Adv Polym Sci (2014) 255: 57–96

DOI: 10.1007/12_2012_171
# Springer-Verlag Berlin Heidelberg 2012
Published online: 8 August 2013

Aggregation of Charged Colloidal Particles

Nikolai I. Lebovka

Abstract This chapter reviews the recent progress in aggregation of colloidal

particles with long-range interactions, including simple colloids and polyelec-
trolytes. The relevant interactions between colloidal particles, including Born
repulsion, van der Waals, electrostatic, structural solvation, hydrophobic hydrody-
namic interactions and attraction between like-charge colloids, charge nonunifor-
mity, and adsorbed polymer, are analyzed. The main types of computer models
used for simulation of cluster morphology and aggregation kinetics of the different
interacting species (similarly and oppositely charged particles and polyelectrolytes)
are reviewed. The main scaling laws for different aggregating kernels that describe
diffusion-limited, reaction-limited, gelling, and retarded aggregations are also
presented and analyzed.

Keywords Aggregation
Charged particles
Colloids
Kinetics
Morphology

Polyelectrolyte complex (PEC)

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2 Interactions Between Colloidal Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.1 van der Waals Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.2 Electrostatic Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.3 Born Repulsion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.4 Structural Solvation Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.5 Hydrophobic Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.6 Effect of Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
2.7 Hydrodynamic Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

N.I. Lebovka (*)

Institute of Biocolloidal Chemistry Named After F.D. Ovcharenko, NAS of Ukraine,
42, blvr. Vernadskogo, 03142 Kiev, Ukraine
e-mail: lebovka@gmail.com
58 N.I. Lebovka

2.8 Interaction Between Colloidal Aggregates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

3 Simulation of Cluster Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.1 Main Types of Computer Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.2 Similarly Charged Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.3 Oppositely Charged Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.4 Effect of Dipolar Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4 Kinetics of Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.1 Aggregation as a Second-Order Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.2 Population Balance Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.3 Popular Kernels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4.4 Classification of Kernels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.5 Dynamic Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

1 Introduction

Polymeric nanoparticles are colloidal particles consisting of macromolecular

compounds. Among these materials the polyelectrolyte complex (PEC) nanoparticles
are of particular interest. These particles with integrated drugs, proteins, vaccines, or
diagnostic agents can be used as carriers in different pharmaceutical and biomedical
applications [1]. The benefits of their use include controlled drug release and limited
toxicity. Moreover, water-soluble and biodegradable PEC nanoparticles may be used
as drug delivery systems in humans [2, 3].
PEC nanoparticles are prepared by mixing of two oppositely charged polyanions
and polycations. Formation of PEC nanoparticles is controlled by the structure of
the polycation/polyanion components, concentration, mixing order, mixing ratio,
ionic strength, pH, temperature, and other factors [4]. Key factors determining
successful applications of PEC nanoparticles are related to good reproducibility
of the formation process, monomodality, and the possibility of obtaining particles
with determined and graded sizes.
The process of PEC nanoparticle formation usually includes the initial diffusion
stage of mutual entanglement between polymers and formation of primary particles
and the further stage of their aggregation and rearrangement of the already formed
aggregates [2, 4]. The macroscopically homogeneous systems containing very
small primary PEC nanoparticles (approximately 5–20 nm) were developed using
special preparation techniques [5]. The process of aggregation of the primary PEC
nanoparticles may result in formation of larger nanoparticles. As a result, a turbid
colloidal, or two-phase system of supernatant liquid and precipitated PEC
nanoparticles is formed [6]. Using of consecutive centrifugation and separation
steps allows significant decrease in the polydispersity of the colloidal system,
elimination of the primary PEC nanoparticles, and, finally, prevailing of the sec-
ondary nanoparticles with radii of about 100–200 nm in dispersion [4]. The
secondary PEC nanoparticles have many particular advantages for applications in
Aggregation of Charged Colloidal Particles 59

biomedicine [7]. Formation of the secondary PEC aggregates and their final size,
stability, and polydispersity are controlled by Coulomb interactions between
charged primary PEC nanoparticles. Computer simulation techniques can be par-
ticularly useful for study of such processes.
This chapter is organized as follows: Section 2 reviews the main equations
describing important interactions between dissimilar colloidal particles. The
contributions of Born repulsion, van der Waals, electrostatic, structural solvation,
and hydrophobic hydrodynamic interactions, as well as the effects of attraction
between like-charge colloids, charge nonuniformity, and adsorbed polymer are
analyzed. Section 3 presents the main types of computer models used for simulation
of cluster morphology and results for different interacting species (similarly and
oppositely charged particles and polyelectrolytes) and different models (DLA-like,
Eden-like). Section 4 considers the classical Smoluchowski model, the theoretical
approach of population balance equations (PBE), classification of kernels, and the
main types of scaling in aggregation behavior. The recent data on aggregations
kinetics of charged PEC particles are also presented.

2 Interactions Between Colloidal Particles

The Derjaguin–Landau–Verwey–Overbeek (DLVO) theory is commonly used to

describe interactions of charged surfaces across liquids [8, 9]. The DLVO theory
models the interparticle interactions by superposing van der Waals attractions and
electrostatic double layer repulsion forces. The direct force measurements have
confirmed this theory down to surface separations of few nanometers [10].
The important characteristic of electrolytes are the Bjerrum length, lB, and
Debye length, lD, defined as:

e2
lB ¼ ; (1)
4pee0 kB T
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ee0 kB T
lD ¼ : (2)
2re2

Here, e and e0 are the dielectric constants of the medium and vacuum, respec-
tively, kB is the Boltzmann constant, T is the absolute temperature, rpisffiffiffiffiffiffiffiffiffiffiffiffiffi
the number 1
density of the added salt, e is the charge of an electron, and lD ¼ 8plB r .
The value of lB is defined as the distance at which the interaction between two
elementary charges equals kBT. At room temperature in water, T ¼ 298 K and
lB ¼ 56/e  0.7 nm. At this condition, the dissociation energy of the ionic pair
with a distance R ffi 0.7 nm between the opposite charges ud ¼ kBTlB/R is of order
60 N.I. Lebovka

10 3

j =0 s , mC/m
1
2

lD = 0.308 / C nm
j =0.2
lD , nm

10 2

j =0.6

10 1
10-5 10-4 10-3 10-2
C, M

Fig. 1 Debye screening length, lD, versus the molar concentration of salt, C, at different volume
fractions ’ and surface charge densities s of colloidal particles. The radius of particles, r, was
85 nm

of kBT and the ionic pair is unstable. However, such an ionic pair may be stable in
less polar solvent. pffiffiffiffi
The Debye length, lD, at room temperature may be estimated as lD ¼ 0:308= C
nm, where C is the molar concentration of salt (1 M ¼ 103 mol m3). Note that when
the volume fraction of colloidal particles ’ is not low, the Debye length lD becomes
dependent on ’, colloidal particle radius, r, and colloidal particle surface charge
density, s [11]:
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1’
lD ð’Þ ¼ lD ð0Þ : (3)
1 þ 1:5s’=ðerrÞ

The lD(’) dependence becomes important when the volume fraction of

particles, ’, is large, or the concentration of salt C is very low. Moreover, the
value of lD passes through a maximum with increase in C (Fig. 1).

2.1 van der Waals Interactions

According to the Derjaguin approximation [12], the energy of van der Waals
interactions between spherical particles of radii ri and rj is:
Aggregation of Charged Colloidal Particles 61

Table 1 Examples of Medium Hamaker constant, A (J/1020)

Hamaker constants for
different materials [13] Polystyrene 7.8
Poly(methyl methacrylate) 7.1
Silica 6.5
Quartz 11.0–18.6
Water 3.3–6.4
Pentane 3.8
Ethanol 4.2
Cyclohexane 5.2

"  2 #!
A 2ri rj 2ri rj R2  r i þ r j
uw ðhÞ ¼    þ þ ln  2
6 R2  ri þ rj 2 R2  ðri  rj Þ2 R2  r i  r j
  2 
A 2 2 S þ 4Sx
¼ þ þ ln ; (4)
6 S2 þ 4Sx S2 þ 4Sx þ 4 S2 þ 4Sx þ 4

where h is the surface-to-surface separation distance, R ¼ h + ri + rj is the dis-

tance between the centers of particles, A is the Hamaker constant,
pffiffiffiffiffiffiffi  S ¼ h=

rg ,
r
g ¼ ri rj is the mean geometrical radius, and x ¼ r
=
rg ; r
¼ ri þ rj =2 is the
mean radius.
At small separation distance, i.e., at h  r
g , the energy of van der Waals
interactions is inversely proportional to the surface-to-surface separation distance:

A r
uw ðhÞ   ; (5)
6 h
 
where r  ¼ 2ri rj = ri þ rj .
The value of A corresponds to the effective Hamaker constant for the interaction
between particles i and j in the dispersion medium. The values of the Hamaker
constant are presented for different materials in Table 1.
The effective value of A123 is related to Hamaker constants of individual
materials A11, A22, and A33 and can be estimated as [14]:
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffi

A123 ¼ A11  A33 A22  A33 : (6)

For similar particles, i.e., at A11  A22, the value of A33 has a positive sign that
corresponds to the attractive van der Waals interactions. However, when the value
of A33 is intermediate between those of A11 and A22, it has a negative sign that
corresponds to the repulsive van der Waals interactions.
In fact, the Hamaker constant A131 is not constant but depends on the concentra-
tion of electrolyte and on retardation [15]:

A123 ¼ A0 ð1  2h=lD Þ expð2h=lD Þ þ A1 FR ðhÞ; (7)

62 N.I. Lebovka

where A0 and A1 are the zero-frequency and high-frequency contributions, respec-

tively, and FR is the retardation function [16]. The zero-frequency contribution
represents the net effect of orientation and induction interactions, and the high
frequency term arises from London dispersion interactions. Note that the retardation
effect can be ignored for a small distance between the particle surfaces (<5–10 nm).
In recent years, many investigators, in order to overcome difficulties with
simulations, have used the simplified forms of attraction potentials, e.g., the
Morse potential [17]:

uM ðhÞ ¼ u0M expðh=lÞðexpðh=lÞ  2Þ; (8)

where l is the range parameter and u0M is the potential well depth.

2.2 Electrostatic Interactions

The most popular potential that captures the essential behavior of electrostatic
interactions between two equal spherical colloids of charge Ze is the Yukawa
potential:
 
Rr r
uY ðRÞ ¼ uaY exp  ; (9)
lD R

where the sign is positive for equally charged particles and negative for oppositely
charged particles, R ¼ h + 2r is the distance between the centers of the particles,
and:

ðZeÞ2 r expðr=lD Þ
u0Y ¼ : (10)
4pee0

2.2.1 DLVO Approximation

The same form of potential follows from Derjaguin, Landau [8], Verwey and
Overbeek [9] (DLVO) theory that invokes the Debye–H€uckel approximation to
linearize the Poisson–Boltzmann equation:

ðZeÞ2 expðh=lD Þ
uY ðhÞ ¼ 
4pee0 R ð1 þ r =lD Þ2
ðZeÞ2 g2
¼ expðR=lD Þ; (11)
4pee0 R
Aggregation of Charged Colloidal Particles 63

where g ¼ expðrlD Þ=ð1 þ rlD Þ is the geometric factor that reflects the absence of
screening inside the particle exclusion region.
Equation 11 is only valid in the weak Coulomb coupling regime. For highly
charged colloidal particles, strong electrostatic coupling between colloidal particles
and ions results in additional screening of Ze. In a general case, the functional
dependence of effective charge Z*e on real charge Ze may be rather complex [18]
and can be determined by geometry of particles, distribution of charges on their
surface, and concentration of ions.
The extension of the Derjaguin approximation for electrostatic interaction
energy between two dissimilarly charged spheres of radii ri and rj was introduced
by Hogg, Healy, and Feurstenau (HHF) [19]. Two expressions that are valid under
both constant charge, s, and constant potential, c, conditions were proposed:
8
9
<  h=lD
 z2i þ z2j
=
1 þ e
ue ðhÞc ¼ u0e 2 ln þ ln 1  e2h=lD (12)
: 1  eh=lD zi zj ;

8
9
<  h=lD
 z2i þ z2j
=
1 þ e
ue ðhÞc ¼ u0e þ2 ln þ ln 1  e2h=lD ; (13)
: 1  eh=lD zi zj ;

 
where u0e ¼ pee0 r  zi zj ; r ¼ 2ri rj ri þ rj ; z is the zeta-potential, and h is the
surface-to-surface separation distance.
Approximations for electrostatic repulsion in Eqs. 12, 13 are valid for h  r*,
relatively small values of z, z/z < kTB/e ffi 25 mV, and r*/lD > 10. Corrections to
the fourth and sixth powers of surface potentials in the HHF formulas have been
made in [20]. Other more general formulas can be found in the literature [21].

