Colloids and Surfaces A: Physicochem. Eng.

Aspects 349 (2009) 4354

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical and

Engineering Aspects
journal homepage: www.elsevier.com/locate/colsurfa

Inuence of pH, high salinity and particle concentration on stability and

rheological properties of aqueous suspensions of fumed silica
Asal Amiri , Gisle ye, Johan Sjblom
Ugelstad Laboratory, Department of Chemical Engineering, Norwegian University of Science and Technology (NTNU), N-7491 Trondheim, Norway

a r t i c l e i n f o a b s t r a c t

Article history: Multiphase light scattering and transient-dynamic rheometry were used in combination in order to con-
Received 16 May 2009 trol the stability, rheology and gelation of aqueous Aerosil 200 suspensions with regard to the inuence of
Received in revised form 28 July 2009 pH, particle and electrolyte concentrations. Effects of these parameters were studied by obtaining the crit-
Accepted 30 July 2009
ical concentration of particles for transition from Newtonian to shear thinning ow in each system. This
Available online 7 August 2009
was attributed to microstructural changes due to the interparticle forces. Further conrmation of results
was obtained by measuring apparent viscosities, yield stresses, time dependent transient rheometry and
Keywords:
light scattering from the samples. All results were described by extended-DLVO theory. Suspensions with
Aerosil
Rheology
acidic pH or at the isoelectric point had high stability even in presence of high concentration of NaCl due
Multiple light scattering to the short-range repulsive hydration forces. On the other hand, suspensions with the highest salt con-
Gelation centration and basic pH showed time dependant rheology, largest yield values and viscosities, smallest
critical particle concentration for ow regime transition and high capability of gel formation. Verication
of gelation kinetics and the gel strength with regard to particle and salt concentration were implemented
by three typical tests: time sweep, strain amplitude and frequency amplitude sweep in the dynamic
oscillation mode at pH 8.5. Strength of the gelled networks was found to be greater at either higher
salt concentration or higher concentration of particles. Strong-link regime of gel was found to govern
the system regarding to increase either in particle volume fraction or in electrolyte. The scaling theory
was tested for the system at 0.6 M NaCl and power-law exponent of the volume fraction dependence of
storage modulus was found to be 3.8. In addition, applying the fractal model for strong gels calculated
fractal dimension of ocs df around 1.85. Based on the exponent value of 3.78 for G and the fractal ocs
dimension of 1.85, it was proposed that cluster-diffusion-limited cluster aggregation (DLCA) was mainly
governing the gel formation mechanism.
2009 Elsevier B.V. All rights reserved.

1. Introduction are reservoir particles (silica, clays), mineral scales (CaCO3 , BaSO4 ,
SrSO4 , CaSO4 ), corrosion products (FeS, oxides), gas hydrates, etc.
Solid suspensions are of large practical interest because of their [14].
widespread occurrence in everyday life. They have important prop- Some typical problems arising from these solids are precipita-
erties that may be desirable in a natural or formulated product tion of compounds from crude oils causing blockage in the process
or undesirable, such as an unwanted suspension in an indus- facilities or sticking the suspended solid particles in the crude oil
trial process. Some important and familiar suspensions include to the wall of the pipes [1,2,57]. Another important effect arising
those occurring in foods, pharmaceuticals, household products, and from inorganic particles like silica and clay released from reservoir
the environment. They are also quite important and widespread is the formation of suspensions and formation of stable W/O or
in the petroleum industry. In fact suspensions may be encoun- O/W emulsions [1,2,8]. Especially during transport of these parti-
tered throughout each of the stages of petroleum recovery such as cles, the control of stability and the ow properties will be most
migrating nes, oil recovery, production and transport, produced important issue. In most cases, silica in its various forms is one of
(well-head) solids, drilling uids (mud), well cementing slurries the most commonly used. Due to the different stability behaviors of
and oil sand slurries [1]. Some of these most conventional solids these particles (both depending on the dispersing media and par-
ticle properties), they can be used as ow modiers or as model
particles for different applications [10,11]. One particular form of
Corresponding author. Tel.: +47 73595878. silica of interest is fumed silica. Fumed silica produced by ame
E-mail addresses: [email protected] (A. Amiri), hydrolysis of chlorosilane (SiCl4 ) [911] is the most hydrophilic
[email protected] (G. ye), [email protected] (J. Sjblom). silica and extremely versatile material which also has unique prop-

0927-7757/$ see front matter 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.colsurfa.2009.07.050
44 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354

erties having nano-size scale, chain-like morphology, high capacity 2. Experimental

for water absorption and capability of forming agglomerate and
3D networks in water due to the weak hydrogen bond [11]. Some 2.1. Materials
of the application of Aerosil involve: ller for rubbers and plas-
tics, coatings, adhesives, cements, sealants and rheology modier Standard, well-characterized fumed Silica (Aerosil 200) was a
[1113]. gift from Degussa (Evonik). These colloidal particles are prepared
In aqueous media, since pure particles are electrostatically sta- by hydrolysis of silicon tetrachloride [10,12] in which silanol groups
bilized, interparticle interactions at a given concentration can be are generated on silica surface. Irreversible aggregates, linked by
changed by varying the ionic strength, particles size, pH of the bonds which are too strong to be broken, are formed during the
solution and the polarity of the media. Stability of fumed silica fabrication process [10,12]. The particles have a quoted surface area
in aqueous solutions has been a topic of numerous studies for of 200 m2 g1 with primary units of 12 nm in diameter, density of
decades, but exact mechanisms describing the stability or insta- 2.2 g cm3 and water contact angle of approximately 13 .
bility of these dispersions in terms of variables such as particle
concentrations, pH and high concentration of electrolyte are not 2.2. Methods
given, always viewed as a multiple parameter approach. In fact, the
purpose of this paper is to determine stability criteria in view of 2.2.1. Sample preparation
the mentioned key parameters and to discuss the ndings accord- Suspensions of different particle concentrations were pre-
ing to updated DLVO theory for the stabilization or destabilization pared by adding ultra-pure water (Millipore milli-Q system,
mechanisms of aqueous solutions of fumed silica. = 18.2 M cm) containing different concentrations of NaCl salt to
It is well known that the classical DLVO theory established the dened amounts of particles. Dispersing was done by means of
by Derjaguin, Landau, Verwey and Overbeek [14,15] accounts for minishaker at 2500 rpm for 5 min. Sonication was applied in order
stability of lyophobic suspensions. According to this continuum to break the aggregates and form homogeneous dispersions before
theory, the stability of colloidal suspensions can be predicted any test. The proper time for sonication was adjusted to 1 h. The
through knowledge of the total energy of interaction between pH of the systems was adjusted by adding NaOH or HCl. Amount of
two particles resulting from the algebraic sum of two interaction added acid or base was small; so that the change in volume of sam-
energies [1418]: the van der Waals attraction (VA ) and the elec- ples was neglected. The pH values were read at room temperature
trostatic double-layer repulsion (VR ). This classical DLVO theory on a 420A plus pH meter (Thermo Orion) which had been calibrated
was found to involve some inherent limitations and failed to ade- with standard buffers before.
quately represent various experimental results [17,19,20]. In fact it
was mentioned by several researches that the DLVO theory in its
2.2.2. Electrokinetic analysis
classic form is just well applicable to lyophobic colloids and apply-
Zeta potential measurement of Aerosil 200 suspensions were
ing it to the lyophilic colloids like silica, requires as assumption
carried out in presence of three salt concentrations at 25 C as a
that the contributions from the electric double layer and from the
function of pH using a Zetasizer Nano-sizer instrument (Malvern
more specic surface structures are additive. Furthermore, many
Instrument Ltd) and the isoelectric point (i.e.p.) of the suspen-
reasons have been invoked to show discrepancies between the the-
sions were determined. The particle concentration was low enough
ory and unexplained experimental ndings, i.e. at high electrolyte
(0.05 vol%) in order to keep the double layer around particles surface
concentrations and small separation distances. One reason for this
in the extended form by prevent particles approaching each other
inconsistency is that DLVO theory considers the ionic species as
and overlapping their double layers. The salt concentrations were 0,
point charges while solventsurface, surfaceion, and solvention
0.1 M and 0.3 M. Prior to the measurements samples were treated by
interactions are neglected [21,22]. However, development of direct
minishaker for 2 min and sonication for 15 min. pH adjustment was
surface force measurement techniques during the past two decades,
accomplished by auto-titration using separate solutions of 0.25 M
with the surface force apparatus (SFA) and the atomic force micro-
NaCl and 0.25 M HCl.
scope (AFM), has allowed the introduction of the short-range forces,
known as solvation (or structural) and steric forces [17,22]. These
forces were found to be either attractive or repulsive, depending 2.2.3. Dynamic light scattering (DLS) measurements
on the nature of the surfaces (or the particles), the solvent, and This technique was used to measure the hydrodynamic diame-
the experimental conditions. Then the total state of occulation is ter of the suspended particles to determine the critical coagulation
described as the sum of the classical DLVO forces and short-range concentration (ccc) of NaCl salt. As it is well known, the critical
forces. However, due to the limitations of instruments, measure- coagulation concentration (ccc) is referred to as the electrolyte
ments of mobility and surface charge of particles in presence of concentration for which the transition from a slow to a rapid
high salt concentration (which is highly interesting in oil industry) aggregation regime occurs for a given suspension [1,15,18]. Several
were not possible with the conventional methods and instruments experiments were done by adding increasing amounts of electrolyte
like zeta-sizer and acousta-sizer. Since the relative magnitude of to several diluted samples (0.05 vol% silica) and measuring the aver-
the above colloidal forces and the stability of suspensions have pro- age aggregate size. The salt concentration corresponding to a sharp
nounced reection on rheological behavior [17,2327] and on the increase in aggregate size was referred to as ccc. It is assumed
other hand, rheological properties of fumed silica particles in water that above this concentration, the fast and irreversible aggregation
has still large importance in many elds, the main objective of this dominates the behavior of the system.
paper is to study stability levels through control of the rheological
properties. Multiphase light scattering and both transient/dynamic 2.2.4. Rheometry
rheometry are applied as complementary techniques. The transient Rheological behavior was monitored using a Physica MCR 301
and dynamic modes of rheometry are used to verify the ow and the rheometer (Anton Paar). Experiments were conducted under both
gelation properties. Information on microstructure and regime of steady state and dynamic mode, depending on purpose of the study.
the gels are obtained by testing consistency of the scaling theory for In the steady state experiments, stability state of suspensions with
the gels systems. Further, relationship between rheology and stabil- regard to the particle concentration, salinity and the pH of sus-
ity is explained by possible particleparticle and particlesolvent pension was compared by determining the critical concentration
interactions. of particles needed at which the suspension regime changed from
 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354 45

