J. Am. Ceram. Soc.
, 88 [9] 2337–2348 (2005)
Journal
Mechanics and Microstructures of Concentrated Particle Gels
Hans M. Wyss,w,z Elena V. Tervoort, and Ludwig J. Gauckler
Nonmetallic Inorganic Materials, Department of Materials, Wolfgang–Pauli–Strasse 10, ETH Zurich, Zurich CH-8093,
It is often assumed that the viscoelastic properties of dense col-
loids are determined by the colloid volume fraction, the inter-
action potential, as well as the particle size distribution and
shape. The dependence of the viscoelastic behavior of particle
suspensions and gels on these parameters has been widely stud-
ied, and is well understood in many cases. In contrast, our
knowledge on the influence of microstructure on mechanical and
rheological properties, in particular for high solid loading sus-
pensions as used in ceramic processing, is much less developed.
This aspect has been the focus of recent experiments, which
show that small changes in microstructure can have dramatic
effects on the mechanics and dynamics of concentrated colloidal
gels. In this article, we attempt to give an overview of the influ-
ence of microstructure on the mechanical and rheological prop-
erties of colloidal systems. Particular attention is given to
colloidal particle gels at high volume fractions.
I. Introduction
C OLLOIDALsuspensions show a wide range of viscoelastic,
dynamic, and structural behavior. Many industrial appli-
cations make use of colloidal suspensions and of the possibility
to change and control their behavior. In ceramic processing, a
variety of different methods are used to exploit the wide-phase
behavior of colloids to produce a dense, homogeneous packing
of particles via ‘‘wet processing’’ routes. For good reviews of the
different processing methods, see Lewis1 and Sigmund et al.2
It is known that the mechanical behavior of a colloidal gel
depends on the size and shape of the particles, the interparticle
forces, the volume fraction, and the spatial arrangement of the
particles or, what is called the microstructure of a particle net-
work. The structural, mechanical, and dynamic properties of a
colloidal suspension can change drastically, when one or several
of these parameters are varied. However, the main focus in the
study of the mechanical behavior of dense colloidal suspensions
D. Green—contributing editor
Manuscript No. 20105. Received February 2, 2005; approved June 2, 2005.
H. W. was supported by the Swiss National Science Foundation under Grant 2002–4 19.
w
Author to whom correspondence should be addressed. e-mail: hwyss@deas.harvard.
edu
z
Division of Engineering and Applied Sciences, Harvard University, Cambridge, MA
02138
DOI: 10.1111/j.1551-2916.2005.00622.x
r 2005 The American Ceramic Society
Switzerland
and gels has been limited to a large extent to the influences of
volume fraction and interparticle forces.
The effect of the interparticle pair potential on the rheological
behavior of suspensions was studied by several authors.3–6 A
variety of other contributions focus mainly on the dependence
on the particle volume fraction.7–9
Microstructural characterization is of tremendous impor-
tance in ceramic processing and technology. The characteriza-
tion of microstructure of sintered ceramic bodies is a standard
procedure that is essential for the optimization of the mechan-
ical behavior of ceramic parts. It is, therefore, surprising that
only a few studies have investigated the microstructures of ce-
ramic materials in the ‘‘wet’’ stage—in a concentrated suspen-
sion or gel of particles.
Because of the difficulties involved in characterizing and
quantifying a microstructure in a colloidal system, it is easier
and more straightforward to study the relation of the macro-
scopic properties to the other relevant parameters. However, an
understanding of the mechanics of flocculated suspensions can
only be achieved when the structure of the particle network is
taken into account10; the interparticle potential alone is not
sufficient.
Even one of the simplest colloidal systems, a suspension of
hard-sphere particles, shows a complex viscoelastic behavior
that is tightly linked to changes in the microstructure. At high
volume fractions, such a system shows pronounced shear thin-
ning, reflecting a shear-induced structural rearrangement (see
Section II). In order to understand the mechanisms that govern
the macroscopic properties of a colloidal system, profound
knowledge of the structural properties is thus essential.
Of particular interest is the influence of the microstructure on
the mechanical behavior of particle gels. At low solid loadings,
theoretical models based on the fractal description of network
structures are successful in linking the structural to the mechan-
ical behavior (see Section III). These models are in good agree-
ment with results from experiments11–13 and from computer
simulations.14 However, the fractal description of the network
structure must break down at some point toward increasing
volume fractions,7,15 and the measurement of fractal dimensions
becomes increasingly difficult at high volume fractions. Never-
theless, the fractal description provides a way to link the micro-
structure, or at least some characteristic quantities that describe
the microstructure, to the viscoelastic behavior directly.
For highly concentrated gels, however, our knowledge on the
relations between structural and mechanical properties is rather
limited. We identify two main problems that make it difficult to
Feature
2338 Journal of the American Ceramic Society—Wyss et al. Vol. 88, No. 9

establish a relation between network structure and mechanical
behavior in concentrated colloidal gels.
Firstly, the reproducible formation of a colloidal gel at high
volume fractions cannot be performed as easily as in a diluted
system. The direct addition of salt or acid/base to stable suspen-
sions will lead to colloidal gels of irreproducible microstructures.
Because of the elevated aggregation rates in concentrated sus-
pensions, the system will generally aggregate locally before the
salt or acid/base is homogeneously distributed. In rheological
studies of concentrated gels, a high shear rate is usually applied
prior to the measurements in order to set reproducible starting
conditions.7,10,16 However, in this case the microstructures of the
resulting gel networks are determined mainly by the shearing
process, and not by the initial aggregation process itself. As a
consequence, the formation of different concentrated gels of sig-
nificantly different microstructure will be difficult to achieve if a
high shear rate has to be applied. It is also important to note that
strong shearing of a particle gel will not always make its struc-
ture more homogeneous. By contrast, strong shearing will lead
to a highly inhomogeneous structure, as it favors the formation
of strong, densely packed aggregates in the system.17,18
Secondly, the characterization of microstructures in concen-
trated colloids proves difficult, as the standard methods used for
the characterization of diluted colloidal suspensions cannot be
easily applied to highly concentrated systems. For example,
static light scattering, the standard method for the characteri-
zation of structure in colloidal gels, fails for systems that exhibit
strong multiple scattering.
II. Role of Microstructure in Stable Suspensions
where fc  0.63 and fc  0.71 for the low- and the high-shear
limit, respectively. Whereas for the low-shear limit, the critical
volume fraction corresponds to the volume fraction of random
close packing, for the high-shear limit it is closer to a crystalline
volume fraction. The value of the critical volume fraction
fc  0.71 reflects the fact that in the high-shear limit, ordering
occurs only in layers and not in a three-dimensional (3D) crys-
talline structure. Thus, although Eq. (2) is purely empirical, the
values for the characteristic volume fractions themselves reflect
the importance of microstructural arrangement. In Fig. 1, we
illustrate the shear thinning effect with a plot of viscosity as a
function of shear stress in a model hard sphere system.
As a model system, we use polymethyl methacrylate particles
of radius a 5 136 nm dispersed in a mixture of cycloheptyle
bromide and decalin; the volume fraction is f 5 0.41. For details
regarding this model system, see Prasad et al.,29 Manley et al.,30
Segre et al.,31 Pham et al.,32 Poon et al., 33 and Yethiraj and Van
Blaaderen.34 The viscosity Z in the transition region between the
low- and the high-shear region is well described by
1
Z ¼ Z1 þ Z0 (3)
1 þ ðs=sc Þn
A fit to the data in Fig. 1, shown as a dotted line, yields n 5 2
and sc 5 0.83 Pa.
The shear-thinning and shear-thickening effects in a concen-
trated suspension illustrate the importance of the microstructure
the viscoelastic properties of colloidal systems. It is obvious that
in a colloidal gel, microstructure could play an even more im-
portant role, while a thorough understanding of these effects will
be harder to achieve than for a stable suspension.

