Surfactant-templated mesostructured materials from inorganic silica

Andreas Berggren, Anders E. C. Palmqvist and Krister Holmberg*

Received 27th May 2005, Accepted 1st July 2005

First published as an Advance Article on the web 25th July 2005

DOI: 10.1039/b507551n

Mesoporous silica materials made by the use of self-assembled surfactants as templates have

attracted a lot of interest in recent years. The number of publications on the topic has increased by

a factor of 100 between 1993 and 2003. The vast majority of the papers deal with syntheses from

organosilicates, such as tetramethylorthosilicate (TMOS) and tetraethylorthosilicate (TEOS), as

the silica source. These alkoxides are convenient as a starting material but they are expensive and

the products made from them can only find use in applications where price is not a major issue. This

review summarizes the published work on the preparation of mesoporous silica from inorganic

sources, such as water glass and silica sol, still using self-assembled surfactants as the template. The

use of such raw materials instead of the organosilicates will radically change the production price of

the products and may open up completely new fields of application. Much less work has been put

into the synthesis of mesoporous silica from these starting materials and relatively little is known

about the reaction mechanism. The current understanding is summarized and discussed and

comparisons are made with the conventional route that originates from the organosilicates.

1. Introduction

The formation of mesostructured silica using surfactants as

the structure-directing agents was reported in the early 1990’s

independently by scientists from the Mobil Oil Corp.,1 and

by Kuroda’s group.2,3 The products obtained had high

surface area pores having tunable dimensions within the range

2–10 nm, and a narrow pore size distribution. Such materials

have several potential applications, such as heterogeneous

catalysis, separation processes, and host–guest chemistry. This

route of synthesis immediately attracted attention in the

materials science community and it is no overstatement to

say that surfactant-templated synthesis of mesoporous mate-

rials has become a new field of research positioned at the

border between surface chemistry and materials chemistry.

Applied Surface Chemistry, Department of Chemical and BiologicalEngineering, Chalmers University of Technology, SE-412 96 Goteborg,Sweden. E-mail: kh@chem.chalmers.se; Fax: +46 31 16 00 62;Tel: +46 31 772 29 69

Andreas Berggren

Andreas Berggren is aPhD student at ChalmersUniversity of Technology,Goteborg, Sweden. He receivedhis MSc in ChemicalEngineering from ChalmersUniversity of Technology in2005. His present researchproject is within the field ofmesoporous silica.

Krister Holmberg

Krister Holmberg is Professorof Surface Chemistry atChalmers University ofTechnology and Head of theDepartment of Chemical andBiological Engineering. Hisresearch interests relate tosurfactants and surfactantapplications, microemulsions,biotechnological surface chemi-stry, and reactions in micro-heterogeneous media. He haspublished 190 papers, writtenor edited six books and is theinventor of 35 patents.

Anders Palmqvist received hisPhD in Inorganic Chemistryfrom the Royal Institute ofTechnology in Stockholm in1997 on the synthesis of col-loidal zeolites and ceria nano-particles. His postdoctoralresearch at University ofCalifornia at Santa Barbarafocused on microporous andhost–guest semiconductors.He joined the faculty ofChalmers Univers i ty ofTechnology in 1999, where hebecame associate professor ofMaterials Chemistry in 2004.

In parallel he held a part time researcher position at VolvoTechnology Corporation between 2001 and 2005. His currentresearch interests span synthesis and characterization of orderedmicro- and mesoporous materials, and nanoparticles, with anemphasis on using surfactants as structure directing agents.

Anders E. C. Palmqvist

REVIEW www.rsc.org/softmatter | Soft Matter

This journal is � The Royal Society of Chemistry 2005 Soft Matter, 2005, 1, 219–226 | 219

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online / Journal Homepage / Table of Contents for this issue

The number of research papers dealing with the topic has

grown tremendously during the last decade, as shown in Fig. 1

which is based on data compiled from a search in the database

SciFinder for the topic mesoporous silica. The majority of

papers deal with mesoporous silica but a substantial number of

other oxides, such as alumina, titania, and zirconia, have also

been synthesized using the same general method of surfactant

templating. A number of recent reviews cover the topic.4–8

The usual procedure is to employ a simple organosilicate

compound, such as tetramethylorthosilicate (TMOS) or

tetraethylorthosilicate (TEOS), as the silica source. These can

be regarded as tetraesters of silicic acid, Si(OH)4. Silicic acid is

a weak acid and it exists only in dilute aqueous solutions.9 It

polymerizes very easily and has never been isolated. The silicic

acid esters are often referred to as alkoxides. They are not

soluble in water, but can be dissolved in a mixture of water and

a water miscible organic solvent. They are hydrolyzed when

the pH is reduced or increased causing the ester bond to cleave,

which generates alcohol and a free silanol group. The silanol

group is reactive and undergoes condensation reactions with

other silanol groups. Depending on the pH and the presence of

salts, the condensation may lead to particle growth and/or

gelation via processes that are relatively well understood.9

The formation of mesoporous silica from alkoxides, such

as TMOS and TEOS, works well and the product can be

controlled with regard to symmetry, pore size and wall

thickness. However, the procedure suffers from one major

drawback: the silica source, the alkoxide, is expensive and

both TMOS and TEOS must be regarded as chemicals for

laboratory rather than industrial use. Thus, mesoporous silica

produced from an alkoxide precursor has a price that limits

large scale applications.

