Colloids and Surfaces B: Biointerfaces 21 (2001) 125– 135
www.elsevier.nl/locate/colsurfb
Specific cation adsorption on protein-covered particles and
its influence on colloidal stability
J.A. Molina-Bolı́var a, F. Galisteo-González b, R. Hidalgo-Alvarez b,*
a
b
Departamento de Fı́sica Aplicada II, Escuela Politécnica, Uni6ersidad de Málaga, Campus de El Ejido, 29013 Málaga, Spain
Grupo de Fı́sica de Fluidos y Biocoloides, Departamento de Fı́sica Aplicada, Uni6ersidad de Granada, 18071 Granada, Spain
Abstract
Protein coated particles present an anomalous colloidal stability at high ionic strength when the classical theory
(DLVO) predicts aggregation. This observed deviation from DLVO behaviour appears for electrolyte concentrations
above some critical bulk value. As we have suggested in previous publications the existence of an additional
short-range repulsive ‘hydration force’ due to specific hydrated cation adsorption could explain this anomalous
stability. The overlap of the hydration layers when two particles approach should provoke this repulsive force. New
evidence of this mechanism has been observed when electrophoretic mobilities of protein-carrying latex particles were
measured at various concentrations of sodium and calcium chloride. In the latter case a sign reversal of zeta-potential
was found, probably due to the specific adsorption of Ca2 + ions on protein molecules. The adsorption increases with
the medium pH. These results have been analyzed following the treatment proposed by Ohshima and co-workers for
large charged colloidal particles coated with a layer of protein. This study shows an increase in the positive
fixed-charge density on the protein caused by the adsorption of cations. © 2001 Elsevier Science B.V. All rights
reserved.
Keywords: Protein colloids; Colloidal stability; Hydration forces; Specific cation adsorption; Electrophoretic mobility
1. Introduction
The study of fluids confined between two solid
surfaces is of great interest in the areas of adhesion, lubrication, stability of colloids, phase behaviour and critical phenomena in porous media,
and the cleaning of semiconductor surfaces. In
recent years interest has also been focused in
* Corresponding author. fax: + 34-9-58243214.
E-mail address: [email protected] (R. Hidalgo-Alvarez).
biological systems such as food colloids and cells,
bacteria and virus adhesion. When the distance of
separation between two solid surfaces immersed
in a fluid medium falls within the range of few
hundred nanometers, it has long been known that
the interaction can be described to a first approximation using the DLVO theory [1,2], which attributes their mutual interaction energy to the
Van der Waals attraction and the electrostatic
repulsion. The resulting force due to these two
long-ranged forces is repulsive at large separations, but become attractive at short distances. As
a consequence, the DLVO theory provides a sim-
0927-7765/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 7 - 7 7 6 5 ( 0 1 ) 0 0 1 6 6 - 7
126
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
ple treatment for the stability of a colloidal suspension, which decreases with electrolyte concentration due to the screening of the repulsive
electrical double layer.
In the last decade, several new forces have been
invoked to explain a variety of short-ranged phenomena. Direct investigations of the interaction
potential between silica surfaces [3,4] and mica
surfaces [5] in aqueous electrolyte solutions have
revealed agreement with DLVO at separations
above a few nanometers but, at smaller separations, a short-range repulsive force appears, often
termed as a ‘hydration’ interaction. This interaction was ascribed by Pashley, for the case of mica
surfaces in aqueous solutions [5,6], as arising from
the removal of hydration water of surface-bound
counterions. Hydration-like forces have also been
reported in surfactant systems [7] and their origin
suggested to be the removal of water of hydration
of the surfactant head groups as well.
The classical DLVO theory treats the intervening medium as continuous, so it is not surprising
that the model breaks down when the liquid
medium between two surfaces is only few molecular diameters in width. It has long been postulated
that a modified water structure exists at solid/water interfaces (in fact, hydrophobicity is a manifestation of water structure at surfaces [8]). For
colloids and interfaces in aqueous media the predominant effect is attributed to the hydration of
adsorbed counterions and ionic functional groups
on the surface. As such surfaces approach each
other closely during interaction, some dehydration of the ions and the surface would have to
occur, leading to an increase in the free energy
and hence a repulsion. These effects are usually
referred to as hydration or structural forces. This
explanation, however, is far from being completely understood (and theoretically modelled)
yet.