As mentioned before, the microstructure has an important in-

fluence on the rheological properties even in a stable colloidal III. Fractal Description of Microstructures for Low Volume
suspension.19–21 At high concentrations, colloidal suspensions Fraction Colloidal Gels
can show pronounced non-Newtonian effects in their flow, in
If the interaction between colloidal particles is strongly attrac-
particular, shear thinning and shear thickening.19,21–25 The vis-
tive, a colloidal system will form aggregates. At low volume
cosity of the suspension is no longer independent of the shear
fractions, the structure of such aggregates is accurately described
rate, but changes with the shear rate. Both phenomena can be
as fractal.13,35 This means that the number of particles contained
explained by a change in microstructure in the suspension, in-
in an aggregate scales with its size as a power law:
duced by the shear. The occurrence of structural rearrangements
upon shear has been experimentally confirmed by both static N / Rdf (4)
light scattering 26 and neutron scattering.8,27 In the following,
we will focus on the shear thinning effect to illustrate the im- where the exponent df, called the fractal dimension, is smaller
portance of microstructure on the viscoelastic behavior of sta- than the Euclidean dimension d. This implies, because the den-
ble particle suspensions. For a review of shear thickening
(dilatancy) in concentrated suspensions, we refer to Barnes.28
By comparing the typical timescale for Brownian motion and
for the shearing process, one can estimate the typical shear rate
at which the structure of the suspension should begin to be per-
turbed by the flow. A characteristic time tD for diffusion is given
by the time it takes a sphere to move a distance equal to its own
size. A characteristic time tshear for the flow is given by the in-
verse of the shear rate. The ratio between these two character-
istic times is known as the Peclet number Pe and is thus given by:
6pZa3
Pe ¼ (1)
kB T
where Z is the viscosity of the liquid phase, and a is the particle
size. At Peclet numbers much smaller than unity, Brownian
motion is able to maintain the structure of the suspension in a
state essentially indistinguishable from the unsheared state. At
shear rates corresponding to a Peclet number greater than unity,
Brownian motion can no longer restore the structure of the sus-
pension to its equilibrium state. Experiments on model systems
have shown that flow curves for suspensions of different particle
sizes fall onto a single master curve, when plotted as a function
of the dimensionless Peclet number.20,25 The high- and low-
shear limits of the viscosity Z are in good agreement with


Z f
2 Fig. 1. Shear thinning effect for a model hard-sphere system at volume
¼ 1
(2) fraction f 5 0.41. Normalized viscosity as a function of the shear stress
Z0 fc s. The dotted line shows a fit to the data using Eq. (3).
September 2005 Concentrated Particle Gels
sity of the fractal clusters thus decreases as their size increases,
that upon formation of fractal aggregates in a suspension, the
system will ultimately form a volume-filling gel. The values of
the fractal dimension df for different aggregation mechanisms
are well established from computer simulations as well as from
experiments at low volume fractions.11–14
Aggregation mechanisms can be divided into two different
categories: diffusion-limited cluster aggregation (DLCA) and
reaction-limited cluster aggregation (RLCA). In DLCA, each
collision between particles during aggregation leads to an irre-
versible bond, whereas in RLCA, particles can escape from a
bond site and penetrate deeper into the aggregated structures.
Therefore, aggregates formed by RLCA are denser and have a
higher fractal dimension than DLCA aggregates.
The size of flocks in a fractal gel network follows directly
from the condition that the density fc of the fractal clusters will
decrease during aggregation until, when the volume of the flocks
fills all space, it reaches the volume fraction f of particles in the
overall suspension. Thus, the characteristic size Rc of fractal
clusters scales as
1=ðdf
dÞ
Rc / f (5)
As a result of the fractal nature of the microstructure, the me-
chanical properties of fractal gels can be directly linked to the
parameters that describe their microstructure. The fractal di-
mension df describes the overall structure of the fractal clusters.
As the elastic constant ke of a cluster depends on the mechanism
for bearing stress, two additional parameters, the parameters e
and db, are needed to describe the geometry of the main stress-
bearing chain of particles within a cluster. The elastic constant ke
of a cluster then scales as36–39
1
ke / (6)
Rdb R2e
where the exponent e, with 0rer1, describes the stress-bearing
mechanism of the chain. The case e 5 0 applies to a straight
chain, where the stress is transmitted by bond stretching.
The case e 5 1 describes a purely isotropic chain, where stress
transmission through a bond-bending mechanism dominates.
These two extreme cases are illustrated in Fig. 2. The geometry
of the chain will determine its elastic properties.
The exponent db is called the bond dimension or chemical
dimension and must be at least 1 to provide a closed path. An
upper bound to db is given by min[df, 5/3], where a value of 5/3
corresponds to a self-avoiding chain. Computer simulations
have suggested that db  1.35 for a 3D percolation network.40
The exponents e and db are not independent. For straight chains
(bond stretching), the parameters are e 5 0 and db 5 1. For a
more detailed description, see De Rooij et al.37 and Potanin.39
The scaling behavior of the shear modulus G0 follows directly
from Eq. (6):
ke
ð1þ2eþdb Þ
G0 / / Rc (7)
Rc
Fig. 2. Schematic representation of different chain geometries. (a) Iso-
tropic chain: dominated by bond bending. (b) Straight chain: pure bond-
stretching mechanism.
2339
For a fractal structure, with Eq. (5), one finally obtains
G0 / fð1þ2eþdb Þ=ðd
df Þ ¼ fm1 (8)
Through similar scaling arguments, de Rooij et al. derive an
expression for the scaling behavior of the yield stress sy of a
particle network as follows 37:
fext
sy  1Rc
ð2þeÞ / fð2þeÞ=ðd
df Þ  fm2 (9)
R2c
The scaling laws for the storage modulus G0 and the yield stress
sy are linked to the stress-bearing mechanism via the para-
meter e:
For a bond-stretching mechanism (e 5 0):
1 þ db 2
m1 ¼ and m2 ¼ (10)
d
df d
df
For the case of bond bending (e 5 1):
3 þ db 3
m1 ¼ and m2 ¼ (11)
d
df d
df
The values of the exponents m1 and m2 are directly linked to
the fractal dimension df of the aggregates. Low values for the
exponents indicate anisotropic aggregate structures, whereas
higher values indicate more isotropic aggregate structures or
higher fractal dimension.37,39
Thus, the fractal description of the network structure and the
above scaling concepts provide a clear-cut relation between the
structural and mechanical properties of fractal particle networks
at low volume fractions. However, at very high volume frac-
tions, the fractal description of network structure is no longer
appropriate. Nevertheless, the scaling concepts summarized in
this section might be useful to some extent in describing the be-
havior of highly concentrated systems as well.
IV. Microstructural Characterization of Gels
Information on microstructures is one of the most useful and
essential measures when it comes to understanding the macro-
scopic properties of materials. In Table I, we list some of the
methods that can be used to access microstructures in particle
suspensions and gels. For a comprehensive review of aggregate
characterization techniques, we refer to a recent article by
Bushell et al.41
In the following, we will briefly discuss the various methods,
as well as their advantages and disadvantages with respect to the
characterization of the microstructure in dense, highly turbid
colloidal suspensions.
The characterization of the structure in dense colloidal sus-
pensions and gels is often an experimental challenge. In partic-
ular for ceramic systems it is hard to access the microstructure
experimentally. One reason for this is that the refractive index
difference between the solid and the liquid phases is usually too
high to enable measurement of their microstructure using stand-
ard light scattering techniques (Table I (1)).42 A meaningful
evaluation of standard light scattering experiments relies on
samples that show only negligible multiple scattering. By using
more advanced light scattering setups such as a multicolor or a
3D setup (Table I (2)),43,44 contributions of multiple scattering
to the detected signal can be suppressed. However, these
methods are still limited to samples of rather low turbidity, as
some measurable amount of single scattering intensity always
has to be present.
Using diffusing-wave spectroscopy (DWS) (Table I (3)),45,46
the dynamics of highly turbid systems can be analyzed. DWS
extends traditional dynamic light scattering to media with strong
multiple scattering by treating the transport of light as a diffu-
sion process. No direct structural information is obtained from
2340 Journal of the American Ceramic Society—Wyss et al. Vol. 88, No. 9

Table I. Different Techniques Applied in Characterizing the Microstructures of Particle Gels

