Marques2013 PDF

S
Sedimentation decreasing exponentially with increase of height

from the bottom of the container. The sedimen-
Tharwat Tadros tation of very dilute suspensions with f 0.01,
Wokingham, Berkshire, UK the rate of sedimentation vo, is given by Stokes’
law: vo ¼ (2/9) R2Drg/Z, where Z is the viscosity
of the medium. At higher volume fractions 0.2 >
Synonyms f > 0.01, one must take into account the hydro-
dynamic interaction between the particles, and
Sedimentation of particles in suspensions the rate of sedimentation v is reduced below the
Stokes’ rate: v ¼ vo (1kf), where k is a constant
that takes into account the hydrodynamic inter-
Definition action, and it is equal to 6.55. At higher volume
fractions f > 0.2, the rate of sedimentation
Sedimentation is a process whereby the particles v becomes a complex function of f.
of a suspension that have density larger than the V decreases exponentially with increase in f,
medium in which they are dispersed settle under and it approaches 0 when f approaches the max-
the influence of gravity. This is particularly the imum packing fraction fp. The latter is equal to
case with particles of radius R > 1 mm when the 0.74 for hexagonal close packing of equal-sized
density difference between the particles and particles, and it is equal to 0.64 for random pack-
medium Dr is > 0.1. When the gravity force ing. For polydisperse systems, fp > 0.74. When
that is (4/3)R3DrgL (where g is the acceleration f fp, the viscosity of the suspension reaches
due to gravity and L is the height of the container) very high values. To reduce sedimentation, one
is higher than the Brownian motion kT (where must add rheology modifiers (“thickeners”) in the
k is the Boltzmann constant and T is the absolute continuous phase that produce a very high
temperature), the particles will settle to the viscosity at very low shear rates (residual or
bottom of the container and form a compact zero shear viscosity).
sediment (“clay” or “cake”) that is dilatant
(shear thickening). For a polydisperse suspension
(that is mostly the case in practice), the larger
particles will settle at a faster rate than the smaller Cross-References
ones. A concentration gradient is produced with
the volume fraction of the suspension f ▶ Suspensions
T. Tadros (ed.), Encyclopedia of Colloid and Interface Science,

DOI 10.1007/978-3-642-20665-8, # Springer-Verlag Berlin Heidelbergk 2013
S 1044 Sedimentation of Particles in Suspensions
structures. The hexagonal phase is built up of

Sedimentation of Particles in (infinitely) long cylindrical micelles arranged in
Suspensions a hexagonal pattern, with each micelle being
surrounded by six other micelles. The radius of
▶ Sedimentation the circular cross section (which may be some-
what deformed) is again close to the surfactant
molecule length. The micellar cubic phase is
built up of regular packing of small micelles,
Self-Assembly of Surfactants which have similar properties of small micelles
in the solution phase. However, the micelles are
Tharwat Tadros short prolates (axial ratio 1:2) rather than
Wokingham, Berkshire, UK spheres since this allows a better packing. The
micellar cubic phase is highly viscous. The
lamellar phase is built of bilayers of surfactant
Synonyms molecules alternating with water layers. The
thickness of the bilayers is somewhat lower
Surfactant self-assembly than twice the surfactant molecule length. The
thickness of the water layer can vary over wide
ranges, depending on the nature of the surfac-
Definition tant. The surfactant bilayer can range from
being stiff and planar to being very flexible
Surfactant self-assembly structures are building and undulating. The discontinuous cubic phases
blocks of surfactant micelles and bilayers. One can be a number of different structures, where
can divide the phase structures into two main the surfactant molecules form aggregates that
groups: (1) those that are built of limited or penetrate space, forming a porous connected
discrete self-assemblies, which may be charac- structure in three dimensions. They can be con-
terized roughly as spherical, prolate, or cylin- sidered as structures formed by connecting rod-
drical, and (2) infinite or unlimited self- like micelles (branched micelles) or bilayer
assemblies whereby the aggregates are structures.
connected over macroscopic distances in one,
two, or three dimensions. The hexagonal phase
is an example of one-dimensional continuity, Cross-References
the lamellar phase of two-dimensional continu-
ity, whereas the bicontinuous cubic phase and ▶ Liquid Crystalline Phase
the sponge phase are examples of three- ▶ Self-Assembly Structures
dimensional continuity. These unlimited self- ▶ Surfactants
assembly structures in 1D, 2D, or 3D are
referred to as liquid crystalline structures. The
latter behave as fluids and are usually highly
viscous. At the same time, X-ray studies of Self-Assembly Structures
these phases yield a small number of relatively
sharp lines which resemble those produced by ▶ Liquid Crystalline Phase
crystals. Since they are fluids, then they are less
ordered than crystals, but because of the X-ray
lines and their high viscosity, it is also apparent
that they are more ordered than ordinary liquids. Self-Organization of Amphiphiles
Thus, the term liquid crystalline phase is very
appropriate for describing these self-assembled ▶ Surfactant Self-Assembly
Smoke 1045 S
Cross-References
Shear Rate
▶ Rheological Techniques
Tharwat Tadros ▶ Rheology
Wokingham, Berkshire, UK
Synonyms Shear Rate Given in Reciprocal

Seconds
Shear rate given in reciprocal seconds
▶ Shear Rate
Definition
The shear rate or strain rate is a measure of the Short-Range Interactions

change of deformation (strain) with time, and it is
given in reciprocal seconds. In simple shear ▶ Specific Surface Interactions
strain, successive layers of material move in
their own planes, relative to a reference layer.
The displacement of a layer is proportional to its
distance from the reference layer (similar to Smoke
a deck of cards that can only be deformed side-
ways). The strain (relative deformation) is given Tharwat Tadros
by the ratio of displacement of a plane dl to the Wokingham, Berkshire, UK
distance between the planes l:
dl Synonyms
Strain g ¼ ; g is dimensionless and
l
dl Smoke or solid in gas dispersion
tan g ¼ :
l
When g is very small, tan g g. The shear Definition

strain is simply the angle of shear in radians. The
shear rate is simply given by the change of strain A smoke is a dispersion of solid particles in gas
with time. It has the unit s1: (mostly air). Smokes are by-products of many S
industrial processes such as coal and oil power
dg stations where the burning fuel leaves particles
g¼ :
dt that can be dispersed in air. A smoke can be
visible at the outlet of a chimney, and it can
Control of shear rate is important in steady state cause pollution and becomes a health hazard.
measurements. Measurement of viscosity as The particles of the smoke can enter the respira-
a function of shear rate is achieved when a steady tory system and become deposited on the lung
state is reached. For a concentric cylinder, the tissue causing in some cases lung cancer. It is,
shear rate can be calculated by assuming that the therefore, important to remove smokes in indus-
fluid elements move in circles about the common trial processes by using filters and/or application
axis of the cylinders with an angular velocity o of an electric field that can aggregate the parti-
which is a function of radius r from the center to cles, thus facilitating their removal. It is essential
the position in the gap (where the stress is s). to understand the structure of the solid/air
S 1046 Smoke or Solid in Gas Dispersion
interface in order to be able to remove the parti- Definition

cles from the air stream. It is also essential to
measure the particle size distribution of the Solubilization is the process of preparation of
smoke using light diffraction techniques. The thermodynamically stable isotropic solution of
particle size, shape, and distribution determine a substance (normally insoluble or sparingly sol-
the possible inhalation of the particles in the uble in a given solvent) by incorporation of an
respiratory system. additional amphiphilic component(s). It is the
incorporation of the compound (referred
to as solubilizate or substrate) within micellar
Cross-References (L1 phase) or reverse micellar (L2 phase) system.
The L1 and L2 can be described by considering
▶ Disperse Systems the phase diagram of a ternary system of water-
▶ Particle Size-Solubility Relation surfactant-cosurfactant system, for example,
water-ionic sulfate-long-chain alcohol system.
The aqueous micellar solution A solubilizes
some alcohol to form normal micelles (L1),
Smoke or Solid in Gas Dispersion whereas the alcohol solution B dissolves large
amounts of water forming inverse micelles (L2).
▶ Smoke
These two phases are not in equilibrium with
each other but are separated by a third
region, namely, the lamellar liquid crystalline
Solid/Liquid Dispersions phase (La phase). Lipophilic (water-insoluble)
substances become incorporated in the L1
▶ Suspension Concentrates (normal micelle) phase. Hydrophilic (water-sol-
uble) substances are incorporated in the L2 phase.
The lamellar liquid crystalline phase is also capa-
ble of incorporation of solubilizates. This also
Solid-Liquid Interface applies to the hexagonal (middle phase) that
may be present in concentrated nonionic surfac-
▶ Interface, Solid-Liquid tant systems. The site of incorporation of the
solubilizate is closely related to its structure:
Nonpolar solubilizate in the hydrocarbon core
Solubilization and semipolar or polar solubilizate oriented
within the micelle (short or deep). The concen-
▶ Solubilization and Solubilized Systems tration of solubilizate that can be incorporated
into a given system with the maintenance of
a single isotropic solution (saturation concen-
tration or maximum additive concentration,
Solubilization and Solubilized MAC) is obtained using the same procedures
Systems for measurement of solubility of any compound
in a given solvent. Since solubilization is tem-
Tharwat Tadros perature sensitive, adequate temperature control
Wokingham, Berkshire, UK is essential. If the refractive indices of solubi-
lizing system and solubilizate are sufficiently
different, saturation is detected by the presence
Synonyms of supracolloidal aggregates with a concomitant
increase on the opacity. Long time may be
Solubilization required to reach equilibrium saturation,
Stability Ratio 1047 S
particularly with highly insoluble drugs. An the particle/particle and particle/interface sys-
excess of solubilizate is shaken up with the tems. They consist of the electrostatic interac-
surfactant solution until equilibrium is reached, tions due to free charges present on interacting
and the two phases could be separated by cen- surfaces and van der Waals interactions. The
trifugation or using Millipore filters. The data latter are the net effects of dipolar attraction
are best expressed as concentration of between all atoms or molecules present in the
solubilizate versus concentration of surfactant macroscopic objects. The specific interactions
or as ratio of solubilizate dissolved per gram of govern aggregation kinetics (colloid suspension
surfactant versus surfactant concentration. It is stability) and deposition phenomena. Contrary to
also possible to express the results using intuition, the surface interactions scale up as the
a ternary phase diagram of solubilizate-sol- size of the particle rather than particle surface
vent-surfactant. area. They also increase with the density of
interacting materials, being especially high for
metallic particles or interfaces. Moreover for the
Cross-References same materials, the specific interactions are twice
as large for the particle/planar interface system
▶ Phase Behavior of Surfactants than the particle/particle system.
▶ Surfactants
Cross-References
Specific Surface Interactions ▶ Deposition

▶ Particle Deposition
Zbigniew Adamczyk
Polish Academy of Sciences, J. Haber Institute of
Catalysis and Surface Chemistry, Kraków,
Poland Stability
▶ Foam
Synonyms
Colloid forces; Colloid interactions; Dispersion

forces; Short-range interactions; Surface Stability Ratio
interactions; van der Waals forces
Tharwat Tadros S
Keywords
Dispersion interactions; electric double-layer Synonyms

interactions; electrostatic interactions; van der
Waals forces Colloid stability
Definition Definition
The specific surface interactions appear univer- The stability ratio W is the ratio of the fast floc-
sally at separations 1–100 nm for all material culation rate ko to the slow flocculation rate k. In
bodies brought into close contact, especially for the absence of an energy barrier, the rate of
S 1048 Stabilization by Polymers
flocculation (Smoluchowski rate) is diffusion

controlled and the process is represented by Stabilization of Dispersions
second-order kinetics. The number of particles
n at any time t may be related to the initial number ▶ Dispersion Stability
(at t ¼ 0) no by n ¼ 1þknoo no t , where k is the rate
constant for fast flocculation that is related
to the diffusion coefficient of the particles
D: k ¼ 8pDR. D is given by Steric Stabilization
kT 4 kT Tharwat Tadros
D¼ k¼ Wokingham, Berkshire, UK
6pZR 3 Z
¼ 5:5 1018 m3 s1 for water at 25 C:
Synonym
The slow flocculation kinetics was treated by
Fuchs who related the rate constant k to the Stabilization by polymers
Smoluchowski rate by the stability constant
W ¼ kko .

W is related to Gmax by W ¼ 12 k exp GkT max
. Definition
Since Gmax is determined by the salt concentra-
tion C and valency Z, one can derive an expression Steric stabilization is the process by which
relating W to C and Z. Plots of log W versus log C adsorbed nonionic surfactants or polymers produce
show the condition log W ¼ 0 (W ¼ 1) that is the strong repulsion between particles and droplets in
onset of fast flocculation. The electrolyte concen- a dispersion. The adsorption of nonionic surfac-
tration at this point defines the critical flocculation tants or polymers on the surface of particles or
concentration c.c.c. Above the c.c.c., W < 1 droplets produces an adsorbed layer with thickness
(due to the contribution of van der Waals attraction d which can be strongly solvated (or hydrated) by
which accelerates the rate above the the molecules of the solvent when the latter is
Smoluchowski value). Below the c.c.c., W > 1 a good solvent for the surfactant or polymer layer.
and it increases with decrease of electrolyte con- When two particles or droplets with radius
centration. The plots show that the c.c.c. decreases R approach to a surface-to-surface distance of sep-
with increase of valency. At low surface potentials, aration h that is smaller than 2d, the adsorbed
c.c.c. 1 1/Z2. This is referred to as the layers can overlap or become compressed,
Schulze-Hardy rule. resulting in strong repulsion as a result of two
main effects: (a) Unfavorable mixing of the
adsorbed layers when these are in good solvent
Cross-References conditions (this is referred to as osmotic repulsion
or mixing interaction, Gmix) and (b) reduction of
▶ Colloid Stability configurational entropy of the adsorbed layers on
▶ Flocculation considerable overlap (this is referred to as entropic
▶ Suspensions or elastic repulsion, Gel). The steric repulsion Gs is
the sum of Gmix and Gel. A plot of Gs versus
separation distance h shows sharp increase when
h < 2d, and this is the origin of steric stabilization.
The criteria for effective steric stabilization are the
Stabilization by Polymers following: (a) The particles should be completely
covered by the polymer (the amount of polymer
▶ Steric Stabilization should correspond to the plateau value). Any bare
Stern-Langmuir Isotherm and Theory 1049 S
patches may cause flocculation either by van der The potential drops linearly in the Stern region
Waals attraction (between the bare patches) or by and then exponentially. Grahame distinguished
bridging flocculation (whereby a polymer two types of ions in the Stern plane, physically
moleculae will become simultaneously adsorbed adsorbed counterions (outer Helmholtz plane)
on two or more particles). (b) The polymer should and chemically adsorbed ions (that lose part of
be strongly “anchored” to the particle surfaces to their hydration shell) (inner Helmholtz plane).
prevent any displacement during particle approach. For the specifically adsorbed ions, the range of
This is particularly important for concentrated sus- interaction is short, that is, these ions must reside
pensions. For this purpose, A-B, A-B-A block, and at the distance of closest approach, possibly
BAn graft copolymers are the most suitable, within the hydration shell. For the indifferent
where the chain B is chosen to be highly insolu- ions, the situation is different, and these ions are
ble in the medium and has a strong affinity to the subjected to an attractive (for the counterions)
surface. Examples of B groups for hydrophobic or repulsive (for the co-ions) potential
particles in aqueous media are polystyrene and (energy ¼ ZFc(x)/RT). The space charge den-
polymethyl methacrylate. (c) The stabilizing sity due to these ions is high near the surface and
chain A should be highly soluble in the medium decreases gradually with distance to its bulk
and strongly solvated by its molecules. Exam- value. Such a layer is the diffuse double layer
ples of A chains in aqueous media are poly(eth- described by Gouy-Chapman. Generally speak-
ylene oxide) and poly(vinyl alcohol). (d) d ing, a double layer contains a part that is specif-
should be sufficiently large (>5 nm) to prevent ically adsorbed and a diffuse part. Because of the
weak flocculation. finite size of the counterions, there is always
a charge-free layer near the surface.
Cross-References
Cross-References
▶ Nanosuspensions
▶ Polymer Adsorption ▶ Electrical Double Layers
▶ Polymeric Surfactant ▶ Suspensions
▶ Surfactants
Stern-Langmuir Isotherm and

Stern Picture Theory
Tharwat Tadros Tharwat Tadros S

Wokingham, Berkshire, UK Wokingham, Berkshire, UK
Synonyms Synonyms
Double layer Stern-Langmuir theory for surfactant adsorption
Definition Definition
Stern introduced the concept of the non- The Stern-Langmuir theory describes the adsorp-
diffuse part of the double layer for specifically tion of surfactants at the solid/liquid interface,
adsorbed ions, the rest being diffuse in nature. and it gives an isotherm relating the surface
S 1050 Stern-Langmuir Theory for Surfactant Adsorption
coverage to surfactant adsorption and free energy Cross-References

of adsorption. The theory applies to the case of
reversible and equilibrium adsorption of surfac- ▶ Surfactant Adsorption
tants at the solid/liquid interface. At equilibrium, ▶ Surfactants
the rate of adsorption is equal to the rate of
desorption, and the ratio of the two is the equilib-
rium constant K: ð1yÞ y
¼ 55:5
C
K, where y is the
fraction of surface coverage that is given by the Stern-Langmuir Theory for
ratio of the amount of surfactant-adsorbed G Surfactant Adsorption
(mole m2) to the number of moles of active
sites Ns (mole m2), C is the surfactant concen- ▶ Stern-Langmuir Isotherm and Theory
tration (mole dm3), C/55.51 is the mole fraction
of surfactant, and K is the equilibrium constant
that is related to the standard free energy of

D Go Stochastic Particle Motion
adsorption: K ¼ exp RTads .
The Langmuir-Stern isotherm is
O
▶ Diffusion of Particles
D G
1y ¼ 55:5 exp RT
y C ads
which applies only at
low surface coverage (y < 0.1) where lateral
interaction between the surfactant ions can be
neglected. Various contributions to the adsorp- Stress Relaxation
tion free energy may be envisaged. To a first
approximation, these contributions may be con- Tharwat Tadros
sidered to be additive. In the first instance, DGads Wokingham, Berkshire, UK
may be taken to consist of two main contribu-
tions: DGelec which accounts for any electrical
interactions and DGspec is a specific adsorption Synonyms
term which contains all contributions to the
adsorption free energy that are dependent on the Relaxation
“specific” (nonelectrical) nature of the system.
Several authors subdivided DGspec into suppos-
edly separate independent interactions: DGcc Definition
a term that accounts for the cohesive chain-
chain interaction between the hydrophobic moi- Stress relaxation is the variation of stress with
eties of the adsorbed ions, DGcs a term for chain/ time after sudden application of strain g. The
substrate interaction, and DGhs a term for the latter is applied within a very short period (i.e.,
head group/substrate interaction. DGcs can be much smaller than the relaxation time of the
accounted for by hydrophobic bonding or van system) and kept constant (shear rate remains
der Waals dispersion interactions. The driving constant). The stress is immediately followed
force for hydrophobic interactions is princi- with time after application of strain. For
pally entropic, arising from the destruction of a viscoelastic liquid (represented by a Maxwell
short-lived structures of water molecules orga- model, mechanical analogue of spring and dash-
nized around apolar substrates, similar to the pot in series), the stress s(0) shows an exponen-
process of surfactant micellization. Other phys- tial decay reaching a value s(t) after time t and it
ical components of DGcs can be envisaged approaches zero at infinite time. The stress that is
depending on the structure of the adsorbent needed to maintain constant strain decreases
and adsorbate. with time due to viscous flow. If the stress is
Sunscreen 1051 S
divided by the applied strain, one obtains the Definition
modulus which shows an exponential decrease
with time: A sunscreen is a material that protects the skin
from the harmful effect of the sun (that may cause
skin cancer), and the active substance must be
sð0Þ sð0Þ t
GðtÞ ¼ ¼ exp capable of absorbing UV radiation of specific
g g tm
wavelengths. The sunscreen agent is usually for-
t
¼ Gð0Þ exp mulated in a “cream” formulation for ease appli-
tm cation to the skin. The actives employed in these
preparations are of two basic types: organics
where G(t) is the stress relaxation modulus, G(0) which can absorb UV radiation of specific wave-
is the value at zero time, t is the experimental lengths due to their chemical structure and inor-
time, and tM is the Maxwell relaxation time that ganics which both absorb and scatter UV
is equal to G(0)/e. radiation. Inorganics have several benefits over
For a viscoelastic solid (represented by organics in that they are capable of absorbing
a Kelvin model, mechanical analogue of spring over a broad spectrum of wavelengths and they
and dashpot in parallel), the modulus does not are mild and nonirritant. Both of these advantages
become zero at infinite time, but it reaches an are becoming increasingly important as the
equilibrium value Ge: demand for daily UV protection against both
UVB and UVA radiation increases. The ability
t of fine particle inorganics to absorb radiation
GðtÞ ¼ Gð0Þ exp þ Ge
tm depends upon their refractive index. For inor-
ganic semiconductors such as titanium dioxide
and zinc oxide, this is a complex number indicat-
ing their ability to absorb light. The band gap in
Cross-References these materials is such that UV light up to around
405 nm can be absorbed. They can also scatter
▶ Rheological Techniques light due to their particulate nature, and their high
▶ Rheology refractive indices make them particularly effec-
tive scatterers. Both scattering and absorption
depend critically on particle size. Particles of
around 250 nm, for example, are very effective
at scattering visible light, and TiO2 of this particle
Submicrometer-Sized Emulsions
size is the most widely used white pigment. At
smaller particle sizes, absorption and scattering S
▶ Nanoemulsions
maxima shift to the UV region, and at 30–50 nm
UV attenuation is maximized. The use of TiO2 as
a UV attenuator in cosmetics was, until recently,
largely limited to baby sun protection products
Sunscreen due to its poor aesthetic properties (viz., scattering
of visible wavelengths results in whitening).
Tharwat Tadros Recent advances in particle size control and coat-
Wokingham, Berkshire, UK ings have enabled formulators to use fine particle
titanium dioxide and zinc oxide in daily skin care
formulations without compromising the cosmetic
Synonyms elegance. The benefits of a pre-dispersion of inor-
ganic sunscreens are widely acknowledged. How-
UV Protection ever, it requires an understanding of the nature of
S 1052 Surface Forces
colloidal stabilization in order to optimize this pre- Definition

dispersion (for both UV attenuation and stability)
and exceed the performance of powder-based for- The surface tension g of a liquid in equilibrium
mulations. Dispersion rheology and its depen- with air (air/liquid interface) in the absence of
dence on interparticle interactions is a key factor any adsorption is the change of the free energy
in this optimization. Optimization of sunscreen Gs with surface area A at constant temperature

actives however does not end there; an apprecia- T and constant composition ni: g ¼ @@A Gs
T;ni
. The
tion of the end application is crucial to maintaining value of g depends on the nature of the liquid. For
performance. Formulators need to incorporate the a polar liquid such as water at 25 C, g is 72
particulate actives into an emulsion, mousse, or gel mNm1; for a hydrocarbon oil such as n-alkane,
with due regard to aesthetics (skin feel and trans- g is 30 mNm1; for a fluorocarbon oil, g is 20
parency), stability, and rheology. mNm1. Any adsorption at the air/liquid
interface, in particular surfactant molecules, will
lower g to an extent depending on the nature of the
Cross-References liquid and surfactant. With most surfactants, g is
reduced from its value of pure water of g 72
▶ Colloid Stability mNm1 to 30–40 mNm1 depending on the
▶ Cosmetics nature of the surfactant. The reduction in g with
▶ Emulsions surfactant addition is important for wetting of
▶ Gel hydrophobic surfaces with water and formation
of foams. Surface tension gradients produced dur-
ing foam formation stabilize the air bubbles, thus
preventing foam collapse. Combination of surfac-
Surface Forces
tant molecules can lower g to lower values below
that of each surfactant molecule, and this can be
▶ Wetting
applied to enhance wetting and foam stability.
Surface Forces and the Equilibrium Cross-References

of Liquids on Solid Substrates
▶ Interfacial Tension
▶ Wetting ▶ Surfactants
Surface-Active Agents
Surface Interactions
▶ Surfactants
▶ Specific Surface Interactions
Surfactant Adsorption
Surface Tension
Tharwat Tadros
Tharwat Tadros Wokingham, Berkshire, UK
Synonyms Synonyms
Surfactant solution Adsorption of surfactants

