chem. Soc. 1992, 114, 18 034.
10.1021/ja047785j CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 10937-10944
9
10937

A R T I C L E S
Ogura et al.
for synthesis can be the phase-determining parameters from a
thermodynamic point of view.
tions, a transformation to hexagonal p6mm was observed. This
transformation has also been rationalized in terms of the packing
parameter. TMB is known to cause swelling at the hydrophobic
regions of liquid crystal structures. Because the TMB molecules
associate with the hydrophobic surfactant tail, the g value is
reduced due to the enhancement of the surfactant volume
without increasing the headgroup area or the tail length. The
swelled geometry results in composites with a lamellar structure,
indicating a larger g value. When the composite is suspended
in water for hydrothermal treatment, a large concentration
gradient is formed initially because a significant amount of TMB
exists in the hydrophobic region of the composite, and there is
essentially no TMB in the aqueous phase. The solubility of TMB
in the aqueous phase increases rapidly with increasing temper-
ature, and the hydrothermal treatment then leads to the release
of TMB from the composite into the aqueous phase. This change
lowers the value of g, resulting in the transformation to the
hexagonal p6mm mesophase.
Charge-Density Matching. Along with the g parameter, a
charge-density matching concept has also been proposed to
7
account for the changes in mesophases. When silicate is highly
charged depending on pH of the solution, a high density of
counter-charged surfactant is required to balance the charges.
This high density is most effectively accomplished by a silicate
structure having a low curvature at the silicate/surfactant
interface. Thus, the surfactant headgroup area determines the
overall composite structure. Because silicate is dehydrated and
condensed to decrease the negative or positive charge density
of the silicate network, this concept involves the effects of
silicate condensation on the mesoporous silica phase. To
maintain the charge-density balance at the interface, the sur-
factant must pack to form a high surface curvature by increasing
the effective headgroup area, which results in a larger a0 value,
and, therefore, a smaller g value. Then, mesoporous silica having
a higher surface curvature becomes more favorable as silicate
condensation proceeds.
The same transformation from lamellar to hexagonal p6mm
has been reported in response to acidification of a strongly basic
1
0
Phase Transformation. With regard to the g parameter and
medium of silicate with a surfactant. The transformation has
been proposed to occur due to the reactivity of the silicate
species, which increases with the increasing acidity of the
reactant gel to further aggregate the uncondensed silicate.
the concept of charge-density matching, most of the phase
transformation observed previously8
sively understood.
-15
has been comprehen-
Hexagonal p6mm to Cubic Ia3d. Landry et al.11 have
reported that hexagonal p6mm is changed into cubic Ia3d, the
direction of which is opposite to the one expected by the packing
parameters, by heating it to 150 °C. Hexagonal p6mm compos-
ites were synthesized using cethyltrimethylammonium cations
Hexagonal p6mm to Lamellar. In situ XRD measurements
have clarified that the hexagonal p6mm structure, which was
synthesized using a 20-carbon surfactant at room temperature,
transforms into a lamellar structure when heated in water. This
phenomenon has been explained with respect to the packing
8
+
parameter. At room temperature, the g value is presumed to be
(CTA ), and an in situ XRD measurement was then performed.
1
/2 because of hexagonal p6mm periodicity. As the material is
Some of the peaks that appear during the phase transformation
can be indexed to a lamellar phase, whereas most of the peaks
correspond to those of a cubic Ia3d structure. This transforma-
tion has been proposed to proceed via a cylinder-merging or
cylinder-branching mechanism.
heated, however, the conformational disorder of the surfactant
tail is enhanced, increasing the effective molecular volume and
a corresponding increase in the g value. Phase transformation
then proceeds from hexagonal to lamellar.