2.2.2 Attraction Between Like-Charge Colloids

Many experimental works have shown that unusual long-range attractive

interactions, which cannot be explained by the DLVO theory [22–25], may exist
for similarly and highly charged colloidal particles. It is interesting that these
interactions were observed only in the presence of charged walls.
The theoretical explanations of this effect are rather controversial [26–28].
A highly charged colloidal particle of charge Ze captures N oppositely charged
counterions of charge ze, which form a very thin shell around the charged colloidal
particle surface, resulting in a very strong screening. Under certain conditions, the
counterions may totally neutralize or even overcharge the colloidal particle [29].
The charge neutrality is fulfilled when the colloidal particle captures N ¼ Nn ¼ Z/z
counterions. In the ground state (i.e. at T ¼ 0), the spherical colloidal particle can
capture even more positionally correlated counterions and the effective charge of
64 N.I. Lebovka

colloidal particle Z*e gains the same sign as that of counterions [26, 30].
Estimations have shown that at Z z [30]:
pffiffiffiffiffiffiffiffi
N  Nn þ 0:8265 Z=z; (14)

and the maximally possible value of Z* is:

pffiffiffiffiffi
Z ¼ ðN  Nn Þz / Zz: (15)

The estimation based on Debye, Huckel, and Bjerrum theory predicts that in the
strong Coulomb coupling regime the counterion correlations can give rise to
attractions at short separations between particles [31, 32]. The analytical
calculations predict that the effective long-range interactions between like-charge
colloids immersed in a confined electrolyte are repulsive [33]. However, the
possibility of long-range attractions arising from charge fluctuations was theoreti-
cally predicted [34]. Recent extensive theoretical and computer simulations have
shown the important role of counterion correlations [32, 35] and charge fluctuations
of either colloidal particles [36], or condensed counterions [37].
The experimentally observed attractive interactions between like-charged col-
loidal spheres [25] were explained by a nonequilibrium hydrodynamic effect [38].
A mechanism was also proposed based on formation of the depletion zone of
counterions between nearly touching like-charged colloidal particles [39].
According to Manning’s theory [40], the rodlike polyelectrolyte can capture N
oppositely charged counterions:

N ¼ ð1  1zxÞZ=z; (16)

where z ¼ lB/b is the coupling strength.

The theory predicts that attraction is possible only in the presence of multivalent
counterions and if the number of counterions condensed on polyions exceeds [41]:

N>Nn =2 ¼ Z=2z 1: (17)

So, attraction is possible only for polyelectrolytes with a high coupling strength,
z > 2/z.
Taking into account the many-body interactions between highly charged
colloids and counterions, Tokuyama proposed the following equation for effective
attractive potential [42, 43]:
n o
uT ðRÞ ¼ u0T ðZ=zÞ3 Fðlm =RÞ  Fðl=RÞ ; (18)

where u0T ¼ 0:5kB T ðZzlB =lD Þ2 ; R ¼ h þ 2r is the distance between the centers of
particles, Ze and ze correspond
pffiffiffiffiffiffiffiffi to the bare charges of colloid particles
Ð 1 and counterions,
respectively, lm ¼ lB z=Z, and FðxÞ ¼ ðx  1Þx expðxÞ  x y1 exp ydy:
Aggregation of Charged Colloidal Particles 65

Note that this type of potential was used in Brownian-dynamics simulations of

charged colloidal suspensions, where gas, liquid-droplet, and crystal-droplet phases
were identified [44].

2.2.3 Effect of Charge Nonuniformity

The charge nonuniformity on the surface of colloidal particles may also signifi-
cantly contribute to the electrostatic interactions. It can arise from selective ion
adsorption on the surface of colloidal particles and distribution of z potential [45,
46]. The surface charge nonuniformity can lead to attractive electrostatic and
hydrophobic interactions between particles and cause suspension instability
[47–49]. An extension of the HHF model for the randomly charged surfaces gives
the following Velegol–Thwar potential [46]:
8
9
<  h=lD
 z2i þ z2j þ s2i þ s2j
=
1 þ e
uV ðhÞc ¼ u0e 2 ln þ ln 1  e2h=lD ; (19)
: 1  eh=lD zi zj ;

where u0e ¼ peeo r  zi zj ; and s is the variance of the surface z potential.

2.3 Born Repulsion

Short-ranged repulsion between the cores of colloidal particles (assuming that the
particles cannot interpenetrate) may be approximated by the Born potential [50]:
!
u0B R2  14R þ 54 2R2 þ 60 R2 þ 14  R þ 54
uB ðhÞ ¼   þ þ ; (20)
R ð R  2Þ 7 R7 ðR þ 2Þ7

where R* ¼ R/2r, R is the distance between the centers of the particles, and value of
u0B determines the primary minimum of potential.

2.4 Structural Solvation Interactions

Direct measurement using atomic force microscopy has shown the presence of
short-range repulsion or attraction solvation forces between colloidal particles.
These forces reflect the finite size of the solvent molecules and are important
only at certain values of pH and electrolyte concentrations. The distance
66 N.I. Lebovka

dependence of the structural solvation potential may be approximated as was first

described by Marcelja and Radic [13, 51]:

us ðhÞ ¼ u0s expðh=ls Þ; (21)

where u0s ¼ rps l2s ; ps is the structural pressure constant, and ls (approximately
0.2–2 nm) is the solvation decay length. Usually, the monotonic exponential decay
superimposes on an oscillatory profile.

2.5 Hydrophobic Interactions

Hydrophobic interactions are rather typical for organic colloidal particles (e.g., the
particles of latex) dissolved in water. They produce a net attractive contribution. The
hydrophobic potential for two equal spherical colloidal particles is expressed as [13]:

uh ðhÞ ¼ u0h expðh=lh Þ; (22)

where u0h ¼ 2prglh ; gð

10  50 mJ=m2 Þ and lh ð
1  2 nmÞ are empirical
parameters.
This contribution is rather short-ranged and reduces only the height of the DLVO
potential barrier without modifying the depth of the secondary minimum [52].

2.6 Effect of Polymers

2.6.1 Adsorbing Polymer

Polymer adsorption may result in bridging attraction or steric repulsion between

two particles. The scaling theory of polymer adsorption that accounts for the
interaction between the polymer and the surface and for variation of polymer
concentration near the surface [53, 54] was used for derivation of the interaction
potential between two spherical colloidal particles coated by polymer [55–57]:
 
pffiffiffi 5=4  1=4  
1=4
uap ðhÞ ¼ u0p  lnð2d=hÞ þ 2G lp =h  lp =2d ; (23)

  9=4
where u0p rnkB T= a3m 16’p l2p G; n is the numerical constant, am is the effective
monomer size, ’p is the polymer volume fraction at a single surface, G is the degree
of surface saturation by the adsorbed polymer (fractional polymer surface cover-
age), d is the thickness of the adsorbed polymer layer, and lp is the scaling length,
which is a measure of the segment surface interactions.
Aggregation of Charged Colloidal Particles 67

The first term within the square brackets in Eq. 23 corresponds to the short-range
bridging attraction and the second term is for steric (excluded volume) repulsion.

2.6.2 Non-adsorbing Polymer

Depletion of attraction caused by the presence of a non-adsorbing polymer may be

calculated as [10]:
 
unp ðhÞ ¼ u0p 1  3t=4 þ t3 =16 ; (24)

where u0p ¼ 4pp0 ðr þ dÞ3 =3; p0 is the osmotic pressure in the bulk, d is the
exclusion thickness, and t ¼ (2r + h)/(a + d).

2.7 Hydrodynamic Interactions

Hydrodynamic interaction mediated by the solvent originates from the movement

of particles and occurs when particles are close to each other. It is determined by the
viscous drag dependence on interparticle distance [16, 58] and results in a decrease
in the particle diffusion coefficient by the factor b, i.e., D ¼ D0b [59, 60]. Here, D0
is the diffusion coefficient of a particle in the infinitely diluted solution. For motion
along the line of the particle centers, the exact expression for b was obtained by
Brenner [61] and was approximated by the following rational function [62]:

6ðh=r Þ2 þ 13ðh=r Þ þ 1
bðh=r Þ ’ : (25)
6ðh=r Þ2 þ 4ðh=r Þ

Formally, hydrodynamic repulsion energy uhd may be expressed as the following

correction to the total energy:

uhd ðhÞ ¼ kB T ln bðh=r Þ: (26)

The effect of hydrodynamic interactions on aggregation of colloidal particles

may be rather essential and simulation results show that they constrain the growth
of aggregates [63]. Computational simulation predicts that many-body hydrody-
namic interactions between colloidal particle are able to reduce the solid fraction
required for percolation or gelation [64, 65]. The merging of clusters into
condensed aggregate was observed at particle volume fracture ’ as low as
0.06–0.12 [64].
In concentrated suspensions of charged colloidal particles, the strong Coulomb
interaction prevents the particles from moving freely and leads to an effective
screening of hydrodynamic interaction [66]. Such a screening may be particularly
68 N.I. Lebovka

important in polyelectrolyte solutions with strong coupling of electrostatic and

hydrodynamic interactions between the polymer chains [67, 68].

2.8 Interaction Between Colloidal Aggregates

Until recently, there was no clear conception about interaction between two
aggregates consisting of many individual particles. Practically all efforts were
devoted to studies of interactions between single particles with an idealized
shape, e.g., spheres, cylinders, and ellipsoids.
The general schema of calculation of the van der Waals interaction energy
between irregularly shaped molecular aggregates was developed in 1998 [69].
More recently, the calculations of van der Waals and double layer interactions
between colloidal aggregates were performed [70, 71]. The direct numerical
calculations of the van der Waals interaction between fractal aggregates of colloidal
particles (based on a pairwise summation of interaction energies between all
particles) have shown that they can be fairly well approximated by the energy of
interaction between the closest pair of primary particles [70]. However, it was noted
that the surface distance between two aggregates is governed by the morphology of
clusters and there may be an apparent impact on the van der Waals interaction.
Similar estimations were done for the double layer interactions between fractal or
hexagonal closed-packed aggregates with considerable overlapping of double layer
inside the aggregates and between two interacting aggregates [71]. It was shown
that for the relatively thin double layer (lD  0.2r for closed-packed aggregates
and lD  r for aggregates with a small fractal dimension), the interaction of
aggregates is close to the interaction of the nearest pair of the primary particles.
However, in thick double layers (lD > r), the overlapping of the double layer
inside the aggregate was noticeable and formation of a spheroidal double layer
around the aggregate was observed.

3 Simulation of Cluster Morphology

A number of theoretical models and computer simulation approaches were devel-

oped for description of the cluster morphology [72], reaction kinetics, and time
dependence of the cluster-size distributions [73].