Newtonian to non-Newtonian behavior. Furthermore, apparent vis- analysis of the transmission or backscattering signals through the
cosities and the values of yield stress were monitored at the xed sample.
particle concentration of 1 vol% for pH 2, 4.5, 8.5 and salinities of
pure water, 0.6 M and 2.75 M NaCl, as well. Experiments were all 3. Results
carried out at 25 C and were allowed to both temperature equili-
brate and the normal forces from sample loading for 2 min before In order to have a better distinction between results we have
each test. Since the rheological properties can be highly dependant divided them in to two parts, i.e. stability and gelation studies.
on the shear treatment history of the sample [28], a constant shear
rate of 50 s1 was applied for 2 min followed by 2 min rest period to
3.1. Stability studies
allow relaxation process to take place and ensure similar starting
conditions for all samples before test.
3.1.1. Zeta potential and aggregates size
For gelation studies, due to the fact that in general structure
Fig. 1a shows zeta potential of Aerosil suspensions in water and
of gels can be easily disrupt, the shear strain had to be kept low
at two different NaCl concentrations within the pH range 210. For
enough. Small amplitude dynamic oscillation measurement tech-
the suspensions in pure water, zeta potential was negative for pH
nique is nondestructive method, compared to the steady shear
>2 indicative of negative charge sites on silica particles. Also, iso-
measurement, since the level of deformation can be kept low.
electric point was found to be around pH 2.5. On the other hand,
Therefore a complete oscillation experiment including time sweep,
values of zeta potential were positive for pH <2 showing the posi-
strain sweep and the frequency sweep test were applied for this
tive value of surface charge in this region. Effect of adding NaCl up
part [29]. In the time sweep test, both the strain and the frequency
to the highest measuring limit of Zetasizer was investigated and is
were xed at a small value just close to the rest condition and fol-
indicated within the narrow range of pH close to the i.e.p. There
lowing evolution of storage and loss modulus were monitored as a
was not a signicant change in zeta potential values up to 0.01 M
function of time. For the strain sweep, the frequency was held xed
NaCl especially at basic pH. However, there was a shift in i.e.p. from
and the strain amplitude increased, and nally in frequency sweep
2 to 4 in presence of 0.3 M NaCl. Shift in i.e.p. in presence of some
performed at a constant strain, with the frequency being increased.
different salts were also reported before [33,34]. It had been also
In general, the purpose of oscillation tests was to investigate the
shown in several works that addition of salt even up to 0.001 M does
effect of particle concentrations and salinity on gelation kinetics
not have signicant effect on values of zeta potential, and the i.e.p.
and its strength. These experiments were all performed at basic pH
as well for 1-1 electrolyte [33]. The ccc of the Aerosil at pH 6 was
8.5. In order to reduce the errors from sample hardening due to
obtained around 2 M for NaCl. Results are depicted in Fig. 1b. Salt
water evaporation a cover was attached to the top of the measuring
concentrations of 0, 0.6 M and 2.75 M were consequently selected
tool.
below and above the ccc.
Since gelation kinetics and therefore equilibrium state of gel are
highly dependent on the pre-shear history imposed on material
before experiment [30], for example, during the loading of the sam-
ple in the rheometer or when sample preparation, a protocol was
dened to assure the desired reproducibility: immediately after
loading the sample, a 1 min pre-shear of 50 s1 was applied to the
sample and following, the time evolution of the elastic modulus
at very small frequency and very small stain% close to rest was
detected and the time required to reach a constant storage modu-
lus was determined. This time, which was different from sample to
sample, could be considered as the time needed to recover an equi-
librium structure. Moreover, we saw that the nonzero normal force,
which appeared during the loading of the sample, had decayed to
zero before starting a test [30].
Two different measuring tools were used (a) concentric cylinder
of double gap bob and cup, for the shear ow tests and (b) classic
concentric cylinder in the oscillation tests. These geometries were
calibrated with the standard oils before and the acceptable range
of shear rates was determined. The unique specication for dou-
ble gap tool was its applicability for very low viscosity uids as
water and no turbulence effect until the high shear rate values of
1000 s1 .

2.2.5. Multiple light scattering analysis

As a guide to investigate the stability, occulation or gelation in
concentrated samples, light scattering of colloidal samples were
measured at 25 C for different salt concentrations and solution
pH using the dispersion optical analyzer Turbiscan (formulac-
tion) [31]. 20 ml of sample was introduced in a glass cylindrical
cell and analyzed using a light beam emitted in near infrared
(850 nm wavelength) which scanned the column [31,32]. The mea-
surements were done for longer than 1 h and each scan lasts
0.5 min. Transmitted and backscattered photons were analyzed
using detectors placed respectively at 0 and 90 from the inci-
dent beam direction. In fact, the evolution of chemical and physical Fig. 1. (a) Zeta potential as a function of pH for 0.1 vol% Aerosil 200 in pure water,
characteristics, during aging of suspensions could be followed by 0.1 M NaCl, 0.3 M NaCl. (b) Aggregate size as a function NaCl salt concentrations.
46 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354

Fig. 2. Effect of concentration increase on ow transition in 0.6 M NaCl and pH 4.5.