Hexagonal p6mm to Cubic Pm3n. Che et al.12 have reported
that cubic Pm3n mesoporous materials are synthesized by
transformation of the once-formed hexagonal mesophase. The
rate of transformation depends on the structure and amount of
additive TMB isomers. Furthermore, using high-resolution TEM,
they have found that the phase transformation is epitaxially
promoted, meaning that the silica mesophase is topologically
restructured along the cylinder axis of the hexagonal and that
the newly generated cubic mesophase is grown on the hexagonal
phase. To account for this transformation, they considered the
charge-density matching mechanism. Polymerization of the
silicate species occurs during the synthesis, causing the positive
charge density of the silicate network to decrease. To maintain
the charge-density matching at the interface, the surfactant packs
to form micelles with a high surface curvature. Therefore,
transformation to the Pm3n cubic phase is considered to occur.
Cubic Pm3n to Hexagonal p6mm. Liu et al.13 have observed
that a phase transformation occurs from cubic Pm3n to
hexagonal p6mm, in the direction of which the packing concept
could not be explained, during drying the precipitates. They
explained this phase transformation based on the nature of the
surfactant micelles. As the solvent evaporates, surfactant is
concentrated, which then favors the formation of rod-shaped
instead of spherical micelles. Concomitantly, the mesoporous
silica is reorganized around the rod-shape micelle, forming a
Very recently, the same mesophase transformation was
observed using in situ scanning microcalorimetry to determine
the thermal energetic process during synthesis. To hydrother-
mally restructure the inorganic/organic composites, high tem-
perature where a melting transition of the surfactant tails and
reorganization of the surfactant occur is needed as an external
force from the system of silicate/surfactant solution.
9
Lamellar to Hexagonal p6mm. Phase transformation from
8
lamellar to hexagonal p6mm has also been reported. When a
lamellar phase synthesized using cethyltrimethylammonium (16-
carbon quaternary ammonium surfactant) cations as well as
TMB as a cosurfactant was heated under hydrothermal condi-
(
7) Monnier, A.; Schuth, F.; Huo, Q.; Kumar, D.; Margolese, D.; Maxwell, R.
S.; Stucky, G. D.; Krishnamurthy, M.; Petroff, P.; Firouzi, A.; Janicke,
M.; Chmelka, B. F. Science 1993, 261, 1299.
(
8) Tolbert, S. H.; Landry, C. C.; Stucky, G. D.; Chmelka, B. F.; Norby, P.;
Hanson, J. C.; Monnier, A. Chem. Mater. 2001, 13, 2247.
(
9) Gross, A. F.; Yang, S.; Navrotsky, A.; Tolbert, S. H. J. Phys. Chem. B
2
003, 107, 2709.
(
(
(
10) Luan, Z.; He, H.; Zhou, W.; Klinowski, J. J. Chem. Soc., Faraday Trans.
1
998, 94, 979.
11) Landry, C. C.; Tolbert, S. H.; Gallis, K. W.; Monnier, A.; Stucky, G. D.;
Norby, P.; Hanson, J. Chem. Mater. 2001, 13, 1600.
12) Che, S.; Kamiya, S.; Terasaki, O.; Tatsumi, T. J. Am. Chem. Soc. 2001,
1
23, 12 089.
(
13) Liu, M.-C.; Sheu, H.-S.; Cheng, S. Chem. Commun. 2002, 2854.
14) Grosso, D.; Babonneau, F.; Soler-Illia, G. J. de A. A.; Albouy, P.-A.;
Amenitsch, H. Chem. Commun. 2002, 748.
(
(
15) Cagnol, F.; Grosso, D.; Soler-Illia, G. J. de A. A.; Crepaldi, E. L.;
Babonneau, F.; Amenitsch, H.; Sanchez, C. J. Mater. Chem. 2003, 13, 61.
10938 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Mesophase Determination
A R T I C L E S
hexagonal mesophase. The cubic Pm3n is considered to be a
thermodynamic product of silica around the spherical micelles.