3.1 Main Types of Computer Models

The cluster morphology may depend on details of colloidal particle interactions,

mechanism of particle attachment to the cluster, and dimensionality of the problem.
The existing models for cluster morphology simulation account for the trajectory of
Aggregation of Charged Colloidal Particles 69

the moving particles (diffusion limited aggregation, DLA; reaction limited aggre-
gation, RLA; ballistic aggregation, BA; and Eden-like aggregation) [74].
In the DLA model, the individual particles or clusters stochastically diffuse via
Brownian trajectories towards one another and every collision between them results
in formation of a larger cluster [75]. The clusters formed display a fractal morphol-
ogy that depends upon the space dimension in the DLA model. In the RLA model,
the probability of attachment is small and only a small fraction of collisions
between clusters leads to formation of larger clusters. Traditionally, DLA and
RLA are called rapid and slow aggregations, respectively.
A similar mechanism of aggregation is used in the BA model; however, here the
trajectories of moving particles are assumed to be linear [76–78]. In the Eden-like
aggregation model, new single particles attach to the cluster at its perimeter [79].
The type of diffusion motion strongly affects the cluster morphology, e.g.,
Brownian motion generates typical fractal DLA clusters, linear trajectories lead
to compact structures, and Levy-flight trajectories allow continuous changes of the
cluster morphology [80].
The above models are irreversible. Finally, it is expected that clusters grow and
merge until formation of a single connected cluster. However, the experimental
studies show that some systems can display the presence of reversible aggregation.
In a rapid aggregation process, the clusters with loose structure arise initially, and
after a certain time they can restructure to more compact clusters with higher fractal
dimensionality [81, 82]. The process of thermal restructuring of fractal polymer
clusters dispersed in water was experimentally observed using small-angle light
scattering [83].
In the absence of aggregation, restructuring resulted in an increase in the fractal
dimension. A simple model of the restructuring kinetics, based on coalescence
theory of liquid droplets, was developed [83]. Reversible-growth models were also
built. These models allow unbinding of particles and imply that later on, during the
restructuring, the ramified clusters become compact [84, 85]. The details of cluster
morphology strongly depend on the interparticle interaction, the presence of
restructuring, and external fields [86]. Impact of interactions on the intrastructure
of colloidal clusters was recently studied by Brownian dynamics simulations [17].
It was shown that an increase in density of suspension results in stronger intercluster
interactions that affect the intrastructure of clusters. However, perturbation of the
intrastructure of clusters was shown to be low in diluted suspension.

3.2 Similarly Charged Particles

For colloidal suspensions, the influence of the combination of short-range

attractions and long-range repulsions on clustering phenomena [87] is extremely
interesting. It is expected for charged particles that cluster accumulates net charge
and may reject association of additional particles at a certain critical size.
70 N.I. Lebovka

The long-range repulsion may result in limitation of the cluster size [87, 88].
Coulomb repulsion tries to break charged spherical clusters or introduce an ellip-
soidal deformation [89, 90]. In 1935, Weizsacker studied the stability of the atomic
nucleus by analyzing the potential energy of the charged spherical cluster ut [91].
In the case when the cluster is filled by N primary particles with radius r and charge
Ze, the value of ut includes the volume uv, surface us, and electrostatic ue, terms:

ut ðNÞ ¼ uv þ us þ ue ¼ u0 N þ aN 2=3 þ bN 5=3 ; (27)

where u0 is the binding energy per particle, a ¼ 4pgr 2 f2=3 is the surface tension
parameter, b ¼ 3kB TlB ’1=3 =5r is the electrostatic repulsion parameter, g is the
surface tension at the interface between the aggregate and solvent, ’ is the volume
fraction of particles in the aggregate, and lB ¼ ðZeÞ2 =ð4pee0 kB T Þ is the Bjerrum
length for primary particles in the aggregate.
The radius of a spherical cluster may be calculated as:

r c ¼ r ðN=’Þ1=3 : (28)

Figure 2 shows examples of dimensionless total potential energy of a spheri-

cal cluster, ut/u0, versus the number of primary particles, N, at a fixed value of
a (a/u0 ¼ 2) and different values of b/u0. The clusters are definitely stable at
ut < 0. At relatively small charge Z of the primary particles, the curve ut goes
through the maximum (see, inset in Fig. 2) at:
 3  
2a 20a3 b
N1max ¼ 1þ ; (29)
3u0 9u30

and bigger clusters are stable (i.e. ut < 0) in the certain range of N values.
A noticeable change in the shape of the ut curve is observed for a charged cluster
(Fig. 2), e.g., at a ¼ 0, i.e. in absence of surface tension, the ut function curve goes
through the minimum and has two zeros at N ¼ 0 and at:
 3=2
3=2 5 r u0
N¼ N2max ¼ ð1=bÞ ¼ 1=3
: (30)
3’ lB kB T

Note that the critical size N2max corresponds to the physical situation when the
next primary particle cannot attach to the compact spherical cluster. In principle,
the cluster grown may continue through formation of noncompact branching
morphology structures with smaller value of ’. Thus, the critical size N2max
corresponds to development of morphological instability.
Stability analysis with respect to ellipsoidal deformation has shown that even
bigger clusters (initially formed and then becoming charged) may lose stability and
dissociate on smaller clusters (Fig. 2) at certain critical size:
Aggregation of Charged Colloidal Particles 71

2
50

Dissociation on

u/uo
40 1
smaller clusters max
N1

30
0

b/uo=5.10-2 N max 0 10 20
3
20 N
u/uo

3.8.10-2
10 unstable
N max
2

0
2.5.10-2
stable
-10
1.86.10-2

0 50 100 150 200 250

N

Fig. 2 Dimensionless total potential energy of a spherical cluster, ut/u0, versus the number of
primary particles, N, at a/u0 ¼ 2 and at different values of b/u0. Inset shows the enlarged part of
this figure at small values of ut/u0 and N

10 r 4pgr 2
N ¼ N3max  2a=b ¼ (31)
3’ lB kB T

that corresponds to development of dissociative instability.

In both Eqs. 30 and 31, the larger cluster corresponds to the larger ratio of r/lB
and smaller value of ’. The stable growth regime is expected for zero value of
lB (b ¼ 0), which corresponds to formation of an infinite cluster or gel-like phase.
However, the completely unstable growth regime is expected when lB exceeds some
critical value. In this case, the value of ut is positive at any value of N > 0 (see, e.g.,
the case of b/u0 ¼ 3.8102 in Fig. 2). Energetically, the critical sizes N1max ; N2max
and N3max are controlled by ratios us/u0, ue/ue and us/u0, respectively (Fig. 3).
Figure 4 compares N2max and N3max versus ratio 2a/b behavior for fixed values of a
(surface tension parameter). The slightly charged clusters (small values of lD) can
grow to larger size and the values of N3max and N2max coincide. In this unstable grown
regime, the morphological and dissociative instabilities develop simultaneously.
For highly charged clusters (large values of lD), the value of N3max may noticeably
exceed the value of N3max at certain critical lD.
In principle, these general considerations are consistent with recent experimental
and computer simulation results. The finite-size clusters were experimentally
realized in suspensions of colloidal particles with interactions induced by a non-
adsorbing polymer [92, 93]. Cluster morphology was dependent on the range of the
72 N.I. Lebovka

Cluster Cluster with Dissociation of

nucleus external branches cluster

3 3/ 2
⎛ us ⎞ ⎛u ⎞ max ⎛u ⎞
N1
max
∝ ⎜⎜ ⎟⎟ N2
max
∝ ⎜⎜ v ⎟⎟ N3 ∝ ⎜⎜ s ⎟⎟
⎝ uv ⎠ ⎝ ue ⎠ ⎝ ue ⎠

Fig. 3 The critical sizes of charged cluster N1max ; N2max and N3max . The values of N1max ; N2max and
N3max, correspond to the formation of initial nucleus, growth of external branches, and dissociation
of charged droplets, respectively

a/ u o= 1 2 5

10 3
N max , N max
3

10 2
2

N max
3
10 1 N max
2
λb

101 102 103

2a/b

Fig. 4 Maximum cluster sizes N2max and N3max versus ratio 2a/b at fixed values of a/u0 ¼ 1, 2, and 5

interaction and was becoming more compact for longer range interactions. The
origin of such behavior is not completely clear so far.
The large charged clusters may be unstable with respect to formation of elon-
gated or linear structures. The theory predicts formation of near-linear chains with
Aggregation of Charged Colloidal Particles 73

small fractal dimension, df ~ 1, in polyelectrolytes in the limit of high electrostatic

interactions (lB ~ 0.1 nm) [94]. Formation of linear structures was observed also in
Monte Carlo simulations of a short polyelectrolyte with various numbers of beads,
N ¼ 8–80 [95]. The value of df was noticeably increasing as Bjerrum length lB
decreased.
The effect of the range of interactions on the structural and kinetic properties of a
computer-simulated two-dimensional aggregating colloidal system was studied
using the repulsive Yukawa potential [96] (Eq. 9). The increase in lD or u0 provokes
arrangement of aggregates into linear structures. The repulsive interactions also
have a strong influence on the kinetic behavior of the coagulation process. The
structure of small clusters composed of up to 80 particles interacting simulta-
neously via attractive and repulsive forces was recently simulated [97]. A short-
range attraction was simulated by generalization of the Lennard–Jones potential
and a long-range repulsion was simulated by the screened electrostatic Yukawa
potential. The competition between attraction and repulsion resulted in formation of
stable clusters, and the ground-state clusters were preferentially growing almost in
one dimension. The extensive numerical simulations were done for suspended
charged colloidal particles at a screening length comparable to the particle radius
[98]. It was shown that at low temperature, particles organize into quasi one-
dimensional aggregates connecting via branching mechanism into a percolating
gel structure. Note that experimental data indicate a more elongated and open
morphology of the fractal-like aerosol agglomerates at larger charge [99].

3.2.1 DLA-Like Model

The impact of long-range interactions on cluster morphology has been intensively

studied for different variants of the DLA model [100–103]. It was shown using a
quasi-deterministic two-dimensional (2D) particle–cluster growth model with
attractive cluster–particle forces proportional to Ra that the growing cluster was
dendritic. Moreover, strong reduction of the fractal dimension, df, with increase in
a was observed [100]. The 2D DLA model with the power-like potential, uðRÞ ¼
u0 Ra , was also studied [101, 104]. Here, u0 and a are parameters. Note that this
model is equivalent to the ordinary DLA in the limits of short range interactions,
a ! 1, or u0 ¼ 0. In this problem, the aggregates were grown on a triangular
lattice and the effects of anisotropy were important.
The detailed structure of clusters was strongly dependent on a, or u0, e.g., in the
case of attractive interactions at large ju0 j and long-range interaction (i.e., for small a),
the clusters grew with stable tips and were of dendritic shape. The observed effects
reflected the anisotropy induced by the underlying lattice that stabilized the tips of the
growing arms. In the limit of small ju0 j and large a, the transition from dendritic to
tip-split aggregates was observed. At fixed value of u0, the estimated fractal dimen-
sion df increased with increase in a, which reflected tip destabilization [101].
The similar cluster–cluster variants of DLA aggregation with either attractive or
repulsive interactions were also studied [102, 103]. These simulations were done
74 N.I. Lebovka

using off-lattice models both for 2D [103] and three dimensional (3D) [102]
systems. The interaction energy uij of two non-overlapping clusters including
i and j particles was approximated by the power law:

X
i X
j
uij ¼ u0 Ra
kl ; (32)
k¼1 i¼1

where Rkl is the distance between the kth particle in the cluster i and lth particle in
the cluster j.
Both the theory and simulation results gave the following estimates for fractal
dimensionality df versus parameter a dependence:
8
< d0 ; a > 2d0
>
df ¼ d0 ða þ 2Þ=ð2ðd0 þ 1ÞÞ; 1<a  2d0 ; (33)
>
:

where d0 is the fractal dimensionality in the cluster–cluster aggregation model

without interactions. It equals d0 ~ 1.6 and d0 ~ 2.1 for 2D and 3D systems,
respectively.
Figure 5 presents df versus a dependencies for 2D and 3D systems. The figure
shows that short range interactions (a > 2d0) do not change the value of df, but long
range (1 < a ; 2d0) interactions lead to substantial changes in df. Moreover, long
range interactions can have an important effect on the local structure of clusters
[102, 103]. The shape and size of the aggregates may be sensitively dependent on
the balance between attraction and repulsion interactions [98].