The ‘hydration forces’ name reflects the proposition that the force is caused by the structure of
water between surfaces, which depends on the
hydrophobic/hydrophilic character of the surface
and on the hydration of adsorbed counterions
and ionic functional groups in the surface [9,10].
Pashley and Israelachvili [5,11,12] have shown
that the hydration repulsion occurring between
two mica surfaces across a monovalent electrolyte
is intimately related to the hydration number of
the cation, i.e. the average number of water
molecules in the first shell. The results showed
clearly that the strength and the range of the force
increased with the series Cs+ B Rb+ B K+ B Na+
B Li+. According to these authors, hydrated
cations adsorb to hydrophilic surfaces, and they
presumably retain some of their water of
hydration.
On the other hand, electrophoresis of colloidal
particles is a convenient method to obtain information about surface structure, specially about
the distribution of ionogenic groups along this
surface. For example, specifically adsorbed ions
can be recognized by their ability to reverse the
sign of the n-potential, whereas indifferent ions
can only reduce it asymptotically to zero [13,14].
Ohshima et al. [15–17] proposed a membrane
model in which the membrane-fixed charges are
distributed at an uniform density through a surface layer of thickness d. The membrane is permeable to electrolyte ions and shows the mentioned
fixed-charged groups in electrolyte solution. With
this model it is possible to calculate the fixedcharge density on a protein layer from the measured n-potential which depends on a weighted
average of the Donnan potential and the potential
at the boundary between the surface charge layer
and the surrounding solution.
Many biocolloids are stabilized by surface
protein molecules. The protein may form an integral part of the particle structure or may have
been adsorbed from solution. As yet, there is no
quantitative theory of colloid stabilization by
proteins, but it seems clear that specific cation
adsorption effects play an important role [18]. In
the important commercial field of dairy products,
it is observed that casein micelle stability does not
conform to the DLVO theory. This system may
tolerate divalent ion concentrations before coagulation [19,20].
In previous publications we have described the
anomalous stability of protein-covered colloidal
systems [21,22]. They can be considered as another exception to the DLVO theory, because
they are stable at high salt concentrations where
the theory predicts aggregation. This stability was
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
attributed to a repulsive hydration force which
prevents protein-covered particles from aggregating. Experimental data suggests that hydrated
cations adsorbed on the protein surface could be
responsible for this behaviour. In the present
work new evidence is presented in relation to this
proposed mechanism of anomalous stabilization,
by analyzing electrophoretic measurements of
protein-covered particles.
2. Experimental
2.1. Materials
The latex sample used in this study was kindly
provided by Joxe Sarobe (Chemical Engineering
Group, University of Basque Country), and prepared by means of a core-shell emulsion polymerization in a batch reactor [23]. The core was a
seed of polystyrene and the shell was a styrene/
chloromethylstyrene copolymer, rendering particles with a chloroactivated functionality at the
surface which can covalently link proteins [24].
The mean diameter of these particles was, as
determined by TEM, 20195 nm.
Latex was cleaned by three cycles of centrifugation and redispersion of the pellet in acidic water,
and then thoroughly washed with deionized water
in a serum replacement cell until constant conductivity was achieved. Surface charged groups were
determined by conductimetric titration (− 3.79
0.2 mC cm − 2, strong acid), and chloride surface
groups by hydrolysis with glycine/NaOH, acidification with HNO3, and subsequent titration of
free Cl− with AgNO3 (59 94 mequiv g − 1, 2.119
0.14 mequiv m − 2).
F(ab%)2 fragments, which were obtained from
polyclonal rabbit IgG, and purified by affinity
chromatography, were kindly supplied by Biokit,
S.A. (Barcelona). The isoelectric points (i.e.p.)
range, determined by isoelectric focussing, was
rather wide (4.7– 5.9) as a consequence of the
polyclonal nature of the IgG. All chemicals were
analytical grade quality. Water was purified by
reverse osmosis, followed by percolation through
charcoal and a mixed bed of ion-exchange resins.
