INTRODUCTION TO GEL-

SCIENCE AND
TECHNOLOGY
Zuhroni Ali Fikri
While it was studied earlier, sol-gel chemistry has been investigated extensively since the mid-
1970s, when sol-gel reactions were shown to produce a variety of inorganic networks that can be
formed from metal alkoxide solutions. Through sol-gel processing, homogeneous, high-purity
inorganic oxide glasses can be made at ambient temperatures rather than at the very high
temperatures required for conventional formation of glasses. Various products, such as molded,
spun fibers, thin films, molecular cages and xerogels have been developed for utility in such
areas as gas separations, elastomers, coatings, and laminates. It is through the inorganic
compound incorporation into organic polymers that a wide variety of property modifications can
be achieved
The sol-gel chemical process is self-described in the definition of a sol, a gel, and a summary of
the process in which a sol evolves into a gel. A sol is a colloidal dispersion of small particles in a
liquid; a gel is usually a substance composed of a continuous network encompassing a
continuous liquid phase. Sol-gel reactions promote the growth of colloidal particles (sol) and
their subsequent network formation (gel) through the hydrolysis and condensation reactions of
inorganic alkoxide monomers. The precursors for synthesizing these colloids consist of a metal
or metalloid element surrounded by carious reactive ligands. Metal alkoxides are most popular
because they react readily with water. The most widely used metal alkoxides are the
alkoxysilanes, such as tetramethoxysilane (TMOS) and tetraethoxysilane (TEOS). However,
other alkoxides such as aluminates, titanates, zirconates, and borates are also commonly used in
the sol-gel process, either alone or in combination with other alkoxides such as TEOS.
To get more understanding in the sol-gel chemical process, the sol-gel reaction must be
examined. Sol-gel reactions are a series of hydrolysis and condensation reactions of an
alkoxysilane, which proceed according to the reaction scheme shown in Figure 1. Hydrolysis is
initiated with the addition of water to the precursor (TEOS) solution under acidic, neutral, or
basic condition.

Figure 1 Sol-gel general reaction scheme

In the first step, hydrolysis, a silanol group is generated. As only acid catalysis is used in this
research, discussions of the sol-gel reactions will refer to acid catalysis only. A positive charge
develops on the alkoxysilane through the attack of an acid catalyst. The alkoxysilane then is
hydrolyzed in an Sn2-type reaction, forming a silanol moiety. The mechanism for this first step is
shown in Figure 2. TEOS isnt soluble in water. Therefore, hydrolysis is promoted by the
addition of organic cosolvents, such as alcohols that aid in the thorough mixing of the alkoxide
molecules with water molecules in solution.

Figure 2 Hydrolysis mechanism for a tetraalkoxysilane.

In this second step, the silanol group can undergo condensation with either an alkoxide or
another silanol group which results in the formation of strong siloxane linkages and produces
either alcohol (ROH) or water in situ. As the number of Si-O-Si bridges increases, the siloxane
particles can aggregate into a sol. This, as shown in Figure 2, consists of the formation of small
silicate clusters dispersed in solution. Gel formation occurs when the sol particles undergo
sufficient condensation reactions such that a network (a gel) is formed, trapping the aqueous and
alcohol by-products. After gel-network formation is complete, the by-products are removed by
heat and vacuum, yielding a vitrified and densified glass network. Below the more detailed
explanation about the process of sol gel reaction would be explained

Hydrolysis
Hydrolysis is the deprotonation of a solvated metal cation; it consists in the loss of a proton by
one or more of the water molecules that surrounds the metal M in the first solvatation shell.
During hydrolysis, the alkoxy groups (OR) are replaced either by hydroxo ligands (OH) or oxo
ligands (O). The more detailed explanation could be seen below
Formation of Hydroxo ligands.
In the first hydrolysis reaction, a first hydroxo ligand substitutes an alkoxy group, as can be seen
in the equation below

This reaction is generally followed by a succession of other similar substitutions.

