Macroporous Ceramics from Particle-Stabilized Wet Foams
Urs T. Gonzenbach,
w
Andre
´
R. Studart,
w
Elena Tervoort, and Ludwig J. Gauckler*
Department of Materials, ETH Zu
¨
rich, Zu
¨
rich, CH-8093, Switzerland
We present a novel direct-foaming method to produce macro-
porous ceramics using particles instead of surfactants as stabil-
izers of the wet foams. This method allows for the fabrication of
ultra-stable wet foams that resist coarsening upon drying and
sintering. Macroporous ceramics of various chemical compos-
itions with open or closed cells, average cell sizes ranging from
10 to 300 lm and porosities within 45% and 95%, can be easily
prepared using this new approach. The sintered foams show high
compressive strengths of up to 16 MPa in alumina foams with
porosities of 88%.
I. Introduction
P
OROUS ceramics are of great interest due to their numerous
potential applications in catalysis, adsorption and separ-
ation, filtration of molten metals or hot gases, refractory insu-
lation of furnaces, as well as hard tissue repair and
engineering.
1,2
The three main processing routes for the fabri-
cation of macroporous ceramics are the replica technique,the
sacrificial template method, and the direct-foaming technique.
The processing route ultimately determines the microstructure
of the final macroporous ceramic.
3
Therefore, the selection of a
given processing method depends strongly on the microstructure
needed in the end application, as well as on the inherent features
of the process such as cost, simplicity, and versatility.
The direct-foaming technique is particularly suitable for the
fabrication of open and closed porous structures with porosities
ranging from 45% to 97% and cell sizes between 30 mm and 1
mm.
3
Direct-foaming methods involve the incorporation of a
gaseous phase into a ceramic suspension consisting of ceramic
powder, solvent, dispersants, surfactants, polymeric binder, and
gelling agents. The incorporation of the gaseous phase is carried
out either by mechanical frothing, injection of a gas stream, gas-
releasing chemical reactions, or solvent evaporation.
The liquid foams obtained upon gas incorporation are
thermodynamically unstable due to their large interfacial area.
Therefore, gas bubbles that initially nucleated as spheres grow as
polyhedral cells. There are mainly three processes responsible
for coarsening
4
: drainage and coalescence and Ostwald ripening.
Drainage of the liquid from the lamellas between the bubbles
results in a close approach of bubble surfaces, which can lead to
their coalescence and to foam collapse. Additionally, due to dif-
ferent Laplace pressures of bubbles with different sizes, gas dif-
fusion occurs from smaller to larger bubbles. This migration of
gas between bubbles leads to coarsening of the foam and to a
broadening of the bubble size distribution. Eventually, the liquid
foam collapses due to the combined action of these destabiliza-
tion mechanisms. Additives are often used to avoid foam col-
lapse by setting the foam structure shortly after air
incorporation. This usually occurs by gelling or cross-linking
organic compounds added to the suspension liquid medium.
5–11
A drawback of these methods is the fact that they cannot avoid
rapid bubble growth before the setting reaction takes place,
which leads to large average bubble sizes (30 mm–1 mm) and a
wide bubble size distribution.
3
We have recently shown
12
that ultra-stable wet foams can be
produced by using particles instead of surfactants as foam sta-
bilizers in the direct-foaming method. The energy gain upon
adsorption of a particle to the air–water interface of fresh gas
bubbles can be as high as thousands of kT’s, as opposed to the
adsorption energy of a few kT’s in the case of surfactants (k is
the Boltzmann constant and T is the temperature). Therefore,
particles can irreversibly adsorb at the surface of gas bubbles, in
contrast to surfactants that adsorb and desorb at relatively short
time scales at the interface.
13
The formation of ultra-stable wet foams requires the adsorp-
tion of partially hydrophobic particles to the air–water interface.
Partial hydrophobization can be achieved through the physical
or chemical adsorption of short-chain amphiphilic molecules on
the particle surface.
