Morphology of silica derived from various ammonium carboxylate templates

Article abstract of DOI:10.1016/j.materresbull.2003.07.006

Synthesis of silica from tetraethyl orthosilicate (TEOS) in the presence of various ammonium carboxylate crystal templates was attempted and the effect of these templates on the shape and size of silica was investigated. The carboxylic acids examined immediately precipitated the crystals of their ammonium salts from ethanol and/or water by ammonia addition. The shape, size, and stability of thus formed crystal templates were varied depending on kinds of carboxylic acids and affected the morphology of TEOS-derived silica products. When the crystal templates formed were needle-like and less soluble, silica deposited not on the end faces but on the side faces of the crystals to form tubes with channels corresponding to the shape and size of the templates. On the other hand, when the crystal templates formed were particle-like with a few micron long and liable to re-dissolve into the solution, silica deposited on the whole surfaces of the templates. In this case, the morphology that silica particles with round edge were connected like twig was observed. It was also found that the size and its distribution of the silica tubes are variable to a certain extent by the reaction conditions such as acid and TEOS concentrations.

Full text of DOI:10.1016/j.materresbull.2003.07.006

Materials Research Bulletin 38 (2003) 1669–1680
Morphology of silica derived from various ammonium
carboxylate templates
*
Fumiaki Miyaji , Yasuhiro Watanabe, Yoko Suyama
Department of Materials Science, Faculty of Science and Engineering, Shimane University,
1
060 Nishikawatsu, Matsue 690-8504, Japan
Received 19 March 2003; received in revised form 28 July 2003; accepted 30 July 2003
Abstract
Synthesis of silica from tetraethyl orthosilicate (TEOS) in the presence of various ammonium carboxylate crystal
templates was attempted and the effect of these templates on the shape and size of silica was investigated. The
carboxylic acids examined immediately precipitated the crystals of their ammonium salts from ethanol and/or
water by ammonia addition. The shape, size, and stability of thus formed crystal templates were varied depending
on kinds of carboxylic acids and affected the morphology of TEOS-derived silica products. When the crystal
templates formed were needle-like and less soluble, silica deposited not on the end faces but on the side faces of
the crystals to form tubes with channels corresponding to the shape and size of the templates. On the other hand,
when the crystal templates formed were particle-like with a few micron long and liable to re-dissolve into the
solution, silica deposited on the whole surfaces of the templates. In this case, the morphology that silica particles
with round edge were connected like twig was observed. It was also found that the size and its distribution of the
silica tubes are variable to a certain extent by the reaction conditions such as acid and TEOS concentrations.
#
2003 Elsevier Ltd. All rights reserved.
Keywords: A. Inorganic compounds; B. Chemical synthesis; C. Electron microscopy; D. Microstructure
1
. Introduction
Since carbon nanotube was found [1], various inorganic hollow tubes have been exploited for boron
nitride [2,3], silica [4–8] and titania [9,10] and so on, although their width and length are widely varied.
Unique tubular structures of inorganic materials are potentially useful for applications such as catalyst
carriers, sensors, electronic devices and storage and release systems [8].
In order to synthesize tubular materials, various organic and inorganic templates have been used so
far. Hollow tubes of amorphous silica is one of the most extensively studied target materials. It has been
*
Corresponding author. Fax: þ81-853-32-8893.
E-mail address: fmiyaji@riko.shimane-u.ac.jp (F. Miyaji).
0025-5408/$ – see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2003.07.006