The mechanism of hydrolysis, illustrated in Figure 5, consist in a nucleophilic substitution in
which a water molecule attacks the alkoxide. Figure 3, consist in a nucleophilic substitution in
which a water molecule attacks the alkoxide. It is independent of the polymerization of the
alcoholate

Figure 3 Nucleophilic substitution in formation of hydroxo ligands

Moreover, experimental studies of its kinetics have determined that if hydrolysis is a very fast
reaction for most cations, it is much slower for Si. This difference, due to the chemical
characteristics of the corresponding alkoxides, is particularly well explained by the partial charge
model.

Condensation
Two mononuclear complexes of M. each comprising only one metal atom M, can react with one
another in a polymerization reaction in order to form a polynuclear complex consisting of two
metal atoms. Such a reaction, also called condensation, can -depending on the metal and if the
conditions are right- keep on going so as to produce bigger polynuclear species.
Polymerization generally occurs if at least one hydroxo ligand (OH) is bounded to the cation M.
This hydroxo ligand belongs either to an aquo-hydroxo complex of the type [M-(OH)(H 2O)N-1](z-
1)+
, or to oxo-hydroxo complex, [M-(OH)(O)N-1](z-2N+1)+. We often simply write it as M-OH. As
indicated in equations (2.4-1) and (2.4-6), those OH ligands are obtained by addition to the
solution either of a base, in the case of metals forming acidic oxides, or of an acid, for metals
forming basic oxide. The condensation occurring afterwards is responsible for the formation of
one of two types of bridges between the two metal atoms.
Condensation by olation
The ftrst step of any condensation reaction always include the construction of an "ol" bridge in
which an hydroxo ligand is caught between the two metal atoms [5]. For the low charge cations,
this is done through a dissociative SN1 mechanism:
A nucleophilic adition reaction (AN) is also possible when the coordination number of metal can
be increased as for [Al(OH)4]-. In this case

As for the transition elements, the mechanism is an associative SN2.

The condensation of the solution continues until a complex [M(OH)h(OH2)N-h](z-h)+ in which h<z
cannot undertake any more condensation, or until, as a result of unlimited polymerization of
[M(OH)h(OH2)N-h](z-h)+, a precipitate appears. Table 2.4-2 gathers the different types of ol
bridges that exist.
Condensation by oxolation
If there is condensation by oxolation, an "ol" bridge is first established between the two metal
atoms before transforming into an "oxo" bridge in an S N2 mechanism. In the intermediate
complex of this mechanism, the maximum coordination number of the metal M is satisfied. As a
consequence, both acids and bases can catalyze this reaction.
During the basic catalysis of the formation of oxo bridges, an OH - anion attacks the partially
charged hydrogen atom, H+, of an hydroxo ligand belonging to the metal complex. This
increases the negative partial charge of the oxygen atom of this ligand that consequently,
becomes more nucleophilic and binds to another solvated hydroxyl group. Two water molecules,
together with an OH- anion, then separates from the complex which is now composed of an
oxo bridge. Similarly, during the acidic catalysis of condensation, an H + cation attacks the
oxygen atom of an hydroxo ligand belonging to the metal complex. Those scheme is presented
by Figure 4.
Hence, the newly formed corresponding H2O ligand gains a positive partial charge +, and
moves toward another OH ligand, thus forming an intermediate complex. The H 3O+ ion then
leaves while the complex is now formed of an oxo bridge. A direct nucleophilic addition is
also possible when the coordination number of a metal can be increased.
 Figure 4 Reaction scheme of oxo-bridge formation