12
The short molecules adsorb with their
polar head group onto the particle,
14
leaving the hydrophobic
tail in contact with the aqueous solution. Owing to their surface
hydrophobicity, particles adsorb onto air–water interfaces and
reduce the foam overall free energy by removing part of the
highly energetic gas–liquid interfacial area. Foam formation is
therefore based on two consecutive assembly processes, namely
the adsorption of short-chain amphiphilic molecules on the par-
ticle surface and the adsorption of these partially hydrophobized
particles onto the air–water interface of freshly incorporated
bubbles. Through these assembly processes, a three-level hier-
archical structure is formed, containing short-chain amphiphilic
molecules in the molecular scale, partially hydrophobized par-
ticles in the colloidal scale, and finally air bubbles in the macro-
scopic scale.
12
The aim of this work is to investigate the stabilization of
foams with colloidal particles and the development of a pro-
cessing route for the fabrication of solid macroporous ceramics
with a tailored microstructure. In order to illustrate the main
features of this new process, we used mainly alumina powder
and short-chain carboxylic acids as particles and amphiphiles,
respectively, for foam stabilization. These molecules show a
high solubility in water and are thus suitable for the surface
modification of a high concentration of particles in the initial
suspension. In this article, the entire processing route to obtain
the macroporous ceramics is outlined, and the final microstruc-
ture and mechanical properties of the porous structures
produced are described.
II. Materials and Methods
(1) Materials
Experiments were carried out using high-purity a-Al
2
O
3
powder
(HPA-0.5w/MgO, Ceralox, Tucson, AZ) with an average par-
ticle diameter (d
50
) of 200 nm, a specific surface area of 10 m
2
/g,
and a density of 3.98 g/cm
3
. Other chemicals used in the experi-
ments were deionized water, hydrochloric acid (2 N, Titrisol,
Merck, Germany), and sodium hydroxide (1 N, Titrisol, Mer-
ck). The short-chain carboxylic acids were propionic, butyric,
valeric, and enanthic acids (Fluka, Buchs, Switzerland). Table I
G. Franks—contributing editor
Supported by CIBA Specialty Chemicals (Switzerland).
*Member, American Ceramic Society.
w
Authors to whom correspondence should be addressed. e-mail: urs.gonzenbach@
mat.ethz.ch; andre.studart@mat.ethz.ch
Manuscript No. 21696. Received April 12, 2006; approved July 19, 2006.
J
ournal
J. Am. Ceram. Soc., 90 [1] 16–22 (2007)
DOI: 10.1111/j.1551-2916.2006.01328.x
r 2006 The American Ceramic Society
16
shows the chemical formulas as well as the pK
a
values of these
amphiphiles. At a pH equal to the pK
a
, 50% of the dissolved
molecules are deprotonated. The amphiphile propyl gallate,
as well as the gelling agents hydroxyaluminum diacetate
((HADA); d
50
B0.7 mm) and sodium alginate (Fluka AG, Buchs,
Switzerland), were later used in this study to produce open-cell
macroporous structures.
(2) Suspension Preparation
Suspensions containing carboxylic acids were produced as fol-
lows: alumina powder was stepwise added to water containing
hydrochloric acid (0.7 wt% to alumina) to obtain a suspension
with a solids loading of 50 vol%. Homogenization and deag-
glomeration were carried out in a ball mill for at least 18 h using
polyethylene milling pots and alumina balls (10 mm diameter,
ratio balls:powderB2:1). An aqueous solution containing
the amphiphile and, if necessary, pH-adjusting agents were
then slowly and dropwise added to the ball-milled suspension
under slight stirring to avoid local particle agglomeration
and coagulation. Afterwards, the pH was set to its desired value
and the amount of water needed to achieve certain solids
contents was added.