1
670
F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
reported that the external surfaces of self-assembled phospholipid fibers [4], organic gel filaments [5],
viroid cylinders [6], anodic alumina nano-array [7], and cylindrical assemblies of surfactant [8] are
effective as templates of the silica tubes.
All the templates for the silica tubes described above are, however, specialized molecules or arrays.
This has the potential drawback that the costs associated with scale-up are likely to be highly
prohibitive. Thus, the process by Nakamura and Matsui on the formation of the silica tubes from
Si(OC H ) (tetraethyl orthosilicate, TEOS)/H O/C H OH/NH aqueous mixtures that contained a
2
5 4
2
2
5
3
small amount of a simple organic acid, i.e. racemic D,L-tartaric acid is particularly interesting [11]. The
size of the silica tubes obtained by this process was 0.1–1.0 mm in width and 200–300 mm in length.
Although they speculated that the chains of hydrogen-bonded D,L-tartaric acid molecules act as a
template of the silica tubes, the present authors clarified that needle-like ammonium D,L-tartrate crystal,
which is formed almost simultaneously with polycondensed silica when aqueous NH is added, play a
3
role of template [12]. The key in this templating process is a dual role of aqueous NH . That is, it acts
3
not only as a source of the crystal template but also as a basic catalyst for TEOS hydrolysis and
polycondensation.
The present authors also synthesized the tubes of transition metal oxides such as TiO and ZrO by
2
2
the modified process using the ammonium D,L-tartrate crystal template [13]. The usefulness of the
ammonium D,L-tartrate crystal template suggests that the other ammonium carboxylate crystals could
also be templates of the silica tubes. In fact, the present authors confirmed that ammonium oxalate is
also useful as a template of the silica tubes [12]. In the case of the formation of worm-like silica
particles in the presence of citric acid [14], ammonium salt of citric acid is assumed to work as a
template of such silica particles, although the authors have not referred to the template. In the present
study, synthesis of silica in the presence of various ammonium carboxylate crystal templates
was attempted and the effect of these templates on the shape and size of silica was investigated.
The effect of reaction conditions on the morphology of silica was also examined for the D,L-tartaric
acid-containing system.
2
. Experimental procedure
2
.1. Template and silica formation
Table 1 shows the carboxylic acids used as templates of silica in this work. First, the solubility of
each carboxylic acid for 100 ml of water or ethanol was examined at 25 8C. Next, 0.5% of the solubility
of each acid for water or ethanol determined above was dissolved into 5 ml of water or ethanol,
respectively, and then 2 ml of 28% aqueous NH (Kanto Chemical Co., Inc.) was added to examine the
3
formation of ammonium carboxylate crystal templates.
Silica synthesis was carried out with the carboxylic acids under the conditions that the acids
precipitate ammonium carboxylate crystals to be templates. Typically, 0.73 g of TEOS (Kanto
Chemical Co., Inc.) was added into 5 ml of C H OH (Kanto Chemical Co., Inc.) containing 0.02 g of
2
5
D,L-tartaric acid (Tokyo Kasei Kogyo Co., Ltd.) and 0.06 g of distilled water, and the mixed solution
was allowed to stand for 30 min. Subsequently, 2 ml of 28% aqueous NH was added to the TEOS
3
solution and the solution was allowed to stand for 30 min. The concentrations of D,L-tartaric acid,
À3
À3
TEOS and NH , and H O/TEOS ratio are 17 mmol dm , 0.45 and 3.7 mol dm , and 1, respectively,
3
2

F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
1671
Table 1
Solubility of carboxylic acids and formation of ammonium carboxylate templates in water or ethanol by ammonia addition
a
b
Carboxylic acid
Solubility (g)
Precipitated crystal
Formed silica
H O
2
C H OH
2
H O
2
C H OH
2 5
5
D,L-Tartaric
L-Tartaric
meso-Tartaric
Oxalic
20.6
139
125
11.8
8.2
5.0
42.5
19.5
54.6
6.8
–
Needle-like
Needle-like
Tube
Tube
–
c
–
Needle-like
Particle
Aggregate
Twig-like
Tube
Tube
Succinic
D,L-Malic
Citric
–
–
–
Needle-like
c
122
194
28.2
80.0
Particle
Particle
Twig-like
Twig-like
c
a
Values as g/100 ml of each solvent at 25 8C.
.5% of the solubility of each carboxylic acid was dissolved into 5 ml of H O or C H OH and then 2 ml of 28% aqueous
2 2 5
b
0
NH was added.
3
c
Formed crystals were re-dissolved into the solution with time.
in this condition. The whole process described above was done at 25 8C. The products were washed
with a large amount of water on a test sieve with 63 mm aperture or a 0.2 mm membrane filter, or
unwashed. In the case of unwashed samples, the products were collected on a 0.2 mm membrane filter.
As for the other carboxylic acids except oxalic acid, similar process was used for silica synthesis. With
respect to oxalic acid, only water was used as a solvent because the crystal formed from the oxalic acid
coagulates in ethanol. In this case, 0.6 g of oxalic acid was dissolved into 5 ml of H O and then 2 ml of
2
2
0
8% aqueous NH was added without stirring to give a turbid suspension of the crystal templates. Next,
3
.365 g of TEOS was added immediately after aqueous NH addition and the mixture was stirred
3
vigorously for 30 min. The products were washed with a large amount of water and collected on the test
sieve with 63 mm aperture, or unwashed.
2
.2. Effect of reaction conditions on the formation and size of the tubes
Effect of various reaction conditions on the formation and morphology of the silica tubes were
examined for the D,L-tartaric acid-containing system. The concentrations of D,L-tartaric acid, TEOS and
aqueous NH , H O/TEOS ratio, reaction temperature, and time allowed to stand after TEOS or aqueous
NH addition were varied. When one condition was changed, the other conditions were unified into the
3
2
3
basic conditions described in Section 2.1.
2
.3. Characterization of silica products
The products were subjected to SEM observation (HITACHI S-2500CX) and powder XRD
measurement (Rigaku RINT-2000). The N adsorption isotherms were measured for water-washed
2
samples (Quantachrome AUTOSORB-1). The specific surface area, pore size distribution, and pore
volume were calculated from the measured isotherm curves. In this work, the strict P/P tolerance
0
(
À0.0001 to þ0.0003) was used in the measurements at lower pressure region (P/P < 0:1) in order to
0
obtain accurate micropores data. The measurements were repeated at least three times for the same
samples to confirm reproducibility.