Method and references Capabilities Limitations

(1) Standard static and
dynamic light
scattering42
(2) Multicolor or 3D setup
for static and dynamic
light scattering43,44
(3) Diffusing wave
spectroscopy45,46
(4) Small-angle X-ray
scattering48–50
(5) Small-angle neutron
scattering
(6) Spin-echo small-angle
neutron scattering52,56
(7) Freeze-fracture
techniques57,58
(8) High-pressure freezing
and fracture49,61
(9) Confocal
microscopy62,64
3D, three-dimensional.
Allows the characterization of structure and
dynamics in colloidal systems of low volume
fraction
Allows suppression of contributions of multiple
scattering in the signal
Allows to extend dynamic light scattering to
highly turbid suspensions and gels. Gives
information on the average dynamics of particles
in the system
Gives structural information in terms of the static
structure factor S(q)
Gives structural information in terms of the static
structure factor S(q)
Gives structural information in terms of the static
structure factor S(q) on length scales between 10
nm and 7 mm
Gives real space structural information
Gives real space structural information
Crystal formation in aqueous systems can be
adequately suppressed
Gives detailed 3D structural information in real
space
Limited to samples of very low turbidity,
where multiple scattering can be neglected
Still limited to rather low turbidity, as some
measurable amount of single scattering always
has to be present
Only limited structural information47
Upper limit to length scales that can be
accessed because of small wavelength
Present detectors set an upper bound to the
accessible length scales at around 200 nm
Facilities not readily available
Crystal formation in the liquid phase during
freezing is likely to distort microstructure
Limited to aqueous systems
Limited to low-turbidity samples or to
measurements near the cover slip (wall
effects). Fluorescent labeling necessary

