Journal of Colloid and Interface Science 249, 44–51 (2002)
doi:10.1006/jcis.2002.8250, available online at http://www.idealibrary.com on
Zeta Potentials and Yield Stresses of Silica Suspensions in Concentrated
Monovalent Electrolytes: Isoelectric Point Shift and Additional Attraction
George V. Franks
Department of Chemical Engineering, University of Newcastle, Callaghan, New South Wales 2308, Australia
E-mail: cggvf@cc.newcastle.edu.au
Received August 8, 2001; revised January 7, 2002; accepted January 24, 2002; published online March 26, 2002
The zeta potentials and yield stresses of silica suspensions were
measured over a wide range of monovalent electrolyte concentra-
tions. The counterions investigated were Li+ , Na+ , K+ , and Cs+ ,
while the co-ion was always Cl− . The poorly hydrated ions (Cs+
and K+ ) adsorb in greater quantity to the silica surface than the
well-hydrated ions (Li+ and Na+ ) and produce lower magnitude
negative zeta potentials at high pH. At high electrolyte concentra-
tions and at low pH the poorly hydrated counterions adsorb in great
enough quantity to reverse the sign of the zeta potential from neg-
ative to positive. This specific adsorption of counterions shifts the
iep to higher values. The shift in the iep is directly related to the
hydration of the counterion with the least hydrated ions creating
the greatest iep shift. The yield stresses of silica suspensions at high
pH were found to increase in the order Li+ < Na+ < K+ < Cs+ . The
magnitude of the yield stress correlates with the amount of adsorbed
ions; i.e., the greatest yield stresses are observed with the least hy-
drated ions. The yield stresses measured at high pH with high salt
concentrations are greater than can be accounted for with just van
der Waals attraction. An attractive ion–ion correlation force is pre-
sumed responsible for the additional attraction necessary to explain
the observed results. C 2002 Elsevier Science (USA)
Key Words: silica; isoelectric point; zeta potentials; yield stress;
monovalent ions; ion correlations; additional attractive force;
Hofmeister series.
INTRODUCTION
The effect of silica surface chemistry on the behavior of col-
loidal suspensions of silica in water have been extensively stud-
ied, yet many aspects of the surface are still poorly understood.
The unusual (or anomalous) behavior of silica suspensions (1–3)
at their isoelectric point (iep) include stability against coagula-
tion and sedimentation (4, 5) and low viscosity (1, 6). A short-
range repulsion not predicted by DLVO theory (7, 8) is usually
observed during direct force measurements of the silica surface
(9, 10). Although there is still debate over the origin of this re-
pulsion, it is most likely due to either structural repulsive forces
due to a hydration layer or steric repulsion due to silicic acid
hairs. The short-range repulsion as well as the low Hamaker
constant of silica in water (relative to other oxides) are the likely
0021-9797/02 $35.00

C 2002 Elsevier Science (USA)
All rights reserved.
cause of the unusual stability of silica at its iep (2). The iep of
silica is usually around pH 2, but will depend on the type and
concentration of electrolyte and type of silica (amorphous or
quartz) (11–15).
The monovalent alkali cations naturally form a sequence,
known as the Hofmeister series, based on the influence of the
ion on the water in its vicinity. The Hofmeister series (Cs+ , K+ ,
Na+ , Li+ ) orders ions from the least hydrated (the so-called
structure breaker) ions to the most hydrated (the so-called struc-
ture maker) ions. Most investigations find that the adsorption
sequence of monovalent cations on to the silica surface follows
the Hofmeister series (Cs+ > K+ > Na+ > Li+ ) (16, 17) with
Cs+ adsorbing in greater quantities than Li+ . Electrophoretic
mobility measurements generally indicate that the magnitude of
the negative zeta potential at high pH increases as the Hofmeister
series (Cs+ < K+ < Na+ < Li+ ) (5, 13, 14) with Cs+ producing
lower magnitude negative zeta potentials than Li+ . On the other
hand, the inverse Hofmeister series is observed for both ion ad-
sorption (18, 19) and zeta potential measurements (20, 21) for
materials such as alumina, hematite, and zirconia, which have
high isoelectric points.