1
672
F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
3
. Results and discussion
3
.1. Template and silica formation
Table 1 summarizes the results of template formation by aqueous NH addition for various
3
carboxylic acids. The oxalic acid precipitated the needle-like crystals from water, whereas formed the
aggregates in ethanol. The other carboxylic acids examined precipitated no crystal from water. The
D,L-, L-tartaric, and succinic acids precipitated needle-like crystals from ethanol. The meso-tartaric,
D,L-malic, and citric acids precipitated particle-like crystals, but the formed crystals gradually
disappeared with time. This indicates that these crystal templates formed are very soluble in water
and therefore re-dissolved into water from aqueous NH . It was found from the preliminary
3
experiments that malonic, glutaric and L-malic acids form crystal templates from neither water nor
ethanol.
Fig. 1 shows the SEM photographs of silicas formed in the presence of various ammonium
carboxylate crystal templates. In the case of the D,L- and L-tartaric acids (Fig. 1a–c), silica tubes
with rectangular-shaped channels were formed. The sizes of the tubes were 50–300 mm in length and
0
.1–1.0 mm in width. Regarding the oxalic acid (Fig. 1d), shorter (50–100 mm in length) and thicker
(
1.0–15 mm in width) tubes were formed, compared to the D,L- and L-tartaric acids-derived tubes. As
for the succinic acid (Fig. 1e), short (ca. 10–30 mm in length) macaroni-like tubes in which the
edges were oblique were aligned to the length direction. The width of the tubes was about 2 mm.
On the other hand, with respect to D,L-malic, meso-tartaric and citric acids (Fig. 1f–h), the
morphology that silica particles with round edge direction were connected like twig was observed. The
sizes of each silica particle were 4–8 mm in length and about 0.5–1.0 mm in width. Although the length
of the silica particles is not so different among the three kinds of acids, the width tended to increase in
the order citric, meso-tartaric and D,L-malic acids. It was reported that worm-like silica with 0.6–1.2 mm
in width and 8–12 mm in length was formed in the presence of citric acid [14]. Although the length of
the worm-like silica particle is a little longer than that of the present twig-like silica particle formed in
the presence of citric acid, both silicas are essentially same in shape. In summary, the silica formed in
the presence of ammonium carboxylate templates are classified into two groups, i.e. end-opened
‘‘tubular’’ silica and end-closed ‘‘twig-like’’ one. This structural difference is explained in terms of the
rates of the template growth along needle direction and the stability of the templates in solution, as
discussed in Section 3.2.
Table 2 shows the results of specific surface area and pore volume of silicas derived from some
ammonium carboxylate templates. As a representative, pore size distribution curve of silica tubes
Table 2
Specific surface area and pore volume of silicas derived from some carboxylic acids
2
Specific surface area (m g
À1
À1
Pore volume (ml g )
Template
)
Ammonium D,L-tartrate
Ammonium L-tartrate
Ammonium oxalate
Ammonium citrate
410
508
0.31
0.32
0.83
0.48
1628
790
The samples were heated at 100 8C in vacuo for 3 h prior to measurements.

F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
1673
Fig. 1. SEM photographs of silicas derived from various ammonium carboxylate crystal templates: (a and b) ammonium D,L-
tartrate; (c) ammonium L-tartrate; (d) ammonium oxalate; (e) ammonium succinate; (f) ammonium D,L-malate; (g) ammonium
meso-tartrate; (h) ammonium citrate.