The economic considerations have recently triggered an

interest in the use of inexpensive inorganic silicate as a starting

material. Aqueous sodium silicate, water glass, is a commodity

chemical with a price that is orders of magnitude lower than

that of the alkoxides. If mesoporous silica could be made from

such a starting material, it would not be an expensive product

and the number of realistic applications for the material would

multiply. Work along this line recent in origin and few papers

dealing with the topic were published before 1999. The present

review aims at compiling and discussing the appropriate

literature concerning the use of inorganic silicates in the

formation of mesostructured materials.

2. Reaction pathways and the source of silica

The pathways described in the literature for the synthesis of

mesoporous silica from an inorganic silica source have here

been divided into four sections. The first three cover the use of

different sodium silicate solutions and the fourth section

makes a short comment on the use of clays as a silica source.

The first two methods are one-step procedures and are

distinguished by the pH of the reaction solution. The third

method is a two-step procedure, where the first step yields a

metastable solution of pH 2.

A silicon atom in a silica or silicate product can be attached

to from zero to four other silicon atoms via siloxane bonds.

TEOS and other simple alkoxides are examples of silica

sources not containing any siloxane bonds and colloidal silica

is an example of a source of silica where the majority of silicon

atoms have four neighbors linked by siloxane bonds. (The

silicon atoms on the periphery of the silica nanoparticle usually

have three siloxane bridges and one silanol group.) The

coordination pattern is usually determined by 29Si-NMR and

it is customary to use Si(Qn) as the nomenclature for signals

from silicon atoms coordinated by n siloxane bonds. Thus,

TEOS gives only the Si(Q0) signal and colloidal silica gives

mostly the Si(Q4) signal. Water glass can show all five signals,

with signals from higher coordination increasing with increas-

ing SiO2 to Na2O ratio, i.e., decreasing pH. It has been shown

that silica sources that display Si(Q4) signals are unsuitable as

starting materials for the synthesis of mesoporous materials.

For instance, colloidal silica and high ratio water glass were

found not to give a mesoporous material under conditions

where low ratio water glass and TEOS gave hexagonal

mesoporous silica.10 However, if the colloidal silica was first

treated with alkali to reach a SiO2 to Na2O ratio of two, it

could be used as a starting material for mesoporous silica. This

silica source gave no Si(Q4) NMR signal. Klotz et al. showed

the importance of the Si(Qn) distribution in the formation of

ordered mesostructured silica and the effects of ageing the

synthesis solution on the Si(Qn) distribution. They found that a

solution with a predominance of Si(Q1) species in the initial

stage of the synthesis resulted in an ordered mesophase and

that the disappearance of the order was related to the presence

of Si(Q3) species in the initial stage of the synthesis.11 It thus

seems that the presence of low-condensed, Si(Qn,3), species in

the initial synthesis solution is a prerequisite for obtaining

ordered mesoporous silica.

2.1 Reaction solution with pH . 6

On the alkaline side, Kim et al. have developed a synthesis

method using nonionic surfactants, such as triblock copoly-

mers and alcohol ethoxylates, and a reaction solution at near

neutral conditions.12–14 The surfactant together with acetic

acid, used in an amount equimolar to the hydroxide content of

the sodium silicate, was added to a sodium silicate solution to

create the reactive solution. At molar ratios of H2O/Si 5 230

Fig. 1 The development of the research field of mesoporous silica

presented as number of articles published per year in journals covered

by SciFinder.

220 | Soft Matter, 2005, 1, 219–226 This journal is � The Royal Society of Chemistry 2005

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online

and surfactant/Si 5 0.008–0.017 the procedure resulted in a

thermally stable 2D hexagonal mesostructured silica, MSU-H.

The reaction time was 20 h at fixed temperatures of 308,

318 and 333 K, where the different temperatures generated

different pore sizes in the hexagonal structure.

Increasing the pH increases the charge density of the silica.

This favors electrostatic interactions between the surfactants

and the silica, which can be taken advantage of when using

cationic surfactants. Han et al. used a mixture of two cationic

surfactants as a template for the synthesis of rod-like meso-

porous silica with hexagonal symmetry.15 In the procedure,

ethyl acetate was added to a dilute solution of a mixture of

sodium silicate and the surfactants to obtain the desired

reactive pH. The reaction was allowed to proceed at 353 K

for 72 h, followed by calcination. The reported specific

BET surface area was about 760 m2 g21. The same group

also synthesized high quality MCM-48 mesoporous silica

by the same method using a cationic gemini surfactant,

[C12H25N+–(CH3)2–(CH3)–N+–C12H25]?2Br2,16 and hollow

spherical silica with a mesoporous shell using a traditional

cationic surfactant (CTAB).17

2.2 Reaction solution with pH , 2

Several papers deal with synthesis under very acidic conditions

and the results obtained are generally good. Setoguchi et al.

reported that their acidic synthesis generated the largest

mesopore volume and the highest surface area with cationic

surfactants.10 Studies on the use of nonionic surfactants

under acidic conditions have been reported by Kim et al.18,19

By using di- and tri-block copolymers together with sodium

metasilicate as the silica source, different mesostructures,

such as 2D hexagonal, 3D hexagonal, and 3D cubic, were

obtained. With blends of different amphiphilic block copoly-

mers and through variations in the hydrophilic–hydrophobic

balance achieved by changing the size of the hydrophilic

blocks, the mesostructured phase was controlled with con-

siderable accuracy.