It has been found that at pH below about 4 or 5 silica is stable
(against aggregation and sedimentation) even when up to about
1.0 M of any monovalent salt is added, but can be coagulated at
pH above about 5 or 6 by the addition of salts. As pH is increased
away from the iep, lower concentrations of salt are required to
coagulate the suspensions. This is contrary to most colloids,
which require less salt to coagulate as the pH is moved toward
the iep. The sequence of coagulation concentration increases
according to the sequence (Cs+ < K+ < Na+ < Li+ ); i.e., lower
concentrations of Cs+ are required for coagulation relative to
Li+ . These unusual coagulation results are hard to rationalize
based on charge neutralization arguments which are success-
ful in explaining most colloid behavior. Details of these and
other unusual behavior of silica can be found in the literature
(4, 5). Conversely, high iep materials (such as alumina, hematite,
and zirconia) tend to coagulate according to the sequence of
least concentration required to greatest Li+ < Na+ < K+ < Cs+
with lower concentrations needed as the iep is approached
(22).
44
 SILICA SUSPENSIONS IN CONCENTRATED MONOVALENT ELECTROLYTES 45

The rheological behavior of silica suspensions (at volume

fractions above the gel point) coagulated with large concen-
trations of monovalent cations has been investigated in more
detail recently (6, 23). At the iep, the silica suspensions have
relatively low and nearly Newtonian viscosities even when high
concentrations of monovalent electrolytes are added. At higher
pH where the silica has a high surface charge (at low elec-
trolyte concentrations) high viscosities and yield stresses oc-
cur when high concentrations (>about 0.1 M) of monovalent
electrolytes are added. The viscosities generally increase as the
electrolyte concentration is further increased. The correspon-
dence of the sequence of viscosities to either the Hoffmeister
or anti-Hoffmeister sequences depends upon the pH, electrolyte
concentration, and type of silica (amorphous or quartz) (6, 23).
The ion adsorption sequences can be explained by a “like
FIG. 1. Particle size distributions of the Geltech and Sigma silica powders.
adsorbs like” concept (21, 22). According to this concept,
surfaces with a high heat of immersion preferentially adsorb
well-hydrated ions while surfaces with a low heat of immersion was measured with a Malvern Mastersizer and is compared to the
preferentially adsorb poorly hydrated ions. The degree of hy- Geltech silica particle size distribution (adapted from Ref. (25))
dration of a surface can be related to its heat of immersion and in Fig. 1. This particle size distribution has been converted to a
thus the iep of the material (24). Accordingly, high-iep materials surface area averaged size ds = 1.98 µm. Powder X-ray diffrac-
preferentially adsorb well-hydrated ions and low-iep materials tion was used to identify the phases present in each powder. As
preferentially adsorb poorly hydrated ions. This concept satis- shown in Fig. 2 the Geltech silica is amorphous and the Sigma
factorily explains the observed adsorption sequences on silica,
alumina, and a number of other oxides (22).
The first objective of this research is to determine the effect
of high levels of monovalent counterions on the zeta potentials
of silica and to determine whether any shift of the iep occurs
at high counterion concentration (as was observed for alumina
in Refs. (20) and (21)). Another objective is to determine the
effect of large quantities of adsorbed ions on the rheological
behavior (yield stress) of silica and determine whether there is
any correlation with the measured zeta potentials. Finally, the
observed rheological results will be discussed in terms of the
interparticle DLVO forces and the possibility of a non-DLVO
additional attraction such as an ion–ion correlation force (which
may exist when large quantities of ions are adsorbed on the
surface).

EXPERIMENTAL

Materials and Suspension Preparation

Two types of silica were investigated: Geltech amorphous sil-
ica, which is high purity, >99.9%, nearly monodisperse, and
nearly spherical, and Sigma ground quartz, which is about 99%
pure, is nonspherical, and has a wide particle size distribution.
The Geltech sample has been extensively characterized and the
particle size distribution and shape characteristics are presented
elsewhere (25). The particles are nearly spherical and the distri-
bution of sizes is quite narrow (80% of the particles between 0.32
and 0.72 µm). This particle size distribution has been converted
(following either Ref. (26) or (27)) to a surface area averaged
size (ds or d(3,2) ) (for use in calculations discussed later) and is
ds = 0.46 µm. The particle size distribution of the Sigma silica FIG. 2. X-ray diffraction patterns of (a) Geltech silica and (b) Sigma silica.