Gelation
Gelation is a process according to which a sol, or a solution, transforms to a gel. It consists of
stablishing links between the sol particles, or the solution molecules, so as to form a 3-
dimensional solid network. However, this is very different from the solidification of a melt, since
the solid structure remains completely impregnated with the liquid of the sol, or solution. This is
also a very general type of transformation and it is often considered that any solid material can
be transformed to a gel.
1. Gels
Inorganic gels constitute a group of materials which cover a large variety of structures. However
any gel is composed of a solid network and a liquid matrix when in the wet state. Quite special
properties result from this double liquid-solid composition. Some of them are reversible
transformations which can be described within equilibrium thermodynamics, such as swelling or
shrinkage inside the liquid in some cases. Other properties are irreversible transformations such
as syneresis, aging and drying. Gels also are unique materials which have very interesting
properties in the wet state as well as in the dry state, in particular new mechanical, optical,
thermal and sound conduction properties. The present chapter addresses successively the
structure of these materials, as well as their properties.
Gel Classification
1.1. Polymeric Gels
1.1.1. Organic Gels
The element most prone to build extended polymeric molecules which are not hydrolyzed by
water, is carbon. Organic gels which comprise a polymeric carbon backbone, constitute a large
class. However, hybrid organic-inorganic gels are very new materials which constitute an
exception. Moreover, it is often useful to refer to the large extent of knowledge gathered on
organic gels, to have an insight in the structures and properties of the inorganic gels.
 1.1.2. Silica Gels
The second element most prone to polymerization is silicon. The covalent nature of the Si-O-Si
bonds is significant, though not as marked as that of C-C bonds. Actually, silica gels which are
termed polymeric have a polymer molecular structure which is not as well defined as in the case
of carbon, so that ceramists have adopted another more practical convention to define polymeric
gels in a "large sense". A gel is here considered as polymeric when the solid backbone is
composed of particles with a size below a given dimension: e.g., 1.5 nm for silica [9].
1.1.3. Borate Gels
Another element easily forming polymeric gels is boron. A possible structural model for borate
gels is based on an infinite network with B-O-B bonds in which tetracoordinated and tri-
coordinated boron alternate. All the non-bonding oxygen are terminated with an alkoxide group.
A structure of this type prevails in the alkali-borate systems xR 2O.(1-x)B2O3 where x must
exceed a minimum value which is 0.2 in alcohol, or 0.15 in tetrahydrofuran (THF)
1.2. Colloidal gels
Gels which can be classified as colloidal, according to the size of the solid particles which makes
the tridimensional network, can differ from each other by:
- The size, shape and crystallographic structure of the colloidal solid particles which
constitute them
- The connectivity between the particles: a gel which macroscopically appears as quite
fibrous may be composed of spherical particles linked in a very linear fashion
- The type of linkage between the particles

1.2.1. Pore texture

One of the best classical way to characterize the solid network in a colloidal gel consists in
describing its porosity. The macroscopic data usually available for this task are:
- The apparent density a : it can be measured for instance by immersion in mercury, which
does not penetrate into the micropores.
- The true density t : it can be measured in a fluid such as helium which penetrates into the
smallest open micropores
- The specific area Sa determined by adsorption of nitrogen by the Brunauer Emmett and
Telling (BET) method.

V por =1 1
a t

r par =3 ( t S a )1

As the pore radius rpor, it can be determined if a shape is assumed. For cylindrical pores this is

r por =2 ( 1
a t ) S a
1 1
The pore volume percent can be extremely high in a gel, up to 98% in silicate gels hydrolyzed in
an excess of solvent and dried in supercritical conditions. The pore structure depends in a large
extent on the fabrication method. Additives named "drying chemical control additives" (DCCA)
can be used to modify it. In the case of alumino-silica gels made from sodium silicate and
sodium aluminate solutions for instance, Snell1 has shown that a sodium ions content> 0.5% by
mass had a marked effect on the pore structure.
The pore size does not only depend on the size of the particles but on the aspect ratio, that is to
say the ratio (Length/thickness) for a section of the solid network located between two branching
point. For instance with TiO2, an aspect ratio = 35 has been reported when the precursor was Ti
ethoxide and = 150 with isopropoxide. The pore size statistical distribution depends on the
aggregation hierarchy of primary solid units in progressively bigger units. The smalles pores are
associated with the smallest primary particles, and so on. A finer description of a network texture
usually requires techniques complementary to the BET method, such as scanning or transmission
electron microscopy, small angles X-rays scattering (SAXS) or light scattering.
The results of all these techniques can make it possible to propose a geometrical model for the
pore network and to compare the resulting specific area and pore volume with the experimental
data measured on a real gel.