Surfactant Adsorption 1053 S
Keywords alkyl chain and the nonpolar surfaces. On
a hydrophilic or charged surface, the interaction
Adsorption isotherms; adsorption; critical is determined by electrostatic and/or polar forces.
micelle concentration; hydrophobic interaction; Different surfactant structures such as bilayers,
electrostatic and polar interactions; interfaces; hemimicelles, or micelles can be identified on the
interfacial tension; surface tension; surfactants solid surface.
Definition Introduction
Adsorption of surfactants is the process of Surfactants play a major role in the formulation
accumulation of the molecules or ions at the of most chemical products. They are used for
interface. Three types of interfaces can be distin- stabilization of emulsions, nano-emulsions,
guished: air/water (A/W), oil/water (O/W) and microemulsions, and suspensions. Secondly,
solid/liquid (S/L). The adsorption of surfactants surfactants are added in emulsifiable concentrates
at the A/L interface results in reduction of the for their spontaneous dispersion on dilution. The
surface tension gA/W of water from 72 mNm1 to surfactant needs to accumulate at the interface,
30–40 mNm1 till the critical micelle concentra- a process that is generally described as
tion (cmc) depending on the nature of surfactant. adsorption. The simplest interface is that of the
At concentrations C > cmc the surface tension air/liquid and, in this case, the surfactant will
remains constant with further increase in C. adsorb with the hydrophilic group pointing
The adsorption of surfactants at the O/W inter- towards the polar liquid (water) leaving the
face follows the same trend as that obtained at the hydrocarbon chain pointing towards the air.
A/W interface. In this case the interfacial tension This process results in lowering of the surface
gO/W decreases from its value in the absence of tension g. Typically, surfactants show a gradual
surfactant (30–50 mNm1) (depending on the reduction in g, till the critical micelle concentra-
nature of the oil) to 1–10 mNm1 till the critical tion (cmc) is reached above which the surface
micelle concentration (cmc) depending on the tension remains virtually constant. Hydrocarbon
nature of the surfactant. At C > cmc, the interfa- surfactants of the ionic, nonionic, or zwitterionic
cial tension remains constant with further ionic type lower the surface tension to limiting
increase in C. The adsorption of surfactants at values reaching 30–40 Nm1 depending on the
the A/W and O/W interfaces can be described nature of the surfactant. Lower values may be
by the Gibbs adsorption isotherm which assumes achieved using fluorocarbon surfactants, typically
reversible adsorption. The amount of surfactant of the order of 20 mNm1. It is, therefore, essential
at full saturation G (mole m2) can be calculated to understand the adsorption and conformation of S
from the slope of the linear portion of the surfactants at the air/liquid interface.
gA/W log C or gO/W log C curves just before With emulsions, nano-emulsions, and
the cmc. From G the area per molecule can be microemulsions, the surfactant adsorbs at the
calculated and this gives information on the oil/water interface, with the hydrophilic head
orientation of the surfactant molecule or ion at group immersed in the aqueous phase, leaving
the interface. The adsorption of surfactants at the the hydrocarbon chain in the oil phase. Again,
S/L interface can be described by the Langmuir the mechanism of stabilization of emulsions,
adsorption isotherm. This shows a gradual nano-emulsions, and microemulsions depends
increase in G with further increase in equilibrium on the adsorption and orientation of the surfactant
concentration and a plateau value G1 is reached molecules at the liquid/liquid interface.
at or above the cmc. The adsorption of ionic and Surfactants consist of a small number of units
nonionic surfactants on hydrophobic surfaces is and they are mostly reversibly adsorbed,
governed by hydrophobic interaction between the allowing one to apply some thermodynamic
S 1054 Surfactant Adsorption
treatments. In this case, it is possible to describe (with A/L interfaces, one of the phase is air
the adsorption in terms of various interaction saturated with the vapor of the liquid).
parameters such as chain-surface, chain-solvent, Using the Gibbs model, it is possible to
and surface-solvent. Moreover, the configuration obtain a definition of the surface or interfacial
of the surfactant molecule can be simply tension g, starting from the Gibbs-Deuhem equa-
described in terms of these possible interactions. tion (Gibbs 1928), i.e.,
The use of surfactants (ionic, nonionic, and
zwitterionic) to control the stability behavior of X
dGs ¼ Ss dT þ Adg þ ni dmi (1)
suspensions is of considerable technological
importance. Surfactants are used in the formula-
tion of dyestuffs, paints, paper coatings, where Gs is the surface free energy, Ss is
agrochemicals, pharmaceuticals, ceramics, print- the entropy, A is the area of the interface, and ni
ing inks, etc. They are particularly robust form of is the number of moles of component i with
stabilization which is useful at high disperse chemical potential mi at the interface. At constant
volume fractions and high electrolyte concentra- temperature and composition of the interface
tions, as well as under extreme conditions of high (i.e., in the absence of any adsorption),
temperature, pressure, and flow. In particular,
surfactants are essential for the stabilization of s
qG
suspensions in nonaqueous media, where electro- g¼ (2)
qA T;ni
static stabilization is less successful. The key to
understanding how surfactants function as
stabilizers is to know their adsorption and It is obvious from Eq. 2 that for a stable
conformation at the solid/liquid interface. interface, g should be positive. In other words,
the free energy should increase if the area of the
interface increases; otherwise, the interface will
Adsorption of Surfactants at the become convoluted, increasing the interfacial
Air/Liquid (A/L) and Liquid/Liquid (L/L) area, until the liquid evaporates (for A/L Case)
Interfaces or the two “immiscible” phases dissolved in each
other (for the L/L case).
Before describing surfactant adsorption at the It is clear from Eq. 2 that surface or interfacial
air/liquid (A/L) and liquid/liquid (L/L) interface, tension, i.e., the force per unit length tangentially
it is essential to define the interface. The surface to the surface measured in units of mNm1, is
of a liquid is the boundary between two bulk dimensionally equivalent to an energy per unit
phases, namely, liquid and air (or the liquid area measured in mJm2. For this reason, it
vapor). Similarly an interface between two has been stated that the excess surface free
immiscible liquids (oil and water) may be defined energy is identical to the surface tension, but
providing a dividing line is introduced since this is true only for a single-component system,
the interfacial region is not a layer that is i.e., a pure liquid (where the total adsorption
one-molecule thick but usually have a thickness is zero).
d with properties that are different from the two There are generally two approaches for
bulk phases a and b (Guggenheim 1967). treating surfactant adsorption at the A/L and
However, Gibbs (1928) introduced the concept L/L interface. The first approach, adopted by
of a mathematical dividing plane Zs in the Gibbs, treats adsorption as an equilibrium phe-
interfacial region (Fig. 1). nomenon whereby the second law of thermody-
In this model the two bulk phases a and b are namics may be applied using surface quantities.
assumed to have uniform thermodynamic The second approach, referred to as the equation
properties up to Zs. This picture applies for of state approach, treats the surfactant film as
both the air/liquid and liquid/liquid interface a two-dimensional layer with a surface pressure
Surfactant Adsorption,
Fig. 1 Gibbs convention
for an interface Uniform
Thermodynamic
Properties
• Mathematical Dividing
• Plane Zs
• (Gibbs Dividing Line)
•
•
•
Uniform
Thermodynamic
Properties
p that may be related the surface excess G dg ¼ Gs1 d m1 þ Gs2 d m2 (4)
(amount of surfactant adsorbed per unit area).
Below, these two approaches are summarized. and if the Gibbs dividing surface is used, G1 ¼ 0
and
The Gibbs Adsorption Isotherm dg ¼ Gs1;2 d m2 (5)
Gibbs (1928) derived a thermodynamic relation- where G2,1s is the relative adsorption of (Eq. 2)
ship between the surface or interfacial tension g with respect to (Eq. 1). Since
and the surface excess G (adsorption per unit
area). The starting point of this equation is the m2 ¼ mo2 þ RT ln aL2 (6)
Gibbs-Deuhem equation given above (Eq. 1). At
equilibrium (where the rate of adsorption is equal or
to the rate of desorption), dGs ¼ 0. At constant
temperature, but in the presence of adsorption, d m2 ¼ RT d ln aL2 (7)
X
dGs ¼ Ss dT þ A dg þ ni d mi ¼ 0 then
or dg ¼ Gs2;1 RT d ln aL2 (8) S

X ns X or
dg ¼ i
d mi ¼ Gi d mi (3)
A
1 dg
Gs2;1 ¼ (9)
where Gi ¼ nis/A is the number of moles of RT d ln aL2
component i and adsorbed per unit area.
Equation 3 is the general form for the Gibbs where a2L is the activity of the surfactant in bulk
adsorption isotherm. The simplest case of this solution that is equal to C2f2 or x2f2, where C2 is
isotherm is a system of two component in which the concentration of the surfactant in moles dm3
the solute (Eq. 2) is the surface-active compo- and x2 is its mole fraction.
nent, i.e., it is adsorbed at the surface of the Equation 9 allows one to obtain the surface
solvent (Eq. 1). For such a case, Eq. 3 may be excess (abbreviated as G2) from the variation of
written as surface or interfacial tension with surfactant
Fig. 2 Variation of surface 72 50
and interfacial tension with
log (CSAA) at the air-water
and oil-water interface g/mNm-1 g/mNm-1
CMC CMC
log C log C
concentration. Note that a2 C2 since in dilute ionic surfactants such as alkyl sulfates, the
solutions f2 1. This approximation is valid area per surfactant is determined by the area
since most surfactants have low cmc (usually occupied by the alkyl chain and head group if
less than 103 mol dm3) and adsorption is com- these molecules lie flat at the interface. In this
plete at or just below the cmc. case the area per molecule increases with
The surface excess G2 can be calculated increase in the alkyl chain length. For vertical
from the linear portion of the g log C2 curves orientation, the area per surfactant ion is deter-
before the cmc. Such g log C curves are illus- mined by that occupied by the charged head
trated in Fig. 2 for the air/water and o/w inter- group, which at low electrolyte concentration
faces; (CSAA) denotes the concentration of will be in the region of 0.40 nm2. Such area is
surface-active agent in bulk solution. It can be larger than the geometrical area occupied by
seen that for the A/W interface, g decreases from a sulfate group, as a result of the lateral repulsion
the value for water (72 mNm1 at 20 C) reaching between the head group. On addition of electro-
about 25–30 mNm1 near the cmc. This is clearly lytes, this lateral repulsion is reduced and the
schematic since the actual values depend on the area/surfactant ion for vertical orientation will
surfactant nature. For the o/w case, g decreases be lower than 0.4 nm2 (reaching in some cases
from a value of about 50 mNm1 (for to 0.2 nm2).
a pure hydrocarbon-water interface) to Another important point can be made from the
1–5 mNm1 near the cmc (again depending g log C curves. At concentration just before the
on the nature of the surfactant). break point, one has the condition of constant
As mentioned above G2 can be calculated slope, which indicates that saturation adsorption
from the slope of the linear position of the curves has been reached.
shown in Fig. 2 just before the cmc is reached.
From G2, the area per surfactant ion or molecule qg
¼ constant (11)
can be calculated since q ln a2 p;T
1 Just above the break point,

Area=molecule ¼ m2
G2 Nav

1018 qg
¼ nm2 (10) ¼0 (12)
G2 Nav q ln a2 p;T
where Nav is the Avogadro’s constant. Determin- indicating the constancy of g with log C above the
ing the area per surfactant molecule is very cmc integration of Eq. 12 gives
useful since it gives information on surfactant
orientation at the interface. For example, for g ¼ constant x ln a2 (13)
Since g is constant in this region, then a2 must adsorbed to maintain neutrally, and dg/dln a is
remain constant. This means that addition of sur- twice as large as for an unionized surfactant.
factant molecules above the cmc must result in If a non-adsorbed electrolyte, such as NaCl, is
association to form units (micellar) with low present in large excess, then any increase in con-
activity. centration of Na+R produces a negligible
As mentioned before, the hydrophilic head increase in Na+ ion concentration and therefore
group may be unionized, e.g., alcohols or poly dmNa becomes negligible. Moreover, dmCl is also
(ethylene oxide) alkane or alkyl phenol com- negligible, so that the Gibbs adsorption equation
pounds; weakly ionized such as carboxylic reduces to
acids; or strongly ionized such as sulfates,
sulfonates, and quaternary ammonium salts. The 1 qg
G2 ¼ (15)
adsorption of these different surfactants at RT q ln CNaR
the air/water and oil/water interface depends
on the nature of the head group. With nonionic i.e., it becomes identical to that for a nonionic
surfactants, repulsion between the head groups is surfactant.
small and these surfactants are usually strongly The above discussion clearly illustrates that
adsorbed at the surface of water from very dilute for calculation of G2 from the g log C curve,
solutions. Nonionic surfactants have much lower one has to consider the nature of the surfactant
cmc values when compared with ionic surfactants and the composition of the medium. For nonionic
with the same alkyl chain length. Typically, the surfactants, the Gibbs adsorption equation (9) can
cmc is in the region of 105–104 mol dm3. be directly used. For ionic surfactant, in the
Such nonionic surfactants form closely packed absence of electrolytes the right-hand side of the
adsorbed layers at concentrations lower than Eq. 9 should be divided by 2 to account for
their cmc values. The activity coefficient of surfactant dissociation. This factor disappears in
such surfactants is close to unity and is only the presence of a high concentration of an
slightly affected by addition of moderate amounts indifferent electrolyte.
of electrolytes (or change in the pH of the solu-
tion). Thus, nonionic surfactant adsorption is
the simplest case since the solutions can Equation of State Approach
be represented by a two-component system and
the adsorption can be accurately calculated In this approach, one relates the surface pressure
using Eq. 9. p with the surface excess G2. The surface pressure
With ionic surfactants, on the other hand, is defined by the equation
the adsorption process is relatively more compli-
cated since one has to consider the repulsion p ¼ go g (16) S
between the head groups and the effect of
presence of any indifferent electrolyte. where go is the surface or interfacial tension
Moreover, the Gibbs adsorption equation has to before adsorption and g that after adsorption.
be solved taking into account the surfactant ions, For an ideal surface film, behaving as
the counterion, and any indifferent electrolyte a two-dimensional gas, the surface pressure
ions present. For a strong surfactant electrolyte p is related to the surface excess G2 by the
such as an Na+ R equation
1 qg pA ¼ n2 RT (17)
G2 ¼ (14)
2RT q ln a
or
The factor of 2 in Eq. 14 arises because
both surfactant ion and counterion must be p ¼ ðn2 =AÞRT ¼ G2 RT (18)
Differentiating Eq. 17 at constant temperature, where A2o is the excluded area or co-area of type
2 molecule in the interface and a is a parameter
dp ¼ RT d G2 (19) which allows for lateral interaction.
Equation 23 leads to the following theoretical
Using the Gibbs equation, adsorption isotherm, using the Gibbs’s equation,
dp ¼ dg ¼ G2 RT d ln a2
y y 2ay
G2 RT d ln C2 (20) Ca2 ¼ K1 exp o (24)
1y 1 y a2 RT
Combining Eqs. 19 and 20,
where y is the surface coverage (y ¼ G2/G2,max),
d ln G2 ¼ d ln C2 (21) K1 is constant that is related to the free energy of
adsorption of surfactant molecules at the
or interface (K1 / exp (DGads/kT)), and ao2 is the
area/molecule.
G2 ¼ K Ca2 (22) For a charged surfactant layer, Eq. 21 has to
be modified to take into account the electrical
Equation 22 is referred to as the Henry’s law contribution from the ionic head groups, i.e.,
isotherm which predicts a linear relationship

between G2 and C2. y y e co
Ca2 ¼ K1 exp exp (25)
It is clear that Eqs. 16 and 19 are based on an 1y 1y kT
idealized model in which the lateral interaction
between the molecules has not been considered. where Co is the surface potential. Equation 47
Moreover, in this model the molecules are shows how the electrical potential energy
considered to be dimensionless. This model can (Co/kT) of adsorbed surfactant ions affects the
only be applied at very low surface coverage surface excess. Assuming that the bulk
where the surfactant molecules are so far apart concentration remains constant, then Co
that lateral interaction may be neglected. More- increases as y increases. This means that
over, under these conditions the total area occu- [y/(1 y)] exp[y /(1 y)] increases less rapidly
pied by the surfactant molecules is relatively with C2, i.e., adsorption is inhibited as a result of
small compared to the total interfacial area. ionization.
At significant surface coverage, the above
equations have to be modified to take into account
both lateral interaction between the molecules The Langmuir, Szyszkowski, and
and the area occupied by them. Lateral interaction Frumkin Equations
may reduce p if there is attraction between the
chains (e.g., with most nonionic surfactant) or it In addition to the Gibbs equation, three other
may increase p as a result of repulsion between equations have been suggested that relate the
the head groups in the case of ionic surfactants. surface excess G1, surface or interfacial tension,
Various equation of state have been proposed, and equilibrium concentration in the liquid phase
taking into account the above two effects, in order C1. The Langmuir equation (Langmuir 1917)
to fit the p A data. The two-dimensional van der relates G1 to C1 by
Waals equation of state is probably the most con-
venient for fitting these adsorption isotherms, i.e., Gm C1
G1 ¼ (26)
C1 þ a
!
ð n2 Þ 2
pþ ðA n2 Ao2 Þ ¼ n2 RT (23) where Gm is the saturation adsorption at
A2
monolayer coverage by surfactant molecules.
a is a constant that is related to the free technique is applied. Below a brief description
energy of adsorption DGoads, of each of these techniques is given.

DGoads
a ¼ 55:3 exp (27)
RT The Wilhelmy Plate Method
where R is the gas constant and T is the absolute In this method (Wilhelmy 1863) a thin plate made
temperature. from glass (e.g., a microscope cover slide) or
A linear form of the Gibbs equation is platinum foil is either detached from the
interface (nonequilibrium condition) or its weight
1 1 a measured statically using an accurate microbal-
¼ þ (28)
G1 Gm Gm C1 ance. In the detachment method, the total force
F is given by the weight of the plate W and the
Equation 28 shows that a plot of 1/G1 versus interfacial tension force,
1/C1 gives a straight line from which Gm and
a can be calculated from the intercept and slope F ¼ W þ gp (31)
of the line.
The Szyszkowski (1908) equation gives where p is the “contact length” of the plate with
a relationship between the surface pressure p the liquid, i.e., the plate perimeter. Provided the
and bulk surfactant concentration C1; it is contact angle of the liquid is zero, no correction is
a form of equation of state, required for Eq. 31. Thus, the Wilhelmy plate
method can be applied in the same manner as
C1 the du Nouy’s technique that is described below.
go g ¼ p ¼ 2:303RTGm log þ1 (29)
a The static technique may be applied for fol-
lowing the interfacial tension as a function of
The Frumkin equation (Frumkin 1925) is time (to follow the kinetics of adsorption) till
another equation of state, equilibrium is reached. In this case, the plate is
suspended from one arm of a microbalance and

G1 allowed to penetrate the upper liquid layer (usu-
go g ¼ p ¼ 2:303RTGm log 1 (30)
Gm ally the oil) into the aqueous phase to ensure
wetting of the plate. The whole vessel is then
lowered to bring the plate in the oil phase. At
this point the microbalance is adjusted to
Interfacial Tension Measurements counteract the weight of the plate (i.e., its weight
now becomes zero). The vessel is then raised S
These methods may be classified into two until the plate touches the interface. The increase
categories: those in which the properties of the in weight DW is given by the following equation:
meniscus is measured at equilibrium, e.g., pen-
dent drop or sessile drop profile and Wilhelmy DW ¼ g p cos y (32)
plate methods, and those where the measurement
is made under nonequilibrium or quasi- where y is the contact angle. If the plate is
equilibrium conditions such as the drop volume completely wetted by the lower liquid as it
(weight) or the du Nouy ring method. The penetrates, y ¼ 0 and g may be calculated directly
latter methods are faster, although they suffer from DW. Care should always be taken that the
from the disadvantage of premature rupture and plate is completely wetted by the aqueous solu-
expansion of the interface, causing adsorption tion. For that purpose, a roughened platinum or
depletion. For measurement of low interfacial glass plate is used to ensure a zero contact angle.
tensions (<0.1 mNm1), the spinning drop However, if the oil is more dense than water,
a hydrophobic plate is used so that when the plate

penetrates through the upper aqueous layer and
touches the interface, it is completely wetted by
the oil phase. ds
The Pendent Drop Method
If a drop of oil is allowed to hang from the end of

a capillary that is immersed in the aqueous phase, d
it will adopt an equilibrium profile shown in dc
Fig. 3 that is a unique function of the tube radius,
the interfacial tension, its density, and the
gravitational field.
The interfacial tension is given by the follow-
ing equation (Bashforth and Adams 1883):
Surfactant Adsorption, Fig. 3 Schematic representa-
Dr g d2e tion of the profile of a pendent drop
g¼ (33)
H
3
where Dr is the density difference between g R R
f¼ ¼f ; (35)
the two phases, de is the equatorial diameter of gideal V r
the drop (see Fig. 3), and H is a function of ds/de,
where ds is the diameter measured at a distance Values of the correction factor f were
d from the bottom of the drop (see Fig. 3). The tabulated by Harkins and Jordan (Eq. 7). Some
relationship between H and the experimental theoretical account of f was given by Freud and
values of ds/de has been obtained empirically Freud (1930).
using pendent drops of water. Accurate values When using the du Nouy method for measure-
of H have been obtained by Nierderhauser and ment of g, one must be sure that the ring is kept
Bartell (1950). horizontal during the measurement. Moreover,
the ring should be free from any contaminant
and this is usually achieved by using a platinum
The Du Nouy’s Ring Method ring that is flamed before use.
Basically one measures the force required

to detach a ring or loop of wire from the liquid/ The Drop Volume (Weight) Method
liquid interface (Du Nouy 1919). As a first
approximation, the detachment force is taken to Here one determines the volume V (or weight W)
be equal to the interfacial tension g multiplied by of a drop of liquid (immersed in the second less
the perimeter of the ring, i.e., dense liquid) which becomes detached from
a vertically mounted capillary tip having
F¼Wþ4pRg (34) a circular cross section of radius r. The ideal
drop weight Wideal is given by the expression
where W is the weight of the ring. Harkins and
Jordan (1930) introduced a correction factor f Wideal ¼ 2 p r g (36)
(that is a function of meniscus volume V and
radius r of the wire) for more accurate calculation In practice, a weight W is obtained which is
of g from F, i.e., less than Wideal because a portion of the drop
a r0 b
r0
l
Prolate ellipsoid Elongated cyclinder
Surfactant Adsorption, Fig. 4 Schematic representation of a spinning drop: (a) prolate ellipsoid; (b) elongated
cylinder
remains attached to the tube tip. Thus, Eq. 36 drop approximates to an elongated cylinder.
should include a correction factor j that is This is schematically shown in Fig. 4.
a function of the tube radius r and some linear When the shape of the drop approximates
dimension of the drop, i.e., V1/3. Thus, a cylinder (Fig. 4b), the interfacial tension is
given by the following expression (Vonnegut
r 1942):
W ¼ 2 p r g f 1=3 (37)
V
o2 Dr r4o
1/3
values of (r/V ) have been tabulated by Harkins g¼ (38)
4
and Brown (1919). Lando and Oakley (1967)
used a quadratic equation to fit the correction where o is the speed of rotation, Dr is the density
function to (r/V1/3). A better fit was provided by difference between the two liquids A and B,
Wilkinson and Kidwell (1971). and ro is the radius of the elongated cylinder.
Equation 38 is valid when the length of the
elongated cylinder is much larger than ro.
The Spinning Drop Method
This method is particularly useful for the Adsortion of Surfactants at the

measurement of very low interfacial tensions Solid/Liquid Interface
(<101 mNm1) which are particularly
important in applications such as spontaneous As mentioned above surfactants consist of a small
emulsification and the formation of number of units, and they mostly are reversibly
microemulsions. Such low interfacial tensions adsorbed, allowing one to apply thermodynamic
may also be reached with emulsions particularly treatments. In this case, it is possible to describe S
when mixed surfactant films are used. A drop of the adsorption in terms of the various
the less dense liquid A is suspended in a tube interaction parameters, namely, chain-surface,
containing the second liquid B. On rotating the chain-solvent, and surface-solvent. Moreover,
whole mass (Fig. 4), the drop of the liquid moves the conformation of the surfactant molecules at
to the center. With increasing speed of revolu- the interface can be deduced from these simple
tion, the drop elongates as the centrifugal force interactions parameters. However, in some cases
opposes the interfacial tension force that tends to these interaction parameters may involve
maintain the spherical shape, i.e., that having ill-defined forces, such as hydrophobic bonding,
minimum surface area. solvation forces, and chemisorption. In addition,
An equilibrium shape is reached at any given the adsorption of ionic surfactants involves
speed of rotation. At moderate speeds of rotation, electrostatic forces particularly with polar
the drop approximates to a prolate ellipsoid, surfaces containing ionogenic groups. For that
whereas at very high speeds of revolution, the reason, the adsorption of ionic and nonionic
surfactants will be treated separately. The When the liquid phase is a dilute solution of
surfaces (substrates) can be also hydrophobic or surfactant (component 1) that much more
hydrophilic and these may be treated separately. strongly adsorbed onto the solid substrate than
Thus, four cases can be considered: (i) adsorption the solvent (component 2), then noDx1 ¼ Dn1
of ionic surfactants on hydrophobic (nonpolar) where Dn1 ¼ change in number of moles of
surfaces, (ii) adsorption of ionic surfactants on component 1 in solution, n2s 0, and x2 1. In
polar (charged) surfaces, (iii) adsorption of this case Eq. 39 reduces to
nonionic surfactants on hydrophobic surfaces,
and (iv) adsorption of nonionic surfactants on Dn1 DC1 V
ns1 ¼ ¼ (40)
polar surfaces. (i) and (iii) are governed by hydro- m m
phobic interaction between the alkyl chain and
hydrophobic surface; the charge plays a minor where DC1 ¼ C1,0 C1, C1,0 is the molar con-
role. (ii) and (iv) are determined by charge and/ centration of component 1 before adsorption, C1
or polar interaction. is the molar concentration of component 1 after
At the solid/liquid interface one is interested adsorption, and V is the volume of the liquid
in determining the following parameters: (i) the phase in liters.
amount of surfactant adsorbed G per unit mass The surface concentration G1 in moles m2
or unit area of the solid adsorbent at a given can be calculated from a knowledge of surface
temperature, (ii) the equilibrium concentration area A (m2 g1),
of the surfactant C (moles dm3 or mole frac-
tion x ¼ C/55.51) in the liquid phase required DC1 V
G1 ¼ (41)
to produce a given value of G at a given tem- mA
perature, (iii) the surfactant concentration at
full saturation of the adsorbent Gsat, (iv) the The adsorption isotherm is represented
orientation of the adsorbed surfactant ion or by a plot of G1 versus C1. In most cases, the
molecule that can be obtained from the area adsorption increases gradually with increase of
occupied by the ion or molecule at full satura- C1 and a plateau G11 is reached at full coverage
tion, and (v) the effect of adsorption on the corresponding to a surfactant monolayer.
properties of the adsorbent (nonpolar, polar, or The area per surfactant molecule or ion at full
charged). saturation can be calculated,
The general equation for calculating the
amount of surfactant adsorbed onto a solid 1018
as1 ¼ nm2 (42)
adsorbent from a binary solution containing two G1
1 Nav
components (surfactant component 1 and solvent
component 2) is given by Aveyard and where Nav is the Avogadro’s number.
Haydon (1973)
no Dx1 Adsorption of Ionic Surfactants on