such a measurement, although measuring the turbidity itself
gives some information on the homogeneity of the microstruc-
ture.47 The single-scattering regime can also be accessed by using
X rays instead of visible light. However, accessing large enough
length scales requires detection at very low angles. Small-angle
X-ray scattering (Table I (4))48–50 is suitable for the character-
ization of microstructures only if the particle size is sufficiently
small.
Also, in small-angle neutron scattering (SANS) (Table
I(5)),48,51 present detectors set an upper bound to the accessible
length scales at around 200 nm.
However, a recently developed technique called spin-echo
SANS (SESANS) (Table I (6))52–56 allows to extend the acces-
sible length scales dramatically. Structural information can be
obtained at length scales spanning three orders of magnitude:
from 10 nm up to 7 mm.56 The technique is based on the Larmor
precession of polarized neutrons in magnetic fields with inclined
faces. The precession of the polarization vector of neutrons in a
magnetic field is proportional to the wavelength, magnetic field,
and path length through this field. It is therefore possible to de-
termine the z component of the wave vector momentum transfer
Qz as a function of the precession angle f. The SESANS signal
can be interpreted in real space; it is possible to determine the
particle density autocorrelation function directly from the meas-
urement. In the first applications of the SESANS technique to
the characterization of colloidal systems, Krouglov et al.55 show
this directly by measuring the structure of hard-sphere suspen-
sions at different concentrations; Bouwman et al.56 show that
the technique is able to directly identify the phase transition of a
colloidal system from a gas via a liquid to a solid. The large
range of accessible length scales combined with the possibility of
measuring samples of very high volume fraction, confers this
method with considerable potential for future studies of the
structure of concentrated colloidal systems.
For aqueous systems, freeze-fracture techniques can be used
to access the microstructure (Table I (7))57,58 This can be per-
formed by applying a coating to the surface of the frozen frac-
ture surface and then examining this templated layer using
transmission electron microscopy. Alternatively, one can exam-
ine the fracture surface directly by using a scanning electron
microscope (SEM) equipped with a Cryo stage. However, the
standard freezing techniques are likely to produce large ice crys-
tals in the liquid phase of the samples and thereby distort the
microstructure. The best method to preserve the structure of
aqueous samples during freezing is high-pressure freezing (Table
I (8)).59 As the crystallization of water requires more energy at
high pressures than at ambient pressure, the formation and
growth of ice crystals in the liquid phase can be adequately sup-
pressed. This technique has recently been applied to ceramic
systems.60,61
The most direct way of accessing 3D microstructure is con-
focal microscopy (Table I (9)).62–64 However, this technique also
requires a low refractive index mismatch between the colloids
and the solvent. It is therefore not suitable for ceramic systems
or for attractive systems that rely on van der Waals forces. This
technique has recently been used to study weakly attractive gels
and colloidal glasses in model systems.62–64
V. Structure and Mechanical Properties of Concentrated Gels
Linking microstructure to macroscopic behavior is of central
interest in materials science and technology. Nevertheless, rela-
tively little is known about the relation between microstructures
and mechanical properties of concentrated colloidal suspensions
and gels. The main difficulties in this endeavor lie in the exper-
imental limitations when it comes to studying and quantifying
the microstructure of these systems.
Channell et al.10 report one of the few studies that elucidate
this relation in a systematic way. The study focuses on the
characterization of the mechanical behavior of gels. The com-
pressive yield stress sy as a function of volume fraction f is ob-
tained by analyzing the density profiles of samples after
subjecting them to a compressive stress in a centrifuge. A de-
tailed discussion of this technique is given elsewhere.65,66 The
advantage of the technique is that a whole curve of yield stress
versus volume fraction is obtained from a single measurement.
The stress at a given point in the sample is given by the load of all
the material on top of this point. Thus, the equilibrium volume
fraction for a whole spectrum of stresses can be obtained by
September 2005 Concentrated Particle Gels
analyzing a single sample. In order to achieve different micro-
structures, the authors use a very simple, yet efficient, technique.
Different levels of homogeneity are achieved by preparing the
samples in different ways, but with identical chemical com-
position. Homogeneous samples are obtained by preparing
suspensions at a given volume fraction f0 and subsequent ho-
mogenization by high-shear mixing. Lower levels of homogene-
ity are achieved by carefully diluting samples gelled at a higher
volume fraction f4f0, and subsequent shearing by either hand
mixing or by ultrasonication. During the measurement of the
compressive yield stress, this low volume fraction structure is
then consolidated in the centrifuge, which leads to a volume
fraction gradient in the sample. The lowest volume fraction is
found at the upper surface of the gel, where the volume fraction
after centrifugation should still be the same as the overall volume
fraction f0 before centrifugation. Figure 3 shows a schematic
illustration of a microstructure obtained from diluting a more
concentrated gel under gentle stirring. The new gel network is
built up from individual clusters that are connected only through
a few weak bonds to form the new volume-spanning network.
The experimental data indicate that lower mixing intensities
lead to lower compressive yield stresses. Figure 4 shows the
compressive yield stress as a function of volume fraction for gels
that were obtained from the different homogenization methods.
If high-shear mixing is used for homogenization of these sam-
ples, the compressive yield stress is independent of the volume
fraction where the gels were initially prepared. Thus, the applied
high-shear mixing seems to be sufficient to achieve a reproduc-
ible microstructure, dependent only on the sample composition,
but not the sample’s preparation history. The lowest compress-
ive yield stresses are observed for samples subjected to the lowest
mixing intensity (hand mixing). Note also that above the volume
fraction where the samples were originally prepared, the yield
stresses are independent of the homogenization used in the di-
lution process.
By using different mixing intensities and starting volume frac-
tions f4f0, samples of different microstructures are obtained.
Even though the authors do not characterize microstructures
experimentally, it is reasonable to assume that the different lev-
els of homogenization lead to different aggregate sizes. The low-
er the mixing intensity, the larger the aggregates and the lower
the compressive yield stresses.
Fig. 3. Schematic illustration of the microstructure obtained by dilut-
ing a more concentrated particle gel with only gentle stirring. The new
gel network is built up from individual clusters that are connected in
weak bonds to form a volume-spanning network (after).10
2341
Fig. 4. Compressive yield stress versus volume fraction for particle gels
formed by centrifugation of a low volume fraction gel. The different
volume fractions correspond to different layers within the centrifuged
sample. The initial low volume fraction gels are obtained by diluting a
gel of volume fraction of f 5 0.45 using different homogenization meth-
ods: hand mixing (open diamonds), ultrasonication (full diamonds), and
high shear mixing (circles). Also shown are results for a gel that was
directly prepared at low volume fraction and homogenized using high
shear mixing (squares) (after).10
A recent study by Franks et al.67 focuses on the relationship
between aggregate size and the rheological properties of a
packed sediment of aggregates. The authors induced the forma-
tion aggregates by flocculation of alumina particles in a dilute
suspension, varying for instance the salt concentration to
achieve different aggregate sizes. The aggregates formed subse-
quently settle to form a dense sediment at the bottom of the
container. By adjusting the salt concentration of the sediment,
the interparticle forces for different samples can be equilibrated
between samples of originally differing salt concentration. Thus,
the presented method enables the separation of the effects of
aggregate size (i.e. microstructure) from those of interparticle
forces. Aggregate sizes can be measured in the dilute suspension,
before the aggregates settle to form a dense sediment.
They find that when other aggregate properties are held con-
stant, the shear and compressive yield stresses of sediments
formed from larger aggregates are generally higher at all vol-
ume fractions. Also, the larger aggregates produce a sediment of
lower maximum packing fractions than smaller aggregates.
Using a different approach to control microstructure, another
recent study 47,60,61,68,69 has focused on both the characterization
of the microstructure as well as the mechanical properties of
concentrated gels. Control over the microstructure is achieved
by using different chemical pathways of coagulating electrostat-
ically stabilized gels, using a destabilization method originally
developed for the processing of high-performance ceramics, di-
rect coagulation casting (DCC).70–72 The method is based on
enzyme-catalyzed chemical reactions, which can be controlled
by both enzyme concentration and temperature. It allows the
formation of gels of highly concentrated particles without per-
turbing the microstructures that develop during the gelation
process. The gels can be produced by two different destabilizat-
ion mechanisms that lead to significantly different microstruc-
tures: either the pH of the suspensions is shifted toward their
isoelectric point (DpH method), or the ionic strength of the sus-
pensions is increased at a constant pH (DI method) (DCC tech-
nique (see Box 1)).
2342 Journal of the American Ceramic Society—Wyss et al.
The authors use these reactions and apply them to concen-
trated particle gels to investigate the relations between the ki-
netics of the destabilization process, the dynamics of particles,
the evolving microstructure, and the resulting mechanical prop-
erties. The dynamics of particles during the aggregation process
is followed by using DWS. This light scattering technique allows
deriving quantitative information about the dynamics of parti-
cles in very turbid suspensions. While in conventional light scat-
tering experiments the samples have to be highly diluted, DWS
extends traditional dynamic light scattering to media with strong
multiple scattering by treating the transport of light as a diffu-
sion process. We derive quantitative information about the sol–
gel transition and the viscoelastic properties of the gels, as well
as a characterization of changes in the microstructure. DI gels
show stronger elastic moduli and exhibit a less homogeneous
microstructure than DpH gels.47
To further characterize the dynamics of the destabiliza-
tion process as well as the evolving microstructures, the authors
performed static light scattering measurements on suspen-
sions of concentrated silica particles.68,74 Here, problems be-
cause of multiple scattering can be avoided by using a very thin
sample thickness. In addition, the refractive index difference be-
tween the aqueous phase and the particles can be reduced by
adding sucrose to the liquid phase of the suspensions. While
during the DpH destabilization, the scattering curve shows sig-
nificant changes only after some characteristic delay time, it
changes continuously during the DI destabilization. This be-
havior is attributed to the formation of a weak pre-gel structure
in the suspensions, as a shallow secondary minimum appears in
the interparticle potential.
The real-space structure of concentrated gels is directly ac-
cessed by combining high-pressure freezing with Cryo-SEM. By
using the high-pressure freezing technique, the formation and
growth of the ice crystals are adequately reduced.59
Microstructures can thus be examined on fracture surfaces of
high-pressure frozen particle gels. In Fig. 5(a), we show the ob-
tained SEM images of systems at a volume fraction of f 5 0.4,
for a DpH-destabilized suspension and a DI-destabilized suspen-
sion. Structural differences are evident and in good agree-
ment with a recent Brownian dynamics simulation study.75,76
A visualization of the structures obtained in this simulation is
shown in Fig. 5(b). Results show a less homogeneous structure
for DI gels, with inhomogeneities on the length scale of a few
particle diameters.
The viscoelastic behavior of concentrated DpH and DI gels is
characterized by using standard macroscopic rheological meas-
urements and, for samples that can support their own weight,
compressive tests. In Fig. 6, we show these results for the two
different destabilization methods as a function of volume frac-
tion. Notable differences in the mechanical behavior of the two
systems are observed. For the inhomogeneous DI gels, signifi-
cantly higher yield stresses are observed. In oscillative rheolog-
ical measurements, the same behavior is observed; both the yield
stresses and the storage moduli G0 are significantly higher for the
less homogeneous DI systems.61,69,77
An analysis of the scaling behavior of yield strains suggests a
fundamental difference in the stress-bearing mechanism of the
two systems, with stresses supported by a bond-bending mech-
anism for DpH gels and a bond-stretching mechanism for DI
gels.69
All results combined strongly indicate that microstructure has
a tremendous influence on the mechanical behavior of these
systems. The two different variants of the in situ destabilization
process naturally lead to gels of significantly different micro-
structures.
A logical step in these experiments is to determine what hap-
pens if the structure formed in situ is perturbed by applying a
large shear rate (as done in many rheological experiments to set
reproducible starting conditions). The result69,78 is surprising
and shows again the strong dependence of mechanical behavior
on the structural arrangement of particles in a gel network. Re-
sults again suggest an increase in mechanical properties of con-
Vol. 88, No. 9
centrated particle gels with decreasing homogeneity of their
microstructures. Shearing a homogeneous DpH gel leads to a
strong increase of the shear modulus and yield stress, while
shearing the more inhomogeneous DI gel instead leads to a
strong decrease of the elastic and yield properties. The influence
of the microstructure is very evident here, as all the other
parameters (including the interparticle potential) remain un-
changed during the shearing process.
Another very direct and simple way to manipulate the micro-
structures of gelling colloids is to introduce voids into the struc-
ture intentionally. This can be achieved by adding alkaline
swellable thickeners (ASTs) to the liquid phase of a colloidal
alumina suspension. Hesselbarth et al.79 use polyacrylic acids,
such as Acusol 820 or Acusol 830, that are slightly internally
cross-linked. These are introduced at low pH (pH 4.5) into high
volume fraction ceramic suspensions. At this pH, these AST
polymers are small (diameter  80 nm) insoluble polymer par-
ticles that have only a minor influence on the viscosity of the
suspensions.
In Fig. 7, we plot the diameter of these polymers, as measured
by dynamic light scattering in dilute aqueous suspension, as a
function of pH. The diameter of the AST polymers increases by
approximately a factor of 10 between pH 4 and pH 9, the iso-
electric point of Al2O3. Thus, increasing the pH in such an alu-
mina suspension from pH 4 to pH 9 by using the DCC
technique (see Box 1) then leads not only to destabilization of
the colloids but also to a drastic increase in the size of these AST
polymers. The AST polymers thus produce inhomogeneities in
the evolving microstructure, which can be controlled by the size
and the concentration of the added AST polymer.
The presence of small amounts of AST polymers (0.4 wt%
with respect to alumina) results in a more than 10-fold increase
in the yield strength and elastic modulus of the wet coagulated
gels. Figure 8 shows the compressive yield stress sy as a function
of the AST polymer content. This drastic change in the me-
chanical behavior is a result of the formation of a less homoge-
neous microstructure, caused by the presence of the swollen
AST polymers. In case higher urea concentrations are used the
reaction first produce a coagulated gel with swollen polymers
and high yield strength but further ions created at the buffer pH
( 5 9) of the urea hydrolysis reaction lead to shielding of the
deprotonated sidegroups of the polymers. This causes the pol-
ymer particles to shrink again and the gel relaxes to a more ho-
mogenous microstructure with a lower yield strength (arrows in
Fig. 8).
Thus, in contrast to the study of Channell et al.,10 the results
of Wyss et al.47,60,61,68,69 and Hesselbarth et al.79 show a signif-
icant increase of the elastic and yield properties of concentrated
particle gels with decreasing homogeneity of their microstruc-
tures.
In the following, we discuss this apparent contradiction and
attempt to provide a qualitative explanation. Figure 9 shows a
schematic representation of the different kinds of microstruc-
ture. Figure 9(a) shows a microstructure obtained by diluting a
more concentrated, homogeneous gel under slow stirring. The
resulting structure consists of dense aggregates that are loosely
packed with only a few bonds between aggregates.
Although the mechanical behavior of the aggregates them-
selves will be much stronger than that of a homogenous network
(Fig. 9(b)), the overall gel structure will be weaker because of the
sparse and weak bonds between the aggregates. Channell et al.10
present a model that takes into account both the inter-aggregate
and intra-aggregate bonds and agrees well with their experimen-
tal data. It is also evident from this model that the overall
mechanical behavior for a packing of large aggregates will be
dominated by the weakness of the inter-aggregate bonds. Alter-
natively, one could view the aggregates as stiff particles and
consider the scaling of the expected mechanical behavior with
the size of these ‘‘aggregate-particles.’’ For packed particles with
fixed coordination number and fixed interaction strength, one
expects the mechanical properties to scale with the number of
bonds per volume. This implies that the mechanical strength, for
September 2005 Concentrated Particle Gels 2343