They indicated that silicate condensation also affects this
transformation.
the mixture was being stirred. The final hydrogel composition was
2 3 2
1.0SiO : 0.13CTEABr: 0.96HNO : 125H O. The mixture was stirred
for 10 min in the ice bath and then allowed to react at 0 °C under
static conditions for a desired period of 3, 6, 12, or 24 h. The resultant
white precipitates were then centrifuged to collect the as-synthesized
mesoporous silica. The obtained samples were dried in air at room
temperature on a glass plate and also under controlled conditions.
Characterization. Powder X-ray diffraction (XRD) patterns were
recorded with an M03X-HF (MAC Science) using CuKR radiation (40
kV, 20 mA) at a rate of 0.15 deg/min over a range of 1.5-6.0 deg (2
theta). The samples were prepared on a glass plate.
Recently, in situ time-resolved small-angle X-ray scattering
was performed during the synthesis of mesoporous silica thin
14
film in an aqueous/alcohol solvent. Lamellar and hexagonal
mesophases were found as intermediates to the final cubic Pm3n
mesophase. At the air interface, lamellar mesophase was formed,
with spherical micelles then being formed underneath the
lamellar phase to organize the 3-D hexagonal mesophase. The
hexagonal phase was subsequently rearranged into discontinuous
cubic mesophase. This process progressed toward the substrate
interface, with the whole process occurring within the water-
rich period range. In their latest report,15 humidity maintained
during the drying process was hypothesized as being an
important factor for the mesophase transformation. The water
content is a critical parameter, and poorly ordered, 2-D
hexagonal, and 3-D cubic final structures are obtained depending
on the humidity, in agreement with the g laws of mesophases.
Thus far, phase transformation of mesoporous silica has been
explained based only on the concept of packing parameters. As
described, however, some phenomena cannot be explained by
these conventional rules alone; therefore, there is an urgent need
for a comprehensive explanation of the phase transformation
of mesoporous silica. There have been much fewer investiga-
tions of the effects of the synthesis process on the formation of
mesoporous silica. Some recent reports as shown above,
however, have indicated that elements of the synthesis process,
including the synthesis period, drying, stirring, and so on,
influence the structure of mesoporous silica. It is therefore
important to understand the effects of such process conditions
on mesophase evolution. In the present study, we focused on
the drying procedure of as-synthesized mesoporous silica, as
drying is one of the most important factors in determining the
structure of mesoporous silica. On the basis of our investigation,
we provide a comprehensive overview of silicate condensation
and drying, and discuss the influence of drying procedures on
the structure of mesoporous silica.
A simultaneous powder XRD (RINT-ULTIMA+, Rigaku)-dif-
ferential scanning calorimetry (DSC, DSC8230, Rigaku) was carried
out with using CuKR (40 kV, 50 mA) radiation at a scanning rate of
1
.0 deg/min over a range of 1.8-4.8 deg (2 theta). The heating rate of
the DSC measurement was 1.5 K/min, and the measurement was
performed in air. Thermogravimetry (TG)-differential thermal analysis
(
(
DTA) and mass spectroscopy (MS) were carried out using a TG8120
Rigaku) equipped with M-QA200TS (ANELVA). Samples were put
on an alumina holder. The heating rate was 5 K/min, and the
measurement was performed in 10%O /He.
2
Results and Discussion
Drying at Room Temperature. The structures of meso-
porous silica have previously been controlled primarily by
changing the initial conditions of synthesis such as the surfac-
tants, pH, or solution compositions. Still, little attention has been
paid to the effects of the drying of obtained precipitates on the
mesophase of silica. In the present study, we observed the
influence of drying on as-synthesized mesophase. XRD patterns
of the obtained samples after drying are summarized in Figure
1
. The XRD measurements were carried out after centrifugation,
which is described as “as-synthesized”, and after drying for
, 2, 3, 6, and 24 h at room temperature. The diffraction of the
1
as-synthesized sample in Figure 1a was assigned to the hexa-
gonal p6mm mesophase. The pattern was obviously changed,
as drying proceeded even at room temperature for 1 h, indicating
that the mesophase of silica is transformed during drying. The
diffraction patterns after drying for 1 h shown in Figure 1a
correspond to the cubic Ia3d mesophase. The main diffraction
peak of the 2 h-drying sample sharpened slightly compared to
the 1 h-drying sample. After this period, the XRD patterns were
hardly changed. The same phenomenon was also observed on
the samples synthesized for 6 h, as shown in Figure 1b.