The possibility of controlling the morphology of meso-

structured particles has been investigated in several

publications using sodium silicate as the silica source.20–23

Experiments using the tri-block copolymer Pluronic P123 as a

nonionic surfactant resulted in a monodisperse rodlike SBA-15

structure with high mesoscopic order. Only a very limited

range of synthesis conditions could be used, however. The

molar ratio SiO2/HCl/P123/H2O was 1/7.84/0.0017/252 and

the reaction run at 303 K for 6 h.21 There are also reports on

the formation of single crystals of mesoporous silica with a

reaction mixture composed of cationic surfactant/SiO2/NaOH/

H2SO4 in a molar ratio of 1/2.15/1.67/4.35–8.90 in a very dilute

water solution under static conditions for 2–4 days at 30uC,

Chao et al. produced SBA-1 mesoporous silica in the form of

single crystals. The crystal shape (spheres, decaoctahedrons

or cubes) was governed by varying the pH between one and

two.22,23 The results were explained by the fact that both

the dilute conditions and the pH used, which is close to the

isoelectric point of SiO2, leads to a slow condensation rate.

From the viewpoint of crystallization kinetics, the crystal

morphology is determined by the relative rate of growth of

different crystal faces, with the slow-growing surface dominat-

ing the final shape.

2.3 Two-step synthesis

Sierra et al. developed a procedure in which they first created a

metastable water solution of silica from sodium silicate and

non-ionic surfactant at pH 2 and then added sodium

hydroxide to generate the reactive solution at pH 2–624 or

pH 3–10.5.25 They found that the regularity and the size of the

pores depended mainly on the pH used during the condensa-

tion and on the sodium ion concentration in the synthesis

mixture. Both high pH and high concentration of sodium ions

favored the formation of large and ordered pores. However, a

disadvantage of the method was that some of the products had

low thermal stability.

Another method to induce condensation in a metastable

silica solution is to add a small amount of a fluoride salt. This

approach was used in a method developed by Boissiere et al.

First they created a metastable solution at pH 2 by dissolving a

sodium silicate solution in an acidified water solution of the

surfactant. Then they induced silica condensation by adding

sodium fluoride. The product obtained was mesoporous

after calcination but with a disordered 3D worm-hole pore

structure, MSU-X. It was comparable to mesoporous silica

obtained in a similar experiment based on TEOS as the silica

source.26

2.4 Clay as source of silica

The first method for making mesoporous silica from layered

silicate clay was presented by Yanagisawa et al. as early as

1990.2 The clay, kanemite (NaHSi2O5?3H2O), was first

prepared from amorphous silica and NaOH, and subsequently

dispersed in a solution containing a cationic surfactant. The

mechanism proposed for this type of templating is believed to

be quite different from the methods discussed earlier. Here the

silica is added as a layered polysilicate and the surfactants

penetrate the layered silica by an ion exchange process

involving the interlayer sodium ions. The surfactants then

swell the clay and create uniform channels in the layered

silica3. Based on subsequent in-situ X-ray powder diffraction,

NMR and TEM studies this mechanism has subsequently been

further developed to also encompass structural changes of

the silicate framework of kanemite such as fragmentation,

intralayer condensation and bending of silicate sheets.27–31

This concept is, however, outside the main scope of this review

and is not further discussed.

3. The role of the surfactant

In this literature survey several different experimental pro-

cedures for the preparation of mesostructured silica are

presented covering a wide variety of different surfactants.

The choice of surfactant is very important because it governs

the size of the pores, as well as the thickness of the walls, and

the symmetry of the mesoporous material. The two main

classes used are cationics and nonionics. Anionic surfactants,

which constitute the largest surfactant class, have rarely been

employed as a template for mesostructured silica,32 whereas

This journal is � The Royal Society of Chemistry 2005 Soft Matter, 2005, 1, 219–226 | 221