Box 1. DCC
70–72 73
DCC is one of the few methods that allow a true in situ destabilization of colloidal suspensions by a direct change of the
interparticle potential. Electrostatically stabilized suspensions are destabilized by shifting the pH of the suspensions toward their
IEP or by increasing the ionic strength in the system. The DCC technique allows for in situ destabilization of the suspensions,
without perturbing the gel formation. The onset and speed of the reaction are controlled by both temperature and enzyme
concentration. This is achieved by the use of time-delayed internal chemical reactions, such as enzyme-catalyzed hydrolysis
reactions, or hydrolysis reactions induced by elevated temperatures. For the destabilization of alumina suspensions, the
hydrolysis of urea,
NH2 CONH2 þ H2 O ! CO2 þ 2NH3 ! HCO
þ
3 þ 2NH4 ðB1Þ
catalyzed by the enzyme urease can be used. The pH of the suspension is shifted because of the products of this reaction to a
buffer pH of 9 in the presence of the enzyme urease. For other oxides with IEPs different from that of alumina, many other
enzymatic destabilization systems are available.54
Furthermore, the ionic strength in a suspension is increased homogeneously and continuously at the buffer pH of 9 by the
products of Eq. (B1).
DpH and DI methods At low ionic strength and at pH values far from the IEP, an oxide particle suspension is electrostatically
stabilized. This is shown in the schematic stability diagram in Fig. B1. Thus, the suspension is destabilized by either shifting the pH
toward the IEP or by increasing the ionic strength to high values. The two different destabilization pathways are referred to as
DpH destabilization and DI destabilization, respectively.
Although in the final stage of destabilization, the interparticle potentials for the two routes might be comparable, the pathway of
destabilization—the way the interparticle potentials change during the continuous destabilization process—is completely different.
For the two different methods, the evolution of DLVO potentials shows characteristic differences. Interaction potentials, as
predicted bythe DLVO theory using parameters typical for alumina suspensions, are plotted in Fig. B2. A secondary minimum in
the DLVO potential appears during the DDI destabilization, whereas for the case of DpH destabilization, no secondary minimum is
observed.

Fig. B1. Schematic stability diagram for a colloidal suspension. At low ionic
strength and at pH values far from the isoelectric point (IEP) the suspension
is electrostatically stabilized. Moving the pH toward the IEP (pathway (a),
DpH destabilization) or increasing the ionic strength to high values (pathway
(b), DI destabilization) leads to aggregation of the suspension as the surface
charge or the Debye screening length, respectively, is reduced.

Fig. B2. Evolution of Derjaguin–Landau–Verwey–Overbeek-potentials for the two destabilization mechanisms. A Hamaker constant of
AH 5 5  10
20 J and a particle radius of 200 nm is used. (a) Decrease of surface charge at a constant ionic strength of I 5 0.001 mol/L (DpH
destabilization). z potential values in the curves are z 5 60 mV (i), 30 mV (ii), 20 mV (iii), 10 mV (iv), and 0 mV (v). (b) DI destabilization: Increase of
ionic strength at constant surface charge (The z potential is kept constant at z 5 60 mV). Ionic strength values in the curves are I 5 0.001 mol/L (i),
0.02 mol/L (ii), 0.1 mol/L (iii), 0.2 mol/L (iv), 1.0 mol/L (v).

DCC, direct coagulation casting; IEP, isoelectric point; DLVO, Derjaguin–Landau–Verwey–Overbeek.

2344 Journal of the American Ceramic Society—Wyss et al. Vol. 88, No. 9

Box 2. DWS
DWS extends DLS to the highly multiple scattering regime.
Traditional DLS relies on samples that show only negligible multiple scattering. In this case, the length scale 1/q that one
observes in an experiment is related to the scattering angle y via

4pn
q¼ sinðy=2Þ ðB2Þ
l
where k is the wavelength of the laser, and n is the refractive index of the sample. Observations of the fluctuations in the detected
intensity then yields information on the dynamics of the system on the length scale 1/q. The fluctuations in intensity are
characterized by the temporal intensity autocorrelation function g2(q, t):

hIðtÞIðt þ tÞi
g2 ðtÞ ¼ D E ðB3Þ
IðtÞ2

The intensities at time t and at time (t1s) are correlated as long as during the lag time s the particles have moved only over a
distance much shorter than the length scale 1/q. When the particle displacements become much larger than this length scale, the
two intensities will no longer be correlated. For a stable suspension of monodisperse particles, the electric field autocorrelation
function
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
g1 ðtÞ ¼ g2 ðtÞ
1 ðB4Þ

will decay as
2
g1 ðtÞ ¼ e
Dq t ðB5Þ

where D is the diffusion constant of the particles. The size of the particles can thus be derived from such a measurement.
In the case of a highly multiple scattering sample, the scenario is different. The detection angle h no longer yields any useful
information about length scales, as each photon passing through the sample would have been scattered by many particles before
it reaches the detector. The movement of each of these particles will contribute to the total loss of temporal correlation in the
detected signal. In DWS, the transport of photons through a highly multiple scattering sample is treated as a diffusion
process.45,46 This allows the theoretical prediction of the distribution of the path length of photons through the sample for
different sample geometries. An important parameter is the transport mean free path l  . It is the distance traveled in the sample
after which the direction of the photon is completely randomized. The electric field autocorrelation function is then given by
Z 1
 2
g1 ðtÞ ¼ PðsÞ e
½ðs=l Þk hDtðtÞi ds ðB6Þ
0

where P(s) is the distribution of photon path lengths, L is the cell thickness, and /Dr2S is the mean square displacement of the
particles. Thus, DWS allows measurement of particle sizes in stable suspensions. The analysis of the particle dynamics can also
be used to characterize the viscoelastic behavior of colloidal gels.84
The main drawback of the technique is its inability to access different length scales. As many particles contribute to the loss of
temporal correlation for each photon path, all measurements represent average properties of the system. On the other hand, this
fact also implies that the technique allows measurement of the dynamics of the system at extremely short length scales. Each
particle only has to move a very short distance (typically sub-nanometers) in order for a photon path length to change by a
wavelength of light.
However, the best advantage of DWS is the possibility to characterize highly turbid colloidal systems that cannot be studied
by means of conventional light scattering. DWS is thus ideally suited for the characterization of ceramic suspensions, whose
strong turbidity excludes them from access by other techniques.
DLS, dynamic light scattering.