Suspensions containing propyl gallate as an amphiphile were
prepared by stepwise adding alumina powder to water contain-
ing 506 mmol/L NaOH and 29 mmol/L propyl gallate. The
suspension solids loading and pH were initially fixed to 50 vol%
and 9.8, respectively. Homogenization of the suspensions was
carried out as described above. Afterwards, the propyl gallate
needed to adjust the amphiphile concentration in the final sus-
pension to 100 mmol/L was dissolved in a NaOH aqueous so-
lution displaying a pH higher than 10. This solution was then
slowly and dropwise added to the ball-milled suspension under
slight stirring to avoid local particle agglomeration. Finally, the
pH was set to 9.9 and the solids loading to 20 vol% by adding
water to the suspension.
(3) Adsorption Measurements
For the adsorption measurements, alumina suspensions were
prepared as described above. After ball milling, the carboxylic
acid was dropwise added to the suspension, the pH was set to
4.75, and the solids loading to 35 vol% alumina. The suspen-
sions were stirred for 2 h to achieve equilibrium conditions. Two
tubes, each filled with 45 mL of suspension, were then centri-
fuged (Z 513 K, Hermle, Wehingen, Germany) for 1 h at a speed
of 4500 rpm. The supernatant obtained was again centrifuged
(5417R Eppendorf, Leipzig, Germany) using 16 small tubes,
each containing 1.5 mL of the supernatant. In this case, the
centrifugation time was set to 20 min and the rotation speed was
15 000 rpm. The pH of the resulting supernatant was set to a
value higher than 10 to ensure complete deprotonation of the
amphiphilic molecules. The amphiphile concentration in the su-
pernatant was measured by titrating the solution to a pH lower
than 2 using either 1 or 0.1 N HNO
3
. An automatic titration
unit (DT1200, Dispersion Technologies Inc., Mount Cisco, NY)
was used for these measurements.
(4) f-Potential Analysis
z-potential measurements (DT1200, Dispersion Technologies
Inc.) were conducted in 2 vol% alumina suspensions at pH
4.75 to assess the effect of the amphiphilic molecules on the
surface charge of the alumina particles. For each amphiphile
concentration, a new suspension was prepared and ultrasonica-
ted for 5 min before the measurement.
(5) Rheological Tests
The rheological behavior of the suspensions was evaluated using
a vane configuration in a stress-controlled rheometer (Model
CS-50, Bohlin Instruments, Cirencester, U.K.). The measure-
ments were performed under steady-shear conditions by apply-
ing a stepwise stress increase until a shear rate of about 200 s
1
was reached. The solids loading of the suspension was 35 vol%
alumina at pH 4.75.
(6) Surface Tension Measurements
The surface tension of the suspensions was measured using the
pendant drop method (PAT1, Sinterface Technologies, Berlin,
Germany). Alumina suspensions were prepared as mentioned
above, by dropwise adding the carboxylic acid, setting the pH to
4.75, and diluting the suspension to a solids loading of 35 vol%
alumina. Depending on the surface tension of the suspension,
the drop volume was set to a constant value within the range 12–
35 mm
3
.
(7) Foaming and Foam Characterization
Foaming of 150 mL suspensions was carried out using a house-
hold mixer (Kenwood, Major Classic, Schumpf AG, Baar,
Switzerland) at full power (800 W) for 3 min. The foam dens-
ity was measured with a custom-built tool, which consisted of a
plastic cylindrical cup with small holes on the bottom and a
massive sliding stamp on top. The foam was carefully filled into
the cup and then slightly compressed with the stamp to remove
possible air pockets introduced during filling. The volume be-
tween the bottom of the stamp and the bottom of the cylinder
was kept constant. Dividing the mass of the foam by its volume
resulted in the foam density. The bubble size distribution of the
wet foams was evaluated using an optical microscope in trans-
mission mode (Polyvar MET, Reichert-Jung, Austria) connect-
ed to a digital camera. The bubble sizes were measured with the
linear intercept method using the software Lince (Linear Inter-
cept, TU Darmstadt, Germany).