In Figure 1c, the diffraction pattern for the as-synthesized
sample indicates a highly ordered hexagonal p6mm mesophase,
and those after drying indicate cubic Ia3d. However, the peak
intensity of the cubic (220) plane was very small although the
intensity of the main peak (211) remained high. The second
peak became more intense with prolonged drying, indicating
that a cubic Ia3d structure was gradually evolving. Such a slow
transformation can be more clearly observed in Figure 1d; in
this case, the hexagonal p6mm mesophase was maintained for
a longer period. It can also be noted that all diffraction peaks
in Figure 1d appear broader than those of the other samples,
indicating that the sample has a disordered structure. These
results obviously suggest that the as-synthesized mesophase is
changed during drying, from the hexagonal p6mm mesophase
transforming to cubic Ia3d. This phase transformation leads to
a larger g value, which cannot be rationalized by the rule of
charge-density matching.
Experimental Section
Materials. The following materials were used for the synthesis:
tetraethyl orthosilicate (TEOS, >96.0%, Tokyo Kasei Kogyo (TCI)),
1
- bromohexadecane (>96%, TCI), triethylamine (>99.0%, TCI),
acetone, and nitric acid (35%, Wako). The chemicals mentioned here
were used without further purification.
Synthesis of Mesoporous Silica. Cetyltriethylammonium bromide
(CTEABr) was synthesized by the reaction of cetylbromide (C16
33
H -
Br) with triethylamine ((C N); C16 33Br was added to (C H ) N
2
H )
5 3
H
2 5 3
dissolved in acetone. The resultant mixture was stirred for 1 week in
an Erlenmeyer flask. The product was separated by filtration and dried
1
3
under vacuum for several hours at 60 °C. C MAS NMR results
+
suggested that CTEA was successfully obtained.
Mesoporous silica was synthesized using CTEABr as the surfactant
micelles and TEOS as a silica source in the presence of HNO
3
and
water, in accordance with the literature. The synthesis was performed
as follows. The surfactant, distilled water, and HNO were mixed to
obtain a homogeneous solution, which was cooled to 0 °C prior to the
addition of TEOS. Typically, 0.25 g of CTEABr, 0.83 g of HNO , and
0.3 g of distilled water were mixed and cooled for at least 1 h in an
12
3
3
1
ice bath. After this cooling period, 1.0 g of TEOS precooled to 0 °C
for at least 1 h in an ice bath was added to the reaction mixture, whereas
Effects of Silica Condensation. In all cases, the samples
dried for more than 3 h included the cubic Ia3d mesophase. As
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10939

A R T I C L E S
Ogura et al.
Figure 1. Mesophase transformation during drying of the samples synthesized at 0 °C for 3 h(a), 6 h(b), 12 h(c), and 24 h(d). The samples were dried at
room temperature for each period noted in the figure.
discussed before, hexagonal mesophases synthesized for 3 and
6
h seem to transform smoothly into the cubic mesophase. On
the other hand, the mesophase transformation in the sample
synthesized for 12 h occurred over a longer period. These results
indicate that the mesophase transformation strongly depends on
the flexibility of the silicate network, as has been observed in
the preparation of mesoporous silica film by a rapid evaporation
method.16 Considering the results for the mesoporous silica
synthesized for 24 h, which consisted of a less-ordered cubic
Ia3d mesophase, the silica wall appears to be too rigid to
transform into another mesophase.