46 GEORGE V. FRANKS
silica is quartz. The density of the powders was determined by
pycnometry. The density of Geltech silica is 2.15 g cm−3 and
the density of the Sigma silica is 2.60 g cm−3 .
Suspensions were prepared at 2 vol% solids for zeta potential
measurements and 40 vol% solids (Geltech silica) or 45 vol%
(Sigma silica) for rheological measurements. The suspensions
were dispersed by ultrasonication at pH 6 for 2 to 5 min with
a Branson 450 sonifier equipped with a 1.9-cm-diameter horn.
The sonifier was operated at a frequency of 20 kHz and the power
output maintained at about 60% of the limiting power (350 W).
Analytical grade LiCl, NaCl, KCl, or CsCl was added in quanti-
ties of between 0.001 and 0.4 M salt to the silica suspensions for
zeta potential measurements and between 0.4 to 4.0 M for rheo-
logical measurements. Adjustments of the suspension pH were
undertaken using analytical grade HCl, LiOH, NaOH, KOH,
and CsOH. The base with the same counterion as the salt was
always used. These reagents were obtained from either BDH or
Aldrich Chemicals. All water used in this study was distilled be-
fore ultra-purification with activated carbon and ion-exchange
resins (Milli-Q: conductivity <1 × 10−6 −1 cm−1 at 20◦ C).
The samples were allowed to age for between 12 and 18 h prior
to measurement.
Zeta Potentials
The zeta potential measurements were performed with a Col-
loidal Dynamics, AcoustoSizer (Warwick, RI). The zeta poten-
tial and size of the particles can be calculated from the dynamic
mobility spectrum and material properties (such as density of the
solid and liquid and liquid viscosity) assuming the thin double-
layer model (O’Brien’s model) (28–30). Another approach is
to assume that the dynamic mobility at the lowest frequency
measured (0.3 MHz) is equal to the static mobility and to use
Smoluchowski’s equation (31) to calculate the zeta potentials. A
concentration of 2 vol% solids was chosen for this investigation
as it is within the range of volume fractions which (1) are high
enough to produce a signal from the particles that is large relative
to the signal from the electrolyte ions and (2) are low enough
that particle–particle interactions do not significantly affect the
signal.
The electrolyte ions will produce an electrokinetic sonic am-
plitude (ESA) signal of their own that can become a significant
portion of the total ESA signal at salt concentrations above about
0.01 to 0.1 M concentrations (depending on the particle concen-
tration and density). For this reason the background electrolyte
signal was measured and subtracted from the total signal so that
accurate zeta potentials could be calculated (32). For each sus-
pension measured, a salt solution with the same salt type and
concentration as the salt type and concentration in the contin-
uous phase of the suspension was prepared. The background
signal was measured on these solutions at pH 5.8 and used for
the subtraction from the total signal measured from the corre-
sponding suspension at all pH values investigated.
The pH of the 2 vol% solids silica suspensions was adjusted
to pH 9 with the appropriate base, and the zeta potential mea-
surements were performed from high to low pH using 1.0 M
HCl to titrate the suspensions. After pH adjustment the suspen-
sion was allowed to equilibrate while stirred for 5 min before
the zeta potential measurement was performed. The suspension
pH was monitored using an epoxy sheathed Sensorex S200C
combination pH electrode coupled to the Acoustosizer. The zeta
potentials of the Geltech silica were measured at 0.001, 0.01, 0.1,
and 0.4 M concentrations for each of LiCl, NaCl, KCl, CsCl. The
zeta potentials of the Sigma silica was measured for the same
four salts but only at 0.1 M concentration.
Yield Stress
The shear yield stress (stress required to initiate flow) is a
convenient measure of the magnitude of the strength of attrac-
tion between particles in a suspension. The yield stress is related
to the solids volume fraction, particle size, and structure of the
particle network as well as the strength of attraction between
particles (21, 26, 33). When the volume fraction, particle size,
and structure of the network are held constant, as is done in
this investigation, the relative magnitude of the strength of at-
traction between particles can be investigated. (Comparisons are
only made between suspensions of the same powder and volume
fraction. In the current work the structure of the particle network
is maintained constant by preshearing the suspension prior to the
yield stress measurement.) Measurement of the shear yield stress
was performed with the vane technique (34, 35), which is quick
and easy and eliminates the possibility of erroneous results due
to slip at the tool/suspension interface.