(8) Foam Setting and Drying
To avoid crack formation during drying, the wet foams were
strengthened before water evaporation by either coagulating the
particles within the foam lamella or by chemically gelling the
foam liquid phase.
In the case of wet foams containing carboxylic acid, the par-
ticles within the foam lamella were coagulated by changing the
pH in situ from 4.75 to 7.5 using the enzyme-catalyzed hydro-
lysis of urea (Sigma-Aldrich, Buchs, Switzerland).
15
The urea
content in the suspensions was 0.05 wt% with respect to alumina
and the concentration of the enzyme urease (Roche Diagnostics
GmbH, Mannheim, Germany) was 1 U/g alumina. The enzy-
matic activity of the urease used was 58 000 U/(g of pure urease).
One unit is defined as the amount of enzyme necessary to release
1 mmol of reaction product per minute from the substrate at
241C and at the pH where the enzymatic activity in water is at its
maximum. Urea was added to the suspension before homogen-
ization, whereas urease was dissolved in water and added to the
suspension before the foaming step.
In the case of wet foams containing propyl gallate as an
amphiphile, the foam liquid phase was gelled using the time-
delayed gelation between sodium alginate and HADA as
described elsewhere in the literature.
16
In order to distribute
homogeneously the gelling agent within the suspension, sodium
alginate was first dissolved in water at 801C. After cooling to
room temperature, the alginate solution was added to the sus-
pension under stirring. The amount of alginate added corre-
sponded to 0.25 wt% with respect to the mass of alumina in the
final suspension. The pH of the suspension was then set to 9.9,
Table I. Carboxylic Acids Used to Partially Hydrophobize the
Alumina Particles
Compound Chemical formula pK
a
Propionic acid CH
3
–CH
2
–COOH 4.86
Butyric acid CH
3
–[CH
2
]
2
–COOH 4.83
Valeric acid CH
3
–[CH
2
]
3
–COOH 4.84
Enanthic acid CH
3
–[CH
2
]
5
–COOH 4.89
January 2007 Macroporous Ceramics from Particle-Stabilized Wet Foams 17
and the amount of water needed to achieve a solids content of 20
vol% was added. In order to slow down the gelation process, the
suspension was cooled in ice before foaming. HADA was then
added as a powder to the suspension under stirring. The weight
ratio HADA:sodium alginate was 7:1. Foaming was finally con-
ducted by mechanical frothing as described.
The resulting wet foams were hand shaped into cylindrical
parts (diameter: 100 mm; height: 50 mm) and subsequently dried
in air at 221–251C for 24–48 h.
(9) Sintering of the Foams
Sintering of the cylindrical dried foams was performed in an
electrical furnace (HT 40/16, Nabertherm, Germany) at 15751C
for 2 h. The heating rate was set to 11C/min and the cooling rate
to 31C/min.
(10) Compressive Strength Measurements
Compressive strength measurements were performed on a uni-
versal testing machine (Instron 8562, model A1477-1003, Nor-
wood, MA). A bulk piece of ceramic foam was ground on both
sides, resulting in parallel opposite surfaces that ensured homo-
geneous sample loading during compression. Cylindrical sam-
ples with diameters of 15 mm and lengths of 30 mm were drilled
out of this bulk piece of foam with a diamond core drill and
crushed under a compression speed of 0.5 mm/min.
III. Results and Discussion
(1) Amphiphile Adsorption and f Potential
The formation of stable foams requires the adsorption of par-
ticles on the surface of freshly incorporated air bubbles.
12
In
order to enable their adsorption at the gas–liquid interface, par-
ticles with a partially hydrophobized surface are needed. In the
case of alumina, hydrophobization can be achieved by modify-
ing the particle surface with short-chain carboxylic acids that
adsorb with the carboxylate group onto alumina,
14
leaving the
hydrophobic tail in contact with the aqueous solution. The sur-
face properties of the resulting modified particles are mainly de-
termined by the concentration of adsorbed amphiphilic
molecules and their tail length.