Gels Phase Transformation

Aging wet gels

A wet gel keeps transforming when it is stored for a long time in its liquid and this evolution is
known as aging. Two fundamental types of transformations underlie aging evolution.
Chemical evolution during aging
A gel has no reason to be composed of the most stable chemical species when it is formed for the
first time, as a result of the hydrolysis-polymerization reactions in liquid medium. A critical
supersaturation of intermediate chemical complexes leading to gelation, may well be due
intermediate species which are not the most stable. The anions in the solution can themselves
have a marked influence as they can be slowly expelled from a gel by re-dissolution.
No universal mechanism can explain the aging evolution of a gel. With Al precursors in acidic
conditions for instance, a pseudoboehmite monohydroxide gel network easily forms, which
comprises ol bridges and oxo bridges. However this boehmite is metastable. Aging, especially in
basic conditions, helps to transform the monohydroxide AIO(OH) gel to a more stable
trihydroxide AI(OH)3 structure. Anything which prevents the deprotonation of hydroxo groups in
monohydroxide, such as the adsorption of glycerol or ethanediol, prevents its aging
transformation to bayerite.

1 Brinker, C. J., & Harrington, M. S. (1981). Sol-gel derived antireflective coatings for silicon. Solar Energy
Materials, 5(2), 159-172.
A similar aging evolution occurs in the case of tungsten oxide gels [39] which are known to be
unstable and crystallize after a few days. They can also be stabilized by foreign colloids such as
MoO3 Even silicic acid gels, which always show an amorphous X-ray diffraction pattern, have a
tendency to crystallize after aging in dehydrating conditions. The reverse can also be true:
crystalline zirconia immersed in water undergoes a surface hydrolysis modification yielding a
thin gel-like coating
Physical evolution during aging
The origin of the physical aging evolution is the same as in syneresis; this is the high specific
area of the solid gel network which tends to decrease slowly.
In silica gels at low pH the polymer species which make the gel network are not very soluble, so
the network is "frozen in"; no modification occurs during aging. However, at higher pH,
colloidal silica gel slowly transform to coarser structures, as indicated by the evolution of water
vapor adsorption desorption isotherms of these gels (Figure 5). These isotherms show a
displacement of the hysteresis loops towards higher pore sizes. That is to say both the pores and
the gel network becomes coarser. The specific surface area of a gel decreases as the aging time
increases.
Sometimes, completely different textures develop, as in gelatinous ZrO 2 precipitates made from
Zr propoxide in cyclohexane, where coalescence of the colloidal particles into rods, with a
periodic swelling along the length of a rod, occurs during aging

Figure 5 Adsorption-desorption isotherms on silica gels after various aging times.

Drying Gels
There are several mechanism of the drying gels. In this part, only one mechanism will be
explained.

Capillary mechanism
The capillary mechanism explains well the reproducible adsorption hysteresis curves of water in
silica gels. This mechanism has been summarized as follows (Figure 6)
(1) Evaporation creates a liquid vapor meniscus at the exit of pores in the gel
(2) This induces a hydrostatic tension in the liquid, which is balanced by an axial
compression on the solid
(3) The latter compression makes the gel shrink
(4) As a result of shrinkage, more liquid is fed to the menisci at the exit of the gel pores,
where it is evaporated and so on.