127
Protein was attached to the latex particles by
incubating the latex (0.4 m2) and protein solution
in phosphate buffer (pH 7.2) at 35°C for 5 h. The
amount of protein adsorbed was determined by
measuring the difference in concentration before
and after adsorption with an spectrophotometer
at 280 nm (Spectronic 601, Milton Roy). After
incubation, the sample was separated from the
supernatant by high-speed centrifugation and the
supernatant filtered using a polyvinyldene difluoride filter (Millipore, pore diameter 0.1 mm)
before measuring the remaining protein concentration. Such a filter has an extremely low affinity
for protein adsorption, so the filtration step does
not interfere with the calculation of the amount of
adsorbed protein.
Protein-covered particles were separated from
the supernatant by high-speed centrifugation, and
redispersed in corresponding buffer (phosphate
for pH 7.2, acetate for pH 5.8). The protein
coverage of the latex-F(ab%)2 complex used in this
work is 3.2 mg/m2.
2.2. Stability
The stability ratio (W) has been used as a
criterion for the stability of the colloidal system:
W=
kr
ks
(1)
in which the rate constant kr describes rapid coagulation, and ks is the rate constant for the slow
coagulation regime. Thus, the inverse of the stability ratio provides a measure of the effectiveness
of collisions leading to coagulation. In this work,
the stability ratio was obtained experimentally
from the rate constant of coagulation of the colloidal particles measured using the low angle light
scattering technique developed by Lips and Willis
[25], where the total scattering intensity for a
dispersion of identical primary particles with a
time varying distribution sizes is [26]:
I(t,q)
=1+ 2knst
Iq (0)
(2)
where Iu (0) is the initial intensity of light scattered at angle q, ns the number of primary particles and k the rate constant. The scattered light
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
128
intensity at low angles increases linearly with time,
and then an absolute coagulation rate can be
obtained from the slope if the number of primary
particles is known.The scattering cell shape is
rectangular, with a 2-mm path length. The cell was
thoroughly cleaned with chromic acid, rinsed with
distilled water and then dried using an infrared lamp.
Equal quantities (1 ml) of salt and protein-carrying
particle solutions were mixed and introduced into
the cell using an automatic mixing device. Dead time
is quite short.
The latex dispersions used for such coagulation
experiments have to be sufficiently dilute to minimize multiple scattering effects, whilst still having
an experimentally convenient coagulation time. For
the protein-carrying particle system used here, a
concentration of 2×1010 particles per ml was
determined to be satisfactory.
2.3. Electrokinetic
The electrophoretic mobility of the coated particles was determined with a Zeta-Sizer IV (Malvern),
by taking the average of five measurements at the
stationary level in a cylindrical cell, and changing
the sample three times. Experimental conditions
were previously worked out to ensure minimal
influence of particle aggregation in the electrophoretic measurements. Standard deviations always fell in the range 2– 4%.
The n-potential was calculated according to the
Smoluchowski equation:
n=
p
v
mrm0 e
(3)
where p is the viscosity of the medium, ve is the
electrophoretic mobility and or and m0 are the relative
permittivity of the medium and the permittivity of
the vacuum, respectively.
2.4. Dynamic light scattering measurements
Photon correlation spectroscopy measurements
were made at a scattering angle of 90° with a
256-channel photon correlator (Malvern 4700c system). Experiments were performed at a particle
concentration of about 109 part/ml and a wavelength
of 488 nm (Argon laser). This technique employs
the laser beam to probe a small volume of the particle
suspension. As they undergo Brownian motion,
interference with scattered light produces a fluctuation in the intensity with time at the detector. This
temporal fluctuation contains information on the
motion of particles and may be analyzed by means
of a correlator that yields, in real time, the autocorrelation function. The nonexponential behaviour of
this function must be analyzed by the method of
cumulants. The first cumulant is used to obtain the
effective translational diffusion coefficient, and the
apparent hydrodynamic radius of the particle obtained from the Stokes–Einstein relation.