Figure 1 shows that carboxylic acids with chain lengths of up
to six carbons adsorb onto alumina particles under acidic
conditions. All measurements were carried out in suspensions
containing 35 vol% alumina at pH 4.75 and carboxylic acid
concentrations typically required to obtain stable foams.
12
Under such conditions, the dissociated carboxylic acid mole-
cules can adsorb electrostatically as counter-ions onto the op-
positely charged alumina surface or through ligand exchange
reactions with the alumina hydroxyl surface groups.
14
We ob-
served that the addition of competing counter-ions such as Cl
to the suspension decreased the amount of adsorbed carboxylate
molecules, indicating that the amphiphiles are predominantly
adsorbed as counter-ions on the particle surface.
The adsorption of the carboxylic acids on alumina was also
influenced by the concentration of particles in the suspension.
For a constant ratio of added propionic acid to particle surface
area of 1.23 mmol/m
2
, we observed a linear increase in the
amount of adsorbed amphiphiles from 0.44 to 1.11 mmol/m
2
by increasing the suspension solids content from 5 to 35 vol%,
respectively. Based on these adsorption data, we estimate an
adsorbed amount of propionic acid of 0.35 mmol/m
2
at a solids
content of 2 vol%. This estimated value is in good agreement
with the data reported by Hidber et al.
14
for this particle con-
centration (0.31 mmol/m
2
). The increase in solids content in the
above experiments was also accompanied by an increase in the
ionic strength of the liquid media. As a result of the increased
ionic strength, higher surface charges are developed on the par-
ticle surface
17
and higher concentrations of counter-ions are
needed for charge neutralization. Therefore, the enhanced ad-
sorption of amphiphilic counter-ions observed at high solids
content might result from the increased surface charge devel-
oped on the particles at high particle concentrations.
The surface modification accomplished through the adsorp-
tion of carboxylic acid molecules also led to a significant reduc-
tion of the particle’s z potential (potential at the shear plane), as
shown in Fig. 2. Surprisingly, this reduction strongly depends on
the tail length of the amphiphilic molecule and is far more pro-
nounced than that achieved with equivalent amounts of the 1:1
electrolyte sodium chloride. The addition of high concentrations
of carboxylic acids did not invert the z potential sign nor
changed the alumina isoelectric point, confirming that these
molecules adsorb as counter-ions rather than as specific adsorb-
ing species around the particles. Therefore, the electrical poten-
tial on the alumina surface remains constant at 45 mV upon
addition of amphiphiles or back electrolyte. On the other hand,
the z potential is strongly influenced by the concentration and
valency of counter-ions in the diffuse layer that screen the par-
ticle surface charge.
18
The stronger screening effect of amphiph-
iles as compared with a standard electrolyte (Fig. 2) might result
from the specific adsorption of additional anions onto the par-
tially hydrophobized particle surfaces. The hydrophobicity im-
parted by the first layer of deprotonated amphiphiles adsorbed
on the surface leads to an energetically unfavorable exposure of
hydrophobic species into the aqueous phase. This favors the
adsorption of additional molecules from the aqueous phase onto
0 50 100 150 200 250
0
2
4
6
8
Propionic Acid
Butyric Acid
Valeric Acid
Enanthic Acid
Adsorbed amphiphile (µmol/m
2
)
Initial amphiphile concentration (mmol/l)
100% adsorption
Fig. 1. Adsorption of short-chain carboxylic acids on alumina parti-
cles: propionic acid (&), butyric acid (J), valeric acid (m), and enanthic
acid (
X). Measurements were obtained from 35 vol% alumina suspen-
sions at pH 4.75.
0 25 50 75 100 12
5
0
10
20
30
40
50
60
NaCl
Propionic Acid
Butyric Acid
Valeric Acid
Enanthic Acid
Zeta-Potential (mV)
Amphiphile concentration (mmol/l)
Fig. 2. z-potential measurements of 2 vol% alumina suspensions con-
taining NaCl (&), propionic acid (&), butyric acid (J), valeric acid (m),
or enanthic acid (
X).