Drying at High Temperature. In the previous reports,
mesophase transformation, or mesophase determination, has
been considered thermodynamically by using the g parameter;
but it is sometimes kinetically hindered by the limited mobility
1
6
of surfactant micelles in the polymerized silicate network. In
such a case, a certain activation energy from the exterior to the
interior system is required to overcome the kinetic barrier.
Mesoporous silica sometimes changes its structure on drying
at a higher temperature, showing that thermal energy promotes
the phase transformation according to thermodynamic rules.
Therefore, temperature variation in the procedures of synthesis
and drying might lead to the different route of transformation
from that observed at room temperature.
Figure 2. Effect of drying temperature on mesophase determination. The
samples were synthesized at 0 °C for each period and then dried at 60
°
C(a) or 100 °C(b) for 24 h.
peaks that are consistent with that of the cubic Ia3d mesophase.
The 24 h-synthesized sample also has sharp peaks, indicating a
highly ordered structure of cubic Ia3d.
In contrast, the diffraction patterns of the samples dried at
We investigated the effects of drying at high temperature,
and the XRD patterns of the samples are illustrated in Figure
100 °C are totally different; the diffraction of the 3 h-synthesized
sample shown in Figure 2b has one broad peak, indicating that
the sample has an almost entirely disordered structure due to
the silicate network being too weak to endure high temperatures.
The pattern for the 6 h-synthesized sample has two peaks at
2
6
, where Figure 2a shows the patterns of the samples dried at
0 °C and Figure 2b shows those of samples dried at 100 °C
just after centrifugation. These two figures also show the effects
of the synthesis period on the mesophase of silica. The
diffraction pattern of the 3 h-synthesized sample has two broad
peaks. Although it is difficult to determine its structure from
these two broad peaks, they are similar to that of cubic Ia3d
mesoporous silica. Accordingly, it can be inferred that this XRD
pattern is derived from a partially disordered cubic Ia3d
structure. The diffraction pattern of the 6 h-synthesized sample
exhibits almost the same character, but a more disordered cubic
structure. In contrast, the 12 h-synthesized sample shows sharp
2
.5 and 4.8 deg with a 1:2 lattice, which is characteristic of a
lamellar mesophase. The sample synthesized for 6 h is expected
to consist of a more rigid silicate network than that synthesized
for 3 h. This result clarifies that, at high temperatures,
mesoporous silica tends to transform into a lamellar mesophase,
which has the largest g value. The appropriate flexibility of the
silicate network allows mesoporous silica to transform from
hexagonal p6mm to lamellar mesophase, presumably through
an intermediate cubic Ia3d mesophase.
Drying at a high temperature of 100 °C after prolonged
synthesis for more than 12 h resulted in the formation of two
(
16) Ogura, M.; Suzuki, Y.; Miyoshi, H.; Naik, S. P.; Okubo, T. Chem. Lett.
2
004, 33, 734.
10940 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Mesophase Determination
A R T I C L E S
Table 1. Mesophase Determined by Synthesis and Drying Conditions
drying
period/h
drying
temperature
synthesis
period/h
3
6
12
24
hexagonal
0
1
2
3
6
as-synthesized
hexagonal
cubic
cubic
cubic
cubic
hexagonal
cubic
cubic
cubic
cubic
hexagonal
hexagonal
cubic
cubic
cubic
room temperature
room temperature
room temperature
room temperature
room temperature
room temperature
hexagonal
hexagonal + cubic
hexagonal + cubic
hexagonal + cubic
hexagonal + cubic
2
2
4
4
cubic
cubic
cubic
+
100 °C
lamellar
lamellar
cubic
cubic(disordered)
Figure 4. Mass spectra of the gaseous compounds released from the as-
synthesized mesophase, which was kept at room temperature.
Ia3d mesophase. The overall tendency for the samples is to
transform their mesophases in the direction of a larger g value,
with various dependencies on the synthesis period and drying
temperature. All of the XRD patterns shown in Figure 2a are
considered to show a cubic Ia3d mesophase, whereas the
samples synthesized for 3 or 6 h exhibit a broad diffraction,
meaning that their structures contain disordered mesophases.