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online

the fourth surfactant class, zwitterionics, do not seem to have

been used at all with inorganic silica sources. There are also

examples of the use of combinations of surfactants from

different classes, such as cationic and nonionic.33

Cationic surfactants of the type CnH2n+1(CH3)3N+X2, with

n being 12–16 and X being Br or Cl, were used in the first

experiments to produce mesoporous silica from an alkoxide

precursor1 and similar surfactants have been used in a large

number of subsequent reports with TEOS or other organo-

silicates as the starting material. Cationic surfactants have

also been employed for the preparation of mesoporous silica

from an inorganic silica source and most of these studies have

been performed under neutral to alkaline conditions,15,16,34–36

although examples from the acidic side have also been

published.10

Shio et al. prepared fine mesoporous silica powders by

adding an acid to a solution of a mixture of sodium meta-

silicate and a surfactant of the type CnH2n+1(CH3)3N+Cl2, with

n being 18 or 22.35 The procedure gave particles with not only

the pore diameter and the specific surface area varying with the

length of the long alkyl chain of the surfactant but, also the

particle size and shape. Stearyl (n 5 18) gave a pore diameter

of 3.0 nm and a specific surface area (BET surface) of

1050 m2 g21 while behenyl (n 5 22) gave a pore diameter of

3.5 nm and a surface area of 900 m2 g21. Since in the liquid

crystal that serves as a template for the mesoporous material

the surfactants assemble in head-out double layers, the pore

diameter should reflect twice the length of the hydrophobic tail

of the surfactant because the pores should equal the size of

the hydrophobic domains of the liquid crystal. The values

obtained, 3.0 and 3.5 nm, are perfectly reasonable for alkyl

chains of 2 6 18 and 2 6 22 carbon atoms. Whereas the

powder particles made with the stearyl-based surfactant were

cubes with a dimension of around 100 nm, the particles made

with the behenyl-based surfactant were rod-like with a length

of 300–500 nm and a diameter of ca. 50 nm.

Setoguchi et al. performed the synthesis of hexagonal

mesoporous silica by adding water glass to a highly acidic

solution of different cationic surfactants.10 Two surfactant

series were tested, CnH2n+1(CH3)3N+X2, with n being 14, 16 or

18 and X being Cl or Br, and C16H33Pyr+X2, with Pyr being

pyridyl and X being Cl or Br. Both the mean pore diameter

and the d-spacing obtained from X-ray diffraction increased

with increasing alkyl chain length of the first series of

surfactants, which is in accordance with expectations since,

as mentioned above, the pore diameter should reflect twice the

length of the hydrophobic tail of the surfactant. The overall

highest quality product was obtained with hexadecylpyridi-

nium chloride as the surfactant.

Recently a cationic Gemini surfactant, (C12H25N+(CH3)2–

(CH2)2–N+(CH3)2 C12H25) 2Br2, was used for making

mesoporous silica with cubic geometry (space group Ia3d)

from sodium silicate.16 Gemini surfactants are known to self-

assemble at much lower concentration than their monomeric

counterparts so they should be interesting as structure-

directing agents. The material obtained was of high quality

as evidenced by small angle X-ray diffraction determination of

specific BET surface area (991 m2 g21) and by assessment of

the pore size distribution. It would have been interesting to

compare the results with corresponding experiments using the

monomeric surfactant, i.e., C12H25N+(CH3)3 Br2.

It is noteworthy that the surfactant counterion seems not to

be of importance. Similar structures are reported for chloride

and bromide as counterion. This is somewhat surprising since

the physical chemistry of a self-assembled cationic surfactant

in water is usually strongly influenced by the choice of

counterion. Bromide is for instance much more strongly bound

than chloride or acetate to micelles and monolayers of cationic

surfactants, which leads to differences in critical micelle

concentration at lower concentration and in phase behavior

at higher concentration. The reason why the surfactant

counterion is not important in the formation of mesoporous

silica is probably that formation of the mesoporous material

involves a cooperative interaction between the surfactant

cation and a growing silicate prepolymer, see the discussion

of the mechanism of formation below. Thus, the type of

counterion is not important since it is being replaced by silicate,

which then serves as an inorganic, polymeric counterion for

the surfactant.

Nonionic surfactants containing polyoxyethylene chains,

either block copolymers of the polyoxyethylene–polyoxy-

propylene–polyoxyethylene type (often referred to as

EO–PO–EO polymers) or fatty alcohol ethoxylates, CnH2n+1–

(EO)m, with n typically being 12–18 and m being 5–8, have

been widely used as structure-directing agents in the formation

of mesoporous silica using TEOS or other alkoxides as the

silica source.4,5,8,37,38 These types of nonionic surfactants

have also been used in the synthesis of mesoporous silica

from inorganic silica sources and both hexagonal and cubic

structures have been prepared.12–14,18–21,24–26 The nonionic

surfactants have most often been used under acidic to neutral

conditions. The hydrophilic–lipophilic balance of these surfac-

tants is strongly affected by the temperature, which makes it

possible to govern the type of liquid crystalline structure

formed in water solutions by temperature adjustments.

Kipkemboi et al. have demonstated how the mesostructure

obtained with EO–PO–EO block copolymers as structure-

directing agents is influenced by the temperature and by the

hydrophilic domain length.39 An increase in temperature leads

to dehydration of the polyoxyethylene chains, causing a

reduction of the size of this block. This results in a change

of the spontaneous curvature of the surfactant film. Increased

temperature will for low surfactant concentrations lead to a

transition from less elongated to more elongated micelles and

for high surfactant concentration to a transition from micellar

cubic, via hexagonal and lamellar, to bicontinuous cubic liquid

crystals. One would expect that the temperature-controlled

mode of self-assembly in solution would be transferred to the

corresponding structures in the mesoporous material. This has

been found to be the case for TEOS as the silica source.39

Ongoing investigations in our laboratory will show if this also

holds true when inorganic silicate solutions are used as the

starting material.