¼ ns1 x2 ns2 x1 (39)
m Hydrophobic Surfaces
no is the number of moles of solution before The adsorption of ionic surfactants on hydropho-
adsorption, Dx1 ¼ x1,0 x1, x1,0 is the mole bic surfaces such as carbon black, polymer sur-
fraction of component 1 before adsorption, x1 faces, and ceramics (silicon carbide or silicon
and x2 are the mole fractions of components 1 nitride) is governed by hydrophobic interaction
and 2 at adsorption equilibrium, m is the mass of between the alkyl chain of the surfactant and the
adsorbent in grams, and n1s and n2s are the num- hydrophobic surface. In this case, electrostatic
ber of components 1 and 2 adsorbed per gram of interaction will play a relatively smaller role.
adsorbent at adsorption equilibrium. However, if the surfactant head group is of the
same sign of charge as that on the substrate Combining Eqs. 45 and 47,
surface, electrostatic repulsion may oppose

adsorption. In contrast, if the head groups are of y C D GO
¼ exp ads
(48)
opposite sign to the surface, adsorption may be 1 y 55:5 RT
enhanced. Since the adsorption depends on the
magnitude of the hydrophobic bonding free Equation 48 applies only at low surface cov-
energy, the amount of surfactant adsorbed erage (y < 0.1) where lateral interaction between
increases directly with increase in the alkyl the surfactant ions can be neglected.
chain length in accordance with Traube’s rule. At high surface coverage (y > 0.1) one should
The adsorption of ionic surfactants on take the lateral interaction between the chains
hydrophobic surfaces may be represented by the into account, by introducing a constant A, e.g.,
Stern-Langmuir isotherm (Hough and Randall using the Frumkin-Fowler-Guggenheim (FFG)
1983). Consider a substrate containing Ns sites equation (Hough and Randall 1983),
(mol m2) on which G moles m2 of surfactant
ions are adsorbed. The surface coverage y is y C D Goads
expðAyÞ ¼ exp (49)
(G/Ns) and the fraction of uncovered surface is ð1 yÞ 55:5 RT
(1 y). The rate of adsorption is proportional to
the surfactant concentration expressed in mole The value of A can be estimated from the
fraction (C/55.5) and the fraction of free surface maximum slope (dy/lnC)max of the isotherm
(1 y), i.e., which occurs at y ¼ 0.5. Furthermore, at
y ¼ 0.5 substitution of A into Eq. 49 gives the
C value of DGoads.
Rate of adsorption ¼ kads ð1 yÞ (43)
55:5 The above treatment using the FFG isotherm
has two limitations. Firstly, it is assumed that
where kads is the rate constant for adsorption. A is constant and independent of surface cover-
The rate of desorption is proportional to the age. In reality A could change in sign as well as
fraction of surface covered y, increase in y. At low coverages, A would reflect
repulsive (electrostatic) interaction between
Rate of desorption ¼ kdes y (44) adsorbed surfactant ions. At higher coverage,
attractive chain-chain interaction becomes more
At equilibrium, the rate of adsorption is equal important. The apparent adsorption energy
to the rate of desorption and the ratio of (kads/kdes) becomes more favorable at high surface coverage
is the equilibrium constant K, i.e., and this could lead to the formation of
“hemimicelles.” Secondly, electrostatic interac-
y C tions are strongly affected by the level of S
¼ K (45)
ð1 yÞ 55:5 supporting electrolyte.
Various authors (Fuerestenau and Healy 1972;
The equilibrium constant K is related to the Somasundaran and Goddard 1979) have used the
standard free energy of adsorption by Stern-Langmuir equation in a simple form to
describe the adsorption of surfactant ions on min-
D Goads ¼ RT ln K (46) eral surfaces,

R is the gas constant and T is the absolute D Goads
G ¼ 2 r C exp (50)
temperature. Equation 46 can be written in the RT
form or
Various contributions to the adsorption free
D Goads energy may be envisaged. To a first approxima-
K ¼ exp (47)
RT tion, these contributions may be considered to be
additive. In the first instance, DGads may be taken whole adsorption process, i.e., from the initial
to consist of two main contributions, i.e., values low to the plateau values. To obtain
accurate results, the solid should have a high
D Gads ¼ D Gelec þD Gspec (51) surface area (usually >1 m2).
Several examples may be quoted from the
where DGelec accounts for any electrical interac- literature to illustrate the adsorption of surfactant
tions (coulombic as well as polar) and DGspec is ions on solid surfaces. For a model hydrophobic
a specific adsorption term which contains all con- surface, carbon black has been chosen
tributions to the adsorption free energy that are (Greenwood et al. 1968; Day et al. 1967). Figure 5
dependent on the “specific” (nonelectrical) nature shows typical results for the adsorption of sodium
of the system (Healy 1974). Several authors dodecyl sulfate (SDS) on two carbon black sur-
subdivided DGspec into supposedly separate inde- faces, namely, Spheron 6 (untreated) and
pendent interactions (Healy 1974; Somasundaran Graphon (graphitized) which also describes the
and Hannah 1979), e.g., effect of surface treatment.
The adsorption of SDS on untreated Spheron 6
D Gspec ¼ D G þD Gcs þD Ghs þ (52) tends to show a maximum that is removed on
washing. This suggests the removal of impurities
where DGcc is a term that accounts for the cohe- from the carbon black which becomes extractable
sive chain-chain interaction between the hydro- at high surfactant concentration. The plateau
phobic moieties of the adsorbed ions and DGcs adsorption value is 2 106 mol m2 (2 m
is the term for chain/substrate interaction, whereas mole m2). This plateau value is reached at 8 m
DGhs is a term for the head group/substrate mole dm3 SDS, i.e., close to the cmc of the
interaction. Several other contributions to DGspec surfactant in the bulk solution. The area per sur-
may be envisaged, e.g., ion-dipole, ion-induced factant ion in this case is 0.7 nm2. Graphitiza-
dipole, or dipole-induced dipole interactions. tion (Graphon) removes the hydrophilic ionizable
Since there is no rigorous theory that can pre- groups (e.g., –C¼O or –COOH), producing
dict adsorption isotherms, the most suitable a surface that is more hydrophobic. The same
method to investigate adsorption of surfactants occurs by heating Spheron 6 to 2,700 C. This
is to determine the adsorption isotherm. Measure- leads to a different adsorption isotherm (Fig. 5)
ment of surfactant adsorption is fairly straightfor- showing a step (inflection point) at a surfactant
ward. A known mass m (g) of the particles concentration in the region of 6 m mole dm3.
(substrate) with known specific surface area As The first plateau value is 2.3 m mole m2,
(m2 g1) is equilibrated at constant temperature whereas the second plateau value (that occurs at
with surfactant solution with initial concentration the cmc of the surfactant) is 4 m mole m2.
C1. The suspension is kept stirred for sufficient It is likely in this case that the surfactant ions
time to reach equilibrium. The particles are then adopt different orientations at the first and second
removed from the suspension by centrifugation plateaus. In the first plateau region, a more “flat”
and the equilibrium concentration C2 is deter- orientation (alkyl chains adsorbing parallel to the
mined using a suitable analytical method. The surface) is obtained, whereas at the second
amount of adsorption G (mole m2) is calculated plateau vertical orientation is more favorable,
as follows: with the polar head groups being directed towards
the solution phase. Addition of electrolyte
ðC1 C2 ÞV (101 mol dm3 NaCl) enhances the surfactant
G¼ (53)
m As adsorption. This increase is due to the reduction
of lateral repulsion between the sulfate head
The adsorption isotherm is represented by groups, and this enhances the adsorption.
plotting G versus C2. A range of surfactant The adsorption of ionic surfactants on
concentrations should be used to cover the hydrophobic polar surfaces resembles that for
Surfactant Adsorption, Graphon in 10-1 mol dm-3 NaCl
Fig. 5 Adsorption 5
isotherms for sodium
dodecyl sulfate (SDS) on
carbon substrates. Graphon 4
in 101 mol dm3 NaCl and Graphon
without added electrolyte.
Spheron 6 and after
G mmolm-2
3
washing and after heat
treatment at 2,700 C Spheron 6 after washing
2
Spheron 6 before
1 washing
0
4 8 12 16 20 24
C/mmol dm-3
carbon black (Saleeb and Kitchener 1965;

10-5 -40
Conner and Ottewill 1971). For example, Saleeb III
-30
and Kitchener (1965) found similar limiting area -6 -20
10
for cetyltrimethyl ammonium bromide on G -10
G Mol m=2
Graphon and polystyrene (0.4 nm2). As with 0
z /mV
II
10-7 z 10
carbon black, the area per molecule depends on
I 20
the nature and amount of added electrolyte. This
-8 30
can be accounted for in terms of reduction of head 10
40
group repulsion and/or counterion binding. 50
Surfactant adsorption close to the cmc may 10-9 -5 -4 -3 -2
10 10 10 10
appear Langmuirian, although this does not
C/mol dm-3
automatically imply a simple orientation. For
example, rearrangement from horizontal to verti- Surfactant Adsorption, Fig. 6 The adsorption isotherm
cal orientation or electrostatic interaction and for sodium dodecyl sulfonate (SDSe) on alumina and the
corresponding z-potential of alumina particles as
counterion binding may be masked by simple a function of the equilibrium surfactant concentration;
adsorption isotherms. It is essential, therefore, to pH ¼ 7.2 and 2 103 mol dm3 ionic strength S
combine the adsorption isotherms with other
techniques such as microcalorimetry and various
spectroscopic methods to obtain a full picture on possible chain-chain interaction. This is best
surfactant adsorption. illustrated by the results of adsorption of sodium
dodecyl sulfonate (SDSe) on alumina at pH ¼ 7.2
obtained by Fuerestenau (Wakamatsu and
Adsorption of Ionic Surfactants on Polar Fuerestenau 1968) and shown in Fig. 6. At the
Surfaces pH value, the alumina is positively charged (the
isoelectric point of alumina is at pH 9) and
The adsorption of ionic surfactants on polar the counter ions are Cl from the added
surfaces that contain ionizable groups may show supporting electrolyte. In Fig. 6, the saturation
characteristic features due to additional interac- adsorption G1 is plotted versus equilibrium
tion between the head group and substrate and/or surfactant concentration C1 in logarithmic scales.
The figure also shows the results of zeta potential i.e. (CH2CH2O)nOH (where n can vary from as
(z) measurements (which are a measure of the little as 2 units to as high as 100 or more units)
magnitude sign of charge on the surface). Both linked either to an alkyl (CxH2x+1) or alkyl phenyl
adsorption and zeta potential results show three (CxH2x+1–C6H4–) group. These surfactants may
distinct regions. The first region which shows be abbreviated as CxEn or CxjEn (where C refers
a gradual increase of adsorption with increase in to the number of C atoms in the alkyl chain, j
concentration, with virtually no change in the denotes C6H4 and E denotes ethylene oxide).
value of the zeta potential corresponds to an These ethoxylated surfactants are characterized
ion-exchange process (Wakamatsu and by a relatively large head group compared to
Fuerestenau 1968). In other words, the surfactant the alkyl chain (when n > 4). However, there
ions simply exchange with the counterions (Cl) are nonionic surfactants with small head group
of the supporting electrolyte in the electrical such as amine oxides (–N –> 0) head group,
double layer. At a critical surfactant concentra- phosphate oxide (–P –> 0), or sulfinyl-alkanol
tion, the desorption increases dramatically with (–SO–(CH2)n–OH). Most adsorption isotherms
further increase in surfactant concentration in the literature are based on the ethoxylated-
(region II). In this region, the positive zeta poten- type surfactants.
tial gradually decreases, reaching a zero value The adsorption isotherm of nonionic
(charge neutralization) after which a negative surfactants is in many cases Langmuirian, like
value is obtained which increases rapidly with those of most other highly surface-active solutes
increase in surfactant concentration. The rapid adsorbing from dilute solutions, and adsorption is
increase in region II was explained in terms of generally reversible. However, several other
“hemimicelle formation” that was originally adsorption types are produced (Clunie and
postulated by Gaudin and Fuerestenau (1955). Ingram 1983) and those are illustrated in Fig. 7.
In other words, at a critical surfactant concentra- The steps in the isotherm may be explained in
tion (to be denoted the cmc of “hemimicelle terms of the various adsorbate-adsorbate,
formation” or better the critical aggregation con- adsorbate-adsorbent, and adsorbate-solvent inter-
centration CAC), the hydrophobic moieties of the actions. These orientations are schematically
adsorbed surfactant chains are “squeezed out” illustrated in Fig. 8. In the first stage of adsorption
from the aqueous solution by forming two- (denoted by I in Figs. 7 and 8), surfactant-
dimensional aggregates on the adsorbent surface. surfactant interaction is negligible (low cover-
This is analogous to the process of micellization age) and adsorption occurs mainly by van der
in bulk solution. However, the CAC is lower than Waals interaction. On a hydrophobic surface,
the cmc, indicating that the substrate promotes the interaction is dominated by the hydrophobic
surfactant aggregation. At a certain surfactant portion of the surfactant molecule. This is
concentration in the hemimicellization process, mostly the case with agrochemicals which have
the isoelectric point is exceeded and, thereafter, hydrophobic surfaces. However, if the chemical
the adsorption is hindered by the electrostatic is hydrophilic in nature, the interaction will be
repulsion between the hemimicelles. dominated by the EO chain. The approach to
Micelles and hence the slope of the adsorption monolayer saturation with the molecules lying
isotherm is reduced (region III). flat is accompanied by a gradual decrease in the
slope of the adsorption isotherm (region II in
Fig. 7). Increase in the size of the surfactant
Adsorption of Nonionic Surfactants molecule, e.g., increasing the length of the
alkyl or EO chain, will decrease adsorption
Several types of nonionic surfactants exist, (when expressed in moles per unit area). On
depending on the nature of the polar (hydrophilic) the other hand, increasing temperature will
group. The most common type is that increase adsorption as a result of desolvation
based on a poly(oxyethylene) glycol group, of the EO chains, thus reducing their size.
Fig. 7 Adsorption
isotherms, corresponding to
the three adsorption
sequences shown in Fig. 8 V
(I–IV), indicating the
different orientation; the G
cmc is indicated by an
arrow IV
IV
III III
I II I II III
I II
cmc cmc cmc
x2
Surfactant Adsorption, I
Fig. 8 Model for the
adsorption of nonionic
II
surfactants showing
orientation of surfactant
molecules at the surface.
I–IV are the successive
A B C
stages of adsorption, and
sequence A–C corresponds
to situations where there are III
relatively weak,
intermediate, and strong
interactions between the IV
adsorbent and the
hydrophilic moiety of the
surfactant
Moreover, increasing temperature reduces the compress the head group and for an EO chain
solubility of the nonionic surfactant and this this will result in a less coiled more extended
enhances adsorption. conformation. The larger the surfactant alkyl
The subsequent stages of adsorption (region III chain the greater will be the cohesive forces and
and IV) are determined by surfactant-surfactant hence the smaller the cross-sectional area. This
interaction, although surfactant-surface interaction may explain why saturation adsorption increases
initially determines adsorption beyond stage II. with increasing alkyl chain length. S
This interaction depends on the nature of the The interaction occurring in the adsorption
surface and the hydrophilic-lipophilic balance of layer during the fourth and subsequent stages of
the surfactant molecule (HLB). For a hydrophobic adsorption is similar to those that occur in bulk
surface, adsorption occurs via the alkyl group of solution. In this case aggregate units, as shown in
the surfactant. For a given EO chain, the adsorp- Fig. 8 V, (hemimicelles or micelles) may be
tion will increase with increase in the alkyl chain formed. This picture was supported by Kleminko
length. On the other hand, for a given alkyl chain (1974) who found close agreement between sat-
length, adsorption increases with decrease of the uration adsorption and adsorption calculated
PEO chain length. based on the assumption that the surface is cov-
As the surfactant concentration approaches the ered with close-packed hemimicelles. Kleminko
cmc, there is a tendency for aggregation of the (1978) developed a theoretical model for the three
alkyl groups. This will cause vertical orientation stages of adsorption of nonionic surfactants. In the
of the surfactant molecules (stage IV). This will first stage (flat orientation) a modified Langmuir
S 1068 Surfactant Adsorption at the Air–Solution and Solid–Liquid Interfaces
adsorption equation was used. In the second stage Gaudin AM, Fuerestenau DW (1955) Trans AIME
of horizontal orientation, the surface concentration 202:958
Gibbs JW (1928) Collected works, vol 1. Longman,
increases by an amount that is determined by the New York, p 219
displacement of the ethoxy chain by the alkyl Greenwood FG, Parfitt GD, Picton NH, Wharton DG
group. Finally, in the region of the hemimicelle (1968) Adv Chem Ser 79:135
formation, the adsorption can be described by Guggenheim EA (1967) Thermodynamics, 5th edn. North
Holland, Amsterdam, p 45
a simple Langmuir equation of the form Harkins WD, Brown FE (1919) J Am Chem Soc 41:499
Harkins WD, Jordan HF (1930) J Am Chem Soc 52:1715
G2 Healy TW (1974) J Macromol Sci Chem 118:603
C2 K
a ¼ infinity
(54) Hough DB, Randall HM (1983) In: Parfitt GD, Rochester
ðG2 G2 Þ
CH (eds) Adsorption from solution at the solid/liquid
interface. Academic, London, p 247
where G21 is the maximum surface excess, i.e., Kleminko NA (1978) Kolloid Zh 40:1105; (1979) 41:78
the surface excess when the surface is covered Kleminko NA (1974) Tryasorukova and Permilouskayan.
with close-packed hemimicelles, Ka* is Kolloid Zh 36:678
Lando JL, Oakley HT (1967) J Colloid Interface Sci
a constant that is inversely proportional to the 25:526
cmc, and C2 is the equilibrium concentration. Langmuir I (1917) J Am Chem Soc 39:1848
Nierderhauser DO, Bartell FE (1950) Report of progress,
fundamental research on occurrence of petroleum.
Publication of the American Petroleum Institute,
Cross-References Lord Baltimore Press, Baltimore, p 114
Saleeb FZ, Kitchener JA (1965) J Chem Soc 911
▶ Dispersions Somasundaran P, Goddard ED (1979) Modern Aspects
▶ Emulsions Electrochem 13:207
Somasundaran P, Hannah HS (1979) In: Shah DO,
▶ Equation of State Approach Schechter RS (eds) Improved oil recovery by surfac-
▶ Microemulsions tant and polymer flooding. Academic, London, p 205
▶ Nanoemulsions Szyszkowski B (1908) Z Phys Chem 64:385
▶ Surfactants Vonnegut B (1942) Rev Sci Instrum 13:6
Wakamatsu T, Fuerestenau DW (1968) Adv Chem Ser
▶ Suspensions 71:161
Wilhelmy L (1863) Ann Phys 119:177
Wilkinson MC, Kidwell RL (1971) J Colloid Interface Sci
35:114
References
Aveyard R, Haydon DA (1973) An introduction to the