18 Journal of the American Ceramic Society—Gonzenbach et al. Vol. 90, No. 1
the particle surface to decrease the system free energy. Hydroxyl
ions are known to adsorb specifically on hydrophobic surfaces
in contact with water.
19
Likewise, deprotonated amphiphiles
might also adsorb as a second layer on the particle surface in a
configuration similar to hemi-micelles. In both cases, a negative
charge would be added to the particle surface, further screening
the positive charges on the particle surface. The effect is more
pronounced for amphiphiles with longer hydrophobic tails, as
these can lead to an increased surface hydrophobicity at the
same concentration in solution (Fig. 2). The hydrophobic inter-
actions between the first layer of adsorbed amphiphiles and the
second layer of anions is apparently not strong enough to either
invert the particle z potential or form a well-defined hemi-micelle
around the surface. Most importantly, the second layer of an-
ions does not cover all the hydrophobic surface sites, keeping the
particles sufficiently hydrophobic to adsorb at the air–water
interface (see Section III (3)).
(2) Rheological Behavior
The screening of the particle’s surface charge upon amphiphile
addition leads to an increase in the suspension viscosity, as
shown in Fig. 3. Owing to the surface charge screening, the
electrical diffuse layer around the particle surface is not suffi-
ciently thick to overcome the attractive van der Waals forces
between particles. This results in particle coagulation and a sub-
sequent increase in the suspension viscosity. Figure 3 shows that
the suspension viscosity increases suddenly above a critical con-
centration of adsorbed carboxylic acid molecules. This critical
concentration decreases with an increase in the amphiphile tail
length due to the stronger screening effect of the longer mole-
cules (Fig. 2). Besides van der Waals attraction, hydrophobic
attractive forces
20–22
might also play a role in the viscosity in-
crease observed in Fig. 3. This hydrophobic attraction is ex-
pected to increase upon an increase of the amphiphile tail length.
(3) Surface Tension
The change in particle hydrophobicity upon adsorption of
short-chain carboxylic acids can also be monitored by surface
tension measurements. The addition of carboxylic acid mole-
cules resulted in a decrease in the surface tension of the suspen-
sions, as illustrated in Fig. 4(a). This decrease is in part due
to the free carboxylic acid molecules present in the suspension
that adsorb at the air–water interface. The concentration
of these free carboxylic acid molecules was determined from
the adsorption measurements (Fig. 1), and their contribution to
the total surface tension reduction is plotted in Fig. 4(b). The
contribution of free amphiphiles was estimated from surface
tension measurements of aqueous solutions containing carbox-
ylic acid concentrations equivalent to that expected for free
amphiphiles in the suspension.
Figure 4(b) shows that the free amphiphiles lead to a gradual
and monotonic decrease of the suspension surface tension. On
the other hand, Fig. 4(a) indicates that above a certain amphip-
hile concentration in suspension, the surface tension declines
more abruptly than one would expect from the reduction given
by the free amphiphilic molecules alone. At this amphiphile
concentration, the particles are hydrophobic enough to adsorb
to the air–water interface. The adsorbed particles replace part of
the highly energetic interface area and lower the overall free en-
ergy of the system, leading to an apparent reduction in surface
tension of the suspension. The amphiphile concentration needed
to render the particles sufficiently hydrophobic to adsorb to the
air–water interface decreases markedly with increasing tail
length. This can be attributed to the more hydrophobic nature
of carboxylic acids with increasing tail length.