In contrast, the samples synthesized for a longer period showed
a sharp diffraction, indicative of a highly ordered structure. This
difference in the degree of mesophase periodicity is caused by
the flexibility of the silicate network. As the synthesis is
prolonged, silicate condensation proceeds, making the silicate
network rigid. When the rigidity in the silicate network is not
so high, the surfactant micelles destroy the silicate wall of the
samples, leading to a relatively disordered structure.
The effect of drying at high temperatures on mesophase
transformation is summarized in Table 1. The results of a 6
h-synthesis are almost the same as those of a 3 h-synthesis.
The sample before drying has a cubic Ia3d mesophase, and that
dried at 60 °C exhibits broad diffractions, while that dried at
100 °C shows a sharp XRD pattern derived from a lamellar
structure.
Figure 3. XRD patterns of the samples synthesized for 3 h and dried at
room temperature for 24 h, and followed by drying at each temperature.
slightly broadened peaks that can be assigned to a partially
disordered cubic Ia3d structure. In this case, similar phenomena
would occur, but rather broader diffractions were observed
compared to those shown in Figure 2a. These results indicate
that a temperature of 100 °C is too high for these samples to
maintain stable mesostructures.
Figure 3 shows the XRD patterns of the samples synthesized
for 3 h and dried at 60 °C or 100 °C after initially drying at
room temperature for 24 h. The XRD pattern of the 100 °C-
dried sample has two sharp peaks whose d spacing are in the
ratio 1:2, indicative of a lamellar mesophase. This figure
indicates that mesoporous silica changes its structure to make
its g value larger even after being dried at room temperature
for 24 h. It should be noted that the sample dried at a higher
temperature exhibits sharper diffractions, although the sample
dried at a lower temperature appears to have mixed cubic and
lamellar mesophases.
It is clear from Figures 2 and 3 that the mesophases are
transformed because they have a hexagonal array before drying.
The mesoporous silica dried at 60 °C has a cubic Ia3d
mesophase. In contrast, the width of the X-ray diffraction peaks,
which gives a measure of mesophase periodicity, depends
largely on the synthesis period. In the case of drying at 100 °C,
the mesophases show an apparently different dependence on
the synthesis period: The sample synthesized for 3 h had a
disordered structure, that synthesized for 6 h showed a lamellar
mesophase, and those synthesized for 12 and 24 h had a cubic
The sample synthesized for 12 h before drying has a cubic
Ia3d mesophase. Although the sample dried at 60 °C has a
disordered structure, the XRD pattern of the sample dried at
100 °C demonstrates a cubic Ia3d mesophase. These results
indicate that no phase transformation occurs on this sample.
The silicate network in the sample synthesized for a longer
period is more rigid than that synthesized for a shorter period
because of the silicate condensation, as suggested by NMR
16
studies. These results therefore indicate that silicate condensa-
tion, which is directly correlated with the flexibility of the
silicate network, is also a dominating factor in the phase
transformation/determination of mesoporous silica. In the case
of a 24 h-synthesis, all of the XRD patterns showed rather broad
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10941

A R T I C L E S
Ogura et al.
Figure 5. XRD-DSC in situ measurement for the mesophase during drying.
Figure 7. Illustration for the surfactant molecules and silicate in the
determined mesophase.
deduced, along with a consideration of the silicate wall
flexibility, that the samples cannot transform into an ordered
structure.
In Situ Detection of Phase Transformation by XRD-DSC-
MS and TG-DTA-MS. To elucidate the details of the phase
transformation mechanism of mesoporous silica during drying,
it is important to determine the compounds released from
mesoporous silica during the drying process. In the present
study, the as-synthesized sample was kept at room temperature
in the TG chamber with flowing air, and released molecules
were detected. The observed mass spectra are shown in Figure
4
. These results clarify that H2O molecules were released from
mesoporous silica during drying at room temperature. Ethanol,
which is generated by TEOS hydrolysis, was also released
during drying. Judging from the mass spectra, the amount of
evaporated ethanol is considered to be less than that of H2O.