3.2.2 Eden-Like Model

The stochastic Eden-like model of aggregation of the charged particles in 2D

systems was developed [105, 106]. In this model, the particles overcome the
electrostatic repulsive barrier created by the aggregate and stick to it due to the
existence of short-range attractions.
By variation of two model parameters, the screening length, l, and the attractive
binding energy per particle, u0, formation of the aggregates with the following
morphologies was studied: linear or near-linear, linear with bending, worm-like,
dense-branching or dense-branching with a core, and compact Eden-like aggregates.
Figure 6 presents an example of a cluster with a charged core and external branches.
The regions of finite and infinite growth of clusters with different morphologies are
presented in the form of a diagram as l versus u0. The structure and fractal properties
of the ramified clusters were studied. It was found that the clusters did not reveal the
fractal properties at any values of l and u0 and a sharp transition between linear
(df ¼ 1) and dense-branching (df ¼ 2) morphologies was observed.
Aggregation of Charged Colloidal Particles 75

2.4

2.2
3d

Short range interactions

1.8
df

ns
2d

io
1.6
ct
ra
te
in
e
ng

1.4
ra
ng
Lo

1.2

1
1 2 3 4 5 6
α

Fig. 5 Fractal dimensionality df versus power exponent a for 2D and 3D cluster–cluster DLA
model

Internal compact
charged core

Fig. 6 Internal structure of cluster obtained using Eden-like model of aggregation of charged
particles (l ¼ 1, u0 ¼ 60). The cluster includes a core (central dense charged part) and external
branches [105]
76 N.I. Lebovka

3.3 Oppositely Charged Particles

3.3.1 Simple Colloids

The particles of a mixture of oppositely charged nanoparticles form charged

aggregates (complexes), in which each charged particle is screened by a shell of
oppositely charged particles [107, 108]. The stable core-and-shell aggregates were
confirmed in experiments with the mixtures of oppositely charged metal
nanoparticles [107] When the net charges of positive and negative species are
equal, the aggregates are neutral and condense into a macroscopic drop, i.e., the
system becomes unstable. The theory developed in [109] predicts a possibility of
enhanced aggregation of larger charged particles induced by the small polyions of
the opposite charge. This phenomenon may reflect the partial “condensation” of the
small particles and reduction of the strong electrostatic repulsion between screened
(larger) colloidal particles.
The Brownian dynamic simulation of aggregation between the oppositely
charged particles has shown that heteroaggregation produces more branched
aggregates than the usual diffusive aggregation [110]. Moreover, the cluster dis-
crimination was observed at the late stages of aggregation, when neutral clusters
were disappearing faster than charged ones [111].
The presence of linear chaining with particle charge alternating down a chain in
aggregates of the oppositely charged polystyrene spheres was revealed by different
experimental techniques and was supported by Brownian dynamics simulation
[112]. It was shown that branching of an aggregate composed of oppositely charged
particles may be varied from a linear chain structure (df ~ 1.2) to a structure of
diffusion-limited aggregates (df ~ 1.7) by an increase in concentration of the
background electrolyte. Formation of linear chains was explained by the short-
range attraction between the oppositely charged particles and long-range repulsion
between the identical particles. The experiments and the Brownian dynamics
simulations of heteroaggregation between the oppositely charged particles
[113–115] has shown that small silica (diameter ~ 25 nm, negatively charged)
particles covered the surface of large alumina (diameter ~ 400 nm, positively
charged) particles. For small amounts of silica particles, agglomeration of the
silica-covered alumina particles was observed, and it was demonstrated that the
agglomerates were of elongated shape.

3.3.2 Polymers

Behavior of solutions of oppositely charged polymers is rather similar to that of

solutions of oppositely charged colloids. However, higher flexibility of the polymer
chains results in a less ordered structure of the polyion solutions. Pairing of the
oppositely charged polyions, their aggregation, and complexation of anionic and
Aggregation of Charged Colloidal Particles 77

cationic polyelectrolytes have been intensively studied in many theoretical

[116–121] and experimental [122–124] works.
The phase diagram of asymmetric positively and negatively charged chains in an
oppositely charged polyelectrolyte was theoretically obtained [125]. It was shown
that cylindrical and lamellar cluster structures with nonzero net charge were formed
at high polymer concentration. Nonuniform distribution of charge along the
annealed polyelectrolyte chains was observed [116, 121]. The theoretical
predictions and Monte Carlo simulations show that charges accumulate at the
ends of chains. [116]. The cluster size distribution functions of the oppositely
charged macroions in solution were studied using the Monte Carlo simulations
[120]. It was demonstrated that increasing electrostatic coupling results in forma-
tion of larger clusters, and a single cluster comprising all the macroions forms at the
strongest coupling. The structure of symmetric polycation–polyanion mixtures
without salt in good solvents was studied using the Langevin sampling simulation
technique and field-theoretic simulation methods [119, 126]. The obtained results
allowed explanation of the phenomenon of complex coacervation, a type of phase
separation in which dense liquid precipitates coexist with supernatant solvent.
Brownian dynamics computer simulations have shown that variation in electro-
static interactions changed the size, shape, and local density distribution of the
complexes formed by terminally charged hyperbranched polymers and oppositely
charged neutralizing linear polyelectrolytes [127]. The ultrasoft core model of
interpenetrating polycations and polyanions was proposed for investigation of the
polyelectrolyte aggregation using different theoretical approaches and molecular
dynamic simulations [117]. The clustering and segregation of the oppositely
charged species was observed. At sufficiently low temperatures and densities, the
oppositely charged polyions tend to form weakly interacting neutral pairs. The
break-up of ion pairs was observed with increase in temperature. This was followed
by the percolation transition from low temperature dielectric (insulator) state to
high temperature ionic (conductor) state [117].
The Monte Carlo simulations were applied to study the complexation, phase
separation, and redissolution of polyelectrolyte–macroion solutions [128]. It was
shown that introduction of the oppositely charged polyelectrolytes into a stable
macroion solution with repelling macroions resulted in a decrease in the solution
stability. The system was unstable at macromolecular charge equivalence when a
large and loose cluster of macroions and polyelectrolytes was forming. Finally,
redissolution of macroions occurred in the excess of polyelectrolyte.
The complexation between a polyampholyte and a charged particle was studied
using Monte Carlo simulations [129, 130]. The increasing charge density and
particle size resulted in change of configuration of the polyampholyte chain
adsorbed on the particle surface. At large charge density and particle size, collapse
of the polyampholyte chain on the particle surface was observed [129]. The effects
of different model parameters (polyampholyte contour length, nanoparticle surface
charge, pH of solvent, ionic concentration, etc.) on possible polyampholyte
conformations at the nanoparticle surface were investigated [130].
78 N.I. Lebovka

The process of aggregation of the oppositely charged polyions (positively

charged liposomes and negatively charged polyelectrolyte sodium polyacrylate)
was studied by means of Monte Carlo simulation [131]. The model accounted for
heterogeneous charge distribution at the particle surface, related with the correlated
polyion adsorption [46]. Formation of long-living clusters of polyelectrolyte-
decorated particles was demonstrated. Molecular dynamics simulations were
performed to study the polyampholyte–polyelectrolyte complexes in solutions
[124, 132]. It was demonstrated that polyampholyte chain binds to a polyelectrolyte
in a way optimizing electrostatic interactions between the ionic groups in both
polymeric chains. Formation of the micellar complex by anionic polyelectrolyte
and cationic surfactants (in monomeric and dimeric forms) was investigated by
molecular dynamics simulation [133]. Results have shown that the dimeric form
interacts more strongly with the polyanion and the size of the micellar complex
becomes larger with an increase in the surfactant concentration.

3.4 Effect of Dipolar Interactions

The effects of dipolar interactions on DLA processes were studied in details for 2D
and 3D off-lattice models [134–137]. The fractal dimensionality df was a monoton-
ically increasing function of the temperature (or decreasing function of dipolar
forces). For example, it varied continuously from about 1.78 for small dipolar
interactions to about 1.35 for large dipolar interactions (3D model) [134, 135].
Therefore, the structure of clusters formed at low temperatures or strong dipolar
forces was less branched and more open (df ~ 1) than in free DLA with no
interactions. On increase in temperature or decrease in dipolar forces, the value
of df reached the limit value of free DLA [136, 137]. The values df ¼ 1.13  0.01
and df ¼ 1.37  0.03 were obtained in the limit of zero temperature for 2D and 3D
systems, respectively. Transitions between an ordered, or quasi-ordered, and a
disordered phase were also observed for high values of the reduced temperature
[138]. The long range correlations between the dipoles were revealed in the low-
temperature ordered phase.

4 Kinetics of Aggregation

Kinetics of aggregation, coalescence, and annihilation or fragmentation are impor-

tant in many physical, chemical and biological processes [21, 80, 139–141]. The
popular mean field Smoluchowski approach [73] gives good description of simple
aggregation systems. However, in the presence of restructuring, long range
interactions, and formation of clusters with fractal geometry, more complicated
approaches based on computer simulation methods are useful.
Aggregation of Charged Colloidal Particles 79

4.1 Aggregation as a Second-Order Reaction

In the simplest case, only primary particles are present at initial time, t ¼ 0, with
initial number density of r0 and volume fraction of ’ ¼ 4pr2r0/3. It can be supposed
that in very diluted systems (1% vol) only two particle collisions are important. The
regime of “fast” aggregation is assumed, i.e., two primary particles form an aggregate
when they “touch” at the distance of 2r between their centers. The disappearance of
primary particles can be considered as a second-order reaction [73]:

dn1
¼ kf n21 ; (34)
dt

where n1 is the dimensionless concentration of the primary particles, i.e. n1 ¼ r1/r0,

and kf (s1) is the fast aggregation rate constant. Note that kf ¼ 1/ta, where ta is the
half-aggregation time corresponding to n1 ¼ 1/2.
The integration of Eq. 34 gives:

n1 ¼ ð1 þ t=ta Þ1 : (35)

Estimations show [73]:

4kB Tr0
kf ¼ 8pDr ¼ ; (36)
3

where D ¼ 6pr
kB T
is the diffusion coefficient, and  is the viscosity of solvent.
The half-aggregation time of colloidal dispersion, ta, may be estimated as:

3
ta ¼ 1=kf ¼ tB =’ ¼ ; (37)
4kB Tr0

where tB ¼ r2/(6D) is the Brownian time. During the Brownian time, the length of
diffusion is equal to the radius of the primary particle, r.
For example, the Brownian time, tB, of 10 nm particles ( ~ 0.001 Pa s) in water
at room temperature (T ¼ 298 K) is equal to 0.76  106 s. Figure 7 presents plots
of ta versus r at different values of the volume fraction of particles, ’.
In the presence of repulsive interactions, the sticking probability becomes
smaller than 1 and the so-called “slow” aggregation regime may be realized. The
slow aggregation rate constant, ks may be estimated as:

kijs ¼ kf wij ; (38)

80 N.I. Lebovka

j=10-4
103 j=10-3

j=10-2
101
j=10-1

10-1
ta, s

10-3

10-5

10-7

100 101 10 2 10 3
r, nm

Fig. 7 Half-aggregation time ta, versus radius of particle r at different values of the volume
fraction of particles, ’. The estimation was done using Eq. 37 for suspension of particles in water,
T ¼ 298 K and  ~ 0.001 Pa s

where Wij ( 1) is the dimensionless Fuchs stability ratio, which can be considered
as the inverse sticking probability [142]:
ð1
expðut ðxÞ=kB T Þ
wij ¼ dx: (39)
0 ð x þ 2Þ 2
 