3.1.2. Steady state rheometry

Steady state rheology measurements were considered here to
predict stability status with regard to particle concentrations at
three different pH namely 2, 4.5, 8.5 and in presence of pure water,
0.6 M and 2.75 M NaCl. In Fig. 2, shear viscosity of Aerosil 200
suspension is plotted as a function of shear rate for some differ-
ent concentrations of particles in presence of 0.6 M NaCl and pH
4.5. Apparent viscosity increased by increasing particles concentra-
tion. In diluted suspensions, Newtonian ow was dominating ow
regime, showing no structure was formed in the system. Increase
in concentration of particles showed a transition in the regime
of suspension from Newtonian to the shear thinning, starting at
concentration of 0.7 vol% particles. In fact below this concentra-
tion the solution behaved more Newtonian in the studied range
of shear rate and above 0.7% there was a tendency for pseudo
plastic (shear thinning) behavior. At concentration around 0.7 vol%
the sample showed Newtonian trend just observed at small shear
rates, whereas shear thinning behavior was found to exist at
intermediate shear rates and higher. The shear thinning behav-
ior showed that the aggregates which were formed by particles
interaction were breaking. The same study was carried out for the
two other systems of pure water and 2.75 M NaCl at three differ- Fig. 3. Effect of salinity and pH on critical particle concentration for ow transition
from Newtonian to non-Newtonian.
ent pH. Similar trend was seen for the ow regime conversion
as a result of increase in particle concentrations. The difference
between the systems was that the corresponding critical concen-
trations for this transition were different in each. It is obvious that
this difference originated from the all colloidal forces and parti-
cle interactions. Fig. 3 illustrates the nal consequence obtained
from the ow curves represented by the critical particle concen-
tration at different salinities and pH. General features of all these
plots were that, by increasing solution salinity at constant pH the
critical particle concentration for ow transition decreased. Fur-
ther, the minimum value of the critical concentration shifted to
larger pH values when salinity increased, i.e., it was observed at pH
2 for pure water, pH 4.5 for 0.6 M NaCl and pH 10 for the 2.75 M
NaCl.
Exact effects of pH and salinity on stability were also investi-
gated by comparing values of apparent viscosities at xed 1 vol%
particles. Results are illustrated by shear ow curves in Fig. 4. The
gure illustrates, at any pH, lowest values of viscosities were all
obtained with pure water as medium. Increase of salinity at any
pH had increasing effect on the viscosities. This was in contrast to
the behavior of when adding salt in small amounts which leads to
decrease in suspension viscosity [24,25,27]. Another quite amaz-
ing nding was that at acidic pH of 2 and up to 4.5, increase Fig. 4. Effect of pH and salinity on ow curves for 1 vol% Aerosil 200.
 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354 47

Fig. 6. Inuence of salinity and pH on transmission plot for 1 vol% Aerosil 200.

3.1.3. Multiple light scattering

Parallel to viscosity measurements at xed particle concen-
trations (1 vol%), measurements of light transmission (which is
proportional to the turbidity) from sample were accomplished by
means of Turbiscan instrument. The turbidity test was used as a
measure of aggregation through the interaction between the par-
ticles. The higher the turbidity, the higher the aggregation level
showed. Data were collected every 30 s for total time of 75 min.
Results are illustrated in Fig. 6. Comparison could be made both with
the light transmitted at any exact time or with evolution of trans-
mission as a function of time. It was observed that for the pure water
system smallest transmission (higher turbidity) could be seen at pH
2 (i.e.p.) and by increase of pH, transmission also increased. On the
other hand, transmission values had a gentle decrease over time
Fig. 5. (a) Effect of rest and experiment time on ow curves for 1 vol% Aerosil 200
in 0.6 M and 2.75 M NaCl at pH 8.5. (b) Effect of rest and experiment time on yield while at other pH the values were mostly constant over the whole
stress in 1 vol% Aerosil 200 at pH 8.5 in 0.6 M and 2.75 M NaCl at pH 8.5. time period. With the knowledge of the i.e.p. of 2 this is easily under-
stood that as a higher degree of aggregation and formation of larger
ocs at this pH in comparison with other pHs. Further, at pH 2,
of salinity even up to rather high value of 0.6 M NaCl, did not the initial transmission (transmission at zero time) did not have
change the viscosities effectively from that of water. The increase very signicant difference for all ionic concentrations (water, 0.6 M
was more sensitive just in presence of very high concentration and 2.75 M). Evolution of transmission, i.e. occulation by time, was
of 2.75 M NaCl. Further, coagulation by salt addition at pH far so small for all these cases indicating high sample stability at this
enough from i.e.p. produced stronger networks than at i.e.p. In acidic pH. One conicting fact observed in Fig. 6 was that increas-
fact, adding salt for coagulating the system was much more ef- ing salt concentration from pure water to very high amounts of salt
cient when increasing the pH from i.e.p. This was opposite to increased the light transmitted. It is opposing the results obtained
results obtained for other types of particles like aluminum [35] for the viscosity values shown in Fig. 4. Another consequence of
and of course was in contrast with the zeta potential increase by Fig. 6 was that an increase in pH from 2 to 4.5 did not change trans-
pH increment. In addition, aggregation process was highly time mission so much in presence of 0.6 M and 2.75 M salt within the
dependant and depended on both the rest period for sample load- short period after test started. In comparison, for time evolution
ing in rheometer and on time interval between the measuring transmission, degree of aggregation was higher for 2.75 M indicated
points. The experiments were done for short and long rest time by decrease of transmission over test time. Sample occulation was
of 2 min and 1 h; short experiment time (5 min), and long one much more pronounced in presence of electrolyte and at basic pH
(25 min). Stronger shear thinning behavior and resistance toward 8.5. It could be recognized both by very low light transmission val-
the hydrodynamic forces was found in ow curves when increasing ues, almost 0% for 2.75 M, and highly decreasing time evolution
either rest or experiment time. It was indicated both by increase transmission for the 0.6 M. Further, zero transmission obviously
in apparent viscosity at any shear rate and also increase in val- showed a highly turbid system. This together with visual observa-
ues of yield stresses. This is shown in Fig. 5a and b. These gures tion of completely solid and non-moving sample could be regarded
revealed that aggregation in presence of electrolyte and at basic as an initial step for studying the gelation. This was previously con-
pH (8.5) was a time dependant process, in which the time evo- rmed by rather high values of viscosity and yield stress.
lution for aggregation was much faster in presence of higher salt
concentration. Another consequence was that none of the stud- 3.2. Gelation studies
ied systems showed yield stress, except the ones in presence of
both high enough electrolyte and high pH 8.5. Surprisingly, no 3.2.1. Multiple light scattering
ow resistance was observed for suspensions of pure water at the Effect of particle concentration on microstructure time evo-
i.e.p. lution was considered as main part. Results of transmission and
48 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354

Fig. 9. Effect of particle concentrations at 0.6 M NaCl and pH 8.5 on time sweep test;
Fig. 7. Effect of particle concentration and time on transmission signal at 0.6 M NaCl lled symbols: G ; empty symbols: G .
and pH 8.5.