example, the compressive yield stress sy, decreases with the size
Raggr of the aggregate as:
sy / R
3
aggr (12)
The structures shown schematically in Figs. 9(c) and 9(d) are
completely different. They were formed by an in situ process,
yielding structural properties that cannot be achieved by mixing
or shearing. In a Gedanken Experiment, we view this kind of
structure as a two-component system, with one phase consisting
of voids (filled with water) and the other phase consisting of a
more concentrated, homogeneous gel phase. The gel phase is
percolating the whole macroscopic volume, while the other
phase consists of isolated pores. This is in contrast to the case
of Fig. 9(a), where the liquid phase between aggregates spans the
overall volume.
VI. Summary and Conclusions
The microstructure of concentrated particle gels strongly influ-
ences their rheological properties. A significant number of
studies have focused on this issue in recent years, aiming at
describing relations between the structural and rheological and
mechanical properties of colloidal gel networks. Nevertheless,
our understanding of the role of microstructure in concentrated
colloids is still rather limited. More work remains to be carried
out to gain a more thorough understanding.
The reasons for this lack of understanding are twofold:
A first important issue is the way microstructure is quantified.
Even if detailed knowledge on the microstructure of a particle
gel is experimentally accessible, for instance as a full set of par-
ticle positions in the network, this information has to be trans-
lated into a useful measure that describes the macroscopic
structure and allows a link to, for instance, the rheological
September 2005 Concentrated Particle Gels
Fig. 5. Top: Scanning electron microscopy images of high-pressure
frozen samples at a volume fraction of f 5 0.4. (a) After DpH des-
tabilization; (b) after DI destabilization. Bottom: Visualizations of
Brownian dynamics simulations74,75 for fast coagulation (c) and slow
coagulation (d).
behavior of the system. It is unclear whether quantities like the
radial pair correlation function g(r), or the structure factor S(q)
are able to capture the essential properties of microstructures
that can be related to the properties of a macroscopic gel net-
work in a meaningful way. Alternative ways of quantifying
microstructures 76,80–82 should be considered in the future. They
might provide us with a much better measure of the essential
structural properties of particle networks.
A second problem is the lack of control on the possible micro-
structures that can be experimentally realized in a reproducible
way. In concentrated and strongly attractive colloidal suspen-
Fig. 6. Compressive yield stress sy as a function of volume fraction f
for DpH- versus DI-destabilized gels.
2345
Fig. 7. Diameter of the alkaline swellable thickener (AST) particles in
H2O, as measured via dynamic light scattering, as a function of pH. The
concentration of the AST was 0.05 wt% of dry polymer, relative to H2O.
sions, the aggregation times are too fast to allow the formation
of a homogeneous and reproducible microstructure by direct
addition of salt or acid/base. In order to reach reproducible
conditions, the aggregated gels have to be strongly sheared prior
to measurement. However, this implies that the microstructure
will be governed mainly by the pre-shearing process and not by
the particle aggregation kinetics themselves. In situ destabilizat-
ion methods offer the possibility of inducing formation gels in a
reproducible way without shearing or mixing.7,9,71 Even differ-
ent kinds of aggregation kinetics, leading to significantly dif-
ferent microstructures, can be realized.47,61,71,83 This allows to
directly study the effect of different kinds of microstructures on
the bulk properties of the formed colloidal gels. Another way to
realize different microstructures is to concentrate dilute suspen-
sions of aggregated clusters into dense gels by centrifugation or
pressure filtration. The morphology and size of particle aggre-
gates in dilute suspension can be controlled by applying different
shear rates to the suspension. A gel of the compacted clusters
will then have properties that will depend on the structure and
Fig. 8. Compressive yield strength of wet green bodies with 57 vol%
alumina, as a function of polymer content and urea concentration (UC).
The yield strength is around 10 kPa for coagulated suspensions without
polymers. High yield strength results for coagulated gels with swollen
polymers inducing inhomogeneities in the gel microstructures upon
changing pH from 4 to 9. In case of much higher urea concentrations
(1.96 mol/L), after shifting pH from 4 to 9, the salt concentration is in-
creased and the swollen polymers shrink as a result of shielding the neg-
ative charges of the deprotonated polymer side groups. Thereby, the gel
microstructures relax, leading again to more homogenous microstruc-
tures and to lower yield strengths (indicated by the arrows in the figure).
2346 Journal of the American Ceramic Society—Wyss et al.
Fig. 9. Schematic illustration of microstructures obtained from differ-
ent gel-forming methods: structure obtained by diluting a more con-
centrated, homogeneous gel under slow stirring (a). Homogeneous
particle network (b) as for instance formed by DCC using the DpH
method (see Box 1). Inhomogeneous, but strongly interconnected net-
work (c) as for instance formed by DCC using the DI method (see Box
1). Particle gel with intentionally introduced voids (d), as achieved by
introducing alkaline swellable thickeners into the suspension. Note that
the homogeneity of the gel structures is not sufficient to explain
differences in mechanical properties.
size of the individual building blocks that constitute the net-
work.10
Future studies are necessary to provide us with a deeper
understanding of the complicated interplays between structur-
al and mechanical properties. The challenge will be to com-
bine new ways of producing different kinds of microstructure
and new, adequate measures to quantify microstructural
differences.
Acknowledgments
The authors thank Theo Tervoort for valuable discussions.
Vol. 88, No. 9
References
1
J. A. Lewis, ‘‘Colloidal Processing of Ceramics,’’ J. Am. Ceram. Soc., 83 [10]
2341–59 (2000).
2
W. M. Sigmund, N. S. Bell, and L. Bergström, ‘‘Novel Powder Process-
ing Methods for Advanced Ceramics,’’ J. Am. Ceram. Soc., 83 [7] 1557–74
(2000).
3
G. Harrison, G. V. Franks, V. Tirtaatmadja, and D. V. Boger, ‘‘Suspensions
and Polymers—Common Links in Rheology,’’ Korea—Aust. Rheol. J., 11 [3] 197–
218 (1999).
4
S. Klein, M. Fisher, G. Franks, M. Colic, and F. Lange, ‘‘Effect of the Inter-
particle Pair Potential on the Rheological Behavior of Zirconia Powders: Ii, the
Influence of Chem-Adsorbed Silanes,’’ J. Am. Ceram. Soc., 84 [5] 991 (2001).
5
W. Y. Shih, W.-H. Shih, and I. A. Aksay, ‘‘Elastic and Yield Behavior of
Strongly Flocculated Colloids,’’ J. Am. Ceram. Soc., 82 [3] 616–24 (1999).
6
F. Stenger and W. Peukert, ‘‘The Role of Particle Interactions on Suspension
Rheology—Application to Submicron Grinding in Stirred Ball Mills,’’ Chem. Eng.
Technol., 26 [2] 177 (2003).
7
C. J. Rueb and C. F. Zukoski, ‘‘Viscoelastic Properties of Colloidal Gels,’’
J. Rheol., 41 [2] 197–218 (1997).
8
M. K. Chow and C. F. Zukoski, ‘‘Nonequilibrium Behaviour of Dense Sus-
pensions of Uniform Particles: Volume Fraction and Size Dependence of Rheo-
logy and Microstructure,’’ J. Rheol., 39 [1] 33 (1995).
9
M. C. Grant and W. B. Russel, ‘‘Volume-Fraction Dependence of Elastic
Moduli and Transition Temperatures for Colloidal Silica Gels,’’ Phys. Rev. E, 47
[4] 2606–14 (1993).
10
G. M. Channell, K. T. Miller, and C. F. Zukoski, ‘‘Effects of Microstructure
on the Compressive Yield Stress,’’ A.I.Ch.E. J., 46 [1] 72–8 (2000).
11
A. A. Potanin, ‘‘On the Model of Colloid Aggregates and Aggregating Col-
loids,’’ J. Chem. Phys., 96 [12] 9191–200 (1992).