Figure 6 Capillary mechanism

One consequence of capillarity is also the existence of a hysteresis loop on adsorption-desorption
isotherms, for instance of silica gels (Figure 7). This comes from the fact that the equilibrium
vapor pressure pv above a meniscus of principal radii r1 and r2, is:

( (
pv = po exp
1 1
)
+ V RT
r1 r 2 m )
Where po is the vapor pressure above a planar surface and Vm is the molar volume of liquid.
Branch IC in Figure 7a corresponds to the initial drying of the gel. At point C, the meniscus is as
in Figure 7b. Further on, it penetrates inside the gel pores which stop shrinking. Along CD, the
meniscus is deeper inside the pores (Figure 7c), but this is a spherical meniscus with two radii of
curvature equal to radius of the pore r por. Its curvature is therefore 2/r por. From D to S, only a thin
layer of water remains on the cylindrical walls of the pores. The cylindrical configuration of
point S maintains if water is readsorbed in the gel and the curve SDP is followed during re-
adsorption. At point F, the gel pores are again full of water. Along FC, during the second
desorption cycle, a spherical meniscus is formed again.
The equilibrium meniscus radius rm at any instant, is given by the compressive stress that the
solid network of the gel can support. The contact wetting angle at the liquid-solid-vapor interface
is undetermined along a sharp solid edge. Higher stresses are required for higher compression
states of the solid, which requires a higher hydrostatic tension in the liquid. It follows therefore
that the menisci are sharper and sharper (smaller radius r m) when contraction increases. This
keeps a gel shrinking until the meniscus radius rm reaches the pores radius rpor' The latter event
defines the drying critical-point. Beyond this critical-point. Capillarity cannot increase the
compressive stresses anymore on the solid network. For instance, for a cylindrical pore of radius
rpor and cross section:
2
A por =r por

The hydrostatic tension T in the pore liquid increases when rpor decreases, as:

cos
T =2
r por

Where is the wetting angle of the liquid over the solid; \ the surface tenson in the
liquid. For a perfect wetting ( =0):

2
T =
r por

However, the tensile force decreases as:

FT = T A por = r por

Hence the compressive force on the gel decreases so that the gel contraction stops at some
moment. For a higher liquid surface tensions y, the gel reaches a higher shrinkage at the critical
point
The capillary mechanism is not able to explain in a straightforward fashion the existence of a
"constant-rate" regime. A mathematical model by Suzuki and Maeda2 has shown that the rate of
drying could be constant when evaporation occurs from menisci, even when these menisci only
occupy a minor relative proportion AL of the gel external surface. This assumes the solvent vapor
transport is the slowest step and controls the kinetics of drying. However, an experimental study

2 Haas, P. A., Clinton, S. D. and Kleinsteuber, A. T. (1966), Preparation of urania and urania-zirconia microspheres
by a sol-gel process. Can. J. Chem. Eng., 44: 348353. doi:10.1002/cjce.5450440610
has shown that the drying rate falls drastically as soon as menisci appear, then it remains constant
while the menisci penetrate the inner part of a gel.
It is also possible to argue that the heat transfer to vaporize the liquid can control the kinetics of
drying. In this case the drying rate should, in a first approximation, be proportional to A L. As
long as a gel is full of liquid, this is a random two phase composite material. In these conditions,
it is known that the proportion of each phase (liquid L and solid S), along a cutting line
(respective linear proportions LL, LS), in a cutting plane (respective area proportions AL and AS),
or in volume (respective volume proportions VL and VS), are independent of the determination
mode:
LL = A L =V L

LS =A S =V S

Hence the drying rate per unit area should decrease linearly with the liquid content V L. The
experimental existence of a "constant-rate" period before the critical-point is therefore generally
interpreted as being due to the evaporation of a continuous liquid film, which covers the entire
surface of the gel. Obviously the existence of such a film can be related to the affinity of oxides
for water, which depends largely on the nature of the oxide. In the case of boehmite and
montmorillonite, this affinity is very important. The latter gels are able to swell back in water
and the constant rate can account for up to 80% of the gel volume shrinkage.

Figure 7 Adorption - desorption isotherms of water vapor in a silica gel.