principles of surface chemistry. Cambridge University
Press, Cambridge
Bashforth F, Adams JC (1883) An attempt to test the
Surfactant Adsorption at the
theories of capillary action. Cambridge University Air–Solution and Solid–Liquid
Press, Cambridge Interfaces Studied by Neutron
Clunie JS, Ingram BT (1983) In: Parfitt GD, Rochester CH Reflectivity
(eds) Adsorption from solution at the solid/liquid inter-
face. Academic, London, p 105
Conner P, Ottewill RH (1971) J Colloid Interface Jeffery Penfold
Sci 37:642 ISIS Facility, Science & Technology Facilities
Day RE, Greenwood FG, Parfitt GD 4th (1967) Int Congr Council, Rutherford Appleton Laboratory,
Surf Active Subst 18:1005
Chilton, Didcot, Oxon, UK
Du Nouy PL (1919) J Gen Physiol 1:521
Freud BB, Freud HZ (1930) J Am Chem Soc 52:1772
Frumkin A (1925) Z Phys Chem 116:466
Fuerestenau DW (1971) In: Hair ML (ed) The chemistry Keywords
of biosurfaces. Marcel Dekker, New York, p 91
Fuerestenau DW, Healy TW (1972) In: Lemlich R (ed)
Adsorptive bubble separation techniques. Academic, Air–water interface; neutron reflectivity; solid–
London, p 91 liquid interface; surfactant adsorption
Surfactant Adsorption at the Air–Solution and Solid–Liquid Interfaces 1069 S
Definition importance (Lu et al. 2000; Penfold and Thomas
2007). The form of the adsorbed layer of surfac-
Surfactants are amphiphilic molecules which tant and the way in which it modifies surface
have both hydrophobic and hydrophilic proper- properties depends upon the nature of surfactant
ties. They adsorb at interfaces lowering surface and its interaction with the surface. Traditionally,
tension, and self-assemble in aqueous solution. surface tension has been extensively used to char-
They are extensively used in a wide range of acterize surfactant adsorption at fluid interfaces
formulations. Their relative surface activities via the Gibbs equation, and to quantify through
(affinity) and the structure of the adsorbed layer the determination of the critical micellar concen-
determine their properties and effectiveness at tration, cmc, and the onset of aggregate formation
lowering interfacial tension. Neutron reflectivity in solution. Dynamic surface tension measure-
is a surface sensitive technique, akin to optical ments are now being increasingly used to provide
reflectivity and ellipsometry, which gives access information about the kinetics of adsorption
to those key properties. Deuterium labeling and (Fainermann et al. 2006). At the solid–solution
partial deuterium labeling of the surfactant mol- interface, a variety of “classical” techniques have
ecules provides detailed structural information been used to characterize surfactant adsorption
about the adsorption. (Levitz 2002), and a wide variety of different
solid surfaces with differing degrees of hydrophi-
licity/hydrophobicity have been studied.
Overview In recent years, a wide range of sophisticated
surface sensitive techniques have been developed
The amphiphilic nature of surfactants, whether and applied to surfactant adsorption, such as
they are synthetic surfactants or naturally pro- ellipsometry (Tiberg et al. 1994), atomic force
duced biosurfactants, gives rise to properties microscopy (Manne and Gaub 1995), and light,
which make their occurrence and application x-ray, and neutron scattering (Penfold and
ubiquitous. The amphiphilic nature arises from Thomas 2007). These new techniques are now
the competition between the hydrophobic alkyl providing a wealth of new and complementary
chain and the hydrophilic headgroup, which can information about surfactant adsorption. In addi-
be anionic, cationic, zwitterionic, or nonionic. This tion to the evaluation of adsorbed amounts, they
competition results in surface activity, adsorption have also provided information about the struc-
to interfaces, and self-assembly to form a variety ture of the adsorbed layer and access to the kinet-
of structures in solution (Wennerstrom and ics of adsorption.
Lindman 1979). The surface activity and In particular, in the past 10–20 years, neutron
self-assembly of surfactants gives rise to their reflectivity has been developed into a powerful
important role in many industrial, technological, technique for the study of surface and interfacial S
and domestic applications, such as detergency, structure, and has been applied to a wide variety
cosmetics lubrication, surface modification (con- of surface problems (Penfold and Thomas 1990;
ditioning), dyeing, colloidal stability, corrosion Penfold 2002; Russell 1990). It is especially pow-
inhibition, mineral floatation, and enhanced oil erful for the study of organic systems in that
recovery (Abe and Scamehorn 2005). hydrogen–deuterium isotopic substitution can be
Surfactant adsorption at interfaces lowers used to manipulate the neutron refractive index
interfacial tension, and this is the basis of their distribution without significantly altering the
ability to modify surface properties, stabilize oil underlying chemistry. Hence, components can
in water emulsions, foams, and colloidal particles be made visible or invisible by refractive index
in solution. Hence, in terms of their widespread matching. This is the basis of its application to the
applications, understanding the nature of surfac- study of surfactant adsorption at the air–solution
tant adsorption at the air–solution, liquid–liquid, and solid–solution interfaces. Neutron reflectiv-
and solid–solution interfaces is of central ity has been used extensively for the study of
surfactant, mixed surfactant adsorption (Lu et al. where l is the neutron wavelength and r is the
2000), and polymer-surfactant adsorption (Taylor scattering length density given by,
et al. 2007) at the air–solution interface, and at the
X
solid–solution interface (Penfold and Thomas r¼ bi ni (2)
2007). The technique provides not only direct i
information about absolute adsorbed amounts, but
also details about the surface structure. where bi is the scattering length and ni is the
In this entry, the phenomena of neutron reflec- number density of species i. The refractive
tivity will be described, and the different ways in index is not a convenient term to use in neutron
which it can be theoretically calculated for model reflectivity as it is 1.0 (in contrast to light), and
surfaces and interfaces presented. Some of the the scattering length density is a more appropriate
key experimental details will be discussed, with quantity.
a particular focus on how surfactant adsorption at For most materials, the neutron scattering
the air–solution and solid–solution interfaces is length, b, is positive (hydrogen is one of the
studied. Examples of how the technique is used to exceptions), and so n is generally < 1.0. This
evaluate adsorbed amounts for a range of surfac- highlights an important difference compared to
tant systems, and how the structure of the light, as in general total external reflection occurs
adsorbed layer is evaluated at the air–solution for neutron as most materials have a refractive
interface are presented. It will be shown how index less than air. So, for example, at the
the different solid surfaces (hydrophobic, hydro- air–D2O interface, there will be total reflection
philic, or surfaces with specific functionalities) from the interface below the critical glancing
and the nature of the surfactant interaction with angle of incidence. The reflectivity beyond the
that surface controls the structure of the surface critical glancing angle contains information
layer and adsorption. about the refractive index distribution, and this
can be simply related to a composition or con-
centration profile to provide detailed information
Neutron Reflectivity about the structure of an adsorbed layer.
In the case of a thin layer of uniform compo-
Theoretical Background sition and density at the interface (which could
Thermal/cold neutrons (neutrons with wave- be a surfactant monolayer) between two
lengths 0.2–20 Å), like x-rays and the light bulk media (say air and a substrate), the neutron
spectrum, exhibit all the classical optical proper- reflectivity can be given exactly as (Heavens
ties, such as refraction, reflection, and interfer- 1955):
ence (Lekner 1987). The focus here is on neutron
reflectivity, and the specular reflectivity of neu- 2
r01 þ r12
2
þ 2r01 r12 cosð2b1 Þ
R ¼ j Rj 2 ¼ (3)
trons, like the analogous light or x-ray reflectiv- 1 þ r01 r12 þ 2r01 r12 cosð2b1 Þ
2 2
ity, from a surface or interface provides

information about the neutron refractive index where the subscripts 0, 1, 2 refer to the air phase,
gradient or distribution in the surface region and the monolayer, and the substrate, b is the optical
in a direction orthogonal to the plane. At an inter- path length, b1 ¼ q1t1, t1 is the monolayer thick-
face between two bulk media it is simply deter- ness, and q1 is the neutron wave vector in the
mined by the refractive index difference across monolayer and normal to the interface. The
the interface and is described by Fresnel’s Law. reflectivity is usually expressed in terms of the
The refractive index, n, for neutrons is related to wave vector transfer, k, also commonly referred
the neutron scattering length density, such that, to as Q, where,
l2 4p sin y0
n2 ¼ 1 r (1) k¼ ¼ 2q0 (4)
p l
and y0 is the grazing angle of incidence. The For n layers, the matrices for each layer are
Fresnel coefficients, rij, are then: multiplied together, [C] ¼ [C1][C2]. . .[Cn+1], to
give the resultant reflectivity from the elements
rj ri of the final 2 2 matrix such that, R ¼ cc*/aa*.
rij ¼ (5)
rj þ ri In practice, recurrence relationships between
the Fresnel coefficients in the successive layers
The critical wave-vector transfer (below can be used to provide an efficient calculation.
which total reflection takes place) is: Furthermore, a Gaussian roughness or diffuse
profile can be included at each interface in the
k2ci ¼ 4q2ci ¼ 16p rj ri (6) stack, such that,
such that the Fresnel coefficients can be rj 0 ¼ rj exp 0:5kj kj1 s2 (11)
expressed in terms of k and kci,
The second method uses the kinematic
1 approximation in which the reflectivity is given
k2 kci
2 2 k
rij ¼ 1 (7) as (Lu et al. 2000; Penfold and Thomas 1990):
k2 kci
2 2 þk
16p2 16p2 0 2
R¼ j ^
r ð k Þ j 2
¼ ^ ðkÞj
jr (12)
For the interface between two bulk media, for k2 k4
example, air and the substrate, Eq. 3 reduces to
Fresnel’s Law: where r^ðkÞ is the one-dimensional Fourier trans-
form of the scattering length density profile nor-
16p2 mal to that interface, r(z),
R ¼ jr01 j2 ¼ ðDrÞ2 (8)
k4 ð1
^ðkÞ ¼
r rðzÞ expðikzÞdz (13)
For a single thin layer (such as a monolayer) at 1
the interface, Eq. 3 becomes:
and r^0 ðkÞ is the corresponding Fourier transform
16p2 h of the derivative, r0 (z).
R ¼ 4 ð r 1 r 0 Þ 2 þ ð r2 r 1 Þ 2
k i (9)
þ 2ðr1 r0 Þðr2 r1 Þ cos kt
Experimental Details
For more complex surface structures than A wide variety of neutron reflectometers exist on
a simple single layer of uniform density, there nuclear reactor based and pulsed neutron sources, S
are two different approaches to calculate the and the data presented in this entry were predom-
associated reflectivity. The first of these uses the inantly obtained on the INTER, SURF, and
optical matrix method (Heavens 1955; Lekner CRISP reflectometers at the ISIS pulsed neutron
1987) developed for thin optical films, and relies source in the UK or on the FIGARO and D17
on a model of the surface structure being described reflectometers at the ILL reactor source in
by a series or stack of thin layers, This assumes that France. A typical experimental arrangement is
in optical terms, an application of Maxwell’s shown in Fig. 1.
equations and the relationship between the electric The essence of a neutron reflectivity measure-
vectors in successive layer leads to a characteristic ment is to measure the specular reflectivity R(k)
matrix per layer, such that, as a function of the wave vector transfer, k.
ib The measurements are made in two different
e j1 reibj1 ways, either at a fixed wavelength, l, and
Cj ¼ (10)
reibj1 eibj1 variable grazing angle of incidence, y, or at
Surfactant Adsorption at the Air–Solution and Solid–Liquid Interfaces Studied by Neutron Reflectivity,
Fig. 1 Schematic layout of the SURF reflectometer at the ISIS pulsed neutron facility
a fixed y and a range of l sorted by time of flight. types of measurements are described in a little
D17 uses the variable y approach, whereas more detail. At the air–solution interface, the
CRISP, SURF, INTER, and FIGARO use the neutron beam is reflected in the horizontal plane
fixed y time-of-flight method. Typical angles of from a free liquid surface, usually contained in an
incidence range from a fraction of a degree to inert container such as Teflon. Typically, the
a few degrees to cover a range of k from 0.002 surface area impinged by the neutron beam at
to 0.5 Å1. Correction of the reflected signal to grazing incidence is 50 to 100 cm2. The partic-
the direct beam, detector efficiencies, and solid ular feature that is exploited is the different scat-
angles subtended by the detector provides R(k) tering powers of hydrogen and deuterium
on an absolute scale (where total reflection is (hydrogen has a negative scattering length of
unity). The minimum size of adsorbed layer 0.374 1012 cm and deuterium a larger pos-
accessible is determined by the maximum k itive value of 0.6764 1012 cm). For a 0.088
value (t 2p/k), and as R(k) decays as k4 this mole fraction D2O/H2O mixture, null reflecting
is usually dictated by the background signal (R(k) water, nrw, the scattering length density is zero (a
106) which arises from incoherent scattering refractive index of unity), and there will be no
from the bulk of the sample. The maximum size specular reflectivity at the interface between that
measurable is determined by the instrument res- and air (in Eq. 8 Dr is zero). An adsorbed layer of
olution (Dk) and is typically 2–3,000 Å. surfactant at the interface which is deuterium
As the focus in this entry is on surfactant labeled will give rise to a reflected signal, making
adsorption at the air–solution and solid–solution the arrangement sensitive only to the adsorbed
interfaces, the experimental details for these layer. As will be described later, this is the essence
of determining adsorbed amounts. More detailed The adsorbed amount, G, can then be deter-
structural information can be obtained by changing mined from the thickness, t, and the scattering
the level of deuteration in the solvent and solute, length density, r, such that,
and this will also be described in more detail.
The same principles can be applied to the rt 1
G¼ ¼ (15)
study of adsorption at the solid–solution inter- Na b Na A
face, although the experimental arrangement is
somewhat different. In this case, the neutron where A is the area/molecule, and Na is
beam is then incident at the solid–solution inter- Avogadro’s number. A similar expression to
face, by transmission through the solid upper Eq. 14 can be obtained if it is assumed that the
phase. The transmissions of thermal/cold neu- monolayer has a Gaussian distribution rather than
trons through single crystal materials such as a uniform form, to give,
silicon and quartz (providing the orientation and
the neutron wavelength combinations does not k4 ps2 r2m k2
give rise to Bragg scattering) are sufficiently 2
R¼ exp k2 s2 (16)
16p 4
high (80 % for a 100 mm path length) that
neutron reflectivity at the solid–solution interface and G ¼ (psrm)1/2/2Na, where s is the 1/e width
is easily tractable. Measurements are made by and rm the scattering length density at the maxi-
transmitting the neutron beam (at normal mum of the distribution. In practice, either func-
incidence) through the crystalline solid upper tional form can be used. This approach has been
phase to grazing incidence at the solid–solution extensively applied to a wide range of surfactant
interface. A typical sample size and hence illu- system (Lu et al. 2000; Penfold and Thomas 2010).
minated area is 50 50 mm. By analogy with Figure 2 shows some typical neutron
the air–solution interface, an appropriate mixture reflectivity data for surfactant adsorption at
of D2O/H2O can be used to index match the the air–water interface, for a deuterium-labeled
solvent to the solid phase, or a specific “contrast,” surfactant in nrw. The data are from a recent
such as solid – D2O, can be engineered; and study on the adsorption and self-assembly
this is often used to maximize the reflectivity of a glycolipid biosurfactants, rhamnolipid
signal. Alternatively a range of different solute (Chen et al. 2010a, b). The rhamnolipids are
and solvent “contrast” are used to maximize rhamnose-containing glycolipid surfactants,
sensitivity. which are produced by different strains of
the bacteria Pseudomonas aeruginosa. They gen-
erally exist as one or two molecules of rhamnose
Surfactant Adsorption linked to one or two molecules of b-
hydroxydecanoic acid. The most common forms S
Air–Water Interface and major components are L-rhamnosyl-L-
As described earlier, the measurements of surfac- rhamnosyl-b-hydroxydecanoyl-b-hydroxydecanoate
tant adsorption at the air–solution interface are (Rha2C10C10, R2) and L-rhamnosyl-b-
predominantly made for deuterium-labeled sur- hydroxydecanoyl-b-hydroxydecanoate (RhaC10C10,
factant in nrw, as outlined earlier. In this case, the R1). As part of the study deuterium-labeled,
reflectivity arises only from the absorbed layer of rhamnolipids (90 % deuterated) were produced,
surfactant. Hence, treating the adsorbed layer as separated, and purified. The data in Fig. 2 are for
a thin layer of uniform density and composition deuterated R1 (d-R1) in nrw at surfactant concen-
Eq. 9 can be simplified to: trations from 7.5 107 to 7.5 105 M.
The data are typical of that measured for sur-
k4 kt factant adsorption. The solid lines in the figure
2
R ¼ 2r2 ð1 cos ktÞ ¼ 4b2 n2 sin
16p 2 are calculated curves for a thin layer of uniform
(14) density and composition, calculated using
Surfactant Adsorption at 1e−3

the Air–Solution and
Solid–Liquid Interfaces 7.5x10−5 M
Studied by Neutron 6x10−6 M
Reflectivity, 3x10−6 M
Fig. 2 Specular 1.5x10−6 M
reflectivity for d-R1 in nrw, 1e−4 7.5x10−7 M
at 7.5 107, 1.6 106,
3 106, 6 106, and
Reflectivity
7.5 105 M. The solid
lines are model calculations
for a single layer of uniform
1e−5
composition (# 2010 ACS
Reproduced by permission
of Chen et al. 2010a, b)
1e−6
0.1
Wave vector transfer, Q (A−1)
Eq. 14, and corresponding to a monolayer of especially as the adsorbed amount decreases, the
adsorbed surfactant. The mean thickness of the error in the product is much smaller. This is
monolayer, t, is about 21 1 Å. Combined with illustrated in Fig. 4 for the adsorption of the
the known value of b (Eq. 15) and values of r nonionic surfactant mono dodecyl octaethylene
obtained from the modeling gives rise to the glycol, C12E8 (alkyl chain deuterated, dC12hE8),
variation in the adsorbed amount with surfactant at the air–solution interface (Lu et al. 1994a).
concentration, as shown in Fig. 3. The figures show the variation in w2, “good-
The solid line is a calculation assuming that ness of fit” with r and t (in Fig. 4a, b) for an
the adsorption has the form of a Langmuir adsorbed layer of alkyl chain deuterated C12E8,
isotherm: dC12hE8 adsorbed at the air–water (nrw) interface
at a surfactant concentration of 9 106 M and
C an area/molecule of 86 Å2. In Fig. 4b, the varia-
G ¼ Gmax (17)
ðC þ k Þ tion in w2, with the product r.t (Г), is plotted, and
although the w2 curves for t and r are quite broad,
where G, Gmax are the adsorbed amounts and the it is much narrower for t.r.
maximum adsorption (the adsorption at satura- Although there are many potential sources of
tion), C is the surfactant concentration, and k is error and importantly impurity, the use of deute-
the adsorption coefficient. For R1, Gmax is 2.7 rium labeling in absolute neutron reflectivity
0.1 1010 mol cm2 and k of 1.83 method of determining adsorbed amount offers
106 mol/l. Typically at an area/molecule a routine and relatively stringent test of surfactant
50 Å2, the error in the measurement is purity. Making the neutron reflectivity measure-
2 Å, and the corresponding error in Г is ments for the deuterated surfactant in nrw and an
0.1 1010 mol cm2, and the sources of equimolar mixture of the deuterated and
error in such measurements have been discussed unlabeled surfactants is a sensitive measure. For
in detail elsewhere (Lu et al. 1993a). Although example, such measurements for the rhamnolipid
the uncertainty in the determination of t and r in R1 gave, at a surfactant concentration of 0.3 mM,
Eq. 14 and used in Eq. 15 to determine the an area/molecule of 62 2 Å2 for d-R1 and 61
adsorbed amount can be relatively larger, 2 Å2 for 50;50 d-R1/h-R1 (Chen et al. 2010a.)
Surfactant Adsorption at 4
Solid–Liquid Interfaces
Studied by Neutron
Adsorbed amount (x10−10 mol cm−2)

Reflectivity,
Fig. 3 Adsorption 3
isotherm for R1 at the air–
water interface. The solid
line is a calculate line
assuming a Langmuir
2
isotherm (# 2010 ACS
of Chen et al. 2010a, b)
1
R1 (no buffer)
Langmuir isotherm
0
1e−6 1e−5 1e−4 1e−3
Solution concentration (M)
Surface tension has been traditionally the This will provide a further verification of the
method of choice to establish the nature of sur- correspondence between the two. This approach
factant adsorption, and below the cmc, the slope was well demonstrated by Hines et al. (1997a, b),
of the surface tension variation with surfactant on measurements of the adsorption of the anionic
concentration is related via the Gibbs equation, surfactant sodium dodecyl sulfate, SDS, the zwit-
terionic surfactant dodecyl dimethylamino
1 dg actetate, dodecyl betaine, and the nonionic surfac-
G¼ (18)
mRT dLnC tant dodecyl-b-D-maltoside, as shown in Fig. 5.
For the zwitterionic and nonionic surfactants,
and where g is the surface tension, to the there is excellent agreement between the surface
adsorbed amount at the interface. Where tension derived from the neutron reflection sur-
accurate and accurately evaluated surface tension face excess values and the direct surface tension
data are available, there is good agreement measurements. For SDS, there is a deviation at
between the isotherms derived from neutron low surfactant concentrations, and this was
reflectivity and surface tension, and this discussed by Hines et al. (1997a, b) and attributed S
has been demonstrated for the cationic to trace impurities of divalent ions such as Ca2+.
surfactant tetradecyltrimethylammonium bromide The value of the prefactor, m, in Eqs. 18 and
(Simister et al. 1992; Lu et al. 1993b) and 19, can be a factor of uncertainty in the interpre-
a range of different nonionic surfactants (Lu et al. tation of surface tension data, and can lead to
1994a, 2000). Given the adsorption isotherm incorrect interpretation. It is normally assumed
from neutron reflectivity measurements, it that m ¼ 1 for nonionic surfactants and m ¼ 2 for
should be possible to use the Gibbs equation to simple ionic surfactants, but there are situations
calculate the expected surface tension behavior where this is not the case. The earlier example of
such that: the rhamnolipid adsorption investigated by Chen
et al. (2010a, b) is an example of how surface
lnðc2
tension data can be misinterpreted. It is generally
Dg ¼ mRT Gd ln C (19) accepted that the rhamnolipids are nonionic in
ln c1 nature at low pH but anionic at high pH due to
a 20 b 10
8
15
6
χ2
χ2
10
4
5
2
0 0
0.15 0.20 0.25 0.30 10 15 20
105 x scattering length density/Å−2 Thickness/Å
c 200
150
χ2
100
50
0
1.6 1.8 2.0 2.2 2.4
1010 x surface excess/mole cm−2
Surfactant Adsorption at the Air–Solution and interface, at a solution concentration of 9 106 M and
Solid–Liquid Interfaces Studied by Neutron Reflec- for a mean area/molecule of 86 Å2. Variation in w2, “good-
tivity, Fig. 4 Least squares analysis of the neutron reflec- ness of fit” with (a) r, (b) t, and (c) r.t (# 1994 ACS
tivity data for dC12hE8 adsorbed at the air–water (nrw) Reproduced by permission of Lu et al. 1994a)
the protonation of the carboxyl groups on the made neutron reflection and surface tension mea-
rhamnose headgroup. At low pH, the surface ten- surements at the air–water interface for a series of
sion and neutron reflection data are in good agree- di-cationic quaternary ammonium Gemini com-
ment, and the mean area/molecule for R1 and R2 pounds, a,o -bis(N-alkyl dimethyl ammonium)
are 62 and 75 Å2 respectively. The interpretation hexane halides (Cn–C6–Cn, for n ¼ 8–16),
of surface tension data has previously implied that -bis(N-alkyl dimethyl ammonium)
a series of a,o
the area/molecule is much larger at high pH, and diethylether halides (Cn–C2OC2–Cn for n ¼
values > 100 Å2 are quoted. Whereas the more 8–16), and for C8–C12–C8, C12–C10–C12. The
direct neutron reflection data show that the variation in the experimentally determined
adsorption is invariant with pH, and that impor- Gibbs prefactors, determined from a comparison
tantly R1 and R2 are only weakly ionic at high pH. of the neutron reflection and surface tension data,
Due to effects such as ion association and is shown in Fig. 6 for Gemini surfactants with
premicellar aggregation, there are situations varying alkyl chain lengths and for different
where it is much more difficult to interpret sur- spacers and spacer lengths.
face tension data. A recent study by Li et al. With increasing alkyl chain length, the Gibbs
(2011), using neutron reflectivity to probe the prefactor varies from 3.0 to 1.5, whereas the
adsorption of quaternary Gemini surfactants, has value is 2.0 for the different C12–Cn–C12 Gem-
highlighted some of these issues. Li et al. (2011) ini surfactants. A model, consistent with the
interpreting surface tension data at concentra-
tions above the cmc are not present for the mea-
surements using neutron reflection. One of the
strengths of the neutron reflection method is that
it can provide unambiguously information about
adsorption above the cmc. For a range of non-
ionic surfactants neutron reflection measure-
ments, showed that (Lu et al. 1997) the
adsorption above the cmc was relatively constant.
In contrast neutron reflection measurements
showed that the adsorption increased above the
cmc for ionic surfactants, and this was demon-
strated for the cationic surfactant tetradecyltri-
methylammonium bromide by Simister et al.
(1992), and for SDS by Lu et al. (1993a, b, c).
The neutron reflectivity measurements give
more information than just adsorbed amounts,
providing access to the structure of the surface
layer (Lu et al. 2000). Using a direct method of
analysis based on the kinematic approximation, the
volume fraction distributions of individually deu-
terium-labeled components can be obtained. The
scattering length density profile, r(z), can be writ-
ten in terms of i discrete components using Eq. 2:
X
rðzÞ ¼ ni ðzÞbi (20)
i
where bi, ni(z) are the scattering lengths and num-

ber density profiles of the different components,
and from Eq. 12, this gives;
Surfactant Adsorption at the Air–Solution and
Solid–Liquid Interfaces Studied by Neutron Reflec- " #
tivity, Fig. 5 Surface tension, g, versus log of surfactant 16p2 X 2 XX
concentration calculated from neutron reflection derived RðQÞ ¼ 2 bi hii þ 2bi bj hij
Q i i j<i
surface excess values and compared with direct measure- S
ments of the surface tension, for (a) SDS, (b) dodecyl (21)
betaine, and (c) dodecyl maltoside (# 1997 ACS
Reproduced by permission of Hines et al. 1997a, b)
The hii factors are the self-partial
structure factors, hii ¼ jnîi ðQÞj2 , the cross- partial
neutron reflectivity and surface tension data, and structure factors, hij, are given by