Here, x ¼ h=rij ; rij ¼ ri þ rj =2; and ut is the total energy, i.e., the sum of the
attractive, ua, and repulsive, ur, parts.
It was shown that the value of Wij is mainly determined by the height of the
energy barrier umax
t located at h ffi lD and the following approximate relation was
obtained for equal sized particles, r ¼ ri ¼ rj, [143]:

lD  
Wij ¼ exp umax
t =kB T : (40)
2r

Usually, it is assumed that fast and slow aggregations are controlled by attractive
and total interactions, respectively, and the equation used for estimation of Wij is
more complicated:
Ð1
dx expðut =kB T Þ=ðx þ 2Þ2
Wij ¼ Ð 01 : (41)
0 dx expðua =kB T Þ=ðx þ 2Þ2
Aggregation of Charged Colloidal Particles 81

4.2 Population Balance Equations

In order to account for formation of dimers, trimers, and larger aggregates, von
Smoluchowski proposed in his pioneering work [73] the following system of
population balance equations (PBE):

dnm 1 X m1 X1


¼ ki;mi ni ðtÞnmj ðtÞ  nm ðtÞ km;i ni ðtÞ; (42)
dt 2 i¼1 i¼1

where nm is the dimensionless concentration of the clusters of size m, ki,mi are

dimensionless reaction constants, or kernels, and t* ¼ kft ¼ t/ta is the normalized
time.
At initial moment of time, t ¼ 0, all clusters are assumed to be monomeric and
n1 ¼ 1. The first term (“birth” term) in Eq. 42 corresponds to the collision between
two clusters (i-mers and m  i-mers) and formation of m-mers. In first summation,
each collision is accounted for twice; hence, a factor of 1/2 is included. The second
summation corresponds to the decrease in concentration of m-mers through aggre-
gation with other clusters.
In fact, the PBE approximation is a mean field approximation valid for dilute
systems. Moreover, this approximation does not account for internal structure and
differences in the spatial configuration of clusters. For the known functional
dependence of km,i, time evolution of the cluster population nm(t) can be calculated
from Eq. 42.
During the aggregation the total mass of clusters:
X
ini ðtÞ ¼ 1; (43)
i 1

is conserved.
The important characteristics of the cluster distributions are dimensionless total
number of clusters n:
X
nðtÞ ¼ ni ðtÞ (44)
i 1

and mean cluster size (mass), s(t):

X
sðtÞ ¼ i2 ni ðtÞ: (45)
i 1
82 N.I. Lebovka

Table 2 Analytical solutions of PBE for constant, sum, and product kernels
Concentration of clusters Number of
Kernel, ki,j of size m, nm clusters, n Size of clusters, s
2 t ðm1Þ 1
1þt
1 + 2t*z, z ¼ 1
ð1þt Þðmþ1Þ

i+j exp(t*) exp(at*), a ¼ 2

ðmaÞm1 expðmat Þ
m! ; a ¼ 1  expðt Þ

2ij, at t*  0.5 1  t*
b
ð2mt Þm1 expð2mt Þ
mm! 1  t =tg ;b ¼ 1

2ij, at t* > 0.5 1/(4t*) 1, gel

ðmÞm1 expðmÞ
2 mm!t

Here, t* ¼ t/ta is the dimensionless aggregation time [73, 144–146]

4.3 Popular Kernels

The analytical solutions of PBE were obtained for some functional forms of
kernels, e.g., for constant kij ¼ 2, sum kij ¼ i + j, and product kij ¼ i + j kernels
(Table 2). An exact solution exists also for linear combination of these three
kernels:

kij ¼ A þ Bði þ jÞ þ Cij; (46)

where A, B, and C are the arbitrary constants, as well as for q-sum kernel:

kij ¼ 4  qi  qj ; (47)

where 0 < q < 1, and for many other kernels (for a review see, [144]).
Figure 8 shows examples of distribution functions f(m) ¼ mnm/(mnm)max
obtained from analytical solutions presented in Table 2 for constant (kij ¼ 2),
sum (kij ¼ i + j), and product (kij ¼ 2ij) kernels at different moments in time. It
is interesting that at long times the distribution function for the constant kernel has a
bell-like shape, whereas for sum and product kernels they monotonically decay
with increasing m.
The case of constant kernel is similar to the situation that was analyzed in
Sect. 4.1. The fast aggregation problem with the constant kernel was exactly solved
by Smoluchowski in 1917 [73]. This model is based on the more complicated kernel
for 3D Brownian aggregation:

1 1=3

kijB ’ i þ j1=3 i1=3 þ j1=3 ¼ 1 þ ði=jÞ1=3 þ ðj=iÞ1=3 : (48)

2
Aggregation of Charged Colloidal Particles 83

1
k ij =2
t*=
0.1
f 0.5
2.0
10.0

0
10 0 101 10 2
m

10 0
k ij =i+j
10 -1 t*=
f

0.1
0.4
10.0
10-2 0
10 101 10 2
m

100
k ij =2ij
t*=
0.1
f

-1
10
0.2
0.49

10-2
10 0 10 1 10 2
m

Fig. 8 Normalized cluster distribution functions f(m) ¼ mnm/(mnm)max for constant (kij ¼ 2),
sum (kij ¼ i + j), and product (kij ¼ 2ij) kernels at different dimensionless times t* ¼ t/ta

Note that Brownian kernel kijB has a constant value of about 2 when i  j. The
kernels for Brownian ðkij Bf Þ or ballistic ðkij bf Þ aggregation of fractal particles are [147]:

1 1=df

kij Bf ’ i þ j1=df i1=df þ j1=df ; (49)

2

1  1 1=2 1=d

kij bf ’ i þ j1 i f þ j1=df ; (50)

2

where df is the fractal dimension of the cluster.

The exact analytical solutions of PBE for Brownian and ballistic kernels have not
yet been obtained. In slow aggregation regime, these kernel may be estimated as:

s;BðbÞ BðbÞ
kij ¼ kij =Wij : (51)

Note that the constant kernel kinetics is the slowest as compared to kinetics of
the sum and product kernels. It is interesting to note that formation of the infinite
84 N.I. Lebovka

Table 3 Classification of kernels among three different classes (I, II, III) and gelling (G) or
non-gelling (N) behavior
Type of Homogeneity Type of cluster Class,
Kernel, ki,j kernel parameter, l union, m behavior
2 Constant 0 0 II, N
i+j Sum l 0 II, N
2ij Product 2 1 I, G

1=3  1=3 Brownian 1/3 0 III, N
1 þ ij þ ji

1=df Fractal 1/df 0 III, N
j1=df
1þ i
j þ i
Brownian

1=2  Fractal 1/df  1/2 1/2 III, N

1
2
1
i þ 1j i1=df þ j1=df ballistic

– DLA 0 0 I, N
– RLA 1 1 I, N
See Fig. 9 for the relationship between l and m and the behavior and class of kernel, respectively

cluster takes place for the product kernel within the finite time at tg ¼ 0.5ts. It
corresponds to the sol–gel transition. At time exceeding tg, the particles belong to
two different populations in the finite (sol phase) and infinite (gel phase) clusters
[145].

4.4 Classification of Kernels

For simple aggregation models, the kernel kij is usually a homogeneous function:

kai;aj ¼ al kij ; (52)

where a is a positive constant and may be presented for large-size clusters as the
product of powers im and jn [148]:

ki;j ’ im jn : (53)

Here, m and n are exponents, n ¼ l  m, i  j, and i 1, and kernels with

l > 2 and l > 1 + m are unphysical. The values of l and m for different types of
kernels are presented in Table 3.
Figure 9 shows the l versus m diagram for different types of kernels. The class of
the kernel is determined by the sign of the m exponent. The big + big and big +
small unions of particles form within the class I (m > 0) and class III (m < 0),
Aggregation of Charged Colloidal Particles 85

k ij=2ij
2

Gelling behaviour

l>1
λ
1 ← RLA →

k ij =i+j
Brownian

Non-gelling behaviour
k ij=2

0 ← DLA →

λ<1
-2 -1 0 1 μ

Ballistic

-1 ↑
← III II I →
Big+small union Big+big union

Fig. 9 Classification of kernels. l versus m diagram is presented. The shaded area corresponds to
the physically plausible limit for l (2) and m (l  1)

respectively. For class II (m ¼ 0), any type of union is probable. The kernels with
l > 1 and l  1 correspond to the gelling and non-gelling behavior, respectively.
Finally, the well-known regimes of DLA and RSA aggregations correspond to the
constant values of l ¼ 0 and l ¼ 1, respectively. The Brownian aggregation
belongs to the DLA-type of aggregation and the sum kernel aggregation belongs
to the RLA-type of aggregation. The simple product kernel kij ¼ 2ij corresponds to
the gelling behavior at the limiting high value of l ¼ 2.

4.5 Dynamic Scaling

It was found in many experimental and theoretical studies of dynamic scaling that
the time dependence of cluster distribution for long times, large m, and non-gelling
system can be expressed as [149]:

nm ðtÞ / sðtÞ2 f ðm=sðtÞÞ; (54)

where f(x) is a scaling function that depends on the sign of m. The f(x) decreases
monotonically for kernels with m 0 (I and II class), whereas it is a bell-shaped
function for m > 0.
86 N.I. Lebovka

103
kij =2ij
kij =i+j

s=(1-2t* )-1 kij =2

102
s = exp(2t* )

s=1+2t*
s

101

Gel point
t=0.5τa
100
10-1 100 101 102
t *=t/τa

Fig. 10 Mean cluster size s(t) versus dimensionless time t* ¼ t/ta for different types of kernels.
Arrow shows the gelling time, tg ¼ 0.5ta for the product kernel, kij ¼ 2ij

The time dependence of mean cluster size s(t) at l < 1 follows:

sðtÞ / tz ; (55)

at l ¼ 1, it follows:

sðtÞ / expðatÞ; (56)

and for gelling systems:

sðtÞ / ð1  t=tg Þb : (57)

Here, a and b are constants and the following is a dynamic exponent:

z ¼ 1=ð1  lÞ: (58)

Note that continuous transition from z ¼ 1 (l ¼ 0) to z ¼ 1 (l ¼ 1)

corresponds to the transition from DLA to RLA model [150].
Figure 10 shows the examples of s versus t/ta obtained from analytical solutions
presented in Table 2 for constant (kij ¼ 2), sum (kij ¼ i + j), and product (kij ¼ 2ij)
kernels. It is important that the three basic kernels can be used for an approximate
description of the main features of DLA (Eq. 55), RLA (Eq. 56), and gelling
(Eq. 57) models.
Aggregation of Charged Colloidal Particles 87

Finally, both theory and experiments have shown a universal long-term behavior
of the cluster-size distribution nm(t) for non-gelling systems, l < 1 [148, 149]:

nm ðtÞ=n1 ðtÞ ! 1; m<1

w t
nm ðtÞ
t m ; m < 0; w > 1; t < l þ 1
nm ðtÞ
tw mt ; m > 1; w ¼ 1; t ¼ l þ 1 (59)

where w and t are the scaling exponents.