Shrinkage of the hydrophilic domains of the templating

surfactant means that the volume where the silica wall is being

formed is reduced. This would be expected to lead to thinner

walls of the mesoporous material. It is now experimentally

verified that this is true, i.e., an increase in temperature for

222 | Soft Matter, 2005, 1, 219–226 This journal is � The Royal Society of Chemistry 2005

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online

systems based on surfactants containing polyoxyethylene

chains leads to a decrease of the wall thickness. The mesopore

diameter is only slightly affected by variations in temperature,

however, because it is governed by the size of the hydrophobic

domain, i.e., the PO block, which is not much influenced by

the temperature.39

4. Formation mechanisms

Already in the first reports by Beck et al. on surfactant-

directed formation of mesoporous silica two different methods

for templating were presented, one starting from a micellar

solution of surfactant and the other from a concentrated

surfactant solution that forms a liquid crystalline phase.1 The

interaction between the silica and the surfactant has been

studied in several reports and different mechanisms have been

proposed. Huo et al. presented four different synthesis routes

(S+I2), (S2I+), (S+X2I+) and (S2M+I2), where S 5 surfactant

(+ cationic, 2 anionic), I 5 soluble inorganic (silica), X 5

halogen ion, and M 5 metal cation. The models are based on

electrostatic interactions and to obtain these conditions the

reaction must occur either at very low pH, where the silica

surface is positively charged, or at pH 5 7–10, where the

surface is negatively charged.40 To these four routes a fifth one

was added, concerning nonionic surfactants (S0I0).41,42 The use

of such surfactants under acid conditions has subsequently

been reported with the proposed mechanism (S0H+)(X2I+).43

Most of the experimental studies of the mechanisms have

been performed with alkoxides as the silica source but the

general concepts should be relevant also for inorganic silica

as the starting material. This statement is supported by a

study where mesoporous silica prepared from an alkoxide

as the starting material was compared with the material

synthesized from an inorganic silica source and found to be

almost identical.44

4.1 Reaction in a micellar solution

The method of performing the templating in a micellar

solution with a relatively low concentration of surfactant is

the only one reported for synthesis with sodium silicate as the

silica source. The principle of how the inorganic precursor

interacts with the organic template, i.e., self-assembled

surfactants, to form the ordered mesoporous silica has

attracted substantial interest. Several mechanistic explanations

have been put forward; however, none of them presents a

definitive answer.8 In 1995 Firouzi et al. presented a

cooperative self-assembly mechanism based on experiments

performed with down to 0.5 wt.% surfactant, still resulting in

long-range organization.45 The concept of cooperative self-

assembly involving silicate prepolymers and cationic surfac-

tants has been further elucidated by Frasch et al.34 Based on

in situ studies using 29Si–NMR to follow the change in the

nature of the silicate species and fluorescence probing to

determine the fraction of micelle-bound counterions (in

particular bromide) that are exchanged by anions, such as

hydroxyl or silicate, the picture illustrated in the left part of

Fig. 2 emerged.

A silicate solution is added to a micellar solution of the

cationic surfactant. Around 80% of the cationic charges at

the micellar surface are initially compensated by bound

counterions (a typical value for bromide as counterion).

Since bromide ions, being highly polarizable, interact strongly

with the micelle, only a small fraction of these are replaced by

a low-molecular weight silicate ion. Addition of acid induces

oligomerization of the silicate and surfactant cations will

interact with the growing prepolymer. With time it is likely

that smaller micelles will form around the silicate polymer.

Such polymer-induced micellization is common and well

understood in organic polymer–surfactant systems, and the

concentration at which such aggregates form, the critical

association concentration, is often one to two orders of

magnitude lower than the critical micelle concentration of the

surfactant in a polymer-free solution. It is likely that the

number of silicate-bound aggregates will grow at the expense

of the number of free micelles and eventually almost all

surfactant will be tied up with silicate species. The siliceous

polymer continues to grow and at some point, when the

polymer has become large enough and when there is charge

Fig. 2 Simplified and schematic representation of the mechanisms

proposed for the formation of mesoporous hexagonal silica from a

micellar solution. The left path represents the use of monomeric silica

as the starting material and the right path represents the use of a silica

source, which is partly condensed. (Redrawn from ref. 8)

This journal is � The Royal Society of Chemistry 2005 Soft Matter, 2005, 1, 219–226 | 223

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online

neutrality in the system, the complex precipitates out. The self-

assembling characteristics of the surfactants, governed by the

surfactant structure and by its interaction with the silicate

polyanion, are responsible for the elaborate structure of the

organic–inorganic composite material. As discussed above in

Section 3 on surfactants, there are powerful tools for fine-

tuning the structure by proper choice of surfactant and/or

reaction conditions.