with conductivity NMR, and fluorescence mea- hij ðQÞ ¼ Re nî ðQÞ^nj ðQÞ , and ni(Q) is the one-
surements was developed, and the prefactor var- dimensional Fourier transform of ni(z). The
iation was explained as a result of a combination self-partial structure factors relate directly to the
of effects due to ion association and dimerization. distributions of the individual components at
The Gibbs equations are usually only applied the interface, whereas the cross-partial structure
to surface tension data below the cmc, and it is factors depend on the relative positions of pairs
generally assumed that adsorption is constant of fragments along the direction normal to the
above the cmc. The difficulties associated with interface. The surfactant self-terms are usually
a 4 b 4
3 3
Gibbs prefactor
Gibbs prefactor
2 2
1 1
0 0
40 35 30 25 20 38 36 34 32 30 28 26
number of CH2 groups number of CH2 groups
Surfactant Adsorption at the Air–Solution and lengths, and (b) different spacer lengths. The solid lines
Solid–Liquid Interfaces Studied by Neutron Reflec- are guides to the eye (# 2011 ACS Reproduced by per-
tivity, Fig. 6 Experimentally determined Gibbs mission of Li et al. 2011)
prefactors for (a) Gemini’s with different alkyl chain
represented in terms of Gaussian distributions and and the cross-terms between two distributions,
the solvent in terms of a tanh distribution such that one of which is even and the other of which is
odd (for example, the solvent), are given by:
ni ðzÞ ¼ ni ðexpð4z2 =s2i ÞÞ (22)
his ¼ ðhi hs Þ1=2 sin Qdis (27)
ns ðzÞ ¼ n0 ½1=2 þ ð1=2Þ tanhðz=zÞ (23)
where dij, dis is the distance between the centers
The surfactant self-partial-structure factors for of the two distributions i, j, or i, s.
a Gaussian distribution are then: From a series of different reflectivity measure-
1 ments, using differently labeled combinations,
hi ðQÞ ¼ expðQ2 s2i =8Þ (24) the different partial structure factors can be
A2i
extracted and quantified as described above.
where Ai is the area/molecule of component i and This approach has been applied successfully to
si is its width. The solvent partial structure factor a range of different systems (Lu et al. 2000).
is given by: Almost the conceptually simplest measure-
2 ment is demonstrated using the alkyltrimethy-
zp zpQ lammonium bromide cationic surfactants,
hs ðQÞ ¼ n20 cos ech2 (25)
2 2 CnTAB, where deuterium labeling of the alkyl
chain, headgroup, and solvent gives rise to:
where n0 is bulk solution number density and z is
the width of the solvent distribution. The cross- rðzÞ ¼ bc nc ðzÞ þ bh nh ðzÞ þ bs ns ðzÞ (28)
partial-structure factors have been shown to be
a good approximation independent of any model and
assumption. The cross-terms between two distri-
16p2 2
butions that are exactly even about their centers Rð k Þ ¼ bc hcc þ b2h hhh þ b2s hss þ 2bc bh hch
are given by: k2
þ 2bc bs hcs þ 2bh bs hhs
hij ¼ ðhi hj Þ1=2 cos Qdij (26) (29)
Making six different neutron reflectivity mea-
1.2
surements for six different combinations of the a C12TAB
isotopically labeled components (alkyl chain, 1.0
headgroup, and solvent) enables the six partial
0.8
structure factors to be extracted, and this was
done for the alkyltrimethylammonuium bromide 0.6
cationic surfactants with dodecyl, tetradecyl,
0.4
hexadecyl, and octadecyl alkyl chains (Lu et al.
1994b; Lyttle et al. 1995). The volume fraction 0.2
distributions of the different components, their rel- 1.2
ative positions at the interface, and the total distri- b C14TAB
bution of surfactant are derived from the modeling 1.0
of the partial structure factors, as described earlier, 0.8
Eqs. 22–27, and are illustrated in Fig. 7.
The volume fraction distributions (for a mean 0.6
area/molecule 44 Å2) represent a rather disor-
0.4
dered structure at the interface, and are typical of Volume fraction
what has been measured for a range of different 0.2
surfactant systems (Lu et al. 2000). The alkyl
1.2
chain, headgroup, and solvent distributions are c C16TAB
relatively broad. The headgroup is totally 1.0
immersed in the solvent, and there is overlap
0.8
between the alkyl chain and headgroup (solvent)
distributions. Although the fully extended chain 0.6
length for C12 to C18 increases from 16.7 to
0.4
24.2 Å, the width of the alkyl chain distribution
hardly varies, even when allowance is made for 0.2
the contribution to the width from capillary 1.2
waves. However, the volume fraction (density) d C18TAB
of the chain region increases markedly from 0.5 1.0
to 0.8. Hence, the alkyl chains are in a disordered 0.8
state with many gauche conformations. The
headgroup region broadens slightly with increas- 0.6
ing alkyl chain length, and this is consistent with
0.4
a change in the conformation of the layer. Lu S
et al. (1995) used a more ambitious labeling 0.2
scheme to study the adsorption of hexadecyltri-
0.0
methylammonium bromide at the air–water inter- −20 −10 0 10 20
face. In that study, the alkyl chain of Distance normal to surface/Å
hexadecyltrimethylammonium bromide was Surfactant Adsorption at the Air–Solution and
labeled in groups of two or four methylene units Solid–Liquid Interfaces Studied by Neutron Reflec-
along the alkyl chain, as illustrated in Fig. 8a. tivity, Fig. 7 Volume fraction distributions for
From the series of neutron reflectivity measure- the alkyl chain (continuous line), headgroup (dotted
line), and solvent (dashed line), for (a) C12TAB,
ments for these different isotopically labeled sur- (b) C14TAB, (c) C16TAB, and (d) C18TAB
factants, the mean conformation of the chain at (# 1995 ACS Reproduced by permission of Lyttle et al.
the interface can be derived, as shown in Fig. 8b. 1995)
a
CH3 CH3
8
CH2 CH2
CH2 CH2
4
7
CH2 CH2
CH2 CH2
6
CH2 CH2
3
CH2 CH2
5
CH2 CH2
CH2 δcsl(8) CH2 δo4h(4)
4
CH2 2 CH2 b
CH2 CH2 15
3
CH2 CH2
Vertical distance/Å
CH2 CH2
2 1 10
CH2 CH2
CH2 CH2
1
CH2 CH2 5
+ +
N CH3 N CH3
h
h CH3 CH3
CH3 CH3
δhs δhs 0
0 5 10 15
solvent plane Horizontal distance/Å
Surfactant Adsorption at the Air–Solution and structure of the alkyl chain derived from the partial struc-
Solid–Liquid Interfaces Studied by Neutron Reflec- ture analysis, determined from the C2 and C4 labeling
tivity, Fig. 8 (a) Representation of the deuterium label- schemes (# 1995 ACS Reproduced by permission of Lu
ing scheme used for C16TAB, and description of the et al. 1995)
parameters used to describe the structure, (b) Mean
The results have provided an unprecedentedly For the ethoxylated nonionic surfactants, it is
detailed description of the alkyl chain conforma- possible to apply deuterium labeling schemes to
tion. The results show that the first part of the both the alkyl chain and ethylene oxide chain, and
alkyl chain (nearest to the headgroup) is oriented so use the partial structure factor approach to
closer to the surface normal than the outer parts of determine the ethylene oxide chain distribution
the chain which are progressively oriented away at the interface. This has been applied by Lu et al.
from the surface normal. They show that the (1993d) to a range of different nonionic surfac-
alkyl chain distribution is well represented by tants. Deuterium labeling the alkyl chains
a Gaussian distribution, consistent with the and ethylene oxide chains provides a complete
assumptions based on other less detailed mea- description of the surfactant structure at the inter-
surements. The roughness parameters extracted face. Figure 9 shows a comparison of the
from the data indicate that the roughness of the volume fraction distributions of different
solvent and surfactant are coupled, and that the ethoxylated nonionic surfactants, C12E2, C12E4,
values are consistent with those calculated from C12E6, and C12E8, from Lu et al. (1993d),
capillary wave theory. Lu et al. (2000).
1.20 adsorb onto hydrophilic surfaces of opposite
a c
charge by electrostatic attraction, and nonionic
0.80 polyoxyethylene surfactants adsorb by hydrogen
bonding between surface hydroxyl groups and
0.40
the ether oxygen on the ethoxylate headgroups.
Volume fraction
The adsorption is cooperative in nature and

results in surface self-assembly at a surface
1.20
b d cmc. At a hydrophobic surface, the adsorption is
in the form of a monolayer, like the air–water
0.80 interface.
Hence unlike the air–water interface, the solid
0.40 surface is generally not inert. Furthermore, it is
generally not possible to make the interface invis-
0.00 ible as there is always an oxide or gel layer on
−20 −10 0 10 20 −10 0 10 20
hydrophilic surfaces such as silicon or quartz, and
Distance normal to surface/Å
hydrophobic surfaces or surfaces with other func-
Surfactant Adsorption at the Air–Solution and tionalities are established via a grafted surface
Solid–Liquid Interfaces Studied by Neutron Reflec- layer. An example of this is shown in Fig. 10,
tivity, Fig. 9 Volume fraction distributions for the for the silicon–water interface (Penfold et al.
nonionic surfactants C12Em for (a) m ¼ 2, (b) m ¼ 4,
(c) m ¼ 6, and (d) m ¼ 8, (solid line) alkyl chain, (dashed 1997).
line) ethylene oxide chain, (dot-dot-dashed line) solvent, The data in Fig. 10 illustrate some important
and (dot-dashed line) total (# 1993 ACS Reproduced by features of measurements at the solid–solution
permission of Lu et al. 1993d) interface. In D2O, the scattering length differ-
ence, Dr, from silicon to D2O is positive and at
A detailed analysis of the parameters that con- sufficiently low k total reflection is observed,
stitute the volume fraction distributions in figure whereas in H2O Dr is negative and as such
reveals some interesting features, and shows that there is no total reflection. For the measurement
the behavior of the ethoxylated nonionic is dif- in cmSi, a mixture of D2O and H2O with the same
ferent to the ionic surfactants. The changes in the scattering length density as silicon, there should
thickness of the alkyl chain and ethylene oxide be no reflected signal, as was observed at the air–
regions and the separation between them are con- nrw interface. For this contrast, the residual sig-
sistent with an increasing intermixing of the alkyl nal arises from a thin silicon oxide layer at the
and ethylene oxide chains as m increases. For the interface. The solid lines in the figure are calcu-
larger ethylene oxide groups, the width of the lated curves for a 20 Å hydrated oxide layer at the
ethylene oxide chain region also implies that, silicon–water interface. Such a characterization S
like the alkyl chains, they are disordered and far is a necessary prerequisite prior to any measure-
from fully extended. Furthermore, it implies that ments of surfactant adsorption.
there is a high occurrence of gauche conforma- Figure 11 shows the neutron reflection for the
tions, and that the ethylene oxide chain behaves adsorption of the cationic surfactant C16TAB, for
broadly similar to that of an anchored polymer the isotopic combination h-C16TAB/D2O in the
chain. concentration range 104 M to 2 mM.
With increasing C16TAB concentration,
Liquid–Solid Interface a pronounced interference fringe is increasingly
Surfactant adsorption at the solid solution inter- visible. The solid lines in the figure are calculated
face depends critically upon the nature of the lines for an adsorbed layer of C16TAB at the
surfactant, the properties of the surface, and the interface, with a thickness 30 Å and an
specific interaction between the surfactant and adsorbed amount, Г, 6 1010 mol cm2 at
the surface. In general terms, ionic surfactants saturation. From the modeling, a thickness and
Surfactant Adsorption at
Studied by Neutron
Reflectivity,
Fig. 10 Neutron
reflectivity from silicon–
water interface for different
isotopic compositions of
the water, (●) D2O, (o)
H2O, and (D). The solid
lines are calculated
reflectivity curves as
described in the text
(# 1997 ACS Reproduced
by permission of Penfold
et al. 1997)
Studied by Neutron
Reflectivity,
Fig. 11 Specular
reflectivity versus wave
vector transfer, k, for
h-C16TAB/D2O at the
hydrophilic silica surface:
(-) bare surface; and (□)
104 M, (△) 2 104 M,
(~) 5 104 M, (○)
103 M, and (●) 2
103 M h-C16TAB. The
solid lines are model fits, as
described in the text. Each
profile is shifted vertically
for clarity (# 2007 ACS
of Penfold et al. 2007)
scattering length density is obtained and the The thickness of the adsorbed layer of surfac-
adsorbed amount, Г, obtained from, tant and the adsorbed amount are not consistent
with a monolayer adsorption. The adsorption
dðrs r2 Þ takes the form of self-assembly at the surface, in
G¼ (30)
Na V ðrs ra Þ the form of a “flattened” micellar structure or
fragmented bilayer at the surface. Indeed the
where d is the layer thickness, rs, ra, and r2 are reflectivity from the adsorbed C16TAB layers
the solvent, adsorbate, and layer scattering length can be equally well modeled using three layers
density and V is the adsorbate volume. which represent a bilayer structure at the
Studied by Neutron
Reflectivity,
Fig. 12 Schematic
representation of the model
use to simultaneously fit the
data for C16TAB adsorption
at the silicon surface at
different isotopic
compositions, as indicated
in the figure (# 1996 ACS
of Fragneto et al. 1996a)
interface, where the initial layer, d1, represents due to the hydrogen bonding between the ether
the headgroup region adjacent to the solid sur- oxygen group of the ethoxylate and the surface
face, the second layer, d2, the interdigitated or hydroxyl groups. The strength of that interaction
partially interdigitated alkyl chains from both and hence the adsorption depends strongly on
sides of the bilayer, and the third layer, d3, the charged state of the surface and hence pH, as S
headgroups adjacent to the solvent. Using was illustrated by Penfold et al. (2002). The
a similar partial labeling scheme on the C16TAB effects of pH on the nonionic surfactant adsorp-
as described earlier for measurements at the air– tion are illustrated in Fig. 13.
water interface, Fragneto et al. (1996a) were able The figure shows the reflectivity for
to probe the structure of the surface layer in more 104 M h-C12E6 adsorbed at the silicon–D2O
detail. By deuterium labeling the headgroup and interface at pH 7 and pH 2.4, and the comparison
blocks of four methylene groups along the C16 with the reflectivity for the bare silicon–D2O
alkyl chain, it was possible to establish that the interface. The departure of the measured reflec-
surface aggregates strongly resemble bilayer tivity from that for the bare interface is
fragments. The resulting structure is shown sche- a measure of the amount absorbed, and
matically in Fig. 12. the marked change between pH 7 and 2.4 is
The adsorption of the ethoxylated nonionic indicative of the pH dependence of the adsorp-
surfactants at the hydrophilic silicon surface is tion. At pH 7, the adsorbed amount is
Studied by Neutron
Reflectivity,
Fig. 13 Neutron
reflectivity profiles for (●)
Si/SiO2/D2O and for
104 M hC12hE6/D2O/
0.1MNaBr at pH 7.0 (△)
and at pH 2.4 (~) (# 1997
ACS Reproduced by
permission of Penfold et al.
1997)
Studied by Neutron
Reflectivity,
Fig. 14 Schematic
representation of the
adsorbed layer of C12E6 at
the silica–aqueous solution
interface. (~) (# 1992
ACS Reproduced by
permission of McDermott
et al. 1992)
0.6 1010 mol cm2 and the fractional cov- a bilayer structure at the interface, the data
erage is 10 %, whereas at pH 2.4, the for C12E6 can only be sensibly modeled using
adsorbed amount is 4.9 1010 mol cm2 a three layer model. This is as a result of the
and the fractional coverage is 50 %. As much larger contribution from the ethylene
discussed by Penfold et al. (2002), these are oxide headgroups and their partial hydration,
typical of the amounts adsorbed for C12E6, and and the values for d1, d2, and d3 in such a three
depend critically upon pH and the surface his- layer model for the data at ph 2.4 in Fig. 13 are
tory. Unlike the modeling of the C16TAB 12.0, 20.0, and 14.0 Å, respectively. A sche-
adsorption described earlier, where the adsorp- matic representation of the surface structure
tion data could be equally well modeled as can be found in the earlier study of McDermott
a single layer or three layers representing et al. (1992), as shown below in Fig. 14.
4.0 layer, it is possible to determine the extent of
penetration of the surfactant layer into the OTS
Coverage x1010/mol cm−2
surface layer.
3.0
Using self-assembled monolayers, such as
OTS, surfaces with other functionalities can be
2.0 produced, and include alkenes and hydroxylated
surfaces. Polymer films and cellulose surface
can be created by adsorption, Langmuir–
1.0 cmc Blodgett (L-B) deposition, or spin coating.
Penfold et al. (2007), and Tucker et al. (2010)
0.0 have studied the nature of hydrophobic and
−3 −2 −1 0 1
hydrophilic cellulose surfaces deposited onto
log (concentration/cmc)
silicon by L-B deposition, and have character-
Surfactant Adsorption at the Air–Solution and ized the nature of nonionic and cationic surfac-
Solid–Liquid Interfaces Studied by Neutron Reflec- tant adsorption onto such surfaces and their
tivity, Fig. 15 Comparison of the adsorption of (■)
C12E4, (●) AOT onto an OTS surface and (○) AOT
impact upon the structure of the cellulose sur-
adsorbed at the air–water interface (# 1996 ACS face. Penfold et al. (2007) produced thin cellu-
Reproduced by permission of Fragneto et al. 1996b) lose films by L-B deposition of trimethylsilyl
cellulose (TMSC) onto silicon with a total
thickness 50 to 100 Å, depending upon the
number of repeated dipping cycles. The films
Increasingly the neutron reflection studies are are initially hydrophobic because of the methyl
being made on an extended range of solid sur- groups on the TMSC, but can be rendered
faces. As described in the introduction, there are hydrophilic by cleaving the methyl groups in
other crystalline solid surfaces that are suitable HCl vapor. Figure 16 shows the characterization
for neutron reflection, and these include quartz, of one such hydrophobic cellulose surface using
sapphire, and calcite. The alternative approach is neutron reflectivity with different solvent con-
to use the silicon as a substrate on which to graft trasts, D2O, H2O, and cmSi.
a layer with specific functionality. This is The cellulose film in Fig. 16 has a total thick-
a convenient way of producing a model hydro- ness 65 Å, and is best described by two layers,
phobic surface, for example, by grafting an inner more dense layer 45 Å and an outer
a hydrocarbon layer, such as octadecyltri- less dense layer 20 Å. The observed structure is
chlorosilane (OTS) onto the hydrophilic silicon typical of what is observed and the exposure to
surface. This also provides the opportunity to HCl vapor to render the surface hydrophilic
deuterium label or partially label the OTS layer, slightly thins the cellulose layer. Cationic surfac- S
and this can provide enhanced sensitivity to the tants, such as C16TAB, and the nonionic surfac-
adsorbed layer and the ability to probe the struc- tants, such as C12E6, adsorb readily onto both the
ture of the OTS layer and the response of that hydrophobic and hydrophilic surfaces. Figure 17
layer to subsequent adsorption by surfactants. shows a comparison of the adsorption of C16TAB
This was illustrated by Fragneto et al. (1996b), onto hydrophilic silica and cellulose and onto
in comparison of Aerosol-OT and C12E4 adsorp- hydrophobic cellulose.
tion at the hydrophilic and hydrophobic surfaces, The adsorption of C16TAB onto hydrophilic
see Fig. 15. cellulose is in the form of a bilayer, and similar to
On the OTS hydrophobic surface, the adsorp- that observed at the hydrophilic silica surface, as
tion is in the form of a monolayer, and there are shown in Fig. 17, whereas the lower adsorption
differences in the C12E4 and AOT adsorption. onto the hydrophobic cellulose surface is associ-
Furthermore, apart from establishing the ated with monolayer adsorption. The results
adsorbed amount and structure of the adsorbed reported by Penfold et al. (2007) further show
1e+0
Studied by Neutron
Reflectivity, 1e−1
Fig. 16 Specular
reflectivity versus wave
vector transfer, Q, for the 1e−2
Reflectivity
hydrophobic cellulose
surface, in (●) D2O, (○)
1e−3
H2O, (△) cmSi. The solid
lines are model fits, as
discussed in the text 1e−4
et al. 2007) 1e−5
1e−6
0.01 0.1
8
Adsorbed amount (x10−10 mol cm−2)
the Air–Solution and 6
Studied by Neutron
Reflectivity,
Fig. 17 Adsorbed amount
of C16TAB, G, 4
( 1010 mol cm2) as
a function of surfactant
concentration onto (●)
hydrophilic silica, (○)
2
hydrophilic cellulose, and
(~) hydrophobic cellulose.
The lines are calculated
curves, assuming
a Langmuir isotherm 0
(# 2007 ACS Reproduced 1e−4 1e−3 1e−2
et al. 2007) Surfactant concentration (M)
that on the hydrophobic cellulose surface, some et al. (2010). The results showed that for
intermixing between the cellulose, surfactant, C12E6 adsorption and C12E6/C16TAB mixed
and solvent occurs, whereas on the hydrophilic adsorption, there was a greater penetration of
surface, no intermixing is observed. The surfactant and solvent into the outer layer of the
adsorption of nonionic surfactants and non- cellulose film than was observed for C16TAB
ionic–cationic surfactant mixtures onto both alone. Furthermore, some of the effects were
hydrophobic and hydrophilic cellulose surfaces not entirely reversible, as highlighted in
is slightly different, as described by Tucker Fig. 18.
Surfactant Adsorption at 1e+1
Studied by Neutron 1e+0 before C16TAB / C12E6 adsorption
Reflectivity, after C16TAB / C12E6 adsorption
Fig. 18 Neutron
1e−1
reflectivity for Silicon/
Cellulose/D2O for
Reflectivity
a hydrophobic cellulose 1e−2
surface, before (○) and
after (●) C16TAB/C12E6
adsorption. The solid lines 1e−3
are model calculations as
described in the text
1e−4
by permission of Tucker
et al. 2010) 1e−5
1e−6
0.01 0.1
The data in Fig. 18 shows the neutron reflec- relative minor changes in the characterization
tivity from a hydrophilic cellulose surface before of the oxide layer or surface roughness. In the
and after exposure to 3 104 M C16TAB/C12E6, strong adsorption to the surface, the polyelectro-
and subsequent rinsing in D2O. The initial surface lyte charge reverses the surface charge; so, for
film of cellulose was 80 Å, modeled in the example, Penfold et al. (2004) showed that the
figure as two layers 50 and 30 Å respectively adsorption of the cationic polyelectrolyte poly
and with a volume fractions of cellulose 70 % (dimethyldiallylammonium bromide) (poly-
in the inner layer, and 20 % in the outer layer. DMDAAC) onto the anionic silica surface
After exposure to the C16TAB/C12E6, the reflec- made the surface cationic. This was illustrated,
tivity is different and implies some irreversible as shown in Fig. 19, by the now strong adsorp-
swelling/adsorption into the cellulose film. tion of SDS onto the surface.
Indeed the detailed modeling, solid line in the The reflectivity data in Fig. 19 show the reflec-
figure, indicates that the total film thickness is tivity from the bare silica surface, and the reflec-
110 Å and that there is an irreversible swelling tivity following adsorption of poly-DMDAAC S
of the outer layer by the incorporation of surfac- onto that surface is indistinguishable from the
tant and/or solvent. bare interface. Furthermore, there is no SDS
It is not always possible to probe the structure adsorption onto the bare surface, and again the
of the surface modification layers in such detail resulting reflectivity is indistinguishable from the
using neutron reflectivity. For example, Penfold bare surface. However, the pronounced interfer-
et al. (2004) have shown that cationic polyelec- ence fringe visible in the reflectivity for the SDS
trolytes adsorb strongly onto the anionic surface adsorption post poly-DMDAAC adsorption is
of silica, but in a flat conformation on the sur- indicative of pronounced SDS adsorption due to
face. Neutron reflectivity measurements and the charge reversal of the surface. Further manip-
subsequent calculations show that polymer ulation of the surface properties is then possible
adsorption such as a thin monolayer ( 5 Å) is by the subsequent adsorption of oppositely
not really visible and indistinguishable from charged polyelectrolytes.
Surfactant Adsorption at 1e+1