To date, the dynamic scaling in heteroaggregation within the PBE approach was
mainly tested for the simplified kernels. Practical applications of Eq. 51 for kernel
estimation is restricted because of the absence of detailed relations for interactions
of real particles with the fractal structure or rough surface.
The realistic kernels of different types were tested for a description of RLA
aggregation [151–153]. The following two types of kernels:

kijB f
kij ¼ ðijÞl (60)
Wij

and:

kijB f ðijÞl
kij ¼ h i ; (61)
Wij 1 þ ðijÞl  1 =Wij

were used for comparison of the theory and experiments [151]. Here, kijB f is the
Brownian kernel of fractal particles and Wij is the stability ratio.
It was assumed that the stability ratio is constant in the course of aggregation,
Wij  Win. The initial value, Win, was experimentally determined for polymer latex
particles in aqueous suspensions. The aggregation rate was measured at the very
initial stage, where the presence of triplets was negligible [151]. The kernels in
Eqs. 60, 62 were found to be appropriate for simulation of experimental results
subject to proper tuning of the exponent l. Moreover, the kernels in Eq. 62 were
suitable for description of the continuous transition from DLA-like aggregation at
W ¼ 1 to RLA-like aggregation at W ! 1 [152].
The Brownian dynamic simulations of aggregation between the oppositely
charged particles were made using classical DLVO potential [111]. The results
were compared with theoretical calculations using a twofold homogeneous kernel,
kij / (i + j)l, where l ( 0) is the homogeneity parameter. The exponent z (see,
Eq. 58) obtained from simulation was an increasing function of r/lD. The continuous
transition from ballistic-like aggregation, at low r/lD (z < 1), to DLA-like aggre-
gation, at high r/lD (z ¼ 1), was observed. Most of the features of charge heteroag-
gregation kinetics obtained using Brownian dynamic simulation were well
described by the PBE theory with dynamic exponent z estimated as ’ 1=ð1  lÞ.
88 N.I. Lebovka

10 8
Classical ‘ fast’
aggregation
107

W in = s =1+ 2t*
10 6 1.00
6.51
60.7
10 5
2000
s ∝ t*z
10 4
s

10 3

s ∝ (logt* ) z
10 2

101

100
100 101 102 103 104 105 106 107 108

t*=t/(W inτa)

Fig. 11 Mean particle size s versus dimensionless time t* ¼ t/Winta for different initial values of
stability ratios Win ¼ 1, Win ¼ 6.51 (lD ¼ 1 nm), Win ¼ 60.7 (lD ¼ 2 nm), and Win ¼ 2,000
(lD ¼ 9 nm). The solid line corresponds to the classical model of fast aggregation, where size
dependence of the aggregate diffusion coefficient is neglected. Dashed lines are results of
approximation by the power law (s / t*z, Win ¼ 1), or logarithmic power law (s / (logt*)z,
Win > 1) The simulations were done assuming A ¼ 1020 J and z ¼ 40 mV. Compiled from
published data [154]

Recently, computer simulations of the aggregation process of PEC nanoparticles

were carried out [154]. The aggregation kinetics was studied accounting for the size
dependence of the aggregate diffusion coefficient, classical DLVO potential for
interactions between two unequal particles, and Fuchs stability ratio with hydrody-
namic corrections. It was assumed that the value of stability factor Wij is not
constant and depends upon the size and charge of aggregating species. Figure 11
presents the calculated time dependencies of mean particle size s for different initial
values of the stability ratio Win ¼ Wij(t ¼ 0). The classical constant rate approxi-
mation results in overestimation of the mean size of a particle compared to the non-
classical model of fast aggregation (Wij ¼ 1) with the size-dependent diffusion
coefficient. Note that both models follow the power law s(t) dependence (Eq. 55),
but with different dynamic exponents, z ¼ 1 and z ¼ 0.745  0.005, for the
classical and non-classical models of fast aggregation, respectively (Fig. 11). The
simulated data show a noticeable deviation of s(t) curves for slow aggregation
(Win > 1) from the same curves for fast aggregation (Win ¼ 1).
Aggregation of Charged Colloidal Particles 89

Moreover, analysis of data presented by Starchenko et al. [154] have shown that
the universal logarithmic scaling:

sðtÞ / ðlog tÞz ; (62)

with the same exponent z ¼ 0.70  0.01 was observed for all the studied values of
Win (>1).
This result reflects a noticeable increase in Wij stability ratios in the course of
aggregation compared to the initial value Win. The data demonstrate also that
enhancement of the electrostatic repulsion, related to the increase in the Debye
length lD or z potential, results in a retardation of the growth. Good correspondence
between theory and experiment was observed at small Debye length, lD < 3–5 nm,
in the so-called colloid regime. However, the effect of electrolyte on the PEC
conformation became important at larger lD and simulation failed to describe
experiments [154].

5 Conclusion

The study of aggregation has been intensively developed during the last 100 years
and is relatively mature. However, aggregation is a very complex problem and this
field still remains an open research area. Colloid stability is basically defined by
interparticle forces of different origin and interaction length. The exact analytical
expressions have not yet been obtained for many important contributions. More-
over, in suspensions of charged particles, the long-range and many-body effects,
collective correlations between charges, and the effects of charges heterogeneously
distributed on the surface of particles (so-called patchy surfaces [155]) may be
particularly important. For polymeric systems, the emphasis should be directed
towards the interaction between core–shell particles containing a compact core and
covered by a charged soft shell. The theory and experiments show realization of the
different morphologies of the aggregates, including compact, linear-like, branched,
fractal, and mixed morphologies. However, the advantages of computer simulations
are still restricted by relatively small and closed systems and, as a rule, the use of
very primitive models for interparticle potentials. The theoretical understanding of
aggregation kinetics is mainly based on original Smoluchowski theory (1917) and
oversimplified assumptions about the size dependence of both the cluster diffusion
coefficient and sticking probability. More complex analysis is required to account
for the restructuring of aggregates, which can be particularly important for clusters
with fractal morphology [156]. Finally, the aggregation processes in PEC
nanoparticles are still far from being fully understood. For these systems, which
behave as suspensions of core–shell particles, the impact of polyelectrolyte
conformations on the final size of nanoparticles was revealed [154]. Future progress
needs the range of experimentally studied types of polyions to be broadened and the
development of more sophisticated theories in order to prepare aggregates of PEC
particles with desired size and properties.
90 N.I. Lebovka

Acknowledgements The author appreciates the financial support from the National Academy of
Sciences of Ukraine and Dr. N. S. Pivovarova for her help with the preparation of the manuscript.

References

1. Mohanraj VYC (2006) Nanoparticles a review. Trop J Pharm Res 5(1):561–573

2. Hartig S, Greene R, Dikov M, Prokop A, Davidson J (2007) Multifunctional nanoparticulate
polyelectrolyte complexes. Pharmaceut Res 24:2353–2369. doi:10.1007/s11095-007-9459-1
3. Lankalapalli S, Kolapalli VRM (2009) Polyelectrolyte complexes: a review of their applica-
bility in drug delivery technology. Indian J Pharm Sci 71(5):481–487. doi:10.4103/0250-
474X.58165
4. Muller M, Kesler B, Frohlich J, Poeschla S, Torger B (2011) Polyelectrolyte complex
nanoparticles of poly(ethyleneimine) and poly(acrylic acid): preparation and applications.
Polymer 3(2):762–778. doi:10.3390/polym3020762
5. M€uller M, Reihs T, Ouyang W (2005) Preparation of monomodal polyelectrolyte complex
nanoparticles of pdadmac/poly(maleic acid-alt-alpha-methylstyrene) by consecutive centri-
fugation. Langmuir 21(1):465–469
6. Dautzenberg H (2000) Light scattering studies on polyelectrolyte complexes. Macromol
Symp 162:1–21
7. Panyam P, Labhasetwar V (2003) Biodegradable nanoparticles for drug and gene delivery to
cells and tissue. Adv Drug Deliv Rev 55:329–347
8. Derjaguin BV, Landau L (1941) Theory of the stability of strongly charged lyophobic sols
and of the adhesion of strongly charged particles in solutions of electrolytes. Acta Phys Chim
USSR 14:633–662
9. Verwey EJW, Overbeek JTG (1948) Theory of the stability of lyophobic colloids. Elsevier,
Amsterdam
10. Quemada D, Berli C (2002) Energy of interaction in colloids and its implications in rheologi-
cal modeling. Adv Colloid Interface Sci 98(1):51–85. doi:10.1016/S0001-8686(01)00093-8
11. Eastman J (2010) Stability of charge-stabilised colloids. In: Colloid science: principles,
methods and applications, 2nd edn. Wiley-Blackwell
12. Derjaguin BV (1934) Untersuchungen ueber die reibung und adhaesion IV. Theorie des
anhaften kleiner teilchen. Kolloid Z 69:155–164
13. Israelachvili J (1991) Intermolecular and surface forces. Academic, New York
14. Shaw DJ (1992) Introduction to colloid and surface chemistry. Butterworth-Heinemann,
Oxford
15. Mahanty J, Ninham B (1976) Dispersion forces. Academic, New York
16. Russel WB, Saville DA, Schowalter WR (1989) Colloidal dispersions. Cambridge University
Press, New York
17. Malins A, Williams SR, Eggers J, Tanaka H, Royall CP (2011) The effect of inter-cluster
interactions on the structure of colloidal clusters. J Non-Cryst Solids 357(2):760–766.
doi:10.1016/j.jnoncrysol.2010.08.021
18. dos Santos AP, Diehl A, Levin Y (2009) Electrostatic correlations in colloidal suspensions:
density profiles and effective charges beyond the Poisson-Boltzmann theory. J Chem Phys
130:124110
19. Hogg R, Healy TW, Fuerstenau DW (1966) Mutual coagulation of colloidal dispersions.
Trans Faraday Soc 62:1638–1651. doi:10.1039/TF9666201638
20. Ohshima H, Healy TW, White LR (1982) Improvement on the Hogg-Healy-Fuerstenau
formulas for the interaction of dissimilar double layers: I. Second and third approximations
for moderate potentials. J Colloid Interface Sci 89(2):484–493. doi:10.1016/0021-9797(82)
90199-0
Aggregation of Charged Colloidal Particles 91

21. Elimelech M, Gregory J, Jia XWR (1995) Particle deposition and aggregation – measure-
ment, modelling and simulation. Elsevier, Amsterdam
22. Boroudjerdi H, Kim YW, Naji A, Netz R, Schlagberger X, Serr A (2005) Statics and
dynamics of strongly charged soft matter. Phys Rep 416(3–4):129–199. doi:10.1016/j.
physrep. 2005.06.006
23. Kékicheff P, Spalla O (1995) Long-range electrostatic attraction between similar, charge-
neutral walls. Phys Rev Lett 75:1851–1854. doi:10.1103/PhysRevLett.75.1851
24. Kjellander R (1996) Ion-ion correlations and effective charges in electrolyte and macroion
systems. Ber Bunsen Phys Chem 100(6):894–904. doi:10.1002/bbpc.19961000635
25. Larsen AE, Grier DG (1997) Like-charge attractions in metastable colloidal crystallites.
Nature 385:230–233
26. Levin Y (2002) Electrostatic correlations: from plasma to biology. Rep Prog Phys
65:1577–1632
27. Bohinc K, Zelko J, Sunil Kumar PB, Iglic A, Kralj-Iglic V (2009) Attraction of like-charged
surfaces mediated by spheroidal nanoparticles with spatially distributed electric charge:
theory and simulation. In: Advances in planar lipid bilayers and liposomes, vol 9. Academic,
Burlington
28. Vlachy V (1999) Ionic effects beyond Poisson-Boltzmann theory. Annu Rev Phys Chem
50:145–165
29. Shklovskii BI (1999) Screening of a macroion by multivalent ions: correlation-induced
inversion of charge. Phys Rev E 60:5802–5811. doi:10.1103/PhysRevE.60.5802
30. Patra M, Patriarca M, Karttunen M (2003) Stability of charge inversion, thomson problem,
and application to electrophoresis. Phys Rev E 67:031402. doi:10.1103/PhysRevE.67.031402
31. Levin Y (1999) When do like charges attract? Physica A 432:432–439
32. Levin Y (2005) Strange electrostatics in physics, chemistry, and biology. Physica A 352(1):
43–52. doi:10.1016/j.physa.2004.12.033
33. Trizac E, Raimbault JL (1999) Long-range electrostatic interactions between like-charged
colloids: steric and confinement effects. Phys Rev 60:6530–6533
34. Kirkwood JG, Shumaker JB (1952) Forces between protein molecules in solution arising
from fluctuations in proton charge and configuration. Proc Natl Acad Sci USA 38:863–871
35. Grønbech-Jensen N, Mashl RJ, Bruinsma RF, Gelbart WM (1997) Counterion-induced
attraction between rigid polyelectrolytes. Phys Rev Lett 78:2477–2480. doi:10.1103/
PhysRevLett.78.2477
36. Belloni L, Spalla O (1997) Attraction of electrostatic origin between colloids. J Chem Phys
107(2):465–480. doi:10.1063/1.474408
37. Podgornik R, Parsegian VA (1998) Charge-fluctuation forces between rodlike polyelec-
trolytes: pairwise summability reexamined. Phys Rev Lett 80:1560–1563. doi:10.1103/
PhysRevLett.80.1560
38. Squires TM, Brenner MP (2000) Like-charge attraction and hydrodynamic interaction. Phys
Rev Lett 85:4976–4979. doi:10.1103/PhysRevLett.85.4976
39. Allahyarov E, D’Amico I, L€ owen H (1998) Attraction between like-charged macroions by
coulomb depletion. Phys Rev Lett 81:1334–1337. doi:10.1103/PhysRevLett.81.1334
40. Manning GS (1969) Limiting laws and counterion condensation in polyelectrolyte solutions
I. Colligative properties. J Chem Phys 51(3):924–933. doi:10.1063/1.1672157
41. Arenzon J, Stilck J, Levin Y (1999) Simple model for attraction between like-charged
polyions. Eur Phys J B 12:79–82
42. Tokuyama M (1998) Theory of slow dynamics in highly charged colloidal suspensions. Phys
Rev E 58:R2729–R2732. doi:10.1103/PhysRevE.58.R2729
43. Tokuyama M (1999) Effective forces between macroions in highly charged colloidal
suspensions. Phys Rev E 59:R2550–R2553. doi:10.1103/PhysRevE.59.R2550
44. Terada Y, Tokuyama M (2004) Novel liquid- and crystal-droplet phases on highly charged
colloidal suspensions. Physica A 334(3–4):327–334. doi:10.1016/j.physa.2003.10.078
45. Fitch RM (1997) Polymer colloids. Academic, New York
92 N.I. Lebovka