process. Such phase separation was not observed when concentra-

backscattering signals from the middle of samples tube are shown
tion of particles exceeded 1.3 vol%. A similar study was repeated
in Fig. 7 in presence of 0.6 M NaCl at pH 8.5. Basic pH 8.5 was
for suspensions in 2.75 M NaCl. The trend was similar for the high
selected according to previous study as the most effective pH for
salinity except that the transmission reached zero at much lower
aggregation. For analysis of the results, transmission was mainly
concentration of particles than 0.5 vol%. To make a comparison
considered and not the back scattering. Constant value of light
between the salinities 0.6 M and 2.75 M in next section, the con-
scattering over time together with its low value revealed that in
centration of 1.3% was picked up as the reference to study the effect
very diluted suspensions of 0.05 vol% and 0.15 vol% high level of
of salt concentration.
stability was dominating. Increase of particle concentration accord-
ingly decreased transmission both at zero time and also over time.
Decrease in transmission at zero time continued until it reached 3.2.2. Dynamic rheometry
zero at a concentration of 2.3 vol%. Here, a completely turbid sys- Results of time evolution storage modulus (G ) and loss mod-
tem with a gel like structure had been formed. In Fig. 7 the kinetics ulus (G ) were studied as a function of particle concentration (at
of aggregation and instability were different at each concentration. xed salinity 0.6 M) and salt concentration (at xed particle con-
Extended time experiments over 24 h were performed to detect centration 1.3 vol%) are depicted in Figs. 9 and 10. Both tests were
both kinetics of aggregation and possible instability. Results are accomplished at basic pH 8.5. This experiment which is also dened
shown in Fig. 8 for two typical concentrations of 0.65% and 1.3 vol%. as time sweep had another important consequence. The onset time
It was recognized that stable structures could form within short required for the gelation, could be determined. This was recog-
time (30 min) for 1.3% in comparison with 6 h for 0.65%. Better nized in the graph where G just >G . In fact, the reciprocal of
packing of the sample and higher possibility of network formation gelling time (1/tg) showed the induction rate for initial connec-
between the ocs originating from smaller separation distances tion of the gel network [36,37]. As the gures show, depending
in more concentrated system can describe such a trend. Another on particle and salt concentration, two or three different regions
result was that although both samples of different concentrations were recognized in the graphs. In the rst region (uid region),
were approaching the same level of equilibrium transmission, but both storage and loss modulus values were very small with storage
in 0.65 vol% particles, water was released 15 h after the test started. modulus lower than the loss. Here, the suspension behaved more
This showed that phase separation occurred within the gelation

Fig. 10. Effect of salt concentration on time sweep at constant particle concentration
Fig. 8. Transmission as a function of time in top of the sample in 0.65% and 1.3 vol%. 1.3 vol%; lled symbols: G ; empty symbols: G .
 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354 49

Fig. 12. Effect of salt concentration on time sweep at constant particle concentration
Fig. 11. Effect of particle concentrations at 2.75 M NaCl and pH 8.5 on time sweep 1.3 vol%; lled symbols: G ; empty symbols: G .
test; lled symbols: G ; empty symbols: G .

of the elastic modulus in response to a strain amplitude sweep

as a uid. This region just existed before the gelation time and was was measured in order to identify regime of gels and network
just observed for the more diluted samples of 1.3% and 1.6%. In the microstructure. Figs. 13 and 14 show the shear strain ampli-
second region both moduli had rapid increase with faster increase tude dependence of the storage modulus and the loss modulus
in storage modulus detected by steeper slope. In this zone, which measured at the angular frequency of 1 rad/s for the studied sys-
was seen for all samples, sample had a solid like behavior, since the tems of Figs. 11 and 12. This viscoelastic behavior is similar to
network between silica particles expanded to the whole volume those obtained for other fumed silica suspensions in polar solvent
in the suspension. Rapid increase in storage modulus implied that [13,17,25,26,30,4043].
extensive three-dimensional network structure developed rapidly. A general trend in Fig. 13: when strain amplitude increased,
This showed more solid like nature of the fumed silica samples in both moduli exhibited a constant value with G > G until a criti-
water solution. In the last region, the storage modulus settled in the cal shear strain amplitude ( c ), which dened extent of the linear
pseudo-plateau third region, where the moduli were nearly invari- viscoelastic regime. The corresponding storage value in this linear
ant, compared with that in the second region. Complete network region is known as shear modulus (G 0 ). In fact, at small ampli-
had been almost formed here. Similar time evolution trend was tude strain, the Brownian motion was able to restore the structure
observed for suspensions of Aerosil 200 in presence of polymer and to the equilibrium value; therefore the elastic modulus remains
sulfuric acid [38]. constant. Above  c , G decreased gradually with increasing strain
As a whole picture, gelation (transition from uid like to solid while G exhibited hardening characterized by a hump with a max-
like) became faster as silica content increased, i.e. at 2% and 2.3 vol%. imum value of  max >  c . This increase in values of loss modulus
The same trend was observed when increasing concentration of with strain is attributed to structural changes through breakdown
NaCl from 0.6 M to 2.75 M. In such high particle concentration or of network to a larger number of smaller clusters which are more
high salinity at which gelation was too fast to be detected and the dissipative. Beyond a  c the structure breakdown became signi-
rst mentioned region within time evolution curve did not exist. cant and both the elastic and loss modulus decreased. A common
Further, the variation in storage modulus from zero time to the equi- trend observed in Figs. 914 was that by increase in either particle
librium state was not high. This indicated that most of networks had or salt concentrations there was an increase in both storage and loss
already been formed during sample preparation. This is observed
in Fig. 9 at concentrations of 2% and 2.3% and further in Fig. 10 at
2.75 M NaCl.
Frequency amplitude sweep was performed following the time
sweep test in order to test if the samples were 3D gels or just sim-
ple aggregates. The experiments were carried out at a small strain
0.05% within the linear viscoelastic region and frequency range of
0.0110 Hz. Results are illustrated in Figs. 11 and 12 with the same
effects mentioned in Figs. 9 and 10 before.
Based on the results in Figs. 11 and 12, we regard such solutions
as solid gels due to the frequency independence. In fact, G () exhib-
ited a plateau extending to whole frequency region. This is the most
acceptable denition to dene a gel state. Solid-like gels are char-
acterized by the absence of an equilibrium modulus, by a storage
modulus, which exhibits a pronounced plateau extending to times
at least of the order of seconds, and by a loss modulus which is con-
siderably smaller than the storage modulus in the plateau region
[39].
The other important feature in Figs. 11 and 12 is that an increase
Fig. 13. Storage and loss modulus as a function of shear strain amplitude: effect of
in either particle or salt concentration increased both storage and particle concentrations at 2.75 M NaCl and pH 8.5; lled symbols: G ; empty symbols:
loss modulus in the studied range. Volume fraction dependence G .
50 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354