Water evaporation can therefore be concluded as one of the
driving forces for the phase transformation.
To determine the relationship between the phase transforma-
tion of mesoporous silica and the evaporation of water, an in
situ XRD-DSC measurement was performed, as shown in Figure
5
. Simultaneous detection of heat flow generated by water
evaporation and mesophase diffraction provided some informa-
tion regarding changes in the mesoporous silica structure. In
the DSC results, two endothermic peaks were observed, one
around 35 °C and the other around 95 °C. The first peak could
be assigned to the heat flow from the evaporation of physisorbed
water. The XRD patterns suggested that the hexagonal p6mm
to cubic Ia3d phase transformation proceeded after the first
endothermic peak. This result indicates that the phase transfor-
Figure 6. Mass spectra of the gaseous compounds released from the
mesophase and DTA measurement for the mesophase during drying.
peaks, likely because the samples already had a relatively
disordered structure before drying. Accordingly, it can be
10942 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004

Mesophase Determination
A R T I C L E S
Figure 8. Comprehensive insight into the phase transformation of silica mesophase.
mation of mesoporous silica is closely related to water evapora-
tion.
mesophase. On the basis of these factors, a speculative illustra-
tion is proposed in Figure 7.
When much water exists around TEOS, the main reaction
would be condensation between hydroxyl groups, originating
from the hydrolysis of TEOS. However, this reaction seems to
be exothermic. It should also be noted that the second endo-
thermic peak did not affect the structure of mesoporous silica.
The evaporation temperature of ethanol is lower than that of
water, so the possibility of vaporization of remaining ethanol
in the mesoporous silica structure can likely be ignored.
Therefore, as water evaporation proceeds, residual -OH and
Mesoporous silica just after separation from the synthesis
solution contains a large amount of H2O. H2O molecules should
exist near the hydrophilic part of the surfactant, that is between
the headgroup of the surfactant micelles and the silicate wall.
The confinement of H2O molecules in the hydrophilic region
of the surfactants results in an enlargement of the effective
headgroup area, leading to a reduced g value. As water
evaporation proceeds, the H2O molecules are taken away from
the spacing area, causing a decrease in the effective headgroup
area and thus a larger g value. The whole view of mesophase
transformation is illustrated in Figure 8. According to our
conventional understanding, the mesophase can be determined
by the charge density of silicate against surfactant micelles,
which is limited in shape by the conventional g parameter. As
for each mesophase determination and transformation, the
mobility of surfactant micelles is an important point, as has been
-OC2H5 would react directly, thus causing the second endo-
thermic peak.
The mass spectra obtained during TG measurement are shown
in Figure 6. H2O was primarily released from mesoporous silica
at low temperature, while ethanol was primarily released around
1
5
00 °C. As described above, phase transformation occurs around
0 °C. No characteristic feature except the start of the release
8
of H2O could be observed around 50 °C from MS spectra. The
endothermic peaks were detected around the same temperatures
as observed in the DSC measurement. This result supports the
hypothesis discussed above.
suggested by Tolbert et al. However, the transformation is
sometimes hindered kinetically because the surfactant micelles
are surrounded by silicate, and the flexible nature of silicate
depends on the degree of silicate condensation; therefore, the
reactivity of silicate is also required for the mesophase
Proposed Comprehensive Mechanism of Phase Transfor-
mation. As described in the Introduction section, Liu et al.
have discovered that cubic Pm3n mesoporous silica trans-
forms into hexagonal p6mm mesoporous silica.13 This trans-
formation also proceeds toward a larger g value. They have
explained that this phase transformation is well in accordance
with the tendency of the micelles to change in shape from
spherical to rod-type as the amount of solvent decreases. This
phase transformation can actually be explained, however, using
the packing parameter. It is possible that the phase transforma-
tion in this study occurs in the same manner as that observed
by Liu et al., in that the g value directs to a large value. As
described above, the phase transformation of mesoporous silica
is closely related to water evaporation. But what favors the
reconstruction of mesoporous silica with a larger g value?