46. Velegol D, Thwar PK (2001) Analytical model for the effect of surface charge nonuniformity
on colloidal interactions. Langmuir 17:7687–7693. doi:10.1021/la010634z
47. Czarnecki J (1985) The effects of surface inhomogeneities on the interactions in colloidal
systems and colloid stability. Adv Colloid Interface Sci 24:283–319. doi:10.1016/0001-8686
(85)80035-X
48. Grant M, Saville D (1995) Electrostatic interactions between a nonuniformly charged sphere
and a charged surface. J Colloid Interface Sci 171(1):35–45. doi:10.1006/jcis.1995.1148
49. Stankovich J, Carnie SL (1999) Interactions between two spherical particles with nonuniform
surface potentials: the linearized poissonboltzmann theory. J Colloid Interface Sci 216
(2):329–347. doi:10.1006/jcis.1999.6326
50. Schowalter WR, Eidsath AB (2001) Brownian flocculation of polymer colloids in the
presence of a secondary minimum. Proc Natl Acad Sci USA 98:3644–3651. doi:10.1073/
pnas.061028498
51. Podgornik R, Harries D, DeRouchey J, Strey HH, Parsegian VA (2008) Interactions in
macromolecular complexes used as nonviral vectors for gene delivery. In: Gene and cell
therapy: therapeutic mechanisms and strategies. CRC, Burlington
52. Morales V, Anta JA, Lago S (2003) Integral equation prediction of reversible coagulation in
charged colloidal suspensions. Langmuir 19:475–482
53. de Gennes P (1981) Polymer solutions near an interface. 1. Adsorption and depletion layers.
Macromolecules 14:1637–1644
54. de Gennes P (1982) Polymer solutions near an interface. 2. Interaction between two plates
carrying adsorbed polymer layers. Macromolecules 15:492–500
55. Runkana V, Somasundaran P (2007) Mathematical modeling of coagulation and flocculation
of colloidal suspensions incorporating the influence of surface forces. In: Colloid stability and
application in pharmacy. Colloid and interface science series, vol 3. Wiley-VCH, Weinheim
56. Runkana V, Somasundaran P, Kapur P (2006) A population balance model for flocculation of
colloidal suspensions by polymer bridging. Chem Eng Sci 61:182–191
57. Somasundaran P, Runkana V (2009) Aggregation of colloids: recent developments in popu-
lation balance modeling. In: Highlights in colloid science. Wiley-VCH, Weinheim
58. Grier DG, Behrens SH (2001) Interactions in colloidal suspensions: electrostatics, hydrody-
namics and their interplay. In: Electrostatic effects in biophysics and soft matter. Kluwer,
Dordrecht
59. Derjaguin B, Muller V (1967) Slow coagulation of hydrophobic colloids. Dokl Akad Nauk
SSSR 176:738–741
60. Spielman LA (1970) Viscous interactions in brownian coagulation. J Colloid Interface Sci
33(4):562–571. doi:10.1016/0021-9797(70)90008-1
61. Brenner H (1961) The slow motion of a sphere through a viscous fluid towards a plane
surface. Chem Eng Sci 16(34):242–251. doi:10.1016/0009-2509(61)80035-3
62. Honig E, Roebersen G, Wiersema P (1971) Effect of hydrodynamic interaction on the
coagulation rate of hydrophobic colloids. J Colloid Interface Sci 36(1):97–109.
doi:10.1016/0021-9797(71)90245-1
63. Kovalchuk N, Starov V (2011) Aggregation in colloidal suspensions: effect of colloidal
forces and hydrodynamic interactions. Adv Colloid Interface Sci. doi:10.1016/j.
cis.2011.05.009
64. Cao X, Cummins H, Morris J (2012) Hydrodynamic and interparticle potential effects on
aggregation of colloidal particles. J Colloid Interface Sci 368(1):86–96. doi:10.1016/j.
jcis.2011.11.050
65. Furukawa A, Tanaka H (2010) Key role of hydrodynamic interactions in colloidal gelation.
Phys Rev Lett 104:245702. doi:10.1103/PhysRevLett. 104.245702
66. Riese DO, Wegdam GH, Vos WL, Sprik R, Fenistein D, Bongaerts JH, Gr€ ubel G (2000)
Effective screening of hydrodynamic interactions in charged colloidal suspensions. Phys Rev
Lett 85(25):5460–5464
Aggregation of Charged Colloidal Particles 93

67. Muthukumar M (1997) Dynamics of polyelectrolyte solutions. J Chem Phys 107

(7):2619–2635. doi:10.1063/1.474573
68. Muthukumar M (2005) Polyelectrolyte dynamics. In: Rice SA (ed) Advances in chemical
physics, vol 131. Wiley, Hoboken
69. Arunachalam V, Marlow WH, Lu JX (1998) Development of a picture of the van der Waals
interaction energy between clusters of nanometer-range particles. Phys Rev E 58:3451–3457
70. Babick F, Schiel K, Stintz M (2011) Van-der-Waals interaction between two fractal
aggregates. Adv Powder Technol 22(2):220–225. doi:10.1016/j.apt.2010.11.014
71. Schiesl K, Babick F, Stintz M (2012) Calculation of double layer interaction between
colloidal aggregates. Adv Powder Technol 23(2):139–147. doi:10.1016/j.apt.2011.01.005
72. Vicsek T (1992) Fractal growth phenomena. Word Scientific, Singapore
73. Smoluchowski M (1917) Uber brownsche molekularbewegung unter einwirkung auserer
kruafte und deren zusammenhang mit der verallgemeinerten diffusions- gleichung. Ann
Phys-Leipzig 48:1103–1112
74. Meakin P (1999) A historical introduction to computer models for fractal aggregates.
J Sol–gel Sci Technol 15:97–117
75. Witten TA, Sander LM (1981) Diffusion-limited aggregation, a kinetic critical phenomenon.
Phys Rev Lett 47:1400–1403
76. Sutherland D (1966) Comment on Vold’s simulation of floc formation. J Colloid Interface Sci
22:300
77. Sutherland DN (1967) A theoretical model of floc structure. J Colloid Interface Sci
25:373–380
78. Vold M (1963) Computer simulation of floe formation in a colloidal suspension. J Colloid Sci
18:684–695
79. Eden M (1961) A two-dimensional growth process. In: Proceedings of the Fourth Berkeley
Symposium on Mathematics, Statistics and Probability, vol 4: Biology and Problems of
Health. University of California Press, Berkeley
80. Jullien R, Botet R (1987) Aggregation and fractal aggregation. World Scientific, Singapore
81. Aubert C, Cannell DS (1986) Restructuring of colloidal silica aggregates. Phys Rev Lett
56:738–741. doi:10.1103/PhysRevLett.56.738
82. Liu J, Shih WY, Sarikaya M, Aksay IA (1990) Fractal colloidal aggregates with finite
interparticle interactions: energy dependence of the fractal dimension. Phys Rev A
41:3206–3213. doi:10.1103/PhysRevA.41.3206
83. Jia Z, Wu H, Morbidelli M (2007) Thermal restructuring of fractal clusters: the case of a
strawberry-like core-shell polymer colloid. Langmuir 23:5713–5721. doi:10.1021/la063254s
84. Jullien R, Meakin P (1989) Simple models for the restructuring of three-dimensional ballistic
aggregates. J Colloid Interface Sci 127(1):265–272. doi:10.1016/0021-9797(89)90027-1
85. Shih WY, Aksay IA, Kikuchi R (1987) Reversible-growth model: cluster-cluster aggregation
with finite binding energies. Phys Rev A 36:5015–5019. doi:10.1103/PhysRevA.36.5015
86. Rioux C, Slobodrian RJ (2012) Experimental discrimination of electrostatic and magnetic
forces in particle-particle aggregation. Adv Space Res 49(10):1408–1414
87. Groenewold J, Kegel WK (2004) Colloidal cluster phases, gelation and nuclear matter. J Phys
Condens Matter 16:S4877–S4886
88. Sciortino F, Mossa S, Zaccarelli E, Tartaglia P (2004) Equilibrium cluster phases and low-
density arrested disordered states: the role of short-range attraction and long-range repulsion.
Phys Rev Lett 93:055701
89. Rayleigh L (1882) On the equilibrium of liquid conducting masses charged with electricity.
Philos Mag 14:184–186
90. Smirnov BM (2006) Cluster processes in gases and plasmas. Distributions, structures,
phenomena, kinetics of atomic systems. Wiley-VCH, Weinheim
91. Weizsacker CFV (1935) Zur theorie der kernmassen. Z Phys 96:431–458
92. Lu PJ, Conrad JC, Wyss HM, Schofield AB, Weitz DA (2006) Fluids of clusters in attractive
colloids. Phys Rev Lett 96:028306. doi:10.1103/PhysRevLett.96.028306
94 N.I. Lebovka