1e+0
Studied by Neutron
Reflectivity,
Fig. 19 Specular 1e−1
reflectivity for
5 104 M h-SDS/D2O (●) 1e−2
Reflectivity
on the poly-DMDAAC-
coated silica surface, and
1e−3
(~) the bare silica surface.
The solid line is
a calculated curve as 1e−4
discussed in the text
(# 2004 ACS Reproduced 1e−5
et al. 2004)
1e−6
1e−7
0.01 0.1
Scattering Vector, Q (ang−1)
Summary and Future Prospects here, and predominantly in the current literature,
represent measurements at equilibrium. There is
In this entry, the use of neutron reflectivity to now the potential to study kinetic effects associated
study surfactant adsorption at the air–solution with surfactant adsorption, and provide informa-
and solid–solution interfaces has been described. tion to parallel, complement, and extend that from
It is shown how at both interfaces, the technique ellipsometry and dynamic surface tension.
provides absolute information about adsorbed
amounts and more detailed structural informa- Cross-References
tion. In the determination of adsorbed amounts,
the technique overcomes many of the difficulties ▶ Adsorption Isotherm
and uncertainties associated with techniques such ▶ Critical Micelle Concentration
as surface tension. It also provides the opportu- ▶ Detergency
nity and additional insights to enable a more ▶ Detergents
detailed interpretation of data from techniques ▶ Gibbs Adsorption Isotherm
such as surface tension. The detailed structural ▶ Interface, Air-Liquid
information that is also available provides new ▶ Interface, Solid-liquid
insights into how the structural factors affect the ▶ Surface Tension
packing and hence adsorption at the interface. ▶ Surfactant Adsorption
The improvements in neutron sources and par- ▶ Surfactant Molecule
ticularly in the neutron reflection instrumentation ▶ Surfactant Monolayers
will provide greater sensitivity and range. This will ▶ Surfactants
enable, at the air–water interface, more dilute
systems to be studied, and reduce the need for
extensive deuterium labeling. At the solid–solution
References
interface, a wider range of substrates, including
noncrystalline materials, will be accessible. It also Abe M, Scamehorn JF (2005) Mixed surfactant systems,
raises the possibility of probing the lateral structure vol 124, Surfactant science series. Marcel Dekker,
and distribution at the interface. The data presented New York
Chen ML, Penfold J, Thomas RK, Smyth TJP, Perfumo A, air-water interface: the configuration of th ethylene
Marchant R, Banat IM, Stevenson P, Parry A, Tucker I, oxide chain. J Phys Chem 97:8012
Grillo I (2010a) Solution self-assembly and adsorption Lu JR, Li ZX, Thomas RK, Staples EJ, Thompson L,
at the air-water interface of the monorhamnose and Tucker I, Penfold J (1994a) Neutron reflection from
dirhamnose rhamnolipids and their mixtures. Lang- alayer of monododecyl octaethylene glycol adsorbed
muir 26:18281 at the air-liquid interface: the structure of the layer and
Chen ML, Penfold J, Thomas RK, Smyth TJP, Perfumo A, the effect of temperature. J Phys Chem 98:6559
Marchant R, Banat IM, Stevenson P, Parry A, Tucker I, Lu JR, Hromadova M, Simister EA, Thomas RK, Penfold
Grillo I (2010b) Mixing behaviour of the biosurfactant, J (1994b) Neutron reflection from hexadecyltrimethy-
rhamnolipid, with a conventional anionic surfactant, lammonium bromide adsorbed at the air/water inter-
sodium dodecyl sulfonate. Langmuir 26:17958 face; the variation of the hydrocarbon chain
Fainermann VB et al (2006) Dynamic surface tension of distribution with surface concentration. J Phys Chem
micellar solutions in the milisecond and sub- 98:11519
millisecond time range. J Colloid Interface Sci 302:40 Lu JR, Li ZX, Smallwood J, Thomas RK, Penfold J (1995)
Fragneto G, Thomas RK, Rennie AR, Penfold J (1996a) Detailed structure of the hydrocarbon chain in
Neutron reflection from hexadecyltrimethy- a surfactant monolayer at the air/water interface: neu-
lammonium bromide adsorbed on smooth and rough tron reflection from hexadecyltrimethylammonium
silicon surfaces. Langmuir 12:6036 bromide. J Phys Chem 99:8233
Fragneto G, Lu JR, McDermott DC, Thomas RK, Rennie Lu JR, Su TJ, Li ZX, Thomas RK, Staples EJ, Tucker I,
AR, Gallagher PD, Satija SK (1996b) Neutron reflec- Penfold J (1997) Structure on monolayers of
tivity study of the adsorption of Aerosol-OT on self- monododecyl dodecaethylene glycol at the air-water
assembled monolayers on silicon. Langmuir 12:477 interface studied by neutron reflection. J Phys Chem
Heavens OS (1955) Optical properties of thin films. B 101:10332
Dover, London Lu JR, Thomas RK, Penfold J (2000) Surfactant layers at
Hines JD, Thomas RK, Garrett PR, Rennie GK, Penfold the air-water interface, structure and composition. Adv
J (1997a) Investigation of mixing in a binary surfactant Colloid Interface Sci 84:143
solution by neutron reflection and surface tension: Lyttle DJ, Lu JR, Su TJ, Thomas RK, Penfold J (1995)
anionic/nonionic and zwitterionic/nonionic mixtures. Structure of a dodecyltrimethylammonium bromide
J Phys Chem B 101:9215 layer at the air-water interface determined by neutron
Hines JD, Garrett PR, Rennie GK, Thomas RK, Penfold reflection: comparison of the monolayer structure of
J (1997b) The structure of an adsorbed layer of n- cationic surfactants with different chain lengths. Lang-
dodecyl-N, N-dimethylamino acetate at the air- muir 11:1001
solution interface, as determined by neutron reflectiv- Manne S, Gaub HE (1995) Molecular organisation of
ity. J Phys Chem B 101:7121 surfactants at solid-liquid interfaces. Science 270:1480
Lekner J (1987) Theory of reflection. Martinus Nijhoff, McDermott DC, Lu JR, Lee EM, Thomas RK, Rennie AR
Dordrecht (1992) Study of the adsorption from aqueous solution
Levitz P (2002) Adsorption of nonionic surfactants at the of hexaethylene glycol monododecyl ether on silica
solid-water interface. Colloid Surf A 205:31 substrates using the technique of neutron reflection.
Li PX, Dong CC, Thomas RK, Penfold J, Wang Y (2011) Langmuir 8:1204
Neutron reflectometry of quaternary Gemini surfac- Penfold J (2002) Neutron reflectivity and soft condensed
tants as a function of alkyl chain length: anomalies matter. Curr Opin Colloid Interface Sci 7:139
arising from ion association and premicellar aggrega- Penfold J, Thomas RK (1990) The application of the
tion. Langmuir 27:2575 specular reflection of neutrons to the study of surfaces
Lu JR, Simister EA, Thomas RK, Penfold J (1993a) Struc- and interfaces. J Phys Condens Matter 2:1369 S
ture of an octadecyltrimethylammonium bromide Penfold J, Thomas RK (2007) Probing surfactant adsorption
layer at the air/water interface determined by neutron at the solid-solution interface by neutron reflectometry.
reflection: systematic errors in reflectivity measure- In: Imae T (ed) Advanced chemistry of monolayers at
ments. J Phys Chem 97:6024 interfaces. Elsevier, Amsterdam, p 87, Chapt 4
Lu JR, Morroco A, Su TJ, Thomas RK, Penfold J (1993b) Penfold J, Thomas RK (2010) Mixed surfactants at the air-
Adsorption of dodecyl sulphate surfactants with mono- water interface. RSC Ann Rep Prog Chem Sect
valent metal counterions at the air-water interface C 106:14
studied by neutron reflection and surface tension mea- Penfold J, Staples EJ, Tucker I, Thompson LJ, Thomas RK
surements. J Colloid Interface Sci 158:303 (1997) The adsorption of mixed surfactants at the
Lu JR, Simister EA, Thomas RK, Penfold J (1993c) Struc- hydrophilic silica surface from aqueous solution: stud-
ture of the surface of a surfactant solution above the ied by specular neutron reflection. Langmuir 13:6638
cmc. J Phys Chem 97:13907 Penfold J, Staples E, Tucker I (2002) On the consequences
Lu JR, Li ZX, Thomas RK, Staples EJ, Trucker I, Penfold of surface treatment on the adsorption of nonionic
J (1993d) Neutron reflection from a layer of surfactants at the hydrophilic solid-solution interface.
monododecyl hexaethylene glycol adsorbed at the Langmuir 18:2967
S 1090 Surfactant Adsorption Kinetics
Penfold J, Tucker I, Staples I, Thomas RK (2004) Manip- Dynamic surface tension; Kinetics of adsorption;
ulation of the adsorption of ionic surfactants onto Surface thermodynamics; Surfactant mixture;
hydrophilic silica using polyelectrolytes. Langmuir
20:7177 Surfactant solution
Penfold J, Tucker I, Petkov JT, Thomas RK (2007) Sur-
factant adsorption onto cellulose surfaces. Langmuir
23:8357 Definition
Russell TP (1990) X-ray and neutron reflectivity for
the investigation of polymers. Mater Sci Rep
5:171 The dynamics of adsorption layers of surfactants
Simister EA, Thomas RK, Penfold J, Aveyard R, Binks and their mixtures is important for many modern
BP, Cooper P, Fletcher PDI, Lu JR, Sokolowski technologies. Their formation follows essentially
A (1992) Comparison of neutron reflection and surface
tension measurements on the surface excess of a diffusion-controlled mechanism. There are spe-
tetradecyltrimethylammonium bromide layers at the cific differences in the layer properties at water/
air-water interface. J Phys Chem 96:1383 air and water/oil interfaces due to the interaction
Taylor DJF, Thomas RK, Penfold J (2007) Adsorption of with the oil molecules. The easiest method of
polymer-surfactant mixtures at the air-water interface.
Adv Colloid Interface Sci 132:69 characterization is the measurement of the
Tiberg F, Jonsson B, Tang J, Lindman B (1994) dynamic surface or interfacial tensions. Most
Ellipsometry studies of the self-assembly of nonionic sensitive to surfactants’ molecular peculiarities
surfactants at the silica-water interface: equilibrium is the frequency dependence of the dilational
aspects. Langmuir 10:2294
Tucker I, Petkov JT, Penfold J, Thomas RK viscoelasticity. The dynamics of adsorption at
(2010) Adsorption of nonionic and mixed nonionic- the water/air interface is essentially quantita-
cationic surfactants onto hydrophilic and hydrophobic tively understood, while at the water/oil interface
cellulose thin films. Langmuir 26:8036 there are yet basic questions.
Wennerstrom H, Lindman B, (1979) Micelles, physical
chemistry of self association. Phys Rep 52:1
Overview
Surfactant Adsorption Kinetics The adsorption of surfactants at liquid interfaces

is essential for many technological processes.
N. Mucic1, A. Javadi1, M. Karbaschi2, Adsorbed surfactants do not only reduce the
A. Sharipova3, V. B. Fainerman4, interfacial tension but provide also other impor-
E. V. Aksenenko5, Nina M. Kovalchuk1,6 and tant properties to the liquid interface, such as
R. Miller1 changes in the interfacial charge, and mechanical
1
Max Planck Institute of Colloids and Interfaces, properties like the dilational viscoelasticity. The
Potsdam/Golm, Germany fundamentals of the thermodynamic of adsorp-
2
Sharif University of Technology, Tehran, Iran tion are discussed in section “Adsorption Iso-
3
International Postgraduate Institute “Excellence therms for Surfactants,” giving an overview of
Polytech” of Kazakh National Technical classical models as well as most recent refine-
University, Almaty, Kazakhstan ments, such as the assumption of orientational
4
Donetsk Medical University, Donetsk, Ukraine changes, formation of 2-D aggregates, or an
5
Institute of Colloid Chemistry and Chemistry of intrinsic compressibility of adsorbed layers.
Water, Kiev, Ukraine The accurate knowledge of the equilibrium
6
Institute of Biocolloid Chemistry, Kiev, Ukraine adsorption state is the basis for a quantitative
understanding of the dynamics of surfactant
adsorption layers. Experimental examples are
Keywords given in section “Adsorption Isotherms of
Standard Surfactants” for two selected homolo-
Diffusion controlled adsorption; Dilational gous series of surfactants, that is, for the homol-
visco-elasticity; Dynamic interfacial tension; ogous series of the cationic surfactants alkyl
Surfactant Adsorption Kinetics 1091 S
trimethyl ammonium bromides CnTAB (n ¼ 10, Section “Dynamic Surface/Interfacial Ten-
12, 14, 16) at the water/air and the homologous sions of Surfactant Solutions” deals mainly with
series of the anionic surfactants sodium alkyl the adsorption dynamics of surfactants at liquid
sulfates NaCnSO4 (n ¼ 10, 12, 14, 16) at the interfaces, with emphasis on the adsorption kinet-
water/hexane interface, respectively. ics measured by tensiometry and on the dilational
First ideas on the dynamics of the adsorption rheology as measured by oscillating drops or
process go back to Milner who has shown in 1907 bubbles. On the basis of selected experimental
that the surface tension of a surfactant solution data, the application of quantitative theories, the
changes with time. Ross later in 1945 was the first main mechanisms and phenomena observed at
who claimed that it must be the diffusion- interfaces between water and gas (mainly air) or
governed transport of surfactants toward the sur- water and an organic solvent, is discussed. This
face that controls the adsorption and hence the data analysis is focused on single surfactants as
change in surface tension. Right after this pro- well as mixtures of two compounds with
posal, in 1946, Ward and Tordai formulated the a respective difference in surface activity. Also
first quantitative model to describe the adsorption an example is given for a surfactant soluble in
process of surfactant molecules on the basis of both liquid phases.
a diffusional transport process. Since then, many Particular experimental efforts are needed to
additional mechanisms were taken into consider- measure data for short adsorption times at the
ation, but finally the pure diffusion-controlled liquid-liquid interface. Some measuring tech-
adsorption model is the one generally used now niques require special attention, mainly from
for describing experimental data quantitatively. a hydrodynamic point of view in order to separate
The main reason for deviations of experimental real adsorption effects from those caused by the
results from the diffusion model was essentially flow of liquid and the deformation of the interfa-
the presence of surface-active impurities as it was cial area. This is true for growing drops and
shown by Lunkenheimer and Miller in 1979. bubbles but also for experiments with harmonic
A brief summary of the history with a detailed area perturbations, such as oscillating drops and
analysis of the diffusion model and its mathemat- bubbles. In section “Dilational Viscoelasticity of
ical solution are given in section “Diffusion Con- Surfactant Solutions,” the state of the art for the
trolled Adsorption.” dilational rheology of surfactant interfacial layers
Since the important work by Aveyard and is summarized, and experimental examples are
Haydon in 1965, it is known that the interaction given.
of adsorbing surfactants with the oil molecules at This entry shows that the adsorption kinetics
a liquid-liquid interface has a significant impact of surfactants at the water/air interface is essen-
on the adsorption layer structure. However, due tially understood quantitatively. From literature
to lack of reliable experimental instruments and data gained for many surfactants at well-defined S
theoretical models, a quantitative picture is yet experimental conditions, we know that so far all
pending. Only recently in 2010, it has been shown “normal” surfactants adsorb according to
by Pradines et al. for the homologous series of a diffusion mechanism. This is true for ionic
alkyl trimethyl ammonium bromides that the and nonionic surfactants, supposed a suitable
chain length of the surfactant and the surface adsorption model has been applied as basis for
coverage are relevant for the incorporation of the quantitative data analysis. Surface active
oil molecules into this type of adsorption layers. impurities have a significant impact on experi-
The remarkable solubility of surfactants, most mental findings and need to be considered care-
of all of nonionic surfactants, in both bulk phases fully in order to avoid misinterpretations.
is another peculiarity for liquid-liquid interfaces. The situation is much more complicated when
In 1997, Liggieri et al. showed that even for such interfaces between two immiscible liquids are
a complex situation, a quantitative description of investigated. There are already some general
the adsorption process is possible. ideas about the impact of the type of oil; however,
we are yet rather far from a quantitative under- molecules on the structure of adsorption layers at
standing. Depending on the length of the surfac- the water/oil interface is rather complex. So far
tant aliphatic chain and the type of oil, different we have no general idea on their effect on the
interfacial structures are discussed which may dynamic interfacial behavior, in particular at very
include or not the oil molecules in the interfacial short adsorption times or very fast perturbations.
layer. An important peculiarity, for example, is One of the very promising tools from the
that remarkable changes in the interfacial tension experimental point of view is the direct incorpo-
can be observed at surfactant bulk concentrations ration of the high-speed video technique into
at which absolutely no change in surface tension measuring processes. Recording, for example,
at the solution/air surface is observed. Experi- of fast deformations of a liquid interface would
ments on the surfactant adsorption at the water/ give access to accurate information on the geom-
air interface in presence of alkane vapor in the air etry of drops or bubble, which are the major
phase represent a kind of bridge between the objects in dynamic interfacial studies.
water/air and water/oil interfaces. This provides A second important extra source of knowledge
additional information needed to understand the for an improved understanding of liquid interfa-
incorporation of oil molecules between adsorbed cial layers will be accessible by the great progress
surfactants at the corresponding water/vapor and in the development of powerful experimental
consequently water/oil interface; however, more tools, such as the capillary pressure technique.
of these investigations are required. Once such technique is supported by advanced
In this entry, we have also shown that the quantitative simulations, based, for example, on
dilational viscoelasticity is very sensitive to computational fluid dynamics (CFD) and molec-
peculiarities in surfactant adsorption layers, ular dynamics (MD) simulations, it will allow to
much more than the equilibrium surface tension access a new level of understanding. With the
isotherm does. The available theoretical models freely accessible OpenFOAM (Open Field Oper-
allow a semiquantitative description of the ation and Manipulation) platform, there is
behavior of mixed surfactant layers. Even tech- a simulation instrument that will allow us solving
nical surfactants can be handled in a way provid- very complex problems, even processes under
ing some effective values for the main parameters highly dynamic conditions.
of the applied theoretical models.
The present state of the art is very favorable
for a deeper understanding of rather complicated Introduction
systems, such as foams and emulsions. Most of
the formulation work in many fields of applica- The adsorption of surfactants at liquid interfaces
tion is yet based on empirical knowledge. Ques- is essential for many technological processes.
tions regarding the most important interfacial Adsorbed surfactants do not only reduce the
properties and how these are correlated to the interfacial tension but provide also various other
behavior of liquid films are still not clearly properties of the liquid interface, such as changes
answered. Moreover, the design of foams or in the interfacial charge and mechanical proper-
emulsions in practical applications based on ties like dilational elasticity and viscosity.
some fundamental interfacial properties of sur- As one of the first, Milner (1907) has clearly
factants and/or their mixtures remains still shown that the surface tension of solutions
a dream for the interfacial and colloid scientists. containing surfactants is a process that requires
There are still a number of open questions. Yet a respective time. Doss (1938) proposed that the
quite unclear, despite the very advanced recent change in surface tension with time is caused by
analysis by Ivanov et al. given in 2007, are the the fact that the surfactant molecules need to
conditions at which the ionic charge of reach the interface in order to adsorb and propose
a surfactant molecule plays a significant role in an empirical transfer mechanism. Later, Ross
the adsorption process. Moreover, the role of oil (1945) was the first who claimed that it must be
the diffusion-governed transport of surfactants layer at such water-oil interfaces. A complete
toward the surface that controls the adsorption picture of the effect of the alkyl chain length of
and hence the change in surface tension. Right the surfactant and the oily bulk phase is yet
after this proposal, the first quantitative model pending.
was proposed by Ward and Tordai (1946) who For liquid-liquid interfaces, there is another
described the adsorption process on the basis of peculiarity worth mentioning, the remarkable
a diffusion process in the solution bulk. In the solubility of surfactants in both bulk phases. This
following years, additional mechanisms were is most important for nonionic surfactants because
taken into consideration as the pure diffusion- ionic surfactants are mainly soluble in water and
controlled adsorption was insufficient for almost insoluble in apolar oils. Some examples in
describing experimental data quantitatively. As literature have demonstrated that even such com-
it was shown only later by Lunkenheimer and plex situation can be described quantitatively, as it
Miller (1979), the deviation from a diffusional was shown by Liggieri et al. (1997).
adsorption is in most cases caused by impurities This entry deals mainly with the dynamic
in the surfactant sample. A good survey of the behavior of adsorption layers at liquid interfaces,
evolution of adsorption kinetics models was with emphasis on the adsorption kinetics mea-
given in a book by Fainerman et al. (2001). sured by tensiometry, and on the dilational rheol-
A brief summary of the history with a very ogy as the method to determine the properties of
detailed analysis of the diffusion model and its adsorption layers to reestablish the adsorption
mathematical solution are given below in section equilibrium state of external perturbations.
“Diffusion controlled adsorption.” The details of the experimental methods used
The basis for the understanding of surfactant’s are given in another entry of this book. Here we
adsorption kinetics is the accurate knowledge of present, on the basis of quantitative theories, the
the equilibrium adsorption state. As it was also main mechanisms and phenomena observed at
discussed by Fainerman et al. (2001), for a long interfaces between water and gas (mainly air) or
period of time, the Langmuir and Frumkin between water and an organic solvent. Some
models were the most frequently applied ones, selected representative experimental examples
before theoretical models were derived which are discussed in more detail to demonstrate the
directly considered molecular details in the procedures of data analysis.
adsorption layer. It was shown that changes in In the further text, we will use “surface” when
orientation and aggregate formation of adsorbed we are dealing with the water-air interface, while
surfactant molecules at the surface are relevant the term “interface” is applied to the case of two
for quite a number of homologous series and its immiscible liquids; here it will be mainly the
consideration improves the quantitative under- water-oil interface.
standing of experimental data significantly. S
Although it is known for a long time that the
interaction of adsorbing surfactants with the oil Adsorption Isotherms for Surfactants
phase at a liquid-liquid interface can be very
significant (see Aveyard and Haydon 1965) for The basis for understanding the dynamics of sur-
the adsorption layer structure, due to lack of factants adsorption layers at liquid/fluid inter-
reliable experimental tools and theoretical faces is the exact knowledge of the equilibrium
models, accurate information has not been state. This state is the target for the interface to
described in literature as yet. Only very recently, reach after it was freshly created, and it is the
it was shown by Pradines et al. (2010) for the state to which the layer relaxes after
homologous series of alkyl trimethyl ammonium a perturbation. The equilibrium adsorption is
bromides that the chain length of the surfactant best described by the thermodynamic equation
and the surface coverage are relevant for the of state, that is, by a relationship between surface
incorporation of oil molecules into the adsorption pressure P and the surface layer composition; for
one adsorbing compound, it is the adsorption G. for all components. For a two-component system,
Also the adsorption isotherm, that is, the depen- Lucassen-Reynders (1966) proposed the position
dence of G on the corresponding bulk concentra- of the dividing surface in such a way that the total
tion c, is a representative relationship. For soluble adsorption of the solvent and surfactant is con-
surfactants, neither P(G) nor G(c) is easily acces- stant 1/o:
sible, but rather the surface or interfacial pressure
(tension) isotherm P(c) can be easily measured. G0 þ G1 ¼ 1=o; (3)
There are various theoretical adsorption models
proposed by different authors. Most of these This approach must be modified for surfactant
attempts use the chemical potentials of surface mixtures with respective molar area values of oi
layer components for deriving such an adsorption from each compound, as done by Lucassen-
model, starting with the general approach by Reynders again (1972):
Butler (1932).
X
n
Basics of Surfactant Adsorption Gi ¼ 1=o: (4)
The chemical potentials msi of the components i in i¼0
the surface layer depend on its composition and