1
674
F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
Fig. 1. (Continued ).
synthesized in the presence of oxalic acid is shown in Fig. 2. A large micropore peak was observed
around 0.6 nm. There was no remarkable difference in pore volume and micropore size distribution
among the silicas derived from different carboxylic acids. The exceptionally large specific surface
area of silica derived from oxalic acid is ascribed to the different synthetic process. In this case,
TEOS is hydrolyzed and polycondensed at the interface between upper TEOS phase and lower
water phase under stirring. Since the size of thus formed silica particles is very small and such fine
silica particles gradually attach onto the ammonium oxalate templates to form the silica tubes, the
specific surface area becomes higher in the oxalic acid-derived silica tubes than in the other silica
tubes.
Fig. 2. Pore size distribution of silica tubes derived from ammonium oxalate template.

F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
1675
3
.2. Formation process of tubular and twig-like silica
In the process of tubular silica formation, the formation of ammonium carboxylate templates and
the deposition of silica onto the side faces of the templates proceeds almost simultaneously when the
aqueous NH is added to the solution [12], although the oxalic acid system is excluded owing to the
3
different processing that the crystal templates were formed prior to silica deposition, as mentioned
in Section 3.1. Since the crystal templates could not grow further when silica starts to deposit on
them, the width of the tube would increase with increasing the ratio of the crystal growth rate along
the width direction to the TEOS deposition rate. As in the succeeding discussion, the width of
the tube changed with some reaction conditions, e.g. the acid and TEOS concentrations. This
phenomenon is also explained in terms of the change in the rates of the template growth and/or silica
deposition.
On the other hand, no silica is deposited on the end faces of the crystal templates, and this is a reason
of formation of the hollow silica tubes in the presence of some carboxylic acids such as D,L- and
L-tartaric acids. The growth rate of the crystal templates along the needle direction is very rapid on
aqueous NH addition. Since the growth along the needle direction is much faster than that of silica
3
deposition onto the end faces, the edges are remained uncoated with silica. One question is why the
edges of the crystal templates were remained uncoated for the ammonium oxalate template, although
the template was formed prior to the TEOS addition. One possible explanation is concerned with
preferential re-dissolution of the template along the needle direction. Although ammonium oxalate
crystal rapidly grows with along the needle direction similarly to ammonium D,L-tartrate, it re-dissolves
in part into water from aqueous NH . If the rate of re-dissolution of the crystal along the needle
3
direction is higher than that of silica deposition, the end faces of the crystal would be remained
uncoated with silica. In fact, the end faces of the ammonium D,L-tartrate templates also remained
uncoated with transition metal oxides in spite of prior formation of the templates [13].
The morphology of the twig-like connection of the end-closed silica particles with a few micron in
length, which were formed in the presence of D,L-malic, meso-tartaric, and citric acids, is clearly
distinguished from that of the silica tubes described above. The crystal templates of ammonium D,L-
malate, meso-tartrate and citrate grow slower along the needle direction than the crystal templates such
as ammonium D,L- and L-tartrates. In this case, silica is deposited on the end faces as well as on the side
faces of the crystal templates, which is responsible for the formation of end-closed morphology. It
should also be noted that the edges of thus formed silica particles are not angular but round. The
formation of such a characteristic morphology is interpreted as follows. Since the crystal templates of
ammonium D,L-malate, meso-tartrate and citrate are highly soluble in water, they rapidly re-dissolve
into the water from aqueous NH and then silica is deposited on the whole surfaces of the templates
3
before complete dissolution of them. This is consistent with the results of powder XRD patterns of the
citric acid-containing system (Fig. 3b) that the unwashed silica product already contains no crystal
template, i.e. ammonium citrate. No presence of the templates was also confirmed for the unwashed
products derived from D,L-malic and meso-tartaric acids. On the contrary, the unwashed product
involving the silica tubes contained the crystal templates, as shown in Fig. 3a as a representative. These
results indicate that the inside of each twig-like silica particle is hollow before rinse with water The
twig-like connection of the silica particles might be appeared in the midway of silica growth. When
one silica particle is contacted to the other, silica is assumed to deposit so as to close the gap between
the particles in order to lower the interfacial energy.