93. Lu PJ, Zaccarelli E, Ciulla F, Schofield AB, Sciortino F, Weitz DA (2008) Gelation of
particles with short-range attraction. Nature 453:499–503. doi:10.1038/nature06931
94. de Gennes P, Pincus P, Velasco R, Brochard F (1976) Remarks on polyelectrolyte conforma-
tion. J Phys-Paris 37(12):1461–1473
95. Brender C, Danino M, Shatz S (1999) Fractals in Monte Carlo simulations of a short
polyelectrolyte. J Phys A Math Gen 32(2):235
96. Fernandez-Toledano JC, Moncho-Jorda A, Martinez-Lopez F, Gonzalez AE, Hidalgo-
Alvarez R (2007) Two-dimensional colloidal aggregation mediated by the range of repulsive
interactions. Phys Rev E 75:041408. doi:10.1103/PhysRevE.75.041408
97. Mossa S, Sciortino F, Tartaglia P, Zaccarelli E (2004) Ground-state clusters for short-range
attractive and long-range repulsive potentials. Langmuir 20(24):10756–10763. doi:10.1021/
la048554t
98. Sciortino F, Tartaglia P, Zaccarelli E (2005) One-dimensional cluster growth and branching
gels in colloidal systems with short-range depletion attraction and screened electrostatic
repulsion. J Phys Chem B 109(46):21942–21953. doi:10.1021/jp052683g
99. Chakrabarty RK, Moosmuller H, Garro MA, Arnott WP, Slowik JG, Cross ES, Jeong-Ho Han
PD, Onasch TB, Worsnop DR (2008) Morphology based particle segregation by electrostatic
charge. J Aerosol Sci 39(9):785–792
100. Block A, van Blah W, Schellnhuber HJ (1991) Aggregation by attractive particle-cluster
interaction. J Phys A Math Gen 24:L1037–L1044
101. Indiveri G, Scalas E, Levi A, Gliozzi A (1999) Morphologies in two-dimensional growth with
attractive long-range interactions. Physica A 273(3–4):217–230. doi:10.1016/S0378-4371
(99)00231-9
102. Meakin P (1990) The effects of attractive and repulsive interactions on three-dimensional
reaction-limited aggregation. J Colloid Interface Sci 134(1):235–244. doi:10.1016/0021-
9797(90)90271-O
103. Meakin P, Muthukumar M (1989) The effects of attractive and repulsive interaction on two-
dimensional reaction-limited aggregation. J Chem Phys 91(5):3212–3221. doi:10.1063/
1.456942
104. Indiveri G, Levi A, Gliozzi A, Scalas E (1996) Cluster growth with long-range interactions.
Thin Solid Films 284–285:106–109
105. Ivanenko Y, Lebovka N, Vygornitskii N (1999) Eden growth model for aggregation of
charged particles. Eur Phys J B 11:469–480
106. Lebovka NI, Ivanenko YV, Vygornitskii NV (1998) Deterministic eden model of charged-
particles aggregation. Europhys Lett 41(1):19
107. Pinchuk AO, Kalsin AM, Kowalczyk B, Schatz GC, Grzybowski BA (2007) Modeling of
electrodynamic interactions between metal nanoparticles aggregated by electrostatic
interactions into closely-packed clusters. J Phys Chem C 111(32):11816–11822.
doi:10.1021/jp073403v
108. Zhang R, Shklovskii B (2005) Phase diagram of solution of oppositely charged polyelec-
trolytes. Physica A 352(1):216–238. doi:10.1016/j.physa.2004.12.037
109. Harnau L, Hansen JP (2002) Colloid aggregation induced by oppositely charged polyions.
J Chem Phys 116(20):9051–9057. doi:10.1063/1.1471550
110. Puertas A, Fernandez-Barbero A, De las Nieves F (2000) Aggregation between oppositely
charged colloidal particles. In: Buckin V (ed) Trends in colloid and interface science XIV.
Progress in colloid and polymer science, vol 115. Springer, Berlin, pp 55–58
111. Puertas A, Fernandez-Barbero A, de las Nieves F (2002) Kinetics of colloidal heteroag-
gregation. Physica A 304(34):340–354. doi:10.1016/S0378-4371(01)00564-7
112. Kim AY, Hauch KD, Berg JC, Martin JE, Anderson RA (2003) Linear chains and chain-like
fractals from electrostatic heteroaggregation. J Colloid Interface Sci 260(1):149–159.
doi:10.1016/S0021-9797(03)00033-X
Aggregation of Charged Colloidal Particles 95

113. Cerbelaud M, Videcoq A, Abelard P, Pagnoux C, Rossignol F, Ferrando R (2008) Heteroag-

gregation between Al2O3 submicrometer particles and SiO2 nanoparticles: experiment and
simulation. Langmuir 24(7):3001–3008. doi:10.1021/la702104u
114. Cerbelaud M, Videcoq A, Abelard P, Ferrando R (2009) Simulation of the heteroagglo-
meration between highly size-asymmetric ceramic particles. J Colloid Interface Sci 332
(2):360–365. doi:10.1016/j.jcis.2008.11.063
115. Piechowiak MA, Videcoq A, Ferrando R, Bochicchio D, Pagnoux C, Rossignol F (2012)
Aggregation kinetics and gel formation in modestly concentrated suspensions of oppositely
charged model ceramic colloids: a numerical study. Phys Chem Chem Phys 14:1431–1439.
doi:10.1039/C1CP22980J
116. Castelnovo M, Sens P, Joanny JF (2000) Charge distribution on annealed polyelectrolytes.
Eur Phys J E 1:115–125
117. Coslovich D, Hansen J, Kahl G (2011) Ultrasoft primitive model of polyionic solutions:
structure, aggregation, and dynamics. J Chem Phys 134(24):244514 (15 pages), www.scopus.
com
118. Coslovich D, Hansen JP, Kahl G (2011) Clustering, conductor-insulator transition and phase
separation of an ultrasoft model of electrolytes. Soft Matter 7:1690–1693. doi:10.1039/
C0SM01090A
119. Lee J, Popov YO, Fredrickson GH (2008) Complex coacervation: a field theoretic simulation
study of polyelectrolyte complexation. J Chem Phys 128(22):224908. doi:10.1063/1.2936834
120. Rydén J, Ullner M, Linse P (2005) Monte Carlo simulations of oppositely charged macroions
in solution. J Chem Phys 123(3):034909. doi:10.1063/1.1949191
121. Zito T, Seidela C (2002) Equilibrium charge distribution on annealed polyelectrolytes. Eur
Phys J E 8:339–346
122. Buchhammer HM, Mende M, Oelmann M (2003) Formation of mono-sized polyelectrolyte
complex dispersions: effects of polymer structure, concentration and mixing conditions.
Colloids Surf A 218(1):151–159. doi:10.1016/S0927-7757(02)00582-4
123. Buchhammer HM, Mende M, Oelmann M (2004) Preparation of monodisperse polyelectro-
lyte complex nanoparticles in dilute aqueous solution. In: Tauer K (ed) Aqueous polymer
dispersions. Progress in colloid and polymer science, vol 124. Springer, Berlin, pp 98–102
124. Dobrynin AV (2008) Theory and simulations of charged polymers: from solution properties
to polymeric nanomaterials. Curr Opin Colloid Interface Sci 13(6):376–388. doi:10.1016/j.
cocis.2008.03.006
125. Oskolkov NN, Potemkin II (2007) Complexation in asymmetric solutions of oppositely
charged polyelectrolytes:phase diagram. Macromolecules 40(23):8423–8429. doi:10.1021/
ma0709304
126. Popov YO, Lee J, Fredrickson G (2007) Field-theoretic simulations of polyelectrolyte
complexation. J Polym Sci Pol Phys 45:3223–3230
127. Dalakoglou G, Karatasos K, Lyulin S, Lyulin A (2008) Brownian dynamics simulations of
complexes of hyperbranched polymers with linear polyelectrolytes: effects of the strength of
electrostatic interactions on static properties. Mat Sci Eng B-Solid 152(1–3):114–118.
doi:10.1016/j.mseb.2008.06.012
128. Skepo M, Linse P (2003) Complexation, phase separation, and redissolution in polyelec-
trolytemacroion solutions. Macromolecules 36:508–519
129. Feng J, Ruckenstein E (2003) Monte Carlo simulation of polyampholyte-nanoparticle com-
plexation. Polymer 44(10):3141–3150. doi:10.1016/S0032-3861(03)00208-8
130. Ulrich S, Seijo M, Carnal F, Stoll S (2011) Formation of complexes between nanoparticles
and weak polyampholyte chains. Monte Carlo simulations. Macromolecules 44:1661–1670
131. Sennato S, Truzzolillo D, Bordi F, Sciortino F, Cametti C (2009) Colloidal particle
aggregates induced by particle surface charge heterogeneity. Colloids Surf A 343
(1–3):34–42. doi:10.1016/j.colsurfa.2009.01.026
132. Jeon J, Dobrynin AV (2005) Molecular dynamics simulations of polyampholyte-
polyelectrolyte complexes in solutions. Macromolecules 38:5300–5312
96 N.I. Lebovka

133. Xu Y, Feng J, Liu H, Hu Y, Jiang J (2007) Molecular dynamics simulation of polyelectrolyte

with oppositely charged monomeric and dimeric surfactants. Mol Simul 33(3):261–268.
doi:10.1080/08927020601158679
134. Jullien R, Botet R, Mors PM (1987) Computer simulations of cluster-cluster aggregation.
Faraday Discuss Chem Soc 83:125–137. doi:10.1039/DC9878300125
135. Mors PM, Botet R, Jullien R (1987) Cluster-cluster aggregation with dipolar interactions.
J Phys A Math Gen 20:L975
136. Pastor-Satorras R, Rubi JM (1995) Particle-cluster aggregation with dipolar interactions.
Phys Rev E 51:5994–6003
137. Pastor-Satorras R, Rubi R (1998) Fractal properties of cluster of colloidal magnetic particles.
Prog Colloid Polym Sci 110:29–33
138. Pastor-Satorras R, Rubi J (2000) Dipolar interactions induced order in assemblies of magnetic
particles. J Magn Magn Mater 221:124–131
139. Family F (1985) Kinetics of aggregation and gelation. Elsevier, Amsterdam
140. Meakin P (1992) Aggregation kinetics. Phys Scripta 46(4):295
141. Meakin P (1998) Fractals, scaling and growth far from equilibrium. Cambridge University
Press, Cambridge
142. Fuchs N (1934) Ueber die stabilitat und aufladung der aerosole. Z Phys 89:736–743
143. Reerink H, Overbeek JTG (1954) The rate of coagulation as a measure of the stability of
silver iodide sols. Faraday Discuss Chem Soc 18:74–84
144. Leyvraz F (2003) Scaling theory and exactly solved models in the kinetics of irreversible
aggregation. Phys Rep 383:95–212
145. McLeod JB (1962) On an infinite set of non-linear differential equations. Q J Math
13:119–128
146. Melzak ZA (1953) The effects of coalescence in certain collision processes. Q J Mech Appl
Math 11:231–234
147. Gmachowski L (2000) Estimation of the dynamic size of fractal aggregates. Colloids Surf A
170(23):209–216. doi:10.1016/S0927-7757(99)00532-4
148. Zift RM, McGrady ED, Meakin P (1985) On the validity of Smoluchowski’s equation for
cluster-cluster aggregation kinetics. J Chem Phys 82:5269–5274
149. Fernandez-Barbero A, Cabrerizo-Vilchez M, Martinez-Garcia R, Hidalgo-Alvarez R (1996)
Effect of the particle surface charge density on the colloidal aggregation mechanism. Phys
Rev E 53:4981–4989. doi:10.1103/PhysRevE.53.4981
150. Asnaghi D, Carpineti M, Giglio M, Sozzi M (1992) Coagulation kinetics and aggregate
morphology in the intermediate regimes between diffusion-limited and reaction-limited
cluster aggregation. Phys Rev A 45:1018–1023. doi:10.1103/PhysRevA.45.1018
151. Lattuada M, Sandkuhler P, Wu H, Sefcik J, Morbidelli M (2003) Aggregation kinetics of
polymer colloids in reaction limited regime: experiments and simulations. Adv Colloid
Interface Sci 103:33–56
152. Odriozola G, Moncho-Jorda A, Schmitt A, Callejas-Fernandez J, Martinez-Garcia R,
Hidalgo-Alvarez R (2001) A probabilistic aggregation kernel for the computer-simulated
transition from DLCA to RLCA. Europhys Lett 53:797803
153. Runkana V, Somasundaran P, Kapur PC (2005) Reaction-limited aggregation in presence of
short-range structural forces. AICHE J 51:1233–1245
154. Starchenko V, Muller M, Lebovka N (2008) Growth of polyelectrolyte complex
nanoparticles: computer simulations and experiments. J Phys Chem C 112:8863–8869
155. Bianchi E, Blaak R, Likos CN (2011) Patchy colloids: state of the art and perspectives. Phys
Chem Chem Phys 13:6397–6410. doi:10.1039/C0CP02296A
156. Gruy F (2011) Population balance for aggregation coupled with morphology changes.
Colloids Surf A 374(13):69–76. doi:10.1016/j.colsurfa.2010.11.010
http://www.springer.com/978-3-642-40733-8