Factors influencing the Sol-Gel reaction

A number of conditions can influence the hydrolysis and condensation reactions. Of these, the
most pertinent include water-to-alkoxide ratio, type and amount of catalyst, type of organic
group(s) attached to the silicon atom, and solvent effects. These several factors will now be
considered
Water-to-Alkoxide ratio
The effect of the water-to-alkoxide ratio for the sol-gel process in such that as the ratio increases,
so does the SiO2 content of the gel. Therefore, for complete hydrolysis, there must be at least one
mole of water for every alkoxide group. Some researchers have gone further to state that if more
than one mole of water per alkoxide group is used, the reverse reaction, reesterification, will
occur faster than the forward reaction. However, in a recent article by McCormick et.al 3, whose
experiments were conducted over a wide range of water-to-TEOS ratios, there was no correlation
between the water:alkoxide ratio and the achievement of complete hydrolysis. The validity of
these experiments on the effect of water:alkoxide ratio are thought to be accurate because water
is generated in the reaction in situ and therefore the reaction, once catalyzed, self-propagates the
hydrolysis.
Type and Amount of Catalyst
One important issue for consideration when deciding the concentration of catalyst is whether a
precise concentration is needed. In the sol-gel process, water is generated in situ through
condensation reactions. This makes calculation and addition of any precise amount of catalyst
difficult. In addition to water:TEOS ratio experiments, McCormick et al. [27] synthesized sol-gel
films, adding a wide range of acid concentrations. Their results indicated no correlation between
the acid concentration and acid initiation of the sol-gel reaction. In the experiments, different
acid concentrations were used with each sample and, as previously discussed, the condensation
reaction caused the in situ generation of water. Consequently, this in situ generation of water
diluted the initial acid concentration in every experiment. However, no differentiation between
the final sol-gel structures was observed by 29SI nuclear magnetic resonance (NMR)
spectroscopy for the various different overall sol-gel reactions. From these experiments, it was
determined that only a catalytic amount of acid was necessary. Therefore, the reaction,
containing this minimum amount of catalyst in all experiments, could self-propagate. This may
cause a change in the kinetics of the reaction but not in the basic structure of the overall network.
Hydrolysis and condensation reactions of most inorganic alkoxides can be carried out without
catalyst because of the extremely fast rates of reaction. However, alkoxysilanes hydrolyze much
more slowly, requiring the addition of either an acid or base catalyst. Acid-catalyzed reactions,
having a particle nucleation rate-determining process, tend to yield more linear-like networks due
to the fast hydrolysis. Therefore, acid catalyzed systems have a less completely formed network

3 Ng, L. V., Thompson, P., Sanchez, J., Macosko, C. W., & McCormick, A. V. (1995). Formation of cagelike
intermediates from nonrandom cyclization during acid-catalyzed sol-gel polymerization of tetraethyl
orthosilicate. Macromolecules, 28(19), 6471-6476
of siloxane bonds with a higher concentration of unreacted silanols. Base-catalyzed reactions, on
the other hand, yield highly and arrange themselves in the most thermodynamically stable
arrangement. This process leaves fewer unreacted silanol groups in the overall network and a
more highly densified network.
Acid catalysis, used in this research, increases the rate of the hydrolysis reaction, as shown in
Figure 8. In the first step, hydrolysis occurs through the protonation of the oxygen atom of the
alkoxy group due to the presence of extra electrons. This causes a shift in the electron cloud of
the Si-O bond toward the oxygen, resulting in a positive charge on the silicon atom 1281. In the
rate-determining step, oxygen from the water attacks the silicon atom, which has low electron
density. This results in the formation of a pentacoordinate transition state in which partial
positive charges are developed. Steric effects, such as the presence of bulky and/or long alkyl
substituents, will affect the rate of the hydrolysis reaction by hindering the inversion of the Sn2
transition state