The above mechanism relates to cationic surfactants, which

will interact with the siliceous polymer by electrostatic

interactions. Expressed differently, the polysilicate will act as

a polycounterion for the surfactant. The mechanistic explana-

tion must be modified when nonionic surfactants are used as

structure-directing agents. Charge neutralization may then be

attained by reducing the pH to such low values that virtually

all remaining Si–O2 groups become protonated and the

surfactant–polymer interaction would then primarily be due

to hydrogen bonding between silanol groups of the inorganic

polymer and ether oxygen atoms of the nonionic surfactant. It

is known from several other studies that there is strong

interaction between surfactants containing polyoxyethylene

segments and silica surfaces,46–49 as well as between such

segments and silica in solution.50 However, it is not clear if the

attraction only can be related to hydrogen bonding. It has been

claimed that in water such an interaction would provide little,

if any, net driving force and the hydrogen bond effect may

therefore be overwhelmed by a hydrophobic attraction.51

Regardless of the nature of the attractive forces between the

nonionic surfactant and the polysilicate it is likely that

the concepts of surfactant–polymer interaction and charge

neutralization hold true also for the case of nonionic

surfactants and the mechanism is supported by a recent study,

in which time-resolved in situ 1H–NMR and transmission

electron microscopy were used to study the formation

of mesoporous silica from TMOS as the silica source and an

EO–PO–EO copolymer as the structure-directing agent.51 The

above mechanistic explanation does not fully comply with the

synthesis performed with nonionics at near neutral conditions

described in Section 2.1. When using inorganic silica as the

starting material there will be a distribution of Si(Qn) species

were the silicon atoms have varying degrees of condensation as

discussed in Section 2. It is then reasonable to believe that the

end product will be less ordered as schematically shown in

right part of Fig. 2.

4.2 Reaction in a liquid crystalline phase

The procedure of using a preformed liquid crystalline phase as

a template for the formation of mesoporous silica was first

presented by Attard et al. and the method is sometimes

referred to as ‘‘the direct templating method’’. To an aqueous

solution containing around 50 wt.% surfactant, TMOS was

added as the silica source. Hydrolysis of the alkoxide generated

methanol, which destroyed the liquid crystalline phase.

Removal of the methanol through gentle vacuum distillation

recreated the phase and an ordered hexagonal mesoporous

structure was obtained.37 The mechanism involved may seem

straight forward and the technique is sometimes referred to as

nanocasting.52 It has mainly been used in combination with

nonionic surfactants as templating agents. One interesting

aspect of the direct templating method is that it does not

require a specific surfactant–silicate interaction, as does the

method of synthesis in a micellar solution. An illustration of

this is the formation of mesoporous alumina with bicontinuous

cubic symmetry, synthesized by direct casting from the

corresponding liquid crystalline phase made up of the nonionic

surfactant monoolein and water.53 Monoolein is uncharged

and does not contain polyoxyethylene chains so it is unlikely to

interact specifically with the dissolved aluminate polyions. The

procedure is not always without complications, however.

Sometimes the polymerisation process affects the phase

behavior such that the composition gradually moves away

from the liquid crystalline phase, which is the intended

template structure. For example, Blin et al. prepared silica

materials in a pre-formed hexagonal liquid crystalline phase

made from nonionic surfactants of alcohol ethoxylate type.54

The reaction gave a disordered worm-like mesoporous

material. It was suggested that the generated methanol and

the interaction between the surfactant and the formed silica

induced displacement of the channels. When the surfactant

concentration was reduced to a value below that required for

the hexagonal liquid crystal to form, the reaction lead to the

formation of a well-ordered mesoporous material. Similar

observations were made by Klotz et al.11 and Alberius et al.,55

who devised predictive routes taking into account the effects of

solvent evaporation on the phase behaviour of the surfactant

templated mesostructure.

The direct templating method seems not to have been used

for the synthesis of mesoporous silica from an inorganic silica

source. The procedure should be well suited for this purpose,

however, since with inorganic silica as the starting material

there will be no generation of alcohol during the course of the

reaction. Thus, the above-mentioned problem of alcohol-

induced distortion of the liquid crystalline phase will be

eliminated and the formation may proceed according to the

schematic shown in Fig. 3.

5. Effect of anions

When using cationic surfactants or acids it is very important to

consider the choice of the counterion. Different anions affect

the resulting mesoporous structure in different ways.

The Hofmeister series is very central to studies of the effect

of anions. The series is based on the anions’ effect on the

solubility of proteins and it orders the ions from those

reducing the solubility the most to those increasing it the most:

SO422, HPO4

22, OH2, F2, HCOO2, Cl2, Br2, NO32, I2,

SCN2, ClO42

The ions to the left of Cl2 are usually small, with a relatively

low degree of polarizability. They have high electric fields at

short distances and bind hard to their hydrated water. These

ions are known to reduce the solubility of proteins. The ions to

the right have the opposite characteristics.

Leontidis has written a review concerning the current

understanding of the mechanisms of the anion effect on the

formation of mesoporous materials. He concludes that

the understanding is limited because of the complexity of the

systems involved. He did, however, arrive at some conclusions.

224 | Soft Matter, 2005, 1, 219–226 This journal is � The Royal Society of Chemistry 2005

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online

When using cationic surfactants faster hydrolysis is achieved

with the ions to the right in the series, probably caused by a

more pronounced lowering of the charge repulsion, which

creates larger and more rod-like micelles. The opposite effect is

reported for the nonionic surfactants of EO–PO–EO type, and

this is probably due to the fact that the anions to the left in the

series lowers the cloud point through dehydration, which also

creates larger and more rod-like micelles.56

6. Outlook

The field of mesoporous silica has been studied for about

15 years but is still an area attracting a lot of attention. The

mechanisms involved are not fully understood and several

issues related to the syntheses are still poorly investigated.

There are some aspects that are of special relevance to the

theme of this review.