12
R. C. Sonntag and W. B. Russel, ‘‘Structure and Breakup of Flocs Subjected
to Fluid Stresses,’’ J. Colloid Interface Sci., 115 [2] 378–90 (1987).
13
D. A. Weitz and M. Oliveria, ‘‘Fractal Structures Formed by Kinetic Aggre-
gation of Aqueous Gold Colloids,’’ Phys. Rev. Lett., 52 [16] 1433–6 (1984).
14
P. Meakin, ‘‘Fractal Aggregates,’’ Adv. Colloid Interface Sci., 28, 249–331
(1998).
15
P. Varadan and M. J. Solomon, ‘‘Shear-Induced Microstructural Evolution of
a Thermoreversible Colloidal Gel,’’ Langmuir, 17 [10] 2918–29 (2001).
16
G. M. Channell and C. F. Zukoski, ‘‘Shear and Compressive Rheology of
Aggregated Alumina Suspensions,’’ A.I.Ch.E. J., 43 [7] 1700–8 (1997).
17
P. D. A. Mills, J. W. Goodwin, and B. W. Grover, ‘‘Shear Field Modification
of Strongly Flocculated Suspensions—Aggregate Morphology,’’ Colloid Polym.
Sci., 269 [9] 949–63 (1991).
18
J. C. Husband and J. M. Adams, ‘‘Shear-Induced Aggregation of Carboxy-
lated Polymer Latices,’’ Colloid Polym. Sci., 270 [12] 1194–200 (1992).
19
J. Bergenholtz, ‘‘Theory of Rheology of Colloidal Dispersions,’’ Curr. Opin.
Colloid Interface Sci., 6 [5–6] 484–8 (2001).
20
W. B. Russel, D. A. Saville, and W. R. Schowalter. ‘‘Chapter 14: Rheology’’;
pp. 456–505 in Colloidal Dispersions, Edited by G. K. Batchelor, Cambridge
University Press, Cambridge, UK.
21
H. Watanabe, L. Yao Ming, K. Osaki, T. Shikata, H. Niwa, Y. Morishima, N.
P. Balsara, and H. Wang, ‘‘Nonlinear Rheology and Flow-Induced Structure in a
Concentrated Spherical Silica Suspension,’’ Rheol. Acta, 37 [1] 1–6 (1998).
22
A. A. Catherall, J. R. Melrose, and R. C. Ball, ‘‘Shear Thickening and Order–
Disorder Effects in Concentrated Colloids at High Shear Rates,’’ J. Rheol., 44 [1]
1–25 (2000).
23
C. G. De Kruif, E. M. F. Van Iersel, A. Vrij, and W. B. Russel, ‘‘Hard Sphere
Colloidal Dispersions: Viscosity as a Function of Shear Rate and Volume
Fraction,’’ J. Chem. Phys., 83, 4717 (1986).
24
G. V. Franks, Z. Zhou, and N. J. Duin, ‘‘Effect of Interparticle Forces on
Shear Thickening of Oxide Suspensions,’’ J. Rheol., 44 [4] 759–79 (2000).
25
I. M. Krieger, ‘‘Rheology of Monodisperse Lattices,’’ Adv. Colloid Interface
Sci., 3, 111 (1972).
26
C. Dux and H. Versmold, ‘‘Light Diffraction From Shear-Ordered Colloidal
Dispersions,’’ Phys. Rev. Lett., 78 [9] 1811 (1997).
27
C. Dux, S. Musa, V. Reus, H. Versmold, D. Schwahn, and P. Lindner, ‘‘Small
Angle Neutron Scattering Experiments From Colloidal Dispersions at Rest and
under Shear Conditions,’’ J. Chem. Phys., 109 [6] 2556 (1998).
28
H. A. Barnes, ‘‘Shear-Thickening (‘‘Dilatancy’’) in Suspensions of Nonaggre-
gating Solid Particles Dispersed in Newtonian Liquids,’’ J. Rheol., 33 [2] 329–66
(1989).
29
V. Prasad, V. Trappe, A. D. Dinsmore, P. N. Segre, L. Cipelletti, and D. A.
Weitz, ‘‘Universal Features of the Fluid to Solid Transition for Attractive Colloi-
dal Particles,’’ Faraday Discuss., 123, 1–12 (2003).
30
S. Manley, H. M. Wyss, K. Miyazaki, J. C. Conrad, V. Trappe, L. J.
Kaufman, D. R. Reichman, and D. A. Weitz, ‘‘Dynamic Arrest in Spinodal De-
composition as a Route to Gelation’’; cond-mat/0412716.
31
P. N. Segre, V. Prasad, A. B. Schofield, and D. A. Weitz, ‘‘Glasslike Kinetic
Arrest at the Colloidal-Gelation Transition,’’ Phys. Rev. Lett., 86 [26] 6042–5
(2001).
32
K. N. Pham, A. M. Puertas, J. Bergenholtz, S. U. Egelhaaf, A. Moussaid, P.
N. Pusey, A. B. Schofield, M. E. Cates, M. Fuchs, and W. C. K. Poon, ‘‘Multiple
Glassy States in a Simple Model System,’’ Science, 296 [5565] 104–6 (2002).
33
W. C. K. Poon, S. U. Egelhaaf, J. Stellbrink, J. Allgaier, A. B. Schofield,
and P. N. Pusey, ‘‘Beyond Simple Depletion: Phase Behaviour of Colloid–Star
Polymer Mixtures,’’ Philos. Trans. R. Soc. London Ser.A, 359 [1782] 897–907
(2001).
34
A. Yethiraj and A. Van Blaaderen, ‘‘A Colloidal Model System with an In-
teraction Tunable From Hard Sphere to Soft and Dipolar,’’ Nature, 421 [6922]
513–7 (2003).
September 2005 Concentrated Particle Gels 2347
35 70
D. A. Weitz, M. Y. Lin, H. M. Lindsay, and J. S. Huang, ‘‘Limits of the L. J. Gauckler, T. J. Graule, F. Baader, and J. Will, ‘‘Enzyme Catalysis of
Fractal Dimension for Irreversible Kinetic Aggregation of Gold Colloids,’’ Phys. Alumina Forming,’’ Key Eng. Mater., 159–160, 135–50 (1999).
71
Rev. Lett., 54 [13] 1416–9 (1985). L. J. Gauckler, T. Graule, and F. Baader, ‘‘Ceramic Forming Using Enzyme
36
P.-G. de Gennes, Scaling Concepts in Polymer Physics. Cornell University, Catalyzed Reactions,’’ Mater. Chem. Phys., 2509, 1–25 (1999).
72
Ithaca, 1979. T. J. Graule, F. H. Baader, and L. J. Gauckler, ‘‘Direct Coagulation Casting
37
R. De Rooij, D. Van Den Ende, M. H. G. Duits, and J. Mellema, ‘‘Elasticity (Dcc)—Principles of a New Green Shaping Technique,’’ Ceram. Mater. Compo-
of Weakly Aggregating Polystyrene Latex Dispersions,’’ Phys. Rev. E, 49 [4] 3038– nents Engines V, 626–31 (1994).
73
49 (1994). L. Bergstrom and E. Sjostrom, ‘‘Temperature Induced Gelation of Concen-
38
Y. Kantor and I. Webman, ‘‘Elastic Properties of Random Percolating Sys- trated Ceramic Suspensions: Rheological Properties,’’ J. Eur. Ceram. Soc., 19 [12]
tems,’’ Phys. Rev. Lett., 52 [21] 1891–4 (1984). 2117–23 (1999).
39 74
A. A. Potanin, ‘‘On the Computer Simulation of the Deformation and Break- J. Innerlohinger, H. M. Wyss, and O. Glatter, ‘‘Colloidal Systems with At-
up of Colloidal Aggregates in Shear Flow,’’ J. Colloid Interface Sci., 157, 399–410 tractive Interactions: Evaluation of Scattering Data Using the Generalized Indirect
(1993). Fourier Transformation Method,’’ J. Phys. Chem. B, 108 [47] 18149–57 (2004).
40 75
H. J. Hermann and H. E. Stanley, ‘‘Building Blocks of Percolation Clusters: M. Hütter, ‘‘Coagulation Rates in Concentrated Colloidal Suspensions Stud-
Volatile Fractals,’’ Phys. Rev. Lett., 53, 1121–4 (1984). ied by Brownian Dynamics Simulation,’’ Phys. Chem. Chem. Phys., 1 [18] 4429–36
41
G. Bushell, Y. Yan, D. Woodfield, J. Raper, and R. Amal, ‘‘On Techniques for (1999).
76
the Measurement of the Mass Fractal Dimension of Aggregates,’’ Adv. Colloid In- M. Hütter, ‘‘Local Structure Evolution in Particle Network Formation Studied
terface Sci., 95 [1] 1–50 (2002). by Brownian Dynamics Simulation,’’ J. Colloid Interface Sci., 231 [2] 337–50 (2000).
42 77
M. Kerker, The Scattering of Light and Other Electromagnetic Radiation. Ac- B. Balzer, M. K. M. Hruschka, and L. J. Gauckler, ‘‘In-Situ Rheological In-
ademic Press, New York, 1969. vestigation of the Coagulation in Aqueous Alumina Suspensions,’’ J. Am. Ceram.
43
K. Schätzel, ‘‘Suppression of Multiple-Scattering by Photon Cross-Correla- Soc., 84 [8] 1733–9 (1999).
78
tion Techniques,’’ J. Mod. Opt., 38, 1849 (1991). H. M. Wyss, ‘‘Microstructure and Mechanical Behavior of Concentrated Par-
44
C. Urban and P. Schurtenberger, ‘‘Characterization of Turbid Colloidal Sus- ticle Gels’’; PhD Thesis, ETH Zurich, 2003.
79
pensions Using Light Scattering Techniques Combined with Cross-Correlation D. Hesselbarth, E. Tervoort, C. Urban, and L. J. Gauckler, ‘‘Mechanical
Methods,’’ J. Colloid Interface Sci, 207 [1] 150–8 (1998). Properties of Coagulated Wet Particle Networks with Alkali-Swellable Thick-
45
G. Maret and P. E. Wolf, ‘‘Multiple Light Scattering From Disordered Media. eners,’’ Mechanical Properties of Coagulated Wet Particle Networks with Alkali-
The Effect of Brownian Motion on Scatterers,’’ Z. Phys. B, 65, 409–13 (1987). Swellable Thickeners,’’ J. Am. Ceram. Soc., 84 [8] 1689–95 (2001).
46 80
D. J. Pine, D. A. Weitz, P. M. Chaikin, and E. Herbolzheimer, ‘‘Diffusing- C. Arns, M. Knackstedt, and K. Mecke, ‘‘Reconstructing Complex Materials
Wave Spectroscopy,’’ Phys. Rev. Lett., 60 [12] 1134–7 (1988). Via Effective Grain Shapes,’’ Phys. Rev. Lett., 91 [21] 215506-1–4 (2003).
47 81
H. M. Wyss, S. Romer, F. Scheffold, P. Schurtenberger, and L. J. Gauckler, K. Mecke, ‘‘The Shape of Parallel Surface: Porous Media, Fluctuating Inter-