3. Theoretical background
Several theoretical studies [16,26–29] have been
made on the electrophoresis of a colloidal particle
covered by an ion-penetrable surface-charged layer
consisting of polymer, in relation to the electrophoresis of biocolloids such as cells or vesicles.
It has been shown [16,30] that when the surfacecharged layer is thicker than 1/s, being s the
Debye –Hückel parameter, the potential within the
surface-charged layer is in practice equal to the
Donnan potential except in the region very near the
boundary between the surface-charged layer and the
surrounding solution.
On the basis of these observations, Ohshima et
al. [15–17] have derived an approximate formula
that directly relates the electrophoretic mobility to
the charge density in the surface region. They assume
an uniform distribution of the fixed charges through
the surface region with a finite thickness, d, and
penetration of the region by electrolyte ions. This
formula has the form:
ve =
mrm0 „DON/u+„(0)/sm zeN
+ 2
pu
p
1/u + 1/sm
+
8mrm0kT
6en0 e − ud/u−e − smd/sm
tan h
pu6e
kT
(1/u)2 − (1/sm)2
(4)
where „DON and „(0) are the Donnan potential and
the potential at the boundary of the surface region
and the medium, respectively, sm is the Debye–
Hückel parameter of the surface region, u is a
parameter whose reciprocal has the dimension
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
of length, and the meaning of the depth of the
flow penetration in the surface region, 6 is the
valence of the electrolyte, z and N are the valence
and the density of charged groups in the region,
respectively, n0 is the zeta potential of bare latex
particle and d is the depth of the protein layer
from the latex surface.
The Donnan potential, the boundary potential,
and the Debye – Hückel parameter of the surface
region are given as:
! " n
! " n
! " n
n
„DON =
„(0) =
kT
zN
ln
+
6e
26n
kT
zN
ln
+
6e
26n
+
26n
1−
zN
sm =s 1+
zN
26n
2
zN
26n
zN
26n
zN
26n
1/2
+1
2
(5)
1/2
+1
2
1/2
+1
(6)
2 1/4
(7)
129
where s is the Debye–Hückel parameter of the
medium and n is the electrolyte concentration.
4. Results and discussion
4.1. Colloidal stability at high ionic strength
The phenomenon of the anomalous stabilization can be observed in Fig. 1 for a protein-covered colloidal system, where the stability ratio is
plotted versus the salt concentration in log scales
for NaCl at two pH values (5.8 and 7.2), and for
CaCl2 at pH 7.2. In the case of NaCl, for example, the stability diagram initially presents the
classical DLVO behaviour, decreasing colloidal
stability with salt concentration until the rapid
aggregation domain where stability is minimal
and independent of salt concentration. However,
when salt is further increased the stability of this
Fig. 1. log W versus log [salt] (M) for a F(ab%)2 –latex particle: , Ca2 + at pH 7.2; , Na+ at pH 5.8; , Na+ at pH 7.2. Open
symbol, DLVO zone; closed symbol, non-DLVO zone.
130
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
Fig. 2. Electrophoretic mobility of F(ab%)2 –latex particles as a function of pH in the presence of two NaCl concentrations: ,
2.1 × 10 − 3 M; , 10 − 2 M.
system increases again, appearing as the so-called
non-classical DLVO region. This anomalous behaviour clearly depends on the pH and counterion
nature. With divalent cations such as Ca2 + the
effect is much more pronounced, whilst the dependence on pH can be attributed to the increase
in net protein negative charge. Both results suggest that the specific adsorption of hydrated
cations on protein molecules may be responsible
for this anomalous stability, creating an energy
barrier opposing aggregation referred to as hydration forces [31].
4.2. Effects of pH and ionic strength on
electrokinetic beha6iour
In order to corroborate this assumption with
new experimental evidences, electrophoretic measurements of protein-covered particles as a function of pH and ionic strength have been carried
out. As can be seen in Fig. 2, for the case of
NaCl, the mobilities depend on pH remarkably,
as expected from the amphoteric nature of
proteins, showing positive values below the i.e.p.
of the protein-covered particle and negative values
above this point. If the salt concentration is increased, the absolute values of the electrophoretic
mobility decrease due to the screening effect of
electrolyte ions on the electrical charge in the
protein-carrying surface region. Nevertheless, Fig.