Figure 8 Effect of catalyst on hydrolysis and condensation

Several pH-dependent rate profiles have also been reported for the hydrolysis and condensation
reactions, where it has been found that reaction rates are largely dependent on pH (see Figure 6)
[24,25,29,30]. For instance, at pH = 7, molecular hydrolysis occurs at a slow rate, while
molecular condensation occurs at a fast rate. It is this inverse relationship between the rates of
the hydrolysis and condensation reactions that must be taken into account in controlling the
kinetics of the reaction and therefore controlling the ultimate network structure.
Type of Network Modifier
When TEOS is commix ed with a silicon alkoxide comonomer containing organic groups
directly bonded to the silicon atom, organically modified silicates (ORMOSILs) are created. The
result of synthesizing ORMOSILs is to modify the network connectivity. Organic groups cause
coordination centers with functionality less than four and influence the reactivity of the alkoxy
groups and therefore the connectivity of the sol-gel network in two ways. (1) Formation of
siloxane bonds requires the diffusion of partially hydrolyzed molecules. The larger the alkyl
group attached to the silicone, the slower the rate of diffusion and therefore the less
interconnection within the network. (2) It is also reported that larger alkyl groups produce
polymers with higher surface area. Larger surface area allows for higher unreacted oxide
concentration in the gels. This produces a branching effect in the sol-gel network. In addition to
these two points, it was previously discussed that steric effects, such as the presence of bulky
and/or long alkyl substituents, will affect the rate of the hydrolysis reaction by hindering the
inversion of the Sn2 transition state. ORMOSILs will be discussed in more detail in a subsequent
technical report on organicinorganic nanocomposites.
Solvents effect
The effect of solvents on hydrolysis and condensation reactions is not usually discussed
primarily because solvents other than water and simple alcohols, which are produced internally,
are not often used. However, addition of external solvents can have an important effect on
controlling hydrolysis and condensation rates. These solvents, excluding the usual water and
alcohol cosolvent, are called drying chemical control additives (DCCAs) and include
tetrahydrofuran, formamide, dimethylformamide, and oxalic acid. These solvents can be used in
addition to alcohol in order to promote slow drying of the silica monolithic gel. In bulk films,
slow drying is necessary to prevent cracking of the film.
Condensation rates may be dramatically affected by solvent type, as can be seen Figure 9.
Condensation, via acid catalysis, is proposed to go through a two-step Sn2 mechanism.

Figure 9 Condensation mechanism for a tetrafunctional silicate.