It is very important to determine if there is a substantial

difference in the mechanisms involved when using inorganic

silica compared to alkoxysilanes as the silica source. If the

differences are small, a lot of the work performed with

alkoxysilanes will be useful when developing procedures for

using inorganic silica. The interest in using inorganic silica as

the silica source has increased in recent years but the procedure

is still in its infancy. The published reports focus on the use of

micellar solutions as the templating agent and the syntheses,

rather than the mechanisms, have been the main objective. To

achieve a better understanding about the formation of the

mesoporous structure is essential in order to be able to control

the result.

Liquid crystalline templating with inorganic silica seems

not to have been reported. A similar procedure to the one

used with alkoxysilanes might be possible. This could be an

interesting way of producing mesoporous silica and, since no

alcohol is produced during the reaction, it seems as better

control of the structure could be achieved. However, there will

certainly be a greater need to control the Si(Qn) distribution in

such syntheses, and perhaps mixed silica sources will be found

to be useful for the purpose.

Control of the mesoporous structure and at the same time

control of the size and the size distribution of the particles

would be desirable for many applications. One example of this

is the report from Kosuge et al..21 This was performed within

a very limited range of synthesis conditions, however an

alternative way of achieving such control might be through the

use of cubosomes or hexosomes. Cubosomes and hexosomes

are cubic and hexagonal liquid crystalline phases respectively,

dispersed in water. If they are produced in a manner that

creates a narrow size distribution this might be a way of

controlling the size of the templated mesoporous silica

particles. This approach seems limited in scope, however,

since only few surfactants are known to form stable water

dispersions of cubosomes or hexosomes.

Acknowledgements

AB wishes to acknowledge financial support from Eka

Chemicals AB and the Knowledge Foundation through the

research school YPK. AECP thanks the Swedish Research

Council for a senior researcher grant.

References

1 J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge,K. D. Schmitt, C. T. W. Chu, D. H. Olson and E. W. Sheppard,et al., J. Am. Chem. Soc., 1992, 114, 10834–10843.

2 T. Yanagisawa, T. Shimizu, K. Kuroda and C. Kato, Bull. Chem.Soc. Jpn., 1990, 63, 988–992.

3 S. Inagaki, Y. Fukushima and K. Kuroda, J. Chem. Soc., Chem.Commun., 1993, 680–682.

4 K. Holmberg, J. Colloid Interface Sci., 2004, 274, 355–365.5 A. E. C. Palmqvist, Curr. Opin. Colloid Interface Sci., 2003, 8,

145–155.6 M. Antonietti, Curr. Opin. Colloid Interface Sci., 2001, 6, 244–248.7 J. Y. M. C. P. Ying and M. S. Wong, Angew. Chem. Int. Ed., 1999,

38, 56–77.8 J. Patarin, B. Lebeau and R. Zana, Curr. Opin. Colloid Interface

Sci., 2002, 7, 107–115.9 R. K. Iler, The Chemistry of Silica: Solubility, Polymerization,

Colloid and Surface Properties and Biochemistry; Wiley, New York,1979.

10 Y. M. Setoguchi, Y. Teraoka, I. Moriguchi, S. Kagawa,N. Tomonaga, A. Yasutake and J. Izumi, J. Porous Mater.,1997, 4, 129–134.

11 M. Klotz, A. Ayral, C. Guizard and L. Cot, J. Mater. Chem., 2000,10, 663–669.

12 S.-S. Kim, T. R. Pauly and T. J. Pinnavaia, Chem. Commun., 2000,1661–1662.

13 S.-S. Kim, T. R. Pauly and T. J. Pinnavaia, Chem. Commun., 2000,835–836.

14 S. S. Kim, A. Karkamkar, T. J. Pinnavaia, M. Kruk andM. Jaroniec, J. Phys. Chem. B, 2001, 105, 7663–7670.

15 S. Han, W. Hou, W. Dang, J. Xu, J. Hu and D. Li, Colloid Polym.Sci., 2004, 282, 761–765.

16 J. Xu, S. Han, W. Hou, W. Dang and X. Yan, Colloids Surf., A,2004, 248, 75–78.

17 S. Han, W. Hou, J. Xu and Z. Li, Colloid Polym. Sci., 2004, 282,1286–1291.

18 J. M. Kim, Y. Sakamoto, Y. K. Hwang, Y.-U. Kwon, O. Terasaki,S.-E. Park and G. D. Stucky, J. Phys. Chem. B, 2002, 106,2552–2558.

19 J. M. Kim and G. D. Stucky, Chem. Commun., 2000, 1159–1160.20 K. K. N. Kosuge and M. Takemori, Chem. Mater., 2004, 16,

4181–4186.21 K. S. T. Kosuge, N. Kikukawa and M. Takemori, Chem. Mater.,

2004, 16, 899–905.

Fig. 3 Suggested schematic description of the direct templating method for synthesis of mesoporous silica from an inorganic silica source.