surface tension g: These models have been less accurate in par-
ticular for mixtures of very different surfactants
msi ¼ m0s
i ðT; P; gÞ þ RT ln f i xi
s s
(1) or even mixtures with polymers/proteins. There-
fore, additional ideas have been used to improve
Here, m0si ðT; P; gÞ is the standard chemical the accuracy of description of adsorption data by
potential of component i (i ¼ 0 refers to a pure theories. This model approach was extensively
solvent) as a function of temperature T, pressure P, discussed by Fainerman and Miller (2009).
and surface tension g; R is the universal gas
constant. The fi and xi ¼ mi / Smi are the activity Classical Isotherms
coefficients and molar fractions, and mi the The simplest adsorption isotherms for surfactants
numbers of moles of component i, respectively; have actually been transferred from other scien-
the superscript “s” refers to the surface (interface). tific fields. Henry (1803), for example, found out
As shown in detail by Fainerman et al. (1998), that the amount of a gas soluble in a liquid is
Eq. 1 transforms into a relationship with a standard proportional to the pressure. From here, in terms
chemical potential independent on g: of adsorption, we obtain the linear Henry adsorp-
tion isotherm:
msi ¼ m0s
i þ RT ln f i xi goi
s s
(2)
G ¼ Kc: (5)
where oi is the partial molar area of component i.
Butler (1932) was the first who derived the equa- The famous Langmuir adsorption isotherm
tion for the chemical potential of the surface layer was derived by Langmuir (1917) more than 100
(Eq. 2), based on an approximate surface layer years later and, as applied to our case of surfac-
model. With this general equation as starting tant adsorption, has the form:
point and few additional assumptions, respective
adsorption models can be obtained. The most y
bc ¼ (6)
important assumption was first given by Gibbs 1y
(1928), who proposed to introduce a so-called
dividing surface at which the adsorptions are where y ¼ G/G1 is the monolayer coverage.
defined. The location of this dividing surface is Here, we have already two important quantities
essential and defines the adsorbed amounts Gi and of the adsorbed surfactant molecules, the maxi-
the corresponding partial molar surface areas oi mum adsorption G1 ¼ 1/o and the adsorption
constant b as a measure for the surface activity. where yX ¼ GoX is the monolayer coverage by
This equation is equivalent to the empirically surfactant ions and counterions and oX is the
derived Szyszkowski’s equation: specific surface area of the complex. The adsorp-
tion isotherm for this model is:
P ¼ RTG1 lnð1 yÞ (7)
yX
where P is the interfacial pressure, that is, the b½cðc þ cC Þ1=2 x ¼ expð2ayX Þ (11)
1 yX
difference between the interfacial tension of pure
solvent and that of surfactant solution. To
improve the agreement between the experimental where c is the concentration of the ionic surfac-
data and theory, Frumkin (1925) introduced tant, cC is the concentration of counterions, and
a so-called interaction coefficient a which con- the activity coefficient x is determined from the
siders the interaction between adsorbed relation:
molecules:
h i 0:0162 ðc þ cC Þ1=2
P ¼ RTG1 lnð1 yÞ þ ay2 ; (8) log10 x ¼
1 þ 0:0416 ðc þ cC Þ1=2 (12)
þ 0:000055 ðc þ cC Þ
which is coupled with the adsorption isotherm:
y
bc ¼ expð2ayÞ (9) In Eq. 12, the concentrations are expressed in
1y mmol/dm3.
Many other isotherms and equations of state were
derived thereafter, however, all of them with The Reorientation Model
emphasis on special systems and peculiar exper- Starting from Butler’s Eq. 2 and the convention
imental conditions. The next real step ahead in for the dividing surface proposed by Lucassen-
the understanding of surfactant adsorption layers Reynders, Eqs. 3 and 4 for mixed adsorption
at fluid interfaces was proposed then by layers, Fainerman derived a model describing
Fainerman et al. (2001). the adsorption of surfactants in different states
at the surface. This so-called reorientation
Frumkin Model for Ionic Surfactants model (Fainerman et al. 1997) assumes that
To account for the formation of complexes by change in molecular orientation results in a vari-
ionic surfactant molecules in the presence of inor- ation of the partial molar area oi. For surfactant
ganic electrolytes, the Frumkin model was applied molecules coexisting at the surface in two orien-
by Fainerman et al. (1998) to explain the experi- tations (e.g., head-on or side-on) with the molar S
mental surface tension of ionic surfactants. It was areas o1 and o2, respectively, the following equa-
shown by Fainerman and Lucassen-Reynders tion of state:
(2002) that the approach by Lucassen-Reynders to
the location of the dividing surface results in the Po10
¼ lnð1 yÞ þ Gðo o10 Þ þ ay2 ;
generalization of Butler’s equation. Then, for RT
nonideal surface layer and nonideal bulk solution (13)
of one unicharged (1:1) ionic surfactant in the
presence of additional unicharged (1:1) non-sur- and adsorption isotherm
face-active electrolyte, the equation of state
becomes:
G1 o10 o1
bc ¼ exp ð2aGoÞ ;
PoX ð1 GoÞo1 =o10 o10
¼ lnð1 yX Þ þ ay2X (10)
2RT (14)
" " n1 !#

were derived. Here, o is the molar area averaged RT G1
over the partial adsorptions G1 and G2 as: P¼ ln 1 o1 G1 1 þ n
o1 Gc
n1 #
G1 1
o ¼ ðo1 G1 þ o2 G2 Þ=ðG1 þ G2 Þ; (15) þ o1 nG1 1 :
Gc n
(17)
and o10 is the minimal molar area the molecule
can occupy at the interface (in this particular Here, G1 is the adsorption of monomers and
case, o10 ¼ o1; see however section Gn ¼ G1(G1/Gc)n 1 is the adsorption of n–mers,
“The 2D Aggregation Model” where the concept where Gc is the critical aggregation adsorption
of the intrinsic compressibility is explained and average o ¼ (G1o1 + Gnon)/(G1 + Gn). The
which assumes the surface pressure dependence corresponding adsorption isotherm for the mono-
of the molar area). For the second orientation, an mers reads:
equivalent equation can be obtained, assuming
that the adsorption constant b increases with G1 o1
increasing molar area according to the power bc ¼ h i : (18)
1 G1 o1 1 þ nðG1 =Gc Þn1
law with a constant exponent B. The ratio of
partial adsorptions in these two adsorption states
has the form: The model was generalized by Vollhardt and
Fainerman (2010) for large aggregates applicable
to structures in insoluble monolayers.
G1 ðo1 =o2 ÞB
¼
G2 ð1 GoÞðo2 o1 Þ=o10 The Intrinsic Compressibility
As it was shown for many surfactant systems,
ðo2 o1 Þ
exp ð2aGoÞ ; (16) packed surface layers can show an intrinsic com-
o10
pressibility e. Synchrotron grazing incidence
X-ray diffraction studies with insoluble mono-
which represents the general physicochemical layers have shown that the molar area o of
principle of Braun-Le Châtelier, applied by a molecule can depend linearly on the surface
Joos (1999) to adsorption layers. This model pressure P (Gehlert and Vollhardt 2002):
described very well the adsorption of nonionic
surfactants, as it will be shown further below. o ¼ o0 ð1 ePÞ; (19)
The 2-D Aggregation Model where o0 is the molar area of the surfactant at
Frumkin (1925) proposed his adsorption model to P ¼ 0. A possible interpretation of this intrinsic
better consider the lateral interaction between compressibility can, for example, be the increase
adsorbed surfactant molecules. Very strong inter- in the tilt angle of adsorbed molecules with
action might lead even to the formation of inter- increasing surface pressure. Eq. 17, and even
facial (two-dimensional, 2-D) aggregates. When more general dependence:
such aggregates are formed in the surface layer,
the molar area in the aggregates is reduced. The o ¼ o0 ð1 eyPÞ; (20)
coexistence between monomers and n-mers of
molar areas o1 ¼ o0 and on ¼ no1, respectively, can be combined with any adsorption model and
in an adsorption layer was first described by by this leads to a general approach, as it was
Fainerman and Miller (1996), further extended shown by Fainerman et al. (2004) for several
to account for the entropy non-ideality surfactant adsorption layers. The strongest effect
(Fainerman et al. 2001) and leads to the following of this compressibility is observed on the values
equation of state: of the dilational viscoelasticity.
Surfactant Mixtures assumed to exist in local equilibrium with the
The model approach to the thermodynamic interface; therefore, the adsorbed amount(s) and
description of surfactant mixtures is based on the surface tension are related to the concentra-
the adsorption isotherms of the individual com- tion(s) of the adsorbed species in this subsurface
ponents (modified as to account for the presence layer via the model adsorption isotherm(s) and
of other components) coupled by the appropriate the equation of state. On the other hand, the
equation of state of the mixed monolayer. The transport of the surface active species between
model equations can also account for the inter- the solution bulk and this subsurface layer, and
species interactions. The mixture of two surfac- thus the concentration(s) of the species in the
tants which obeys the Frumkin model, the subsurface layer, is governed by the Fick’s diffu-
mixture of a reorientable surfactant with the sur- sion laws. Therefore, any qualitative description
factant which obeys the Frumkin model, and the of the temporal evolution of the interfacial ten-
mixture of two reorientable surfactants were con- sion g(t) and adsorption(s) Gi(t) of each ith
sidered in Aksenenko et al. (2007), Fainerman adsorbed species is based on the solution of
et al. (2010a), and Fainerman et al. (2010b), these differential equations, for which the model
respectively. equations should be treated as boundary condi-
With the increase of the number of compo- tions at the interface.
nents, the calculations become more involved. While the first attempt to relate the diffusion
Therefore, a general approach was developed by process to the time dependence of surface tension
Fainerman and Miller (2001) for the thermody- was made more than a century ago (Milner 1907),
namic description of surfactant mixtures which it was not until the 1950s that sound approximate
also results in a quite complicated set of equa- solutions were obtained. This is because even for
tions. Assuming an ideal surface behavior, it is rather simple (while reasonable from practical
however possible to simplify the equations which point of view) model isotherms, the solution
yield a very easy relationship, which for a binary of the corresponding mathematical problem
mixture reads: requires cumbersome numerical calculations
which became possible only recently, due
¼ exp P
exp P 1 þ exp P
2 1; (21) to rapid progress of computation facilities.
However, several approximate treatments were
The functions P ¼ Po=RT, P 1 ¼ P1 o=RT, proposed, some of which are briefly summarized
and P 2 ¼ P2 o=RT are the dimensionless sur- below in section “Approximate Solutions.”
face pressures of the surfactant mixture and
respective individual surfactant solutions at the
same surfactant concentrations as in the mixed Classical Diffusion Model – Adsorption on
system. A generalization of this equation was a Stationary Surface S
proposed also by Fainerman et al. (2010c) for For diffusion-governed temporal variation of the
the description of adsorption layers formed from concentration of each ith component, the second
multiple surfactant mixtures and applied success- Fick’s law reads:
fully to surface tension data of a ternary surfac-
tant mixture. @ci
¼ Di H2 ci (22)
@t
Diffusion-Controlled Adsorption where Di is the diffusion coefficient of ith com-

ponent in the solution, and t is the time. The
For the understanding of diffusion-controlled approach for the case of a single surfactant was
adsorption at the interface, the concept of the first proposed by Ward and Tordai (1946) who
subsurface layer is essential. This layer, which transformed Fick’s equation into an integral
is immediately adjacent to the interface, is equation. While the original development refers
to a plane interface only, one can account for the Short-time approximation follows immedi-
surface curvature using the Ward-Tordai equa- ately from (23) in which the convolution
tion in the form (Joos 1999): integral is neglected. For the simplest case of
rffiffiffiffiffi single surfactant with G(0) ¼ 0 (and hence
Di cS/S(0) ¼ 0):
Gi ðtÞ ¼ 2
p
2 pffi 3
ðt GðtÞ ¼ 2c0 ðDt=pÞ1=2 (25)
6 pffi pffiffiffi7
4ci t ci;S=S ðt t0 Þd t0 5
where c0 is the bulk concentration. To obtain the
0 time dependence of surface tension, Eq. 25
2 3
ðt should be coupled with appropriate model equa-
Di 4
ci t ci;S=S ðt t0 Þdt0 5 (23) tion(s).
r Further approximations for small surface cov-
0
erage and small and large adsorption times were
Here, ci,S/S(t) is the subsurface concentration given by Hansen (1961), expressed as a function
of surfactant i, and the “” or “+” sign before the of the dimensionless time Y ¼ Dt/(G0/c0)2, where
second term in the right-hand side corresponds to G0 is the equilibrium adsorption:
a diffusion to the interface impenetrable for the
surfactant from the drop bulk and from the solu- rffiffiffiffi pffiffiffiffiffiffiffi !
tion outside of the bubble, respectively. Equa- GðtÞ Y pY
2 1 ... for
tion 23 should be coupled with the model Go p 2 (26)
equations which determine the dependence of pffiffiffiffi
Y 0:2
adsorptions and surface tension on the subsurface
concentrations for the particular adsorbed layer

model. GðtÞ 1 1
1 pffiffiffiffiffiffiffi 1 ... for
Go pY 2Y (27)
Approximate Solutions pffiffiffiffi
Y 5:0
Based on the Ward-Tordai equation, several
approximations were proposed. One example is
the equation derived by Sutherland (1952) for Fainerman et al. (1994) and Miller et al.
small deviations from equilibrium. This can be (1994) published several approximate solutions
used as a good approximation, however, for useful for easy data interpretation.
rather limited range of application (Miller
1990). Sutherland considered a linear adsorption Numerical Calculations
isotherm Eq. 5, valid also in nonequilibrium con- A numerical solution of Ward-Tordai Eq. 23
ditions; however, the development can involve coupled with the model equations yields the
any adsorption isotherm (Joos 1999) provided time dependence of surface tension and other
that the diffusion relaxation time tD ¼ (1/D)∙ related quantities. For the models described in
(dG/dcS/S)2 can be calculated from model section “Adsorption Isotherms for Surfactants”
equations. The final expression for the deviation above, this approach was applied to the adsorp-
of surface tension from its equilibrium tion from single surfactant solutions (Aksenenko
value ge is: et al. 1998; Miller et al. 2004a) and protein-/
surfactant-mixed solutions (Alahverdjieva et al.
g ge ¼ ð g0 ge Þ 2008; Kotsmar et al. 2009) where it is described
pffiffiffiffiffiffiffiffiffi in more details.
expðt=tD Þerfc t=tD (24) Ward-Tordai equation can be reformulated for
the case of spherical geometry of the interface
where g0 is the surface tension of pure solvent. Eq. 23 (see, e.g., Joos 1999) which conforms to
the configuration of the instruments generally The boundary conditions at r ¼ 0 and r ¼ RE
used in experimental studies. However, the follow from the symmetry of the system and the
approach based on the direct numerical solu- fact that the system is closed:
tion of Fick’s equations seems to be the most
appropriate, because it provides for more @ci @ci
¼ ¼0 (30)
straightforward account of the interfacial dr r¼0þ dr r¼R
E
geometry and for the initial distribution of the
surfactant(s) both over the interface and The initial conditions are determined by the
throughout the solution bulk. More specifically, initial distribution of the surfactants over the
for a spherical cell of radius RE which consists region ci(r,0). Also for nonzero initial adsorp-
of two regions (the internal spherical region (I) tions, the initial subsurface concentrations should
of radius RI, the center of which coincides be determined via the model equations.
with the center of the cell, and the external To solve Eqs. 28-30 numerically, a finite dif-
region (E)), Fick’s equations in spherical ference scheme was implemented. As the maxi-
coordinates read: mum gradients of surfactant(s) concentration
exist in the vicinity of the interface, a scheme
2 with non-equidistant spatial nodes concentrating
@ci @ ci 2 @ci
¼ DIi þ ; 0 < r < RI ; toward the interface should be used. To enable
@t @r2 r @r
2 the simulation of the temporal evolution of the
@ci @ ci 2 @ci
¼ DEi þ ; RI < r < RE ; system throughout a large time range, the vari-
@t @r2 r @r able time discretization interval also should be
(28) used, with the increasing interval between the
nodes. Then, the Crank-Nicolson scheme is
Here, DIi and DEi are the diffusion coefficients of implemented at each temporal step, resulting in
the ith surfactant in the regions I and E, respec- a solution of a tridiagonal set of linear
tively, and ci ¼ ci(r,t) are the ith surfactant con- nonhomogeneous equations. The approach
centrations at time t and distance r from the center outlined above is extensively explained in
of the cell. Then, assuming a diffusion-controlled Fainerman et al. (2009a) and Fainerman et al.
adsorption at the interface (r ¼ RI) and allowing (2010a), where the details of numerical calcula-
also for a transfer of the surfactants between the tions are presented.
two regions, the time evolutions of adsorption of
each surfactant caused by diffusive fluxes are: Diffusion-Governed Adsorption on
a Continuously Deformed Interface
dGi
I @ci

E @ci A continuous surface deformation results in
¼ Di þ Di (29)
dt dr r¼R dr r¼Rþ a convection current which should be accounted S
I I
for in the diffusion equation. In this case, instead
which are the boundary conditions at the inter- of Eq. 22 one has:
face. Note that, as the dependencies of Gi on
subsurface concentrations of the ith surfactant @ci
_ ~
ar Hci ¼ Di H2 ci (31)
cIi;S=S ci R
I ;t or cEi;S=S ci ðRþ
I ; tÞ (which @t
are interrelated via the distribution coefficients
of the surfactants across the interface, where a_ ¼ d ln A=dt is the rate of surface defor-
. mation (dilation) and A is the surface area. The
i ¼ ci;S=S ci;S=S ) involve the equation of
KI;E I E
mass balance equation is the boundary condition
state and adsorption isotherms of the surfactants at the interface:
for any specific model explained in section
“Adsorption Isotherms for Surfactants” dGi
_ i ¼ Di ½Hn ci 0
þ aG (32)
above, Eq. 29 couples the set of Eq. 28. dt
where the derivative in the right-hand side of constant rate of surface area expansion, constant
Eq. 32 is taken at the interface along the normal rate of surface dilation, and expanding drop at
to the interface. a constant volume flow rate. This latter case is
In Eq. 31, the “” or “+” sign corresponds to important with regard to the drop volume tech-
the expanded or compressed surface, respec- nique and is described here in more details.
tively. It can be shown, see Joos (1999), that for For a radial flow of a non-aggregating, nonionic
both these cases the diffusion equation can be surfactant in the internal (I) and external (E) region,
transformed to the form similar to the Ward- that is, inside and outside of the spherical drop,
Tordai equation; however, the resulting func- the convective-diffusion Eq. 31 in spherical coor-
tional dependencies are somewhat different; dinates was obtained by MacLeod and Radke
therefore, these cases are considered separately. (1994):
In what follows, the systems are considered for !
which the symmetry provides for the reduction to @cðkÞ @cðkÞ 2 ðkÞ
ðkÞ @ c 2 @cðkÞ
the one-dimensional formulation, namely, the þ vr ¼D þ ;
@t @r @r2 r @r (34)
plane or spherical interface.
k ¼ I; E
Diffusion-Governed Adsorption at an Expanding
where r is the radial distance from the drop cen-
Surface
ter, c(k) is the concentration of the surfactant in
Assuming that at t ¼ 0, that is, before the expan-
the corresponding region, and vr ¼ (RI/r)2(dRI/dt)
sion started, the surface exists in the equilibrium
is the fluid radial velocity. The initial and bound-
with the bulk, and the equation similar to the
ary conditions are similar to those applicable to
Ward-Tordai equation for the adsorption of
the case of stationary interface, except that the
a single surfactant onto a planar interface reads:
boundary condition at the drop interface should
account also for the depletion/accumulation of
Gf ðtÞ ¼ Ge
" pffiffi # surfactant due to the stretching/contracting of
pffiffiffiffiffiffiffiffiffi pffiffiffi ð t 0
pffiffiffiffi the interface:
þ 2 D =p c t cS=S ðt t Þd t 0
0
dG @cðIÞ ðEÞ
ðEÞ @c
(33) ¼ DðIÞ þ D Ga_ (35)
dt @r R @r Rþ
Ðt I I
where the time-scaling variable t ¼ 0 ½ f ðtÞ2 dt;

here, f(t) ¼ A(t)/A0 and A0 is the surface area Assuming that the thickness of the diffusion
before the expansion. Several regimes were boundary layer is small as compared with the
extensively considered previously (Joos 1999): drop radius, the Eqs. 34 and 35 are reduced to:
8sffiffiffiffiffiffiffiffi 9
>
< ð ðkÞ >
=
dG X D ð k Þ t dcS=S dt 0
¼ ½RI ðtÞ2 q ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ga_ ð36Þ
dt >
k¼I; E:
p 0 dt Ðt 00 4 00 > ;
0 0 ½RI ðt Þ dt
t
t
which should be solved simultaneously with the radius and dilation rate on time. The numerical
appropriate model dependence of G (and hence solution of this problem by MacLeod and Radke
P) on the subsurface concentration in either of the (1994) involved a suitable discretization of Eq. 31
regions, taking into account the distribution coeffi- and the Frumkin adsorption model both for
cient KI,E and model dependencies of the interface diffusion and diffusion-kinetic-governed transport.
Diffusion-Governed Adsorption at a Contracting – The two-dimensional surface layer intrinsic
Surface compressibility coefficient e which has
For the continuously compressed surface, Eq. 31 a stronger impact at high surface coverage
with boundary condition (32) also yields the ana- Consideration of these parameters gives an
logue of the Ward-Tordai equation: idea of the state and interactions of molecules in
the adsorption layer. The values of the parameters
can be found from fitting the particular adsorption
GgðtÞ 2G½gðtÞ 1 ¼ Ge model to experimental data. As an illustrative
" pffiffi #
pffiffiffiffiffiffiffiffiffi pffiffiffi ð k pffiffiffiffi example, the experimental data on the equilib-
þ 2 D =p c k cS=S ðk k0 Þd k0 ; rium adsorption properties of the series of cationic
0
surfactants alkyl trimethyl ammonium bromides
(37)
CnTAB (n ¼ 10, 12, 14, 16) at the water/air and
the series of anionic surfactants sodium
Ðt
Here, k ¼ 0 ½gðtÞ2 dt, g(t)¼ A0/A, and alkyl sulfates NaCnSO4 (n ¼ 10, 12, 14, 16)