The experimental apparatus consists of a small four-bladed
vane with dimensions, length = 30 mm, diameter = 20 mm, and
a blade thickness of 1.15 mm attached to a Haake VT-550 con-
stant rate viscometer. The vane is immersed in the suspension,
and the torque is measured as a function of time as the tool is
rotated at a slow constant rate (0.04 revolutions per minute). The
torque response is monitored as a function of time and is simply
related to the shear yield stress as detailed elsewhere (34–36).
The Geltech suspensions were measured at 0.40 volume frac-
tion while the Sigma suspensions were tested at 0.45 volume
fraction. The yield stresses were measured at between 0.4 and
4.0 M concentration for all salts investigated. All shear yield
stress measurements were performed at ambient temperature
(21–24◦ C). The suspension pH was monitored using an epoxy
sheathed Sensorex S200C combination pH electrode coupled
with a Horiba Model D-24 pH meter. All experiments were con-
ducted with the suspension pH progressing from acidic to basic
conditions. The suspensions were allowed to stand for approxi-
mately 60 min between each pH adjustment and the subsequent
measurement of the shear yield stress. The sample was pres-
heared by shaking the bottle vigorously for about 30 s prior
to inserting the vane tool. The sample was allowed to set for
1 min after the tool was inserted prior to initiating the measure-
ment. For some suspensions it was noticed that the yield stress
increased with setting time and this effect was investigated as
noted.
 SILICA SUSPENSIONS IN CONCENTRATED MONOVALENT ELECTROLYTES 47

RESULTS
Zeta Potentials
The zeta potential measurements at low salt concentration
(0.001 and 0.01 M) were consistent with most other low salt
zeta potential measurements (5, 11, 12, 15) and produced nearly
the same iep for each salt investigated. (See Figs. 3a and 3b.)
The Acoustosizer was able to fit all the data to the thin double-
layer model of O’Brien and co-workers (28–30). These results
were nearly identical to the zeta potentials produced by apply-
ing Smoluchowski’s equation to the 0.3-MHz dynamic mobil-
ity. Background electrolyte corrections were performed for the
0.01 M concentration series, and at most a 3-mV change (for
a suspension with 50 mV) was found from the uncorrected re-
sults. At higher salt concentrations, in some cases the Acousto-
sizer was not able to fit a zeta potential and size to the dynamic
mobility spectrum using O’Brien’s model. In these cases only a
zeta potential produced by applying Smoluchowski’s equation
(31) was obtained. When both results were obtained at most only
about 20% difference in magnitude was found. Figures 3c and 3d
show the zeta potentials for the 0.1 and 0.4 M concentration se-
ries produced by the Smoluchowski approximation. Figures 4a
and 4b show the Acoustosizer fit to O’Brien’s model (28–30) for
the same two sets of experiments. Even at these high salt concen-
trations there is very little difference between the zeta potentials
produced with these two different methods. The data shown in
Figs. 3c, 3d, 4a, and 4b have been obtained after the background
electrolyte corrections have been made. These corrections have
a significant effect on both the magnitude and sign of the zeta
potential results, and must be performed to produce accurate re-
sults. The high salt concentration results are in good agreement
with those of Kosmulski and co-workers (13, 14). The 0.1 and
0.4 M results indicate that the monovalent alkali cations specif-
ically adsorb (at high salt concentrations) and shift the iep of
the Geltech (amorphous) silica to higher pH values. The Sigma
silica was investigated at only 0.1 M salt concentrations and
a somewhat different result occurs. As shown in Fig. 5 the Sigma
silica has a very low iep (<pH 2) even at this relatively high salt
concentration of 0.1 M. In all cases (both Geltech and Sigma
silica) at pH above the iep the magnitude of the zeta potentials
was found to increase in the order (Cs+ < K+ < Na+ < Li+ ),
consistent with the results of other investigators (5, 13, 14).
Yield Stresses
At pH below about 3 or 4 there was no measurable yield stress
for any of the suspensions investigated. At higher pH values the

FIG. 3. Zeta potentials of Geltech silica calculated using the Smoluchowski approximation. (a) 0.001 M added chlorides, (b) 0.01 M added chlorides, (c) 0.1 M
added chlorides, and (d) 0.4 M added chlorides. Counterions are Li+ , Na+ , K+ , and Cs+ , as indicated in the figures. The lines are drawn to guide the eye.