Fig. 16. Logarithmic plot of critical strain as a function of volume fraction of Aerosil
Fig. 14. Storage and loss modulus as a function of shear strain amplitude: effect of 200 at 0.6 M NaCl and pH 8.5.
electrolyte concentrations at 1.3 vol% particles and pH 8.5; lled symbols: G ; empty
symbols: G .
isotropic structure behaves like Newtonian uids. This originates
from the Brownian motion of the particles which are signicant
in diluted systems and especially for such small nano-scale Aerosil
particles [52]. This entails a tendency toward more random radial
and orientation distribution functions, which in turn counteracts
the effect of ow. In fact in diluted enough systems, the separa-
tion distance between particles is large enough for accounting the
effect of interparticle forces. Meanwhile, when the particle load-
ing reaches some critical levels, particles get closer to each other
and one cannot neglect the inuence of particle interactions that
results in formation of a microstructure. For these structures again
at low shear rates, Brownian motion (random particles collisions)
dominates which can counteract the effect of hydrodynamic forces
and prevent alignment of micrometric aggregates in the direction
of shear. This can result in formation of equilibrium structure at
low shear rates and the Newtonian ow behavior is observed [52].
On the contrary at higher shear rates the convective motion of the
ow dominates [27,53] and affects the structure and results in non-
Newtonian, shear thinning behavior which appears in a situation
Fig. 15. Logarithmic plot of storage modulus as a function of particle concentration similar to that of the large particles [54]. Further, when the concen-
at 0.6 M NaCl and pH 8.5.
tration exceeds this critical value, strong attraction between the
particles can form tight packed microstructures (aggregates) [55]
modulus within the linear viscoelastic domain. On the other hand, which are rather strong but can be broken trough the imposed shear
linear range ( c ) gets shorter. Scaling theory [4447] was applied to rates.
the system and both storage modulus and resulting  c values were Results obtained for electrophoresis measurements of fumed sil-
double-logarithmical plotted as a function of particle volume frac- ica dispersions gained the same feature obtained by all other oxides
tion. Plots are displayed in Figs. 15 and 16. Resulting plots were well showing an increasing negative zeta potential with increasing pH
tted with the power law exponents as 3.8, and as 2.16. According for pH above the i.e.p. This is consistent with the previous results
to the scaling theory high positive exponent for G together with obtained by others [56]. Other conventional facts observed here
negative exponent for  c revealed that the studied gel in presence is that the magnitude of the zeta potential decreases uniformly
of 0.6 M NaCl attributed to a regime of strong gel. Further, applying at each pH as the salt concentration increases. It has also been
the fractal gel model [45,46,48,49] for strong gels concluded fractal demonstrated previously by [57,58]. This normal trend can also be
dimension of ocs (df ) to be around 1.85. Based on the exponent val- predicted by the classical DLVO theory [14,15] which mostly gives a
ues of 3.78 for G and fractal oc dimension of 1.85, we proposed that good prediction on stability of the oxides in aqueous solutions. On
governing gel formation mechanism is the cluster-diffusion-limited the other hand, some abnormal stability behavior of fumed silica
cluster aggregation (DLCA) [45,48,50,51]. suspensions in aqueous phase has been demonstrated in this work
and by several other authors [5759]. This is documented as high
4. Discussions stability of suspensions in presence of high amounts of chloride ions
at low acidic pH, surprisingly at i.e.p., and pronounced instability
Particle volume fraction dependence of silica aqueous sus- with salt within basic pH regions of having higher zeta potentials. It
pensions was investigated in the rst part of the work. At xed has been observed before that the diameters of silica Ludox HS sols
adjusted pH and ionic strength, signicant deviation from Newto- were remained considerably constant within the acidic pH range
nian behavior was induced by the combined contributions from the 24 in presence of a high salinity as 5 M NaCl [59]. The same trend
hydrodynamic and non-hydrodynamic interparticle interactions has been found [60] for fumed silica suspensions. Further, in con-
and also random Brownian motions under a uniform shear eld. At trast to the DLVO calculation which accounts for an increase in of ccc
low particle volume fractions, the suspension is nearly random and with pH, an opposite trend was measured experimentally for those
 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354 51

silica sols [60]. Possible explanations for the enhanced stability of stability mechanisms at any of acisic and basic pH. Gel formation,
silica at low pH have been postulated. However, exact mechanisms low critical concentration of particles for ow transition from New-
for stability has not mentioned clearly in view of combined pH and tonian to shear thinning and existence of yield stresses at basic pH at
high salinity. 8.5 are evidences of these different mechanisms when comparing
The abnormally enhanced stability at low acidic pH together to oppositely very weak occulation, high particle concentration
with direct force measurements of the silica surface provide a for ow transition and zero values of yield stress at pH 2 and 4.5 in
strong evidence for a presence of short-range repulsive forces not presence of same electrolyte concentrations. At acidic pH and i.e.p.
predicted by DLVO theory [61,62]. The presence of such forces is the following reaction is governing at the particle surface [24,25]:
more pronounced at the i.e.p. of silica. Since at i.e.p. there is no con-
tribution of the repulsive EDL. These short-range repulsive forces SiOH + H3 O+ SiOH2 + + H2 O (5)
are found to be strong and dominating over the attractive van der
It is assumed that there is a rather higher ability of water to form
Waals Force. Although there is debate over the origin of this repul-
hydrated layer as above on fumed silica (SiOH) surface compared
sion, it is most likely due to either structural repulsive forces of
to the granular silica (SiO2 ). This is due to existence of OH bound
a hydration layer or steric repulsion due to silicic acidic groups
to the silanol groups of fumed silica. Such hydrated layer between
[61,62]. The most plausible explanation for silica is that stability
the particles keeps particle surfaces apart from each other and
of silica in acidic conditions is caused by particle hydration, with
prevents their bridging. This is understood to be the main rea-
the silanol groups being capable of hydrogen bonding with water
son for abnormal stability of these suspensions at acidic pH and
[62,63]. This water layer between the silica surfaces is acting as
particularly at i.e.p. On Addition of electrolyte to the above sys-
steric repulsion and stabilizes the silica sols in water. This is so pro-
tem, the electrolyte cation ions can extract water from the silica
moting with regard to small value of the Hamakar constant of silica
surface and decrease the thickness of the hydration layer formed
particles in water which consequentially induces weak attractive
by water and reduce the extent of such short-range repulsion.
van der Waals Force [59,64,65]. Considering this hydration force,
According to this hypothesis, the size of the cations and its afnity
total interparticle energy accounting for the stability level of silica
to the water are essentially important [22,64,65]. Smaller cations
can be obtained by addition of the hydration term to the classical
like Na+ (known as structure making) have been also found to
DLVO forces:
have stronger afnity to water in comparison with larger poorly
V = VDLVO + Vhydration (1) hydrated cations (known as structure making) [19,22]. The former
can diffuse easier in the hydrated layer between silica surfaces and
With [1418]: remove more hydrated water from the surface and therefore pro-
 2 duce more disabilities and occulating the slurries. Since the ability
aA123 e 0
VDLVO = + 32akT tanh e2 exp(kH) (2) of fumed silica for water structuring is very high in comparison
12H 4kT
with other minerals and also in comparison with many cations, in
 1/2 order to occulate the system, high enough concentrations of salt is
2e2 NA ce1 z 2 needed to generate competition for water between the silica surface
k= (3)
kT and cation ions in solution. Obviously, salt concentration is much
where A123 is the Hamakar constant for the system lower with more structure breaking cations like Na+ compared
particlewaterparticle, e is elementary charge, NA is Avogadros to the structure breaking cations. Removal of structured water
number, ce1 is the bulk concentration of electrolyte of valency z, molecules from the silica surface produces a thinner hydration
k is Boltzmanns constant, T is temperature, is the permittivity layer and subsequently much weaker short-range repulsive forces
of water, 0 is the surface potential (often substituted by the zeta [22,65].
potential), H is the surface-to-surface separation between the At basic pH, we believe that two different mechanisms are
particles and k is the inverse Debye length (or the reciprocal of simultaneously responsible for destabilization of Aerosil in term
double layer thickness). of electrolyte addition. The rst one is explained by ion-exchange
And hydration force as [59,65]: mechanism in which cations in electrolyte displace the protons of
the silanol groups (OH), thereby preventing the hydrogen-bonding
Vhydration = aNA Ch ce1 2 exp(H) (4) of water molecules to silica. This can be understood by the following
interaction at high pH [24,25]:
where Ch is the hydration constant and  is the decay length. It can
be seen that hydration force is also dependent on the salinity and SiOH + OH  SiO + H2 O (6)
the pH inuencing ce1 and indirectly H.
It was mentioned in the results that addition of high concen- On the other hand it was shown in Fig. 5 that the occulation
tration of salt reduces the stability of Aerosil 200 suspensions. by salt at basic pH is highly time dependent, opposite to the acidic
Comparing this result with previous works on addition of small condition. This implies that other but time dependant destabiliza-
amounts of salt to the silica suspensions [24,25,27,53] shows a con- tion mechanism also contributes to the overall destabilization at
tradiction. For the latter case, it has been indicated that in presence basic pH. However, there is always an equilibrium exist in reaction
of small amounts of salt, there is almost no signicant change in (6), hence there can be an interaction between the associated and
values of zeta potential in comparison with silica suspensions in disassociated silanols meaning (SiO and SiOH). Therefore, coag-
pure water. Instead, it has been shown that as salt concentration ulation occurs because the particles become linked by acidbase
increases, the shear viscosity reduces. This has been explained by interactions by silanol groups on the particle and dissociated SiO
the ability of salt in charge screening and compression of electrical groups on another. The binding of dissociated and associated sur-
double layer which leads to decrease in effective volume fraction face silanols is found to be implemented through the hydrogen
and viscosity [24,27,53]. For the case of high concentration of salt bound water of hydrated cation in the vicinity of the surface [65,66].
as in this work, salt concentration is so high that particles in sus- This is possible that hydrated cations of salt act as a bridging agent
pension cannot be electrostatically stabilized. However, the results between two silica particles by having at least two of its hydration
are consistent with what has been reported by [26,40]. Further, waters displaced hydroxyls of the surface silanol groups [59,45,66].
high stability at acidic pH and pronounced instability at basic pH in This is known to be as bridging mechanism which is pronounced in
presence of same concentration of salt shows existence of different suspensions of fumed silica having the OH group at their surfaces.
52 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354