1
676
F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
Fig. 3. Powder XRD patterns of ammonium carboxylate crystal templates, unwashed and washed silica products:
a) ammonium D,L-tartrate; (b) ammonium citrate.
(

F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
1677
Fig. 4. Width of ammonium D,L-tartrate-derived silica tubes as a function of D,L-tartaric acid concentration. Closed square and
vertical bars indicate the average width and the observed range, respectively.
3
3
.3. Effects of synthetic conditions on the silica tube formation
.3.1. Effect of acid concentration
First of all, the yield of the silica tube was about 20% at most under every synthetic condition. The
silica other than tubular form was the colloidal particles which are same as ones formed in the
presence of basic ammonia catalyst. The yield of the tubes increased with increasing the concentration
À3
À3
of D,L-tartaric acid up to 17 mmol dm . Over 17 mmol dm , the yield of the formed tube gradually
decreased, and in contrast the amount of the aggregates of silica increased. In this case, the
aggregation of the crystal templates would also give rise to the aggregation of silica. In addition, the
amount of the colloidal silica particles attached on the outer surfaces of the tubes increased with
increasing the acid concentration. The increase in the width of the tubes due to increase in the acid
concentration, as discussed below, might be favorable for the attachment of the colloidal particles to
the tube surfaces.
Fig. 4 shows the width of the silica tube as a function of D,L-tartaric acid concentration. There was a
tendency that the average and distribution of the width increased with increasing the acid concentration.
This is explained as follows. The growth rate of the crystal templates increases with increasing the acid
concentration. This causes the increase in the width of the crystal templates at which silica starts to
deposit on them, resulting in the formation of thicker tube. While the curved tubes were mainly
observed when their widths were below 1 mm, the straight tubes were majority when the widths were
over 1 mm. The formation of the curved tubes is attributed to the stress by the deposited silica to the
crystal template during its growth. When the crystal growth is quite rapid and the resultant template is
thicker (>1 mm), the crystal template is strong enough not to be curved by the stress due to the
deposited silica. In this case, the thick straight tubes are formed.
3
.3.2. Effect of TEOS concentration
Fig. 5 shows the width of the silica tube as a function of TEOS concentration. No tube was formed at
À3
the smallest concentration (0.08 mol dm ). In this case, TEOS concentration is too low to deposit
À3
silica on the surface of the templates. On the other hand, at the largest concentration (0.8 mol dm ) a
large amount of colloidal particles was attached on the surfaces of the tubes, although the total amount

1
678
F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
Fig. 5. Width of ammonium D,L-tartrate-derived silica tubes as a function of TEOS concentration. Closed square and vertical
bars indicate the average width and the observed range, respectively. The cross means no formation of tube.
À3
of the tubes was larger than that for lower concentration (0.45 mol dm ). This indicates that the
fraction of silica precipitated not on the templates but in the solution increased with increasing TEOS
À3
concentration. The distribution of the width of the tubes was wider in 0.16–0.35 mol dm than in
À3
0
.45 mol dm . Since the decrease in TEOS concentration retarded the silica deposition on the
templates, the period of the crystal templates growth was prolonged. As a result, the distribution of the
widths of the templates and the resultant tubes became wider.
3
.3.3. Effect of ammonia concentration
Fig. 6 shows the width of the silica tube as a function of ammonia concentration. The average and
maximum width of the tube and the amount of particles attached onto the tube increased with
decreasing concentration of ammonia. This is explained as follows. The hydrolysis and
polycondensation of TEOS is drastically retarded by the decrease in concentration of ammonia,
whereas the rate of the crystal templates formation is not changed so much. This causes the increase in
Fig. 6. Width of ammonium D,L-tartrate-derived silica tubes as a function of ammonia concentration. Closed square and
vertical bars indicate the average width and the observed range, respectively.