This journal is � The Royal Society of Chemistry 2005 Soft Matter, 2005, 1, 219–226 | 225

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online

22 M.-C. Chao, H.-P. Lin, D.-S. Wang and C.-Y. Mou, Chem. Lett.,2004, 33, 374–375.

23 M.-C. Chao, D.-S. Wang, H.-P. Lin and C.-Y. Mou, J. Mater.Chem., 2003, 13, 2853–2854.

24 L. Sierra, B. Lopez and J. L. Guth, Microporous MesoporousMater., 2000, 39, 519–527.

25 L. Sierra and J.-L. Guth, Microporous Mesoporous Mater., 1999,27, 243–253.

26 C. Boissiere, A. Larbot and E. Prouzet, Chem. Mater., 2000, 12,1937–1940.

27 S. O’Brien, R. J. Francis, A. Fogg, D. O’Hare, N. Okazaki andK. Kuroda, Chem. Mater., 1999, 11, 1822–1832.

28 T. Kimura, D. Itoh, N. Okazaki, M. Kaneda, Y. Sakamoto,O. Terasaki, Y. Sugahara and K. Kuroda, Langmuir, 2000, 16,7624–7628.

29 T. Kimura, D. Itoh, T. Shigeno and K. Kuroda, Langmuir, 2002,18, 9574–9577.

30 T. Kimura, T. Kamata, M. Fuziwara, Y. Takano, M. Kaneda,Y. Sakamoto, O. Terqasaki, Y. Sugahara and K. Kuroda, Angew.Chem. Int. Ed., 2000, 39, 3855–3859.

31 K. Kuroda, Stud. Surf. Sci. Catal., 2004, 148, 73–108.32 S. Che, A. E. Garcia-Bennett, T. Yokoi, K. Sakamoto, H. Kunieda,

O. Terasaki and T. Tatsumi, Nat. Mater., 2003, 2, 801–805.33 B. L. Newalkar, H. Katsuki and S. Komarneni, Microporous

Mesoporous Mater., 2004, 73, 161–170.34 J. Frasch, B. Lebeau, M. Soulard, J. Patarin and R. Zana,

Langmuir, 2000, 16, 9049–9057.35 S. Shio, A. Kimura, M. Yamaguchi, K. Yoshida and K. Kuroda,

Chem. Commun., 1998, 2461–2462.36 S.-E. Park, D. S. Kim, J.-S. Chang and W. Y. Kim, Catal. Today,

1998, 44, 301–308.37 G. S. Attard, J. C. Glyde and C. G. Goltner, Nature, 1995, 378,

366–368.38 D. Zhao, Q. Huo, J. Feng, B. F. Chmelka and G. D. Stucky, J. Am.

Chem. Soc., 1998, 120, 6024–6036.

39 P. Kipkemboi, A. Fogden, V. Alfredsson and K. Flodstrom,Langmuir, 2001, 17, 5398–5402.

40 Q. Huo, D. I. Margolese, U. Ciesla, D. G. Demuth, P. Feng,T. E. Gier, P. Sieger, A. Firouzi, B. F. Chmelka, F. Schuth andG. D. Stucky, Chem. Mater., 1994, 6, 1176–1191.

41 S. A. Bagshaw, E. Prouzet and T. J. Pinnavaia, Science, 1995, 269,1242–1244.

42 P. T. Tanev and T. J. Pinnavaia, Science, 1995, 267, 865–867.43 D. Zhao, J. Feng, Q. Huo, N. Melosh, G. H. Frederickson,

B. F. Chmelka and G. D. Stucky, Science, 1998, 279, 548–552.44 C. Boissiere, A. Larbot, A. van der Lee, P. J. Kooyman and

E. Prouzet, Chem. Mater., 2000, 12, 2902–2913.45 A. Firouzi, D. Kumar, L. M. Bull, T. Besier, P. Sieger, Q. Huo,

S. A. Walker, J. A. Zasadzinski, C. Glinka, J. Nicol, D. I. Margolese,G. D. Stucky and B. F. Chmelka, Science, 1995, 267, 1138–1143.

46 G. J. Howard and P. McConnell, J. Phys. Chem., 1967, 71,2974–2981.

47 J. C. Dijt, M. A. C. Stuart, J. E. Hofman and G. J. Fleer, J. ColloidsSurf., 1990, 51, 141–158.

48 M. Malmsten, P. Linse and T. Cosgrove, Macromolecules, 1992,25, 2474–2481.

49 L. M. Grant, F. Tiberg and W. A. Ducker, J. Phys. Chem. B, 1998,102, 4288–4294.

50 R. Takahashi, K. Nakanishi and N. Soga, J. Sol-Gel Sci. Technol.,2000, 17, 7–18.

51 K. Flodstrom, H. Wennerstrom and V. Alfredsson, Langmuir,2004, 20, 680–688.

52 S. Polarz and M. Antonietti, Chem. Commun., 2002, 2593–2604.53 N. Cruise, K. Jansson and K. Holmberg, J. Colloid Interface Sci.,

2001, 241, 527–529.54 J. L. Blin, A. Leonard and B. L. Su, Chem. Mater., 2001, 13,

3542–3553.55 P. C. A. Alberius, K. L. Frindell, R. C. Hayward, E. J. Kramer,

G. D. Stucky and B. F. Chmelka, Chem. Mater., 2002, 14,3284–3294.

56 E. Leontidis, Curr. Opin. Colloid Interface Sci., 2002, 7, 81–91.

226 | Soft Matter, 2005, 1, 219–226 This journal is � The Royal Society of Chemistry 2005

Publ

ishe

d on

25

July

200

5. D

ownl

oade

d by

Uni

vers

ity o

f M

iam

i on

20/0

9/20

13 1

4:10

:08.

View Article Online