However, the bridging mechanism requires some primary con- and obtained the proper particle concentration to form stable gel
ditions. First high enough OH concentration is needed to prevent regime. A typical graph is shown in Fig. 8.
hydration layer. Second, the surface of the silica particles must Two approaches, i.e. scaling theory and fractal gel model were
be covered with acidic hydrogen pertaining to the silanol groups applied to determine regime of the gel in view of storage and loss
or to the water of hydration of the adsorbed cations [66]. This modulus and their evolution by strain and frequency. Storage mod-
is mainly the reason for time dependent destabilization at high ulus is a property representing the elasticity of a material and
pH which almost takes some minutes. Here one can deduce that interaction between the particles: higher storage modulus reects
according to this hypothesis the more the hydration energy of the enhancement of the elasticity, extensive interactions between par-
cation, the more difcult to adsorb on silica. Therefore, in contrast ticles and formation of stronger network structure. Increase of
to the acidic pH, poorly hydrated cations (Cs+ and K+ ) can adsorb the loss modulus indicates an increase of the aggregates size
in greater quantity to the silica surface than the well-hydrated ions and of polydispersity [1,48]. According to the fractal gel model
(Li+ and Na+ ) at high pH. This was also shown by [66,67]. This is the [45,46,48,49], the magnitude of elastic modulus can follow the
background that highly hydrated cations regain higher amounts of power law dependence of volume fraction ():
salt coagulate the sol. This is in contrary to acidic pH which the
G  A (7)
most hydrated cations have stronger ability to coagulate the sys-
tem [66,67] and is a strong proof to the proposed hypothesis. Based where magnitude of exponent A is related to network structure
on the results obtained by rheometry and dynamic light scatter- through the fractal dimension.
ing, gel formation in fumed silica suspensions can be realized as a The value  c corresponds to a measure of the brittleness of the
consequence of combined ion-exchange and bridging destabiliza- gel structure. Moreover, the  c is dependant on volume fraction
tion mechanism which is implemented in presence of electrolyte through the power-law relation [45,46,48,49]:
at high basic pH.
c m (8)
In general, the process of gelation is an aggregation of particles
or polymers into fractal clusters and consequent interpenetration of According to the theory of gelation, gel network is considered
the clusters linking together to form an innite network [39,48,49]. to be a collection of fractal ocs which are closely packed through-
There is usually no sharp distinction between the occulation and out the sample. Depending on the strength of the links inside the
the gelation as the concentration of colloid or electrolyte increases ocs and between the ocs, there can be two types of behavior,
or the pH of the system varies. So, if the charge and hydration namely, strong (interoc) and weak-link (intraoc) gels [4446].
forces are gradually reduced by addition of electrolyte, a point In the strong gel regime, the elastic modulus has a high increase
would be reached were adhesions will occur if two particles come with increasing particle concentration but the limit of linearity
together at one particular orientation. Also, during gelation, Brown- decreases.
ian motion of the particles slows down and eventually cease. About Exponents A and m are attributed to such a network structure,
the reversibility of the gels it is good to know when regular ori- each through different expressions depending on the regime of gel
ented structures will be built up through out the system, since only with strong or weak bonds. In fact, in the strong gel regime, links
very limited areas of the surfaces are in contact, any stress such between ocs (interoc links) have a higher elastic constant than
as shaking tends to destroy the orientation and results in destroy- those in the ocs (intraoc links) [46]. This is exactly opposite to
ing the structure. When the stress is ceased, the structure will that of the weak gel regime. Exponents A and m obtained 3.8 and
form again and even faster due to disorientation of the structure 2.16 are well located within the range 4 0.5 and 2 0.2 previ-
through the stress applied [30,68]. With high enough electrolyte ously obtained [17,40,42] for silica suspensions. A negative power
or particle concentration or both, random coagulation will occur law exponent, m, is consistent with that predicted by the fractal gel.
and more compact irregular ocs will be formed that some times It shows that regime of gel here is strong one. In addition, for any
even cannot be dispersed again by applying stress with shaking. It regime of gels there exist two different methods to obtain the frac-
is also worth saying that Similar to the other instable systems, if tal dimension of ocs in order to give the proper mechanism for the
the concentration of the solid or electrolyte is not enough, the oc- gelation. The fractal dimension can be obtained whether through
culation, coagulation and consequently phase separation occurs in its relationship with the exponent A [42] or the exponent m [40].
the system. But on the other hand, with just a smooth increase in For the strong gel regime the exponents A and m are related to the
particle concentration or electrolyte, this instability transforms to fractal structure of the gel according to A = (d + dchem )/(3 df ); and
the stable-gel system [40,45]. In another word, packing of the sys- m = (1 + dchem )/(3 df ), where dchem is fractal dimension of back-
tem should be enough in order to reach gelation threshold. After bone and df the fractal dimension of ocs [40,46]. According to [69]
achieving this point, further increase in concentration would just the ratio df /dchem can be considered as invariant with a value of
enhance the kinetics and gel strength. 3/2 for three-dimensional networks of silica gels. By substation of
The gelation process of silica sols is found to be similar to above exponents 3.8 and 2.16 as for A and m the fractal dimension of
mechanism to good extent [45]. Some studies have been done to ocs is estimated as df = 1.85 and df = 1.87, respectively, which are
investigate inuence of ion specicity and concentrations on silica found to be well superimposed applying the both methods.
gelations. It is mostly found that depending on these parameters, The exponent A was predicted before to be 4.5 0.2 for
four different stages can exist in phase diagram of suspensions reaction limited cluster aggregation (RLCA) [17,45] and 3.5 0.2
of silica in water as: sol (S), pregelled (PG), gelled (G), and two- for diffusion-limited cluster aggregation (DLCA) [25,45,48,50,51].
phase separated (TP) silica suspensions [26,40]. When talking about Accordingly, for these two limiting regimes RLCA and DLCA, the
gel, often the specic region as Gelled (G) is under consideration fractal dimension calculated from the lattice models have been
which almost gives the most stable, homogenous and stronger gel reported to be respectively 2 and 1.8 [51].
properties. In order to determine if the system parameters are lead- The exponent value of 3.78 and fractal ocs dimension of
ing to formation of gel region, multiple light scattering analysis 1.85 which is obtained here, is coincident with the results of the
with measuring transmission or backscattering is a good tool for cluster-cluster aggregation model [25,45,48,50,51]. This indicates
this purpose. A simultaneous follow-up of gel formation with both that diffusion-limited cluster aggregation (DLCA) is mainly gov-
Turbiscan and small deformation oscillatory by rheometery is a erning gel formation mechanism. The clustercluster aggregation
strong and complete approach especially when the gelation kinet- model introduced by Meakinl [51] suggests that colloidal aggre-
ics is most demanded. We also applied this complementary method gates should behave as stochastic mass-fractals on a scale large
 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354 53