‘‘Diffusing-Wave Spectroscopy of Concentrated Alumina Suspensions During faces and Complex Fluids,’’ Physica A, 314 [1–4] 655–62 (2002).
82
Gelation,’’ J. Colloid Interface Sci., 241, 89–97 (2001). K. Mecke, T. Buchert, and H. Wagner, ‘‘Robust Morphological Measures for
48
M. Ballauff, ‘‘Saxs and Sans Studies of Polymer Colloids,’’ Curr. Opin. Colloid Large-Scale Structure in the Universe,’’ Astron. Astrophys, 288 [3] 697–704 (1994).
83
Interface Sci., 6 [2] 132–9 (2001). B. Balzer, M. K. M. Hruschka, and L. J. Gauckler, ‘‘Coagulation Kinetics and
49
A. Guinier and G. Fournet, Small Angle Scattering of X-rays. Wiley, New Mechanical Behavior of Wet Alumina Green Bodies Produced Via Dcc,’’ J. Col-
York, 1955. loid Interface Sci., 216 [2] 379–86 (1999) FTXT Academic Press IDEAL (European
50
O. Glatter and O. Kratky, Small Angle X-ray Scattering. Academic Press, Mirror) http://www.europe.idealibrary.com/links/doi/10.1006/jcis.1999.6293.
84
London, 1982. A. H. Krall and D. A. Weitz, ‘‘Internal Dynamics and Elasticity of Fractal
51
J. S. Higgins and H. C. Benoit, Polymers and Neutron Scattering. Academic Colloidal Gels,’’ Phys. Rev. Lett., 80 [4] 778–81 (1998). &
Press, Oxford, 1994.
52
R. Gahler, R. Golub, K. Habicht, T. Keller, and J. Felber, ‘‘Space–Time De-
scription of Neutron Spin Echo Spectrometry,’’ Physica B, 229 [1] 1–17 (1996).
53
M. Rekveldt, ‘‘Novel Sans Instrument Using Neutron Spin Echo,’’ Nucl.
Hans Wyss studied experimental
Instrum. Methods B, 114 [3–4] 366–70 (1996). physics at the Swiss Federal Insti-
54
W. Bouwman and M. Rekveldt, ‘‘Spin-Echo Small-Angle Neutron Scattering tute of Technology (ETH) in Zurich,
Calculations,’’ Physica B, 276, 126–7 (2000). Switzerland. For his graduate stud-
55
T. Krouglov, W. G. Bouwman, J. Plomp, M. T. Rekveldt, G. J. Vroege, A. V.
Petukhovb, and D. M. E. Thies-Weesieb, ‘‘Structural Transitions of Hard-Sphere
ies, he joined Prof. Ludwig Gauck-
Colloids Studied by Spin-Echo Small-Angle Neutron Scattering,’’ J. Appl. Cryst., ler’s group in the Department of
36, 1417 (2003). Materials at ETH, where he started
56
W. Bouwman, T. Krouglov, J. Plomp, S. Grigoriev, W. Kraan, and M. to work on ceramics processing and
Rekveldt, ‘‘Sesans Studies of Colloid Phase Transitions, Dairy Products and Pol-
ymer Fibres,’’ Physica B, 350 [1–3] 140–6 (2004).
colloid chemistry. His PhD thesis fo-
57
C. C. Donaldson, J. Mcmahon, R. F. Stewart, and D. Sutton, ‘‘The Charac- cused on the experimental character-
terisation of Structure in Suspensions,’’ Colloids Surface, 18, 2–4 (1986). ization of microstructure in dense
58
R. Menold, B. Lüttge, and W. Kaiser, ‘‘Freeze Fracturing: A New Method for ceramic suspensions and gels, as
the Investigation of Dispersions by Electron Microscopy,’’ Adv. Colloid Interface
Sci., 5, 281–335 (1976).
well as on the relation between the microstructure and the me-
59
E. Shimoni and M. Müller, ‘‘On Optimizing High-Pressure Freezing: From Heat chanical behavior of these systems.
Transfer Theory to a New Microbiopsy Device,’’ J. Microsc., 192 [3] 236–47 (1998). He is currently working as a post-doctoral fellow at Harvard
60
H. M. Wyss, M. Hütter, M. Müller, L. P. Meier, and L. J. Gauckler, ‘‘Quan- University in the lab of Prof. David Weitz. His general research
tification of Microstructures in Stable and Gelated Suspensions From Cryo-Sem,’’
J. Colloid Interface Sci., 248 [2] 340–6 (2002).
interest lies on the physics of soft materials such as colloids,
61
H. M. Wyss, E. Tervoort, L. P. Meier, M. Müller, and L. J. Gauckler, ‘‘Re- emulsions, or polymers. He is particularly interested in under-
lation Between Microstructure and Mechanical Behavior of Concentrated Silica standing the relations between the dynamics, structure, and
Gels,’’ J. Colloid Interface Sci., 273 [2] 455–62 (2004). macroscopic properties of these systems.
62
A. D. Dinsmore, E. R. Weeks, V. Prasad, A. C. Levitt, and D. A. Weitz,
‘‘Three-Dimensional Confocal Microscopy of Colloids,’’ Appl. Opt., 40 [24] 4152–
9 (2001). Elena V. Tervoort received her ac-
63
W. K. Kegel and A. Van Blaaderen, ‘‘Direct Observation of Dynamical Hetero- ademic degrees at the Moscow In-
geneities in Colloidal Hard-Sphere Suspensions,’’ Science, 287 [5451] 290–3 (2000).
64
E. Weeks, J. C. Crocker, A. Levitt, A. Schofield, and D. Weitz, ‘‘Three-Di-
stitute of Chemical Technology
mensional Direct Imaging of Structural Relaxation Near the Colloidal Glass and at the Department of Poly-
Transition,’’ Science, 287 [5453] 627–31 (2000). mers and Composites in the Rus-
65
L. Bergström, ‘‘Sedimentation of Flocculated Alumina Suspensions: Gamma- sian Academy of Science in
Ray Measurements and Comparison with Model Predictions,’’ J. Chem. Soc. Far-
aday Trans, 88 [21] 3201–11 (1992).
Moscow, Russia, where she
66
K. T. Miller, R. M. Melant, and C. F. Zukoski, ‘‘Comparison of the Com- worked as a research associate
pressive Yield Response of Aggregated Suspensions: Pressure Filtration, Centrif- from 1985 until 1992. Her PhD
ugation, and Osmotic Consolidation,’’ J. Am. Ceram. Soc., 79 [10] 2545–56 (1996). thesis was on the modification of
67
G. V. Franks, Y. Zhou, Y. Yan, G. Jameson, and S. Biggs, ‘‘Effect of Ag-
gregate Size on Sediment Bed Rheological Properties,’’ Phys. Chem. Chem. Phys.,
polyolefins with siloxane additives
6 [18] 4490–8 (2004). and anorganic fillers. After a com-
68
H. M. Wyss, J. Innerlohinger, L. P. Meier, L. J. Gauckler, and O. Glatter, bined post-doctoral stay at DSM research in Geleen, and the
‘‘Small-Angle Static Light Scattering of Concentrated Silica Suspensions During Polymer Technology group of the Eindhoven University of
In Situ Destabilization,’’ J. Colloid Interface Sci., 271 [2] 388–99 (2004).
69
H. M. Wyss, A. Deliormanli, E. Tervoort, and L. J. Gauckler, ‘‘Influence of
Technology, both in the Netherlands, she accepted a position
Microstructure on the Rheological Behavior of Dense Particle Gels,’’ A.I.Ch.E. J., as Research Associate in Prof. Gauckler’s Nonmetallic Inor-
51 [1] 134–41 (2005). ganic Materials group at the Department of Material Science at
2348 Journal of the American Ceramic Society—Wyss et al.
the ETH-Zurich in Switzerland. Her research interests include
colloidal aspects of ceramic processing and ceramic foams.
Ludwig J. Gauckler received his de-
gree in physics at the University of
Stuttgart and his PhD in materials
science in 1977 with a Thesis on
SiAlON that he completed at the
Max Planck Institute for Metals
and Materials Research in Stutt-
gart. He then carried out research
in the area of high-perform-
ance structural and functional ce-
ramics as research associate at the
University of Michigan, Ann Arbor
in 1977. From 1979 to 1988, he was responsible for the inorganic
non-metallic materials development in the central laboratories
of Alusuisse-Lonza AG. Since 1988, he is Professor for ‘‘Non-
metallic Inorganic Materials’’ in the Department of Material
Vol. 88, No. 9
Science at the ETH-Zurich. He served as head of the Depart-
ment from 1991 to 1993. He is guest professor at the Tsinghua
University in Beijing since 1993 and a guest professor at the
MIT, Boston in 2001.
Ludwig Gauckler and his co-workers received several nation-
al and international awards for their work on colloid chemistry
for ceramic processing and high-temperature solid oxide fuel
cells, rapid prototyping of ceramics (Cercon), among them, the
Award for Real Advances in Materials from NASTS (National
Association for Science, Technology, and Society) and the Fed-
eration of Materials Societies in 1994. He is a Fellow of
the American Ceramic Society and served as President of the
scientific advisory board of the Swiss Academy of Technical
Sciences, and is a member of the Academy of Ceramics. His re-
search interests include structural ceramics, high Tc-supercon-
ductors, computer-assisted modeling of thermodynamics, mixed
ionic–electronic conducting oxides and colloidal chemistry for
ceramic processing. Prof. Gauckler has published more than 180
scientific papers, and holds 15 patents.