48 GEORGE V. FRANKS
FIG. 4. Zeta potentials of Geltech silica calculated by the AcoustoSizer
using O’Brien’s model. (a) 0.1 M added chlorides, (b) 0.4 M chlorides, and
counterions are as indicated in the figures. Same raw data as used in Figs. 3c and
3d. The lines are drawn to guide the eye. The dashed portion of the line indicates
the expected trend at pH values where the AcoustoSizer could not produce a
good fit to O’Brien’s model.
yield stress depended upon the type of powder, salt concentra-
tion, volume fraction, and time after preshearing. Many of the
suspensions were noted to increase in yield stress with time of
sitting undisturbed. Although no detailed study of how the yield
FIG. 5. Zeta potentials of Sigma silica calculated using the Smoluchowski
approximation in 0.1 M chlorides, and counterions are as indicated in the figure.
The lines are drawn to guide the eye.
FIG. 6. Yield Stresses of 40 vol% Geltech silica in 0.4 M chloride solutions.
The counterions are as indicated in the figure. The lines are drawn to guide the
eye.
stress increases with time was performed, it was found that a
great fraction of the long time yield stress developed within the
first minute. The relative magnitudes of the yield stress of dif-
ferent suspensions never changed their sequence as a function
of time. For these reasons all results presented in this paper were
measured after one minute of resting after the preshearing treat-
ment. The yield stresses of the Geltech silica suspensions for
0.4, 1.0, and 4.0 M are shown in Figs. 6, 7, and 8. The high-
est maximum yield stresses are obtained with Cs+ at all salt
concentrations. The magnitude of the maximum yield stress in-
creases with increasing salt concentration (75 Pa for 0.4 M CsCl,
95 Pa for 1.0 M CsCl, and 200 Pa for 4.0 M CsCl). Aside from
a few transgressions, the yield stress at high pH (near where
the maximum Cs+ yield stress occurs) decreases in the order
Cs+ > K+ > Na+ > Li+ . At about pH 5 (for 0.4 M), pH 6 (for
1.0 M), or pH 7.5 (for 4.0 M) the yield stresses of the suspensions
containing different salts are similar in magnitude to each other.
At pH below these pH values the sequence of highest to lowest
yield stresses inverts to Li+ > Na+ > K+ > Cs+ . These results
are in good agreement with the rheological behavior observed
FIG. 7. Yield Stresses of 40 vol% Geltech silica in 1.0 M chloride solu-
tions. The data presented are from two separate batches. The counterions are as
indicated in the figure. The lines are drawn to guide the eye.
 SILICA SUSPENSIONS IN CONCENTRATED MONOVALENT ELECTROLYTES
FIG. 8. Yield Stresses of 40 vol% Geltech silica in 4.0 M chloride solutions.
The counterions are as indicated in the figure. The lines are drawn to guide the
eye.
by Colic et al. (6). In some cases (particularly 0.4 M concen-
trations), the typical “bell-shaped” yield stress curve (usually
associated with a material that has an iep at the same pH as the
peak in the yield stress) is observed.
Figure 9 shows the yield stresses of the Sigma silica suspen-
sions with 1.0 M salts. The yield stresses increase monotonically
from low to high pH. The magnitude of the maximum yield stress
of each salt is significantly greater than that for the correspond-
ing Geltech 1.0 M yield stresses. This is contrary to what one
would expect from a larger-particle-sized powder (26, 37, 38).
At all pH values the sequence of highest to lowest yield stress
follows the order Cs+ ≈ K+ > Na+ > Li+ (except at pH above
about 8 where Cs+ begins to tail off).
DISCUSSION
It is difficult to reconcile all of the suspension (rheological,
coagulation, and zeta potential observations, presented here and
FIG. 9. Yield stresses of 45 vol% Sigma silica in 1.0 M chloride solutions.
The counterions are as indicated in the figure. The lines are drawn to guide the
eye.
49
in Refs. 1–6, 11–17, 22, 23) with direct force measurements
(9, 10, 39, 40) at the same ionic strength because very different
total surface areas of silica are investigated in the different types
of experiments. Thus, the amount of ions adsorbed per unit area
may be very different for the same nominal salt concentration
in the three different types of experiments. None the less, mean-
ingful comparisons can be made at least qualitatively.