Protic solvents seemingly have little or no effect on condensation rates. Polar aprotic solvents,
such as those previously mentioned, solvate the protonated silanol (Si-OH2+). The cation is
solvated by orienting the negative end of the polar aprotic around the cation and by donating
unshared electron pairs to the vacant orbitals of the cation. This process stabilizes the transition
state and increases the rate of condensation. Use of DCCAs can be an advantage for solgel
reactions within a membrane template. A DCCA will allow molecular diffusion but will delay the
reaction, resulting in inorganic uniformity through the template. Without a DCCA, the overall
reaction rate is high in relation to the slow diffusion process; therefore, the ultimate geometrical
distribution of the silicate phase will be highly nonuniform.
Advantages and limitations of Sol-Gel Processing
There are many advantages to sol-gel processing. It not only allows for materials to have any
oxide composition, but it also permits the production of new hybrid organic-inorganic materials
which do not exist naturally. Sol-gel processing also has many other significant advantages. Very
pure products are obtained by simply purifying the precursors either by distillation,
crystallization, or electrolysis. Moreover, the chemical processes of the first steps are always
carried out at low temperatures. By comparison with the classical high temperature synthesis of
conventional ceramics, this minimizes considerably the chemical interactions between the
material and the container walls. Another advantage is the association of the solid colloidal state
with a liquid medium, thus avoiding any pollution by the eventual dispersion of dust. This
explains why the biggest industrial application of sol-gel synthesis of ceramics is for nuclear
fuels; pollution is very critical in this domain.
There are other more fundamental advantages to sol-gel processing. For instance, the kinetics of
the various chemical reactions can be easily controlled by the low processing temperatures and
by the often dilute conditions. The nucleation and growth of the primary colloidal particles can
also be controlled in order to give particles with a given shape, size, and size distribution. This
size domain is of course largely submicronic. However, in order to do that, the sintering
temperatures of both the amorphous and the crystalline material must be lowered of several
hundred degrees Celsius. Apart from this energetic disadvantage, this cooling may be critical for
some applications. This is the case of ceramic-ceramic composites for which the processing
temperature must be lowered below the temperature needed for the fiber-matrix reaction. This is
also the case of hybrid organic-inorganic materials.
The structure of sol-gel ceramics can be as easily controlled as the size of the particles. An
amorphous or semi-vitreous state is, for example, quite easy to obtain and many new types of
glasses that were not feasible by the conventional quench techniques can now be synthesized. As
a matter of fact, the two problems that the conventional quench techniques encountered, that is
the very high fluidity of the liquidus of the material that required a very fast cooling rate for the
glass to form and the existence of a miscibility gap in some domains of composition making it
impossible to obtain a homogeneous glass, were completely solved by the sol-gel process. The
distribution of the pores and crystalline or amorphous phases can also be tailored to a large
extent. This is particularly true if the phase present at grain boundaries can be either controlled or
eliminated. Eliminating this phase can be important in some cases, as for the ionic conductors
composed of sodium and silicon (Nasicon) for which the segregation of ZrO2 at grain boundaries
must be avoided.
Sol-gel processing offers the most outstanding advantages for mixed oxides systems in which the
chemical homogeneity of the various elements can be controlled down to the atomic level. This
is the case of the Lead Lanthanum Zirconium Titanates (PLZT). This condition is, however,
virtually impossible to achieve when solid powders are mechanically mixed such as in the
conventional processes; and hence the optical transparencies obtained are, in this case, not as
high as those obtained by sol-gel processing.
The greatest limitation to the synthesis of ceramic by sol-gel processing is still the cost of the
precursors, and especially that of alkoxides. Most of these alkoxides are nonetheless quite easy to
make; especially if they do not tend to polymerize. A few of them such as Zr and Ti are even
used industrially by the Schott company for coating applications and are thus quite affordable.
Moreover, alkoxides can also be mixed with much cheaper metal salts provided that a
purification step is included in the procedure.
The sol-gel synthesis of ceramics will never be able to compete for the mass production of some
large scale materials such as window glass for which the conventional processes can rely on
much cheaper raw materials. Sol-gel processing becomes, however, much more interesting for
highly advanced ceramics.
Sol-gel processing is not the only chemical route that leads to better ceramics. Another procedure
includes the use of organometallic compounds in which an organic group is directly bonded to a
metal without any intermediate like oxygen. The chemistry of organometallics essentially
concerns the synthesis of complex polymers and especially those containing carbides and
nitrides. If the chemistry is carried out in a liquid medium, then it can be considered as a sol-gel
process. Precipitation and co-precipitation techniques are also used and are sometimes even
considered as a side branch of sol-gel processing. The chemical reactions concerned in this case
are the same as those occurring in sol-gel synthesis; they often lead to the production of colloidal
particles, but can also be dispersed into a stable sol.
Two other techniques different from sol-gel processing are:
- The thermal decomposition of precursors in the vapor phase. This technique is also known as
chemical vapor decomposition (C.V.D). Some of those methods use quite sophisticated
nucleation and growth procedures such as heating by laser. As in sol-gel processing, very pure
products can be obtained if certain precursors such as, again, alkoxides are used.
- The hydrothermal process. It is carried out on solutions in an autoclave and at higher
temperatures than those required in sol-gel processing. This technique also produces crystals
with a size ranging in the micrometer and beyond. Hydrothermal synthesis is not within the
scope of this book, although the wet chemistry aspect of it is essentially the same as that of sol-
gel processing.