3 shows that this usual influence of electrolyte
concentration on mobility is not observed when
Ca2 + is employed. For the highest concentration
presented, electrophoretic mobility is positive all
over the pH range, whereas net charge is negative
over the i.e.p. of the protein-covered particle. This
sign reversal in the mobility may be attributable
to the specific adsorption of Ca2 + ions onto the
surrounding protein surface [13,14,32,33]. In the
case of Na+ (see Fig. 2) there exists no sign
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
reversal, but some specific adsorption may also be
suspected from the shift in the point of zero
charge (p.z.c.) that happens when the NaCl concentration is increased [14,32]. Although Ca2 +
ions are clearly much more effective than Na+
ions in reducing the electrokinetic potential of the
colloidal particle, this major effect of Ca2 + is a
general divalent counterion phenomenon, and is
not related only to direct ion-binding on the
protein.
The specific adsorption of cations should depend, in principle, on the electrolyte concentration
in solution [34]. This fact can be observed from
the experimental data shown in Figs. 4 and 5,
where the n-potential of the protein-covered particles is plotted against NaCl and CaCl2, respectively, for two pH values, namely 5.8 and 7.2.
With regard to the NaCl electrolyte, the n-potential at pH 7.2 is higher, in absolute values, than
at pH 5.8. This feature can be readily explained
131
by considering that at pH 5.8, closer to the i.e.p.
range of this protein, the average charge density
must be lower than at pH 7.2. This clearly indicates an important contribution of the charge of
the bound protein molecule to the n-potential,
correlating with the highest colloidal stability in
the DLVO region at pH 7.2 (see Fig. 1).
As shown also in Fig. 4, the ionic strength of
the medium affects the n-potential of the proteincovered particle, but this influence depends on the
pH. At pH 5.8, the behaviour is the expected from
theory [14], with the n-potential decreasing first in
absolute value, and then levelling off, converging
to a non-zero value, with increasing ionic
strength. This trend arises from the screening
effect of electrolyte ions on the protein charge
groups. Nevertheless, in the case of pH 7.2, after
an initial screening effect similar to that occurring
at pH 5.8, instead of reaching a plateau in the
n-potential, it continues decreasing probably due
Fig. 3. Electrophoretic mobility of F(ab%)2 –latex particles as a function of pH in the presence of two CaCl2 concentrations: ,
2 × 10 − 3 M; , 3.8× 10 − 2 M.
132
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
Fig. 4. n potential of F(ab%)2 –latex particles as a function of NaCl concentration at two pHs: , 7.2; , 5.8.
to adsorption of Na+ ions on protein surface. The
effect is so pronounced that the curves at both
pHs tend to converge. This greater tendency for
cation adsorption at pH 7.2 may justify the
highest colloidal stability of the system in the
non-DLVO region shown in Fig. 1.
In the case of Ca2 + ions, it can be observed in
Fig. 5 that at both pH values the plateau in the
n-potential is not reached, but it presents a continuous derive that even passes from negative to
positive values. The only possible explanation for
these positive values of n-potential, when the net
charge of the protein is negative, is the assumption of specific binding of cations to the surface.
As in the case of Na+, this phenomenon is more
extended the higher the pH. By comparison of
these two cations, it can be inferred that Ca2 +
ions, due to their improved specific adsorption,
are more likely to stabilize the protein-covered
particle in the non-DLVO region, as shown in
Fig. 1.
The influence of ionic strength on the electrophoretic mobility of casein-coated polystyrene
particles was investigated by Douglas et al. [35]
who also showed the Ca2 + -binding properties of
casein molecules.
4.3. Analysis of electrokinetic data
In order to calculate the charge distribution in
the protein-carrying particle from the experimentally obtained n-potential values, the membrane
model proposed by Ohshima et al. [15– 17] was
employed. This model revealed that the n-potential loses its meaning for colloidal particles with a
structured surface, since the electrophoretic mobility is insensitive to the precise position of the
slipping plane.