Adsorption Sequences, Zeta Potentials, and IEP Shift
The results presented in Figs. 3, 4, and 5 confirm that for
both amorphous silica and quartz the least hydrated ions pref-
erentially adsorb relative to the most hydrated ions as indicated
by the increase in the magnitude of the negative zeta potentials
at high pH according to the sequence Cs+ < K+ < Na+ < Li+ .
These results support the premise of Dumont et al. (22) con-
cerning ion adsorption sequences and the pH of the iep. The
Geltech amorphous silica investigated shows a shift in the iep
to higher pH values as the salt concentration is increased. This
indicates that the alkali cations are adsorbing to the silica surface
in quantities greater than required for charge neutralization. The
greatest shift in iep is found to follow the sequence from greatest
iep shift to least iep shift Cs+ > K+ > Na+ > Li+ . These results
are consistent with the high salt zeta potential measurements
of Kosmulski and co-workers (13, 14). The shift noted in the
iep of silica is in the opposite sequence observed for alumina
reported recently utilizing the same equipment (20, 21). The op-
posite sequences noted for the two different materials rule out
any possible iep shift due to improper background electrolyte
correction, or other systemic experimental error. When greater
than 0.1 M salts are used, there are positive zeta potentials ob-
served at pH below the iep. These positive zeta potentials tend
to increase in the order Li+ < Na+ < K+ < Cs+ . Up to about +5
to +10 mV are observed for the least hydrated ions.
Yield Stress and IEP Shift
The yield stress corresponds directly to the magnitude of at-
traction between particles when the particle size, shape, and vol-
ume fraction as well as the suspension structure are maintained
constant. The results presented in Figs. 6 through 9 indicate that
at moderate to high pH the sequence of yield stresses generally
decreases according to the sequence Cs+ > K+ > Na+ > Li+ .
This is due to two factors. First, at higher pH there is a residual
negative charge on the particles that is not neutralized by the
adsorbed ions. (See Figs. 3 through 5.) The particles in the solu-
tions of poorly hydrated counterions are neutralized to a greater
extent than the particles in the well-hydrated ion solutions and
thus have a lower magnitude zeta potential and thus less electri-
cal double-layer repulsion. Also, the least hydrated ions adsorb
in greater quantities (since they neutralize more of the surface
charge). As described in the next section it is believed that there is
an attraction created by adsorbed ions. Both these factors (small
electrical double-layer repulsion and additional attraction) will
contribute to higher yield stresses due to the resulting greater
overall attraction.
50 GEORGE V. FRANKS
On the other hand at low pH the sequence of yield stresses is
reversed to Li+ > Na+ > K+ > Cs+ . This can be explained by
the charge reversal of the surface to positive zeta potentials at low
pH with the poorly hydrated ions. The small magnitude positive
zeta potentials observed with the poorly hydrated ions produce
a larger electrical double-layer repulsion than that of particles
coagulated with well-hydrated ions which continue to have their
iep’s at low pH and thus very little electrical double-layer repul-
sion at low pH.
The pH at which the maximum yield stress occurs corresponds
to the iep in most low electrolyte concentration colloidal systems
(33, 41, 42) since this is the condition that produces the greatest
attraction. The shift in maximum yield stress to higher pH values
with increasing salt concentration observed for the amorphous
silica seems to correlate reasonably well with the shift of the iep
to higher pH values with increased salt concentration.
Additional Attraction
Based upon recently developed models (26, 42, 43) it is now
possible to estimate the yield stress of a suspension if the parti-
cle and solution properties are known. Using these models, the
volume fraction of particles, the surface area average size of the
particles, and the assumptions stated below, the maximum yield
stress for silica suspensions at their iep was estimated. Calcula-
tions were performed based on the following assumptions: the
silica suspension structural characteristics are similar to the alu-
mina suspensions structures used to develop the model (26), and
the silica Hamaker constant appropriate for use in high salt con-
centrations is 0.5 × 10−20 J. The assumption of similar structural
characteristics in the silica suspensions relative to the alumina
suspensions previously investigated (20, 21, 26) is likely valid
since the suspensions were both prepared similarly and vigor-
ously presheared prior to measurement. Only the dispersive part
of the silica Hamaker constant is used since the static component
will be screened by the high salt concentration. It is also assumed
that the interaction distance of the particles is 25 Å, as was the
case for the alumina (26). The results of these calculations pre-
dict that the maximum yield stress possible due to van der Waals
attraction alone is 52 Pa for the Geltech silica and 10 Pa for the
Sigma silica. In fact much greater yield stresses were measured
for these powders in high-salt-concentration solutions at high
pH values. The maximum measured yield stress for the Geltech
silica is about 200 Pa at pH 10 with 4.0 M added CsCl. For
the Sigma silica the maximum measured yield stress measured
is about 700 Pa at pH 10 with 1.0 M KCl. The magnitide of
yield stresses measured (much greater than that predicted due
to van der Waals attraction strongly) suggests that an additional
attraction not predicted by DLVO theory (7, 8) is operating un-
der conditions when large quantities of ions are adsorbed to the
surface of the particles. It may also be possible that the very
high yield stresses observed for the quartz are due to either a
greater van der Waals attraction (due to the higher density of the
quartz), the wider particle size distribution of the quartz, or the
rough nonspherical shape of the quartz particles.