compared with the primary particle size. In fact, there are small [2] S. Puskas, j. Balzs, A. Farkas, I. Regdon, O. Berkesi, The signicance of col-
aggregates in the fresh sample which grow slowly with time and loidal hydrocarbons in crude oil production, Part 1. New aspects of the stability
and rheological properties of water-crude oil emulsions, Colloids Surfaces A:
form the clusters and eventually gel networks. This is also probed Physicochem. Eng. Aspect 123 (1996) 279293.
here by the Turbiscan. Bigger average sizes of the ocs (smaller [3] B. Edmonds, R.A.S. Moorwood, R. Szczepanski, A Unied Model of Solid Deposi-
fractal ocs dimension) for DLCA process in comparison with those tion from Petroleum Fluids Including Waxes, Asphaltenes, Hydrates and Scales,
Infochem Computer Services Ltd., 2000.
of RLCA are responsible for a strong regime gel compared to the [4] P. Ungerer, B. Faissat, C. Leibovici, H. Zhou, E. Behar, G. Moracchini, J.P. Courcy,
weak-link gel, respectively. It was mentioned earlier that in strong High pressure-high temperature reservoir uids: investigation of synthetic
regime of gel, the links between the ocs (interoc links) have a condensate gases containing a solid hydrocarbon, Fluid Phase Equilib. 111
(1995) 287311.
higher elasticity than those of the ocs which is completely oppo- [5] J.R. Payne, Oil/suspended particulate material interactions and sedimentation,
site to the weak gels. This is attributed to the fact that ocs which Spill Sci. Technol. Bull. 8 (2003) 201221.
are formed either at increasing particle or salt concentrations are [6] K. Lohne, Separation of solids from produced water using hydrocyclone tech-
nology, Trans. IChemE 72A (1994) 169175.
blocked within the microstructure of the gel and have less exible
[7] G.A. Mansoori, Asphaltene, resin, and wax deposition from petroleum uids:
structure compared to small ocs. Therefore structure is disrupted mechanisms and modeling, Arabian J. Sci. Eng. 21 (1996) 707723.
just by a very small imposed strain [37,46]. Higher elastic modulus [8] L.L. Schramm, Emulsions, Foams, and Suspensions, Wiley-VCH, Weinheim,
in higher particle or salt concentration is also attributed to higher 2005.
[9] R.K. Iler, The Chemistry of Silica: Solubility, Polymerization, Colloid and Surface
extent of interparticle interaction with regard to strengthening the Properties and Biochemistry, Wiley, New York, 1979.
interoc links. Same experiment for gelation was done on this study [10] Degussa technical bulletins, Basic characterization of Aerosol (No. 11); Aerosol
at i.e.p. pH 2 with pure water. No gel was which there reveals that as a thickening agent for liquid systems (No. 23), Deggusa corp. (1993).
[11] H. Barthel, L. Rosch, J. Weis, Organic Chemistry: From Molecules to Materials,
hydration forces are opposing interparticle bindings even at i.e.p. VCH, Weinheim, 1996, pp. 761777.
[12] Degussa, Technical Bulletin Pigments, vol. 6, 1989, pp. 1223.
[13] A. Chougnet, A. Audibert, M. Moan, Linear and non-linear rheological behavior
5. Conclusions of cement and silica suspensions: effect of polymer addition, Rheol. Acta 46
(2007) 793802.
The total particleparticle and particlesolvent interactions in [14] B.V. Derjaguin, L. Landau, Theory of the stability of strongly charged lyophobic
sols and of the adhesion of strongly charged particles in solutions of electrolytes,
suspensions of fumed silica have been controlled by salt concen- Acta Physicochim. URSS 14 (1941), 663662.
tration, particle concentration and pH of solution in aqueous phase. [15] E.J.W. Verwey, J.T.G. Overbeek, Theory of the Stability of Lyophobic Colloids,
Complementary method of multiple light scattering and transient- Elsevier, Amsterdam, the Netherlands, 1948.
[16] D.J. Shaw, Introduction to Colloid and Surface Chemistry, fourth ed., Butter-
dynamic rheometry has been successful in predicting the stability worths, London, 1992.
levels in view of the extended-DLVO theory for the aqueous fumed [17] J.L. Trompette, M.J. Clifton, Inuence of ionic specicity on the microstructure
silica suspensions. Experiments revealed that suspensions in water and the strength of gelled colloidal silica suspensions, J. Colloid Interface Sci.
276 (2004) 475482.
are very stable at their i.e.p. and other acidic pH even in presence
[18] P.C. Hiemenz, Principles of Colloid and Surface Chemistry, second ed., Marcel
of high amounts of electrolyte. This has been attributed to pres- Dekker, New York, 1986.
ence of strong short-range hydration forces. Intensive coagulation [19] B.W. Ninham, V. Yaminsky, Ion binding and ion specicity: the Hofmeister effect
and Onsager and Lifshitz theories, Langmuir 13 (1997) 20972108.
has been well described by both ion adsorption and silica binding
[20] J.N. Israelachvili, Intermolecular and Surface Chemistry, second ed., Academic
mechanisms at basic pH and in presence of high enough electrolyte. press, San Diego, 1992.
Highest yield stress and apparent viscosity, highly time dependant [21] J.A. Greathous, D.A. McQuarrie, Conventional hypernetted chain force calcula-
rheometry and light transmission, and smallest critical particle con- tions for charged plates with adsorbing counterions, J. Colloid Interface Sci. 181
(1996) 319.
centration for ow transition from Newtonian to shear-thinning are [22] M. Colic, G.V. Franks, M.L. Fisher, F.F. Lange, Inuence of ion size on short-range
all correlated with the highest salt amount at basic pH. repulsive forces between silica surfaces, Langmuir 14 (1998) 61076112.
Network structure forms as gel at pH 8.5 and depends on both [23] A.A. Zaman, A.L. Fricke, B.M. Moudgi, Rheological behavior of highly con-
centrated aqueous silica suspensions in the presence of sodium nitrate and
concentration of salt and particles and it strongly inuences both polyethylene oxide, J. Rheol. 40 (1996) 11911210.
the critical strain and the storage modulus. The critical strain in [24] S. Savarmand, P.J. Carreau, F. Bertrand, D.J.E. Vidal, M. Moan, Rheological proper-
presence of higher salt or higher particle concentration has been ties of concentrated aqueous silica suspensions: effects of pH and ions content,
J. Rheol. 47 (2003) 11331149.
maintained to much less values suggesting that less open struc- [25] Sh. Chen, G. ye, J. Sjblom, Effect of pH and salt on rheological properties of
tures with less exible network have been formed. The higher Aerosil suspensions, J. Dispers. Sci. Technol. 28 (2007) 852860.
storage here is attributed to the larger interparticle attractive inter- [26] K. Yokoyama, Y. Koike, A. Masuda, M. Kawaguchi, Rheological properties of
fumed silica suspensions in the presence of potassium chloride, J. Appl. Phys.
actions and stronger interocs compared to the intraocs in view
46 (2007) 328332.
of strong-link regime of the gel. Fractal dimension of ocs as 1.85 [27] J.-H. So, S.-M. Yang, Ch. Kim, J. Chun, Hyun, Microstructure and rheological
together with the volume fraction dependence of storage modulus behavior of electrosterically stabilized silica particle suspensions, Colloids Sur-
faces A: Physicochem. Eng. Aspects 190 (2001) 8998.
with power low exponent 3.8 reveals that mechanism of gelation
[28] F.J. Galindo-Rosales, F.J. Rubio-Hernndez, J.F. Velzquez-Navarro, A.I. Gmez-
is mainly governed by diffusion-limited cluster aggregation (DLCA) Merino, Structural level of silica-fumed aqueous suspensions, J. Am. Ceram. Soc.
mechanism. 90 (2007) 16411643.
[29] T. Mezger, The Rheology Handbook, second ed., Ronnenberg, Germany, 2006.
[30] S.R. Raghavan, S.A. Khan, Shear-induced microstructural changes in occulated
Acknowledgements suspensions of fumed silica, J. Rheol. 39 (1995) 13111325.
[31] O. Mengual, G. Meunier, I. Cayre, K. Peuch, P. Snabre, TURBISCAN MA 2000, Mul-
tiple light scattering measurement for concentrated emulsion and suspension
This work is performed by the FACE centre research coop- instability analysis, Talanta 50 (1999) 445456.
eration between IFE, NTNU and SINTEF. The centre is funded by [32] R. Vie, N. Azema, J.C. Quantin, E. Touraud, M. Fouletier, Study of suspension set-
The Research Council of Norway, and by the following industrial tling: approach to determine suspension classication and particle interactions,
Colloids Surfaces A: Physicochem. Eng. Aspects 298 (2007) 192200.
partners: StatoilHydro AS, Norske ConocoPhillips AS, Vetco Gray [33] G.V. Franks, Zeta Potentials and yield stresses of silica suspensions in concen-
Scandinavia AS, Scandpower Petroleum Technology AS, FMC, CD- trated monovalent electrolytes: isoelectric point shift and additional attraction,
adapco, ENI Norge AS, Shell Technology Norway AS. J. Colloid Interface Sci. 249 (2002) 4451.
[34] M. Kosmulski, Positive electrokinetic charge of silica in the presence of chlo-
rides, J. Colloid Interface Sci. 208 (1998) 543545.
References [35] B.V. Velamakanni, J.C. Chang, F.F. Lange, D.S. Pearson, New method for efcient
colloidal particle packing via modulation of repulsive lubricating hydration
[1] L.L. Schramm, Suspensions: Fundamentals and Applications in the Petroleum forces, Langmuir 6 (1990) 1323.
Industry, Advances in Chemistry Series 251, American Chemical Society, [36] V.M.-F. Lai, C.-Y. Lii, Gelation kinetics of agars from pterocladia capillacea exam-
1996. ined by dynamic rheometry, J. Food Sci. 67 (2002) 672781.
54 A. Amiri et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 349 (2009) 4354