(4) Foaming and Foam Stability
The concentration of amphiphilic molecules in the initial sus-
pension and the length of their hydrophobic tail can be used to
tailor the degree of surface hydrophobization of the alumina
particles in water (see Fig. 4(a)). By providing the proper hy-
drophobicity on the particle surface, suspensions were homoge-
neously foamed throughout the whole volume upon mechanical
frothing. The mixing speed significantly influences the shear
rates and shear stresses applied around the air bubbles during
0 50 100 150 200
0 50 100 150 200
40
50
60
70
80
Amphiphile concentration in suspension (mmol/l)
Am
p
hi
p
hile concentration in sus
p
ension
(
mmol/l
)
Surface tension of
suspension (mN/m)
40
50
60
70
80
90.870.856.928.1
Surface tension
due to free amphiphile (mN/m)
(a)
(b)
Free propionic acid in suspension (mmol/l)
2.5
Propionic Acid
Butyric Acid
Valeric Acid
Enanthic Acid
Fig. 4. (a) Surface tension of 35 vol% alumina suspensions at pH 4.75
for different concentrations of carboxylic acids with different tail length
(propionic acid (&), butyric acid (J), valeric acid (m), enanthic acid
(
X)). (b) Surface tension of free carboxylic acids in water at pH 4.75.
The concentration of free amphiphiles was determined from adsorption
measurements of 35 vol% alumina suspensions at pH 4.75 and plotted
as top x-scale for propionic acid as an example.
0 50 100 150 200 250 300
10
−2
10
−1
10
0
10
1
Propionic Acid
Butyric Acid
Valeric Acid
Enanthic Acid
Viscosity (Pa s)
Am
p
hi
p
hile concentration
(
mmol/l
)
Fig. 3. Viscosity of 35 vol% alumina suspensions at pH 4.75 containing
short-chain carboxylic acids with different tail lengths: propionic acid
(&), butyric acid (J), valeric acid (m), and enanthic acid (
X).
January 2007 Macroporous Ceramics from Particle-Stabilized Wet Foams 19
frothing. Even though the mixing conditions were not investi-
gated here, an increase of the shear stresses applied during
frothing should facilitate the mechanical rupture of freshly in-
corporated bubbles and thus lead to foams with smaller average
bubble sizes and narrower bubble size distributions.
23,24
The wet foams obtained after mixing feature air contents up
to 85% (Fig. 5) and possess a high stiffness in the wet state. By
increasing the concentration of a given carboxylic acid, the air
content of the foam first increases rapidly and then reaches a
plateau before decreasing sharply at high amphiphile additions.
This sharp decrease at high amphiphile concentrations is attrib-
uted to an increase in the viscosity of the initial suspension,
caused by the screening effect of counter-ions on the surface
charge of the particles (Section III (2)). An increased viscosity
hinders the incorporation of air into the initial suspension and
therefore results in foams with lower air contents.
The as-prepared wet foams exhibit a pronounced yield stress,
which allows for shaping of parts using extrusion, injection
molding, pressure filtration, or related techniques. The foams
can also be sprayed or easily poured into molds after dilution of
the as-prepared foams with water. Dilution does not affect the
stability of the air bubbles and can be used to adjust the foam
viscosity according to the shaping method desired.
The stability of the obtained particle-stabilized foams was
compared with that of shaving foam known to be very resistant
against destabilization (Gillette Foam, Regular, Gillette Co.,
London, U.K.). Figure 6 shows that the particle-stabilized
foams are stable against bubble growth and drainage over
days, whereas the shaving foam shows a rapid bubble growth
within the first hours.
The enhanced stability of the foams obtained here is based on
the different mechanism used to stabilize the air–water interface
compared with that applied in the conventional shaving foam. In
the ceramic foams, particles are used as stabilizers, whereas con-
ventional amphiphilic surfactants are used to stabilize the air–
water interface in the shaving foam. The energy required to de-
sorb a particle from an air–water interface is orders of magnitude
higher than the few kT’s needed to desorb a surfactant molecule
from the interface.
13
The irreversible adsorption of particles to
the interface results in a percolating interfacial armor that mech-
anically impedes bubble growth, shrinkage, and coalescence.