F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
1679
the width of the template at which silica starts to deposit onto them. Consequently, thicker tubes are
formed as in the case of increase in acid concentration.
3
.3.4. Effect of H O/TEOS ratio
2
The effect of H O/TEOS ratio on the formation and morphology of the silica tube was examined for
2
the H O/TEOS ¼ 0, 1, and 4. When the H O/TEOS ratios were 0 and 1, silica tubes with a relatively
2
2
small amount of colloidal particles were formed. When the H O/TEOS ratio was 4, colloidal silica
2
particles were preferentially formed, and the yield of the silica tube was much lower than that for the
H O/TEOS ¼ 1. The H O/TEOS ratio did not affect the shape and size of the silica tubes.
2
2
It is known that acid-catalyzed TEOS solution with low H O/TEOS ratio is comprised of linear
2
polymer, whereas solution with high H O/TEOS ratio causes the formation of non-linear or network
2
polymers in hydrolysis–polycondensation [15]. It is therefore assumed that silica linear polymer is
mainly present in the solutions for the lower H O/TEOS ratios (0 and 1) before ammonia addition.
2
Since flexible silica linear polymer may quickly deposit on the templates on ammonia addition, the
yield of the silica tube increases. On the contrary, in the case of higher H O/TEOS ratio (4), non-linear
2
or network polymers of silica would preferentially form colloidal silica particles. In other words,
less-hydrolyzed silica linear polymer would be favorable for deposition on the surfaces of the crystal
templates.
3
.3.5. Effect of reaction temperature
The average width of the tube increased and the distribution of the width widened when the
temperature was raised over 40 8C. This reflects the increase in the rate of growth of the crystal
templates. The width of the tube was not changed in the range 5–25 8C. This indicates that both rates of
the crystal growth and silica deposition did not change so much relatively in this temperature range.
3
.3.6. Effect of time allowed to stand after TEOS addition
The width and its distribution were not changed with variation of time allowed to stand after TEOS
addition. The colloidal silica particles attached onto the tubes, however, increased when the time was
longer than 30 min. Moreover, too long standing time, 24 h, gave not silica tubes but gelatin-like silica
gel aggregates. The extremely prolonged time develops a cross-linked network of silica sol and it
immediately aggregates on ammonia addition.
3
.3.7. Effect of time allowed to stand after ammonia addition
Almost no tube was formed when the time allowed to stand after ammonia addition was below 1 min.
If the time is too short, the polycondensation of TEOS is insufficient and silica cannot deposit onto the
surfaces of the crystal templates. In the case of the tube formation, the width of the tubes and its
distribution were hardly changed depending on the time.
4
. Conclusions
Silica was synthesized from TEOS in the presence of various ammonium carboxylate crystal
templates. The morphology of silica formed was classified into two groups. When the crystal templates
formed were needle-like and less soluble, silica tubes with channels corresponding to the shape and size

1
680
F. Miyaji et al. / Materials Research Bulletin 38 (2003) 1669–1680
of the template were formed. When the crystal templates formed were particle-like with a few micron
long and liable to re-dissolve into the solution, silica particles with round edge were connected like
twig. The size and its distribution of the silica tubes varied to a certain extent depending on the reaction
conditions such as acid and TEOS concentrations.
Acknowledgements
This work was supported by the Grant-in-Aid for Scientific Research from the Ministry of Education,
Culture, Sports, Science, and Technology of Japan (No. 12750603) and Electric Technology Research
Foundation of Chugoku.
References
[
[
[
[
[
[
[
[
[
1] S. Iijima, Nature 354 (1991) 56.
2] E.J.M. Hamilton, S.E. Dolan, C.M. Mann, H.O. Coloijin, C.A. Mcdonald, S.G. Shore, Science 260 (1993) 659.
3] P. Gleize, M.C. Shouler, P. Gadelle, M. Gaillet, J. Mater. Sci. 29 (1994) 1575.
4] S. Baral, S. Schoen, Chem. Mater. 5 (1992) 145.
5] Y. Ono, K. Nakashima, M. Sano, Y. Kanekiyo, K. Inoue, J. Hojo, S. Shinkai, Chem. Commun. (1998) 1477.
6] W. Shenton, T. Douglas, M. Young, G. Stubbs, S. Mann, Adv. Mater. 11 (1999) 253.
7] M. Zhang, Y. Bando, K. Wada, J. Mater. Res. 15 (2000) 387.
8] M. Harada, M. Adachi, Adv. Mater. 12 (2000) 839.
9] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Langmuir 14 (1998) 3160.
[
[
[
[
[
[
10] H. Imai, Y. Takei, K. Shimizu, M. Matsuda, H. Hirashima, J. Mater. Chem. 9 (1999) 2971.
11] H. Nakamura, Y. Matsui, J. Am. Chem. Soc. 117 (1995) 2651.
12] F. Miyaji, S.A. Davis, J.P.H. Charmant, S. Mann, Chem. Mater. 11 (1999) 3021.
13] F. Miyaji, Y. Tatematsu, Y. Suyama, J. Ceram. Soc. Jpn. 109 (2001) 924.
14] H. Nakamura, Y. Matsui, Adv. Mater. 7 (1995) 871.
15] S. Sakka, K. Kamiya, J. Non Cryst. Solids 48 (1982) 31.