[37] S.A. Khan, R.K. Prudhomme, E.M. Rabinovich, M.J. Sammon, Rheology of the [53] S. Jae-Hyun, O. Kyoung, Y. Seung-Man, Microstructure and phase behavior
gelation of uorine doped silica sols, J. Non-Crystalline Solids 110 (1989) of concentrated silica particle suspensions, Korean J. Chem. Eng. 21 (2004)
153162. 921928.
[38] J. Parka, S.B. Parka, S.M. Yanga, W.H. Honga, C.R. Choia, J.H. Kim, Rheological [54] D.R. Foss, J.F. Brady, Structure, diffusion and rheology of Brownian suspensions
characterization and optimization of gelled electrolyte for sealed leadacid bat- by Stokesian dynamic simulation, J. Fluid Mechanics 407 (2000) 167200.
teries by small amplitude dynamic oscillation measurement, J. Non-Crystalline [55] E. Windhab, Fluid immobilization a structure-related key mechanism for the
Solids 351 (2005) 23522357. viscous ow behavior of concentrated systems, Appl. Rheol. 10 (2000) 134144.
[39] K. Almdal, J. Dyre, S. Hvidt, O. Kramer, Towards a phenomenological denition [56] J. Scales, F. Grieser, W.H. Thomas, L.R. White, Y.C. Chan, Electrokinetics of the
of the term Gel, Polym. Gels Netw. 1 (1993) 517. silica-solution interface: a at plate streaming potential study, Langmuir 8
[40] K. Okazaki, M. Kawaguchi, Inuence of monovalent electrolytes on rheologi- (1992) 965974.
cal properties of gelled colloidal silica suspensions, J. Dispers. Sci. Technol. 29 [57] L.H. Allen, E. Matijevic, Stability of colloidal silica: I. Effect of Simple Electrolytes,
(2008) 7782. J. Colloid Interface Sci. 31 (1969) 287.
[41] F. Bossard, Linear and nonlinear viscoelastic behavior of very concentrated [58] J. Depasse, A. Watillon, The stability of amorphous colloidal silica, J. Colloid
plate-like kaolin suspensions, J. Rheol. 51 (2007) 12531270. Interface Sci. 33 (1970) 430.
[42] F. Yziquel, P.J. Carreau, P.A. Tanguy, Non-linear viscoelastic behavior of fumed [59] B.P. Binks, S.O. Lumsdon, Stability of oil-in-water emulsions stabilized by silica
silica suspensions, Rheol. Acta 38 (1999) 1425. particles, Phys. Chem. Chem. Phys. 1 (1999) 30073016.
[43] S.A. Khan, N.J. Zoellera, Dynamic rheological behavior of occulated fumed [60] J. Rubio, Stability of colloidal silica in the presence of quaternary ammonium
silica suspensions, J. Rheol. 37 (1993) 12251235. salts, Colloid Interface Sci. 36 (1971) 289.
[44] P.G. De Gennes, Scaling Concepts of Polymer Physics, Cornell University Press, [61] H. Yotsumoto, R.H. Yoon, Application of extended DLVO theory: stability of silica
Ithaca, New York, 1979. suspensions, J. Colloids Interface Sci. 157 (1993) 434442.
[45] R. Buscall, P. Mills, Scaling behaviour of the rheology of aggregate networks [62] H. Yotsumoto, R.H. Yoon, Application of extended DLVO theory: stability of rutile
formed from colloidal particles, J. Chem. Soc., Faraday Trans. I 84 (1988) suspensions, J. Colloids Interface Sci. 157 (1993) 426433.
42494260. [63] G. Vigil, Z. Xu, S. Steinberg, J. Israelachvili, Interaction of silica surfaces, J. Colloid
[46] Sh. Wei-Heng, et al., Scaling behavior of the elastic properties of colloidal gels, Interface Sci. 165 (1994) 367385.
Phys. Rev. A 42 (8) (1990) 44724480. [64] M. Colic, G.V. Franks, M.L. Fisher, F.F. Lange, Effect of counterion size on short
[47] M. Chen, W.B. Russel, Characteristic of occulated silica dispersion, J. Colloid range repulsive forces at high ionic strengths, Langmuir 13 (1997) 31293135.
Interface Sci. 141 (1991) 564577. [65] J.P. Chapel, Electrolyte species dependent hydration forces between silica sur-
[48] F. Family, D.P. Landau, Kinetics of Aggregation and Gelation, NorthHolland, faces, Langmuir 10 (1994) 42374243.
Amsterdam, 1984. [66] J. Depasse, Coagulation of colloidal silica by alkaline cations: surface dehydra-
[49] E. Dickinson, Structure and rheology of simulated gels formed from aggregated tion or interparticle bridging? J. Colloid Interface Sci. 194 (1997) 260262.
colloidal particles, J. Colloid Interface Sci. 225 (2000) 215. [67] S.K. Milonjic, A relation between the amounts of sorbed alkali cations and the
[50] P. Meakin, Formation of fractal clusters and networks by irreversible diffusion- stability of colloidal silica, Colloids Surf. 63 (1992) 113.
limited aggregation, Phys. Rev. Lett. 51 (1993) 11191122. [68] M. Dolz, F. Gonzalez, J. Delegido, M.J. Hernandez, J. Pellicer, A Time-dependent
[51] P. Meakin, A historical introduction to computer models for fractal aggregates, expression for thixotropic areas. Application to Aerosil 200 hydrogels, J. Phar-
J. Sol-Gel Sci. Technol. 15 (1999) 97117. maceut. Sci. 89 (2000) 790797.
[52] T. Aubry, B. Largenton, M. Moan, Study of aggregate density in fumed silica [69] A.A. Potanin, On the mechanism of aggregation in shear ow of suspensions, J.
suspensions, J. Colloid Interface Sci. 202 (1998) 551553. Colloid Interface Sci. 145 (1991) 140157.