A similar additional attraction was observed in alumina sus-
pensions when large quantities of ions are adsorbed to the alu-
mina surface (20, 21). (In the case of alumina it is the well-
hydrated counterions that strongly adsorb, in contrast to the
poorly hydrated ions that adsorb strongly to silica.) As pos-
tulated previously, it is believed that this additional attraction is
due to a van der Waals-like attraction caused by dipoles created
by the positioning of ions on the surface of the particles (21).
This sort of attraction is known as an ion–ion correlation force
(44, 45). An additional attraction, due to large quantities of ad-
sorbed ions, helps to explain the rheological results presented in
this paper as well as some of the previously unexplained aspects
of silica behavior.
CONCLUSIONS
An explanation for some of the anomalous behavior of silica
is likely to be related to the shifting iep and the additional attrac-
tion found at high salt concentrations. The zeta potential results
indicate that the least hydrated ions adsorb in greater quantities
than well-hydrated ions and shift the iep to higher pH values.
The yield stress results indicate that the strongest attraction be-
tween silica particles occurs at high pH when in concentrated
solutions of poorly hydrated ions. These conditions correspond
to the situation where large quantities of ions are adsorbed to
the surface of the silica. It is believed that the ions on the sur-
face correlate their positions across the mid-plane between the
surfaces in such a way as to minimize their free energy. The
resulting configuration of dipoles creates a van der Waals-like
attraction not predicted by the DLVO theory.
ACKNOWLEDGMENTS
The zeta potential measurements were performed while GVF was a member
of the Advanced Mineral Products Research Centre (now the Particulate Fluids
Processing Centre), a Special Research Centre of the Australian Research Coun-
cil at the University of Melbourne. Thanks to Stephen Johnson, Peter Scales,
and Tom Healy for the enlightening discussions. Thanks to Angry Russell, Steve
Johnson, and Nanda Duin for help with the Acoustosizer measurements. Thanks
to L. L. Muk for performing some of the yield stress measurements.
REFERENCES
1. Iler, R. K., “The Chemistry of Silica.” Wiley Interscience, NY, 1979.
2. Dumont, F., in “ACS Advances in Chemistry Series No. 234” (H. E. Bergna,
Ed.), p. 143. Am. Chem. Soc., Washington DC, 1994.
3. Healy, T. W., in “ACS Advances in Chemistry Series No. 234” (H. E. Bergna,
Ed.), p. 147. Am. Chem. Soc., Washington DC, 1994.
4. Allen, L. H., and Matijevı́c, E., J. Colloid Interface Sci. 31, 287 (1969).
5. Depasse, J., and Watillon, A, J. Colloid Interface Sci. 33, 430 (1970).
6. Colic, M., Fisher, M. L., and Franks, G. V., Langmuir 14, 6107 (1998).
7. Derjaguin, B., and Landau, L., Acta Physiochem. URSS 14, 633 (1941).
8. Verwey, E. J. W., and Overbeek, J. Th. G., “Theory of the Stability of
Lyphobic Colloids.” Elsevier, Amsterdam, 1948.
9. Israelachvili, J. N., “Intermolecular and Surface Forces,” 2nd ed. Academic
Press, San Diego, 1992.