12
Particles are also expected to form a network throughout the
foam lamella, which further prevents bubble coarsening.
(5) Microstructure and Mechanical Strength of Sintered
Foams
The high stability of the wet foams allows for drying directly in
air at room temperature. However, in order to avoid crack for-
mation during water removal, the wet foam had to be slightly
strengthened to overcome the capillary stresses developed dur-
ing drying and to avoid differential shrinkage within the drying
foam. In this work, the strengthening effect was achieved either
by coagulating the particles within the foam lamella or by gelling
the foam liquid phase. Particles were coagulated in situ by shift-
ing the pH from 4.75 to 7.5 using an enzyme-catalyzed decom-
position reaction of urea,
15
whereas the gelation of the liquid
phase was achieved by cross-linking sodium alginate macromol-
ecules with an ion-releasing agent (see Section II (8)).
Figure 7(a) shows the microstructure of a macroporous
alumina obtained after drying and sintering a wet foam
prepared with butyric acid. In this case, crack formation was
avoided by coagulating the particles with the internal pH-shift
reaction. The macroporous structure obtained exhibits a total
porosity of 88%. Cells are mostly closed with an average size
of approximately 35 mm and a standard deviation of 15 mm.
Single cells are separated by walls with minimum thicknesses
below 1 mm (inset Fig. 7(a)). The foam cells are, in this case,
predominantly closed due to the fact that the air bubbles of the
original wet foams are completely covered with the surface-
modified particles. Such coverage remains through the gelling,
drying, and sintering procedure, resulting in the closed cells
depicted in Fig. 7(a).
Alternatively, the particle layer that covers the air bubbles
of wet foams can also be disrupted during the gelling process
to produce macroporous structures with interconnected cells,
as shown in Fig. 7(b). Here, the average cell size is in the range
of 100–150 mm. In this case, the foam was prepared with
propyl gallate and gelled with sodium alginate. The gelation
process led to a significant shrinkage of the wet foam. Cell
interconnectivity is most likely formed by a local differential
shrinkage of the particle layer around the air bubbles during
the gelation process. This shrinkage favors the rupture of the
particle coating around the bubbles, leading to interconnecting
cells after drying and sintering.
The microstructure of our particle-stabilized foams can be
tailored to render average cell sizes within the range of 10–300
mm at porosities between 40% and 95%.
3
Compared with sur-
factant-stabilized foams, our foams can reach smaller average
cell sizes and exhibit either open or closed cells even at high
0 50 100 150 200 250 300
0
20
40
60
80
100
Air content in foam (%)
Amphiphile concentration in suspension (mmol/l)
Propionic Acid
Butyric Acid
Valeric Acid
Enanthic Acid
Fig. 5. Air content of foamed suspensions containing 35 vol% alumina
at pH 4.75 and different short-chain carboxylic acids: propionic acid
(&), butyric acid (J), valeric acid (m), enanthic acid (
X).
0.1 1 10 100
0
1
2
3
4
5
6
0.01 0.1 1
Shaving Foam
Propionic Acid
Butyric Acid
Valeric Acid
Enanthic Acid
Relative average bubble size (−)
Time after foaming (hours)
Time after foaming (days)
Fig. 6. Relative average bubble size as a function of time after foaming
for particle-stabilized foams containing propionic acid (176 mmol/L,
&), butyric acid (50 mmol/L,J), valeric acid (30 mmol/L, m), and en-
anthic acid (10 mmol/L,
X) (35 vol% alumina, pH 4.75). The relative
average bubble size corresponds to the ratio between actual and initial
average bubble size. All data for the particle-stabilized foams were ob-
tained from 35 vol% alumina suspensions at pH 4.75. The stability of a
surfactant-stabilized shaving foam (Gillettet)(&)isalsoincludedinthe
graph for comparison.
20 Journal of the American Ceramic Society—Gonzenbach et al. Vol. 90, No. 1