Sakuma et al. [36,37] have indicated that the
n-potential in protein-covered particles is approximately equal to the surface potential relative to
the external bulk solution (Eq. (5)). It is possible
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
to calculate the charge density in the protein layer
per area unit from the measured n-potential (|p =
zeNd). For this purpose, the layer thickness of
protein adsorbed on the particle (d) can be determined by Photon Correlation Spectroscopy (PCS)
from the difference between the mean diameter of
bare latex and protein–latex particles. In this
case, the value of d may be estimated as 591 nm.
In Fig. 6 the charge density (|p) calculated from
the n-potential as a function of the CaCl2 concentration of the medium at two different pH values,
5.8 and 7.2, is shown. Here, the charge density is
a net value, which is given by the sum of positive
and negative charges per area unit of the protein
layer. These charges should proceed only from the
ionic groups on the protein, which depend exclusively on pH. This fact is observed at low electrolyte concentration, but with increasing
concentration the value of |p increases, reaching
positive values due to the adsorption of cations to
the surface.
133
From the data shown in Fig. 6, the molecular
size of F(ab%)2 [38] and the adsorbed amount of
protein per unit area, the average number of
charges per protein molecule may be estimated
(Table 1). The values obtained are within expectations, with a net charge slightly more negative at
pH 7.2 than at pH 5.8 at low electrolyte concentrations, but changing to positive values as the
salt concentration increases, due to the specific
adsorption of cations. Moreover, this effect is
more pronounced at pH 7.2. This method seems
to provide a good way for calculating molecular
effective charges in adsorbed systems.
5. Conclusions
Colloidal particles covered with protein are stable at high ionic strength, when the classical
DLVO theory predicts aggregation. In previous
Fig. 5. n potential of F(ab%)2 –latex particles as a function of CaCl2 concentration at two pHs: , 7.2; , 5.8.
134
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
Fig. 6. Charge density on protein layer (|p) for F(ab%)2 – latex particles as a function of CaCl2 concentration at two pHs: , 7.2; ,
5.8.
publications we have suggested the existence of a
repulsive hydration force originated from the specific adsorption of cations in the protein surface.
This assumption has been further corroborated by
studying the electrokinetic behaviour of proteincovered particles at different pH and ionic
strength conditions, and using two chloride salts,
namely NaCl and CaCl2. Experimental data suggest that both cations are adsorbed specifically at
the protein– aqueous interface, although in the
case of Ca2 + the adsorption may even provoke a
sign reversal in the electrokinetic charge density.
This effect is more pronounced the higher the pH.
The analysis of the electrokinetic data by the
Ohshima et al. model has allowed us to obtain the
average number of charges per protein molecule.
It can be observed that this parameter changes
from a negative to a positive value as the Ca2 +
concentration increases and is then probably adsorbed onto the protein surface.
Acknowledgements
Financial support from ‘Comisión Interministerial de Ciencia y Tecnologı́a’, project no
MAT1999-0662-03-C02 y FEDER project no.
1FD97-1366 are gratefully acknowledged. The authors are very grateful to Joxe Sarobe and Jacqueline Forcada for providing the latex sample.
Table 1
Dependence of net charge per protein molecule (zp) on pH and
CaCl2 concentration
CaCl2 (M)
zp at pH 7.2
zp at pH 5.8
1.16×10−3
5.1×10−3
1×10−2
5×10−2
7.5×10−2
−3.0
−2.7
−2.5
+5.1
+10.5
−1.9
−1.8
−1.7
−1.3
+8.2
J.A. Molina-Bolı́6ar et al. / Colloids and Surfaces B: Biointerfaces 21 (2001) 125–135
References
[1] B.V. Derjaguin, L. Landau, D. Acta Phys. Chim.
U.R.S.S. 14, (1941) 633; JETP1. (USSR) 15 (1945) 633.
[2] E.J.W. Verwey, J.Th.G. Overbeek, Theory of the Stability
of Lyophobic Colloids, Elsevier, Amsterdam, 1948.
[3] J.P. Chapel, J. Colloid Interface Sci. 162 (1994) 517.