10. Israelachvili, J. N., and Wennerstrom, H., Nature 379, 219 (1996).
11. Parks, G. A., Chem. Rev. 65, 177 (1965).
 SILICA SUSPENSIONS IN CONCENTRATED MONOVALENT ELECTROLYTES
12. Lyklema, J., “Fundamentals of Interface and Colloid Science,” Vol. II.
Academic Press, London, 1995.
13. Kosmulski, M., and Matijevic, E., Colloid Polym. Sci. 270, 1046 (1992).
14. Kosmulski, M., J. Colloid Interface Sci. 208, 543 (1998).
15. Schwarz, S., Lunkwitz, K., Kessler, B., Spiegler, U., Killmann, E., and
Jaeger, W., Colloids Surf. A 163, 17 (2000).
16. Tadros, Th. F., and Lyklema, J., J. Electroanal. Chem. 17, 267 (1968).
17. Sonnefeld, J., Gobel, A., and Vogelsberger, W., Colloid Polym. Sci. 273,
926 (1995).
18. Sprycha, R., J. Colloid Interface Sci. 127, 1 (1989).
19. Tschapek, M., Wasowski, C., and Torres Sanchez, R. M., J. Electroanal.
Chem. 74, 167 (1976).
20. Johnson, S. B., Scales, P. J., and Healy, T. W., Langmuir 15, 2836
(1999).
21. Franks, G. V., Johnson, S. B., Scales, P. J., Boger, D. V., and Healy, T. W.,
Langmuir 15, 4411 (1999).
22. Dumont, F., Warlus, J., and Watillon, A., J. Colloid Interface Sci. 138, 543
(1990).
23. Farrow, J. B., Horsley, R. R., Meagher, L., and Warren, L. J., J. Rheol. 33,
1213 (1989).
24. Healy, T. W., and Fuerstenau, D. W., J. Colloid Sci. 20, 376 (1965).
25. Franks, G. V., Zhou, Z., Duin, N. J., and Boger, D. V., J. Rheol. 44, 759
(2000).
26. Zhou, Z., Solomon, M. J., Scales, P. J., and Boger, D. V., J. Rheol. 43, 651
(1999).
27. Kelly, E. G., and Spottiswood, D. J., “Introduction to Mineral Processing.”
Australian Mineral Foundation, 1995.
51
28. O’Brien, R. W., J. Fluid Mech. 212, 81 (1990).
29. O’Brien, R. W., J. Fluid Mech. 190, 71 (1988).
30. O’Brien, R. W., Cannon, D. W., and Rowlands, W. N., J. Colloid Interface
Sci. 173, 406 (1995).
31. Shaw, D. J., “Introduction to Colloid and Surface Chemistry,” 4th ed., p. 157.
Butterworths, London, 1992.
32. Desai, F. N., Hammad, H. R., and Hayes, K. F., Langmuir 9, 2888 (1993).
33. Johnson, S. B., Franks, G. V., Scales, P. J., Boger, D. V., and Healy, T. W.,
Int. J. Miner. Process. 58, 267 (2000).
34. Nguyen, Q. D., and Boger, D. V., J. Rheol. 27, 321 (1983).
35. Nguyen, Q. D., and Boger, D. V., J. Rheol. 29, 335 (1985).
36. Johnson, S. B., Franks, G. V., Scales, P. J., and Healy, T. W., Langmuir 15,
2836 (1999).
37. Franks, G. V., and Lange, F. F., Colloids Surf. A 156, 5 (1999).
38. Franks, G. V., and Lange, F. F., J. Am. Ceram. Soc. 82, 1595 (1999).
39. Vakarelski, I. U., Ishimura, K., and Higashitani, K., J. Colloid Interface Sci.
227, 111 (2000).
40. Chapel, J. P., Langmuir 10, 4237 (1994).
41. Leong, Y. K., Boger, D. V., Scales, P. J., Healy, T. W., and Buscall, R.,
J. Chem. Soc. Chem. Commun. 7, 639 (1993).
42. Scales, P. J., Johnson, S. B., Healy, T. W., and Kapur, P. C., AIChE J. 44,
538 (1998).
43. Kapur, P. C., Scales, P. J., Boger, D. V., and Healy, T. W., AIChE J. 43, 1171
(1997).
44. Guldbrand, L., Jonsson, B., Wennerstrom, H., and Linse, P., J. Chem. Phys.
80, 2221 (1984).
45. Kjellander, R., Ber. Bunsen-Ges. Phys. Chem. 100, 894 (1996).