Totally from the perspective of phase transformation, two key
factors are involved; one is the flexibility of silicate, and the
other is the amount of H2O molecules remaining in the silicate
10-12
transformation, as has been suggested.
During the synthesis
of mesoporous silica thin film, rapid solvent evaporation leads
to a similar transformation, with silicate condensation being
1
4,15
controlled by humidity,
or by the quenching of silicate
1
6
species as they are synthesized. In such a case, an energetic
process is required to transform such a rigid structure. Me-
9
sophase transformation is then induced by drying or heating at
high temperatures. Force from an external system of silicate/
surfactant aqueous solution affects the kinetically hindered
transition, leading to a different route of mesophase transforma-
tion, which is sometimes in opposition to the direction suggested
by the conventional packing rule. The critical criteria for
transformation are silicate flexibility and the number of H2O
molecules involved at the hydrophilic headgroup area of
surfactant for the confinement of surfactant micelles.
The results of the present study elucidate the effects of the
drying process on the determination of silica mesophases. The
J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004 10943

A R T I C L E S
Ogura et al.
results obtained here provide us a comprehensive understanding
of the effects of synthesis and drying procedures on the final
structure of mesoporous silica based on packing parameters and
charge-density matching rules with introducing kinetic hindrance
caused by silicate condensation. These results will enable us to
control the whole synthesis process to obtain desired silica
mesophases.
During drying, mesophase transformation was observed im-
mediately after the release of water. The as-isolated mesoporous
silica just after separation from the synthesis solution contains
a large amount of water, which exists near the hydrophilic part
of the surfactant. This arrangement causes a larger effective
headgroup area, which in turn produces a smaller g value. As
water evaporation proceeds, the H2O molecules are taken away
from the area, and a smaller effective headgroup area is attained,
leading to a larger g value. This is the thermodynamic aspect
of drying process. Water content in as-synthesized mesoporous
silica, as well as the degree of silicate condensation, is an
important factor in determining the direction of the phase
transformation. Comprehensively, confinement of silicate and
surfactant micelles should thermodynamically determine the
final mesophase, but the final form of mesophase is also
influenced by the flexibility of silicate based on kinetic factors.
Conclusions
Phase transformation of mesoporous silica during the drying
process was found to proceed even at room temperature. As-
synthesized hexagonal p6mm obtained under the conditions
employed was transformed to cubic Ia3d as the drying period
increased. Prolonged synthesis resulted in the formation of a
well-ordered hexagonal p6mm mesophase; phase transformation
was hardly observed. The flexibility of silicate, which correlates
well with the degree of silicate condensation, plays an important
role in determining the final mesophase. Drying at higher
temperatures promotes the phase transformation not only from
hexagonal p6mm to cubic Ia3d, but also from cubic Ia3d to a
lamellar mesophase.
The direction of the phase transformation observed here, with
the mesophase tending toward a larger g value, is in opposition
to the one suggested by the packing-parameter rule. XRD-DSC
and DTA-MS observations indicate that the phase transformation
is closely related to water evaporation from the mesophase.
Acknowledgment. This work was supported in part by grants-
in-aid for Scientific Research from Japan Society for the
Promotion of Science and by PRESTO-JST. The authors show
great appreciation to Prof. Takashi Tatsumi, Yokohama National
University, for his fruitful suggestions on the comprehension
to the mechanism suggested in this study. The authors also thank
Dr. Akira Kishi and Mr. Toshiaki Hayashi, Rigaku Corp., for
their cooperation to XRD-DSC measurements.
JA047785J
10944 J. AM. CHEM. SOC.
9
VOL. 126, NO. 35, 2004