[4] G. Vigil, Z. Xu, S. Steinberg, J. Israelachvili, J. Colloid
Interface Sci. 165 (1994) 367.
[5] R.M. Pashley, J. Israelachvili, J. Colloid Interface Sci. 97
(1984) 446.
[6] R.M. Pashley, J. Colloid Interfac Sci. 80 (1981) 153.
[7] D.M. LeNeveu, R.P. Rand, V.A. Parsegian, Nature 259
(1976) 601.
[8] E.A. Vogler, Adv. Colloid Interface Sci. 74 (1998) 69.
[9] J.N. Israelachvili, Chem. Scr. 25 (1985) 7.
[10] H.K. Christenson, J. Dispersion Sci. Technol. 9 (1988)
171.
[11] R.M. Pashley, J. Colloid Interface Sci. 83 (1981) 531.
[12] R.M. Pashley, Adv. Colloid Interface Sci. 16 (1982) 57.
[13] H.J. Modi, D.W. Fuerstenau, J. Phys. Chem. 61 (1957)
640.
[14] R.J. Hunter, Zeta Potential in Colloid Science. Principles
and Applications, Academic Press, London, 1981.
[15] H. Ohshima, S. Ohki, Biophys. J. 47 (1985) 673.
[16] H. Ohshima, T. Kondo, Colloid Polym. Sci. 264 (1986)
1080.
[17] H. Ohshima, M. Nakamura, T. Kondo, Colloid Polym.
Sci. 270 (1992) 873.
[18] H. Tamai, A. Fujii, T. Suzawa, J. Colloid Interface Sci.
118 (1987) 176.
[19] T.A.J. Payens, J. Dairy Res. 46 (1979) 291.
[20] Z. Saito, Neth. Milk Dairy J. 27 (1973) 143.
.
135
[21] J.A. Molina-Bolı́var, F. Galisteo-González, R. HidalgoÁlvarez, Phys. Rev. E 55 (1997) 4522.
[22] J.A. Molina-Bolı́var, F. Galisteo-González, R. HidalgoÁlvarez, J. Chem. Phys. 110 (1999) 5410.
[23] J. Sarobe, J. Forcada, Colloid Polym. Sci. 274 (1996) 8.
[24] J. Sarobe, J.A. Molina-Bolı́var, J. Forcada, F. Galisteo,
R. Hidalgo-Álvarez, Macromolecules 31 (13) (1998) 4282.
[25] A. Lips, E.J. Willis, J. Chem. Soc. Faraday I 67 (1971)
2979.
[26] A. Lips, E.J. Willis, J. Chem. Soc. Faraday I 69 (1973)
1226.
[27] R.W. Wunderlich, J. Colloid Interface Sci. 88 (1982) 385.
[28] S. Levine, M. Levine, K.A. Sharp, D.E. Brooks, Biophys.
J. 42 (1983) 127.
[29] K.A. Sharp, D.E. Brooks, Biophys. J. 47 (1985) 563.
[30] H. Ohshima, T. Kondo, J. Colloid Interface Sci. 116
(1987) 305.
[31] J.A. Molina-Bolı́var, F. Galisteo-González, R. HidalgoÁlvarez, Colloids Surf. B: Biointerfaces 14 (1999) 3.
[32] T.H.F. Tadros, Solid/Liquid Dispersions, Academic
Press, New York, 1987.
[33] R.K. Mishra, S. Shander, D.W. Fuerstenau, Colloids
Surfaces 1 (1980) 105.
[34] K.E. van Holde, Physical Biochemistry, 2nd edn, Prentice-Hall, Englewood Cliffs, New Jersey, 1985.
[35] G.D. Douglas, D. Eric, H.W. Richard, J. Colloid Interface Sci. 108 (1985) 174.
[36] H Sukuma, T. Ohshima, J. Kondo, J. Colloid Interface
Sci. 133 (1989) 252.
[37] H. Sukuma, T. Ohshima, J. Kondo, J. Colloid Interface
Sci. 135 (1990) 455.
[38] J. Buijs, J.W.T. Lichtenbelt, W. Norde, J. Lyklema, Colloids Surf. B: Biointerfaces 5 (1995) 11.