Ðwhere
k
the average value G ¼ 0 Gd½gðtÞ =½gðtÞ 1. at the water/hexane interface are discussed in
Similarly to the case of adsorption to an detail here.
expanding surface, various regimes of surface
compression can be distinguished: compression Water/Air Interface
with a constant speed dA/dt, compression with Systematic surface tension measurements have
a constant (negative) dilation rate a_ ¼ d ln A=dt, been performed for a series of standard, widely
and compression at a constant surface pressure. used surfactants CnTAB (n ¼ 10, 12, 14, 16)
These regimes were comprehensively explained dissolved either in pure water (Bergeron 1997;
by Joos (1999). Stubenrauch et al. 2005) or in phosphate buffer
(10 mM, pH 7) (Pradines et al. 2010).
The phosphate buffer contains two electro-
Adsorption Isotherms of Standard lytes dissociating into Na+ and H2PO4 ions in
Surfactants the ratio 1:1 and Na+ and HPO42 in the ratio 2:1.
Equations 10–12 are valid only for the case of 1:1
The understanding of dynamic interfacial prop- electrolytes and, therefore, cannot be directly
erties is based mainly on the equilibrium thermo- applied to buffer solutions. It was proposed in
dynamics of the surfactants’ adsorption layers. (Pradines et al. 2010) to calculate the effective
The equilibrium state of an adsorption layer is concentration of the 1:1 electrolyte providing the
fully described by the adsorption isotherm and best coincidence of the data for all four surfac-
equation of state, as it is shown above in section tants with the corresponding isotherms for pure
“Adsorption Isotherms for Surfactants,” where water by using the mean ionic activities, cf. S
the most often used adsorption models are Eq. 11:
discussed in detail. The characteristic parameters
of each adsorption model are: 1=2
– The adsorption equilibrium constant b which c
¼ cðc þ cC Þ x: (38)
essentially defines the location of the isotherm
along the concentration axis
– The partial molar area o which defines the An electrolyte concentration of cC ¼ 2 mM
final slope of the isotherm was found to describe the data in an optimum
– The intermolecular interaction constant way.
a which has its strongest effect on the isotherm Figure 1 represents the equilibrium surface
shape intermediate surface coverage tension plotted versus the mean ionic activities
Surfactant Adsorption 75
Kinetics,
Fig. 1 Adsorption 70
isotherms at pure water/air
interface (open symbols) 65
and buffer/air interface
(filled symbols): (□, ▪) 60
C10TAB, (D, ~) C12TAB,
γ [mN/m]
(e, ♦) C14TAB, and 55
(○, ●) C16TAB. All data
50
are plotted as a function of
the mean ionic activity c*
45
(According to Pradines
et al. 2010) 40
35
30
1.00E–05 1.00E–04 1.00E–03 1.00E–02 1.00E–01 1.00E+00
c* [mol/l]
Surfactant Adsorption Kinetics, Table 1 Model parameters of the adsorption isotherm obtained with the FI model for
the buffer/air interface (FI, according to Pradines et al. 2010) and the pure water/air interface (FIH2O according to
Stubenrauch et al. 2005)
o0 [m2/mol] a b [l/mol]
CnTAB FIH2O FI FIH2O FI FIH2O FI
n ¼ 10 3.4∙105 0.7 9.6∙101
n ¼ 12 2.58∙105 3.2∙105 0.96 0.2 1.74∙102 5.3∙102
n ¼ 14 2.76∙105 3.6∙105 0.99 0.9 7.16∙102 1.5∙103
n ¼ 16 2.92∙105 2.8∙105 1 1.6 3.11∙103 1.8∙103
for solutions prepared in pure water as well for is shown for FI model taking into account the
those in buffer. The solid lines correspond to the intrinsic compressibility (e ¼ 0.01 m/mN).
theoretical curves obtained with the Frumkin With increased bulk concentration, the molar
model for ionic surfactants (FI), Eqs. 10–12. area gradually decreases because of the surface
The model parameters are given in Table 1. layer compressibility, while the surfactant
Small difference between the parameters for the adsorption increases. The surfactant adsorption
Frumkin model for ionic surfactants applied to curve of C16TAB shows a steep slope that might
the buffer/air (FI) and pure water/air interface indicate a phase transition from a liquid-
(FIH2O) can be caused by the fact that an effective expanded to a liquid-condensed state (Ferri and
salt concentration cc in the model calculations Stebe 1999).
was used.
After fitting the adsorption isotherms with the Water/Oil Interface
corresponding model, a clearer picture of the In this Section, the adsorption behavior of sodium
molecules at the interface can be obtained alkyl sulfates at the interface between water and
by calculating the surfactant adsorption G and hexane is considered. This topic attracted much
partial molar area o. In Fig. 2, the change of interest recently, mainly because the impurity
G and o with increasing bulk concentration effect plays an important role at the water/air
4.00E+05 7.00E−06
3.50E+05 6.00E−06
3.00E+05
5.00E−06
2.50E+05
4.00E−06
ω [m2/mol]
Γ [mol/m2]
2.00E+05
3.00E−06
1.50E+05
2.00E−06
1.00E+05
5.00E+04 1.00E−06
0.00E+00 0.00E+00
1.00E−07 1.00E−06 1.00E−05 1.00E−04 1.00E−03 1.00E−02 1.00E−01 1.00E+00
c [mol/l]
Surfactant Adsorption Kinetics, Fig. 2 Variation of the curves correspond to the FI model taking into account the
adsorbed amount G (open symbols, right scale) and molar intrinsic compressibility, according to Pradines et al.
area o (filled symbols, left scale) of the C10TAB (○, ●), (2010)
C12TAB (D, ~), C14TAB (e, ♦), and C16TAB (□, ▪); all
interface, while on the contrary, at the water/oil deviates from that at the water/air interface hav-
interface, the effect of impurities (homologous ing a value of about 10 (Fainerman et al. 2001).
alcohols) is negligible due to their high solubility The theoretical parameters obtained after
in the oil phase. An additional effect on NaCnSO4 applying the FI and R models at the water/hexane
adsorption layers is manifested by incorporation interface are shown in Table 2. It is seen that both
of hexane molecules into the layer structure. This models show similar values for the comparable
effect has an essential impact at low interfacial parameters.
coverage and especially for shorter alkyl chain For the FI model, the intrinsic compressibility
surfactants. The adsorption characteristics are coefficient was fixed at e ¼ 0.008 m/mN for all S
discussed in the framework of the two well- surfactants. On the other hand, for the
known thermodynamic models: the Frumkin reorientation model, the used intrinsic compress-
model for ionic surfactants (FI) and the ibility coefficient was e ¼ 0.003 m/mN. In this
reorientation model (R), both explained in sec- case, the e value was taken smaller because the
tion “Adsorption Isotherms for Surfactants.” change in the average molar area is already taken
For the interpretation of the thermodynamic into account by the two states of a molecule at the
data presented in Fig. 3, the calculated isotherms interface. At low surface coverage, the surfactant
for both adsorption models are shown. The iso- can use the maximum free area at the interface
therms are shifted toward lower concentrations leading to a tilt of the alkyl chain (Fainerman
with increasing alkyl chain length. With et al. 2002). Due to the possibility of changing
each additional CH2–CH2 group, the shift the orientation, the surfactant’s molar area
becomes smaller and less than 10. Hence, the changes as a function of the surface pressure,
Traube factor for the water/hexane interface that is, of the surface coverage.
Kinetics, Fig. 3 Interfacial
tension isotherms of
NaCnSO4 plotted versus 50
bulk concentration; (♦)
n ¼ 10, (~) n ¼ 12, (▪) 40
n ¼ 14, and (●) n ¼ 16.
γ [mN/m]
Solid line corresponds to
the Frumkin model for 30
ionic surfactants and
dashed line to the 20
reorientation model
(According to Sharipova
et al. 2011) 10
0
1.00E–06 1.00E–05 1.00E–04 1.00E–03 1.00E–02 1.00E–01
c [mol/l]
Surfactant Adsorption Kinetics, Table 2 Parameters of adsorption layers obtained with the reorientation and Frumkin
models for ionic surfactant models at the water/hexane interface (According to Sharipova et al. 2011)
o0 [m2/mol] o10 [m2/mol] o2 [m2/mol] a b [l/mol]
NaCnSO4 FI R R FI FI R
n ¼ 10 6.5∙105 4∙105 1.5∙106 0 6.8∙103 8.8∙103
n ¼ 12 6∙105 4∙105 2∙106 0 3.1∙104 5.6∙104
n ¼ 14 6∙105 3.9∙105 3∙106 0.5 9∙104 1.6∙105
n ¼ 16 5.6∙105 3.6∙105 3∙106 0.6 2.2∙105 3.4∙105
The changes of the partial molar area o and screens the charges of the adsorbed ions. The
surfactant adsorption G with increasing bulk con- larger molar area can better be explained by the
centration calculated from FI and R models are presence of intercalated solvent molecules that
plotted in Figs. 4 and 5, respectively. Both graphs can be seen as a kind of competitive adsorption
show the decrease of partial moral area, while between the surfactants and the hexane mole-
the surfactant adsorption increases with the cules. Due to the higher flexibility of the shorter
increase of the surfactant concentration. Compar- chains, the resulting mean area is consequently
ing Figs. 4 and 5, one can see that for the larger than for the better-organized long-chain
reorientation model, the changes of the mean homologues, for which due to the interchain
molar area become much less at a certain con- interactions, the short hexane molecules are
centration indicating that all molecules are squeezed out of the interfacial layer. The struc-
already in the state of smaller area. ture of surface films at the water/air interface
Among all samples, NaC10SO4 has the largest formed in systems containing alkanes and cat-
molar area o0, while adsorbed amount G and ionic surfactants was studied by Lu et al. (1992,
adsorption equilibrium constant b are the lowest. 1995, 1998), and it was found that the partial
The physical picture for this finding cannot be penetration of oil molecules in between surfac-
electrostatic repulsion between the charges tant chains occurs.
because the bulk concentration of the surfactant, The affinity of the surfactants to the oil phase
and hence the ionic strength, is rather high and increases with increasing alkyl chain length.
Surfactant Adsorption 7.00E+05 3.00E−06
Kinetics, Fig. 4 Variation
of the total adsorbed 6.00E+05
amount G (dashed lines, 2.50E−06
right scale) and molar area
o (solid lines, left scale) of 5.00E+05
2.00E−06
the NaC10SO4 (●),
NaC12SO4 (~), NaC14SO4
ω [m2/mol]
Γ [mol/m2]
4.00E+05
(♦), and NaC16SO4 (▪); all 1.50E−06
curves are obtained from
3.00E+05
the FI model (According to
Sharipova et al. 2011) 1.00E−06
2.00E+05
5.00E−07
1.00E+05
0.00E+00 0.00E+00
1.00E−06 1.00E−05 1.00E−04 1.00E−03 1.00E−02 1.00E−01
c [mol/l]
1.20E+06 4.00E−06
3.50E−06
1.00E+06
3.00E−06
Kinetics, Fig. 5 Variation 8.00E+05
2.50E−06
of the total adsorbed
ω [m2/mol]
Γ [mol/m2]
amount G (dashed lines,
6.00E+05 2.00E−06
right scale) and the
averaged (over the two
1.50E−06
states) molar area o (solid 4.00E+05
lines, left scale) of the
1.00E−06
NaC10SO4 (●), NaC12SO4
(~), NaC14SO4 (♦), and 2.00E+05
5.00E−07
NaC16SO4 (▪); all curves
are obtained from the
0.00E+00 0.00E+00
reorientation model
1.00E−06 1.00E−05 1.00E−04 1.00E−03 1.00E−02 1.00E−01
(According to Sharipova
et al. 2011) c [mol/l]
S
Therefore, NaC12SO4 and NaC14SO4 have larger over the electrostatic repulsion forces. This leads
adsorption constants b than NaC10SO4. Electro- also to an increased interaction (attraction)
static repulsion forces are still essential for parameter a > 0.
NaC12SO4 and NaC14SO4, but the decreasing
molar area in comparison to NaC10SO4 means
that surfactant-surfactant interactions between Dynamic Surface/Interfacial Tensions of
the adsorbed molecules dominate. NaC16SO4 Surfactant Solutions
shows the lowest molar area o0 and the largest
adsorbed amount G and adsorption constant b. The easiest accessible quantity for characterizing
Here, the alkyl chain length becomes long the interfacial dynamics is the surface/interfacial
enough to let the hydrophobic forces dominate tension which is directly linked to the amount of
adsorbed molecules via the respective adsorption solution of the nonionic surfactant dodecyl
isotherm and equation of state, as mentioned dimethyl phosphine oxide (C12DMPO) at
above. For regular surfactants and assuming a medium concentration (2∙105 mol/l) is a very
a diffusion-controlled adsorption process (when suitable sample. Figure 6 shows such experimen-
the kinetic barrier is not an issue), the dynamic tal data measured by capillary pressure tensiom-
surface tension is directly linked to the bulk con- eter ODBA using the stop growing drop protocol
centration within the sublayer via the adsorption (SGD, see Javadi et al. 2010a). The theoretical
isotherm. results are calculated by the numerical solution of
Accurate determination of dynamic tension the diffusion equation with relevant boundary
requires a lot of efforts in order to have reliable conditions (28)–(30) for a spherical interface
experimental techniques at hand for their mea- between two finite liquid phases (Liggieri et al.
surement. In the chapter on ▶ Interfacial Dynam- 1997; Aksenenko et al. 1998; Fainerman et al.
ics Methods, different instrumental approaches 2009a). For the relation between subsurface con-
together with some examples of measured data centration and adsorption, the Langmuir model
obtained under slow and fast dynamic conditions was applied with the following parameters for
are presented. In this section, selected experimen- the C12DMPO, as obtained in Mucic et al.
tal data are presented and compared with ade- (2011): b ¼ 176 m3/mol and G1 ¼ 4.3∙106
quate theoretical models. mol/m2. The experimental conditions were bulk
concentration – c ¼ 2.5∙105 mol/l, drop radius
Surfactant Solutions Against Air RI ¼ 0.65 mm. As optimum diffusion coefficient,
Surfactant solutions against air are the most a value of D ¼ 2∙1010 m2/s was estimated.
common investigated system mainly because The theoretical fitting is shown in Fig. 6 by thick
of the numerous available tensiometry tech- red curve passing through the experimental points
niques as discussed in the chapter on ▶ Interfa- in an optimum way. Additional curves
cial Dynamics Methods. However, in literature, are presented in this figure to demonstrate the
there are not many publications which present influence of main parameters on the quality
experimental data together with a quantitative of the theoretical prediction. For example, the
analysis based on theoretical models for the variation in bulk concentration leads to a strong
prediction of respective adsorption mechanism change in the shape of the calculated dependency.
and other peculiarities of dynamic interfacial Hence, any error in the concentration of the
layers. For this purpose, surfactants of known droplet solution can result in significant errors.
chemistry and very high purity are required. The effect of changes in the drop radius is
The adsorption isotherm parameters should significant only for larger times and when the
be available or obtained using relevant experi- size is really very small. For small droplets, the
mental techniques. Then, depending on the sur- amount of adsorbed surfactant is comparable with
face activity and concentration, a feasible the amount of surfactant present in the bulk.
experimental method should be considered for Thus, the adsorption in this case results in the
performing the dynamic surface tension mea- considerable decrease of the bulk concentration
surements (Mucic et al. 2011). During a and therefore the equilibrium interfacial tension
dynamic adsorption process, large changes in (depletion effect). For sufficiently large droplets,
surface tension can happen within a very short the effects become smaller throughout the whole
period of time. adsorption process because the volume to surface
area ratio increases and depletion effect can be
Principle of Dynamic Surface Tension Analysis neglected. The diffusion coefficient has a signif-
Based on a Simple Adsorption Isotherm icant impact on dynamic curve over the whole
For observing a real experimental run with easily adsorption time. The adsorption activity parame-
visible surface tension changes during ter (b) shows effects similar to the influence of the
a reasonable time of few minutes, aqueous bulk concentration. The maximum adsorption
1/2Cexp
1/10D
1/2 Rexp
Cexp 1/2D
Concentration effects Rexp
2D D
2Cexp Diff. coeff.
Drop radius effects effects
10D
2Rexp
71 1/2b
69
Surface activity parameter b
67 Theory :
Cexp=2.5e−5 mol/l
65
γ [mN/m]
Rexp=0.65 mm
D=2e−10 m 2/s 2b
63
b=176 m3/mol
Γinf=4.310e−6 mol/m 2
61 Theory
59 1/2Γinf
Exp.
57
Maximum adsorption
55 Γinf
0.001 0.01 0.1 1 10 100 1000
t [s]
2Γinf
Surfactant Adsorption Kinetics, Fig. 6 Dynamic sur- model parameters on the theoretical curve in comparison
face tension of a 2.5∙105 mol/l C12DMPO solution drop with the experimental data. The experimental conditions
in air: comparison of experimental data and theoretical are given by c ¼ 2.5∙105 mol/l, drop radius RI ¼ 0.65
results calculated by the diffusion-controlled adsorption mm, b ¼ 176 m3/mol, and G1 ¼ 4.3∙106 mol/m2,
given by Eqs. 28–30 and considering a Langmuir adsorp- according to Mucic et al. (2011)
tion model. The extra graphs show the effects of various
parameter G1 shows also remarkable effects in of the reorientation model to describe the adsorp-
particular at long adsorption times, as expected. tion kinetics of a 6.25∙106 mol/l tetradecyl
dimethyl phosphine oxide (C14DMPO) solution
Analysis of Dynamic Surface Tensions for Specific (Aksenenko et al. 1998). This model is also
Adsorption Models superior for the nonionic surfactants n-alkyl
For the prediction of adsorption mechanisms polyoxyethylene glycol ethers (CiEj). Several
from dynamic surface tension, as in general studies have shown that for these surfactants,
explained above, it is also very essential to the Langmuir model fails to describe the adsorp- S
apply the correct adsorption model. Clearly, tion behavior in particular at very low concentra-
as the first step, this requires a quantitative tions (Liggieri et al. 1999). Figure 8 illustrates the
examination of available models with respect to improvement of the theoretical fitting of dynamic
experimental equilibrium data. An example is surface tensions for some aqueous solutions of
presented in following section for SDS adsorbed C10EO8 at the water/air interface by the using of
at the aqueous solution – hexane interface (Javadi reorientation isotherm instead of the Langmuir
et al. 2010b) which compares the respective isotherm (Miller et al. 1999).
results for the Langmuir and Frumkin models. One could expect that also the change in
The model most frequently discussed in recent molecular orientation during the adsorption
studies is the two-state reorientation model kinetics can become an additional limiting step
presented in section “The Reorientation Model.” in the whole adsorption process. However, for
Figure 7 shows an example of the improvement example, the dynamic data for aqueous solutions
of the theoretical prediction by the employment of C10E4, C10E5 and C10E8 have shown that the
Kinetics, Fig. 7 Dynamic
surface tension of
a 6.25∙106 mol/l
C14DMPO solution against
air measure by maximum
bubble pressure (e), drop
volume (▪), and ring (D)
methods, as compared with
the diffusion model
neglecting reorientation
effects (curve 1) and
consideration of
reorientations (curve 2)
(Aksenenko et al. 1998)
73
70
67
γ [mN/m]
Kinetics, Fig. 8 Dynamic
surface tension for aqueous
64
C10EO8 solutions at
different concentrations:
4 106 (D), 6 106 (□),
and 4 105 (○) mol/l as 61
compared with theoretical
calculation: Langmuir
model (dotted line) and
two-state reorientation 58
0 5 10 15 20
model (solid lines) (Miller
et al. 1999) t1/2 [s1/2]
diffusion mechanism alone gives already the best Dynamic Surface Tension of Surfactant Mixtures
description of the experimental data (see Fig. 9), In most technological applications, mixtures of
and taking into account the dynamics of the ori- different surface-active components are used.
entation process itself does not improve the Therefore, studies of the behavior of surfactant
agreement (Liggieri et al. 1999). This underlines mixtures are important, especially under dynamic
that the reorientation process is much faster than conditions. The adsorption equilibrium models
the diffusion process. For particular conditions, for surfactant mixtures were discussed above.
however, the reorientation kinetics itself, in addi- For the dynamic conditions, a diffusion-
tion to the transport by diffusion, can become an controlled mechanism together with the relevant
important issue. adsorption model can be used for data analysis.
30 C12EO5 and C14EO8 as measured by bubble pro-
file tensiometry. The theoretical predictions by
25 a diffusion-controlled adsorption model using
an isotherm for reorientable surfactant mixtures
20 reflect the experimental data very well. For the
calculation of the dynamic surface tensions, only
Π [mN/m]
15
the isotherm parameters obtained from the anal-
ysis of equilibrium surface tensions of the indi-
10
vidual solutions C12EO5 and C14EO8 and of their
mixture were used, and no additional free fitting
5
parameter was required.
0 Figure 12 shows recent results for mixed solu-
tions of SDS and C14EO8 (Fainerman et al.
2010a), which follow different adsorption
10–1 100 101 102 103 104
t [s]
models: the nonionic C14EO8 is best described
by the two-state reorientation model, while the
Surfactant Adsorption Kinetics, Fig. 9 Dynamic sur- anionic SDS follows a Frumkin-type model. The
face pressure of C10E8 solutions for different concentra- experimental results for the equilibrium and
tions: 2∙105 (●), 4∙105 (▪), and 8∙105 (~) mol/l, as
compared with theoretical curves for the diffusion-
dynamic surface tensions of all studied mixed
controlled adsorption model (D ¼ 3.7∙1010 m2/s) assum- SDS/CnEOm solutions are described satisfacto-
ing the two-state reorientation isotherm without (solid rily by the theoretical model.
line) and with reorientation kinetics for rate constant
k12 ¼ 0.1 s1 (dashed line) (Liggieri et al. 1999)
Aqueous Surfactant Solutions Against Oil
In comparison with the surface of aqueous solu-
tions to air, one of the main extra challenges for
Miller et al. (2004b) presented a simple theoret- water/oil interfaces is the question about the con-
ical approximation for studies of some mixed tribution of the oil phase on the structure of sur-
solutions of two alkyl dimethyl phosphine oxides, factant adsorption layer at the interface. In
C10DMPO and C14DMPO. In Fig. 10, the addition to this, while evaporation effects can be
dynamic surface tension of different mixed important for gas/liquid interfaces, for liquid/liq-
solutions is shown: at a fixed concentration of uid systems, the solubility of the solute compo-
0.3 mmol/l C10DMPO, different amounts of nents in both phases might be an important issue.
C14DMPO were added: 0.001 (□), 0.003 (e), Any data analysis is easier when the solubility is
and 0.01 mmol/1 (D). The solid and dotted to be considered only in one phase, and transfer of
lines refer to theoretical calculations using surfactant across the interface is negligible. In S
a generalized Ward and Tordai diffusion model this case, the set of diffusion equations,
and the Frumkin adsorption model. In addition, Eqs. 28–30, has to be solved just in one phase
a correction factor proportional to P10/P20 was together with relevant boundary condition at the
introduced, where Pi0 is the equilibrium surface interface.
pressure of component i.
An exact description of the adsorption kinetics System with the Surfactant Soluble in One Liquid
for the surfactant mixtures is possible when the Phase Only
coadsorbing species follow the same adsorption An example of such a system is presented in
model. Such a study is presented for mixtures of Fig. 13 which shows the experimental results
oxyethylated alcohols C12EO5, C14EO8, and Tri- obtained with a water drop in a solution of
tons X-n, considering the reorientation model Span80 in hexane by using the capillary pressure
(Aksenenko et al. 2010). Figure 11 shows the tensiometer ODBA and the stop growing
dynamic surface tensions for mixtures of drop (SGD) protocol (Javadi et al. 2010a).
Kinetics,
Fig. 10 Dynamic surface
tension of aqueous
solutions of 0.3 mmol/l
C10DMPO mixed with
C14DMPO at
concentrations of: 0.001
(□), 0.003 (e), and 0.01
mmol/1 (D). Solid lines:
theoretical modeling using
modified Ward and Tordai
equation and Frumkin
adsorption model for the
mixture system, dotted
lines: corrected by a factor
proportional to P10/P20,
where Pi0 is the
equilibrium surface
pressure of component
i (Miller et al. 2004b)
70
0.05
0.2
Kinetics,
γ [mN/m]
Fig. 11 Dynamic 1
surface tensions of mixed 2
C12EO5/C14EO8 solutions 50
at a mixing ratio of 3:1 at 5
different total
concentrations (numbers 40 10
correspond to mmol/l);
black solid curves, 20
experimental data; red
dashed curves, simulated 30
values (Aksenenko et al. 0 3000 6000 9000 12000 15000
2010) t [s]
The solubility of Span80 in water is negligible as with G1 ¼ 4.22∙106 mol/m2 and b ¼ 0.94∙103
compared to the oil phase: The ratio of concen- m3/mol for Span80 at the water/hexane interface
trations in water versus oil is 5.5∙104 (cf. was considered for describing the equilibrium
Peltonen and Yliruusi 2000). For a fresh droplet adsorption (Javadi et al. 2010a). The theoretical
at the capillary tip after the liquid flow was values in the figure were obtained for a diffusion
stopped, in absence of hydrodynamic effects, coefficient of 6∙1010 m2/s for Span80 in hexane.
these data can be described by the set of Another example is presented in Fig. 14 for
Eqs. 28–30. In this study, the radius of the droplet different SDS solutions of 105, 104, and 5∙104
RI was about 1 mm, and RE 50∙RI. The numer- mol/l in aqueous buffer (pH 7) against hexane.
ical solution (see section “Numerical Calcula- The data are also measured by using the capillary
tions” for details) shows good agreement with pressure tensiometer ODBA and the SGD proto-
the experimental data. A Langmuir isotherm col (Javadi et al. 2010b). The theoretical analysis
Kinetics,
tension for individual
solutions of C14EO8
and SDS and of three
SDS/C14EO8 mixed
solutions with molar ratios
of 5:1, 30:1 and 200:1, all at
a total concentration of 50
mmol/L; symbols are
experimental data and lines
refer to the theoretical
calculations (Fainerman
et al. 2010a)
55 experimental data. Indeed, it is one of the

most important purposes of any dynamic surface/
50
interfacial tension experiments – to provide the
45 information about the kinetics of adsorption and
transport coefficients of the studied surfactant
γ [mN/m]
40 systems.
35 System with the Surfactant Soluble in Both Liquid

Phases
30
The data analysis for surfactants soluble in both
phases is generally more complex as the transfer
25
of surfactant molecules across the interface must
10–2 10–1 100 101 102
t [s]
be considered and a set of diffusion equations for
both phases, for example, Eqs. 28–30, has to be
Surfactant Adsorption Kinetics, Fig. 13 Comparison solved. As example of such a system in Fig. 15,
of experimental data g(t) (filled symbols) measured for the dynamic interfacial tension is shown for
105 mol/l Span80 in hexane at the interface to water
and the numerical solution (open symbols) of the diffu-
C13DMPO adsorbed from an aqueous solution
sion-controlled adsorption kinetics (of different initial concentrations 1, 2, 3, and S
5∙105 mol/l) at a fresh water/hexane interface
of the data was performed with the Frumkin and (Ravera et al. 2000). According to the experimen-
Langmuir adsorption models using the following tal protocol, a drop of the aqueous surfactant
parameters (obtained from the equilibrium adsorp- solution is formed in hexane, with the drop
tion data): volume by a factor of 1,000 smaller than that of
For the Frumkin model: o0 ¼ 2.15∙105 m2/mol, the hexane. The solid curves are the theoretical
a ¼ 3, b ¼ 2.03∙102 m3/mol. results obtained from a numerical solution of the
For the Langmuir model: o0 ¼ 4∙105 m2/mol, diffusion equations for a spherical geometry and
b ¼ 1.91∙102 m3/mol. considering a Langmuir model for the nonionic
Both models are capable to predict the C13DMPO. The excellent agreement between
experimental data well; however, the Frumkin theory and experiment was achieved by selecting
model is slightly better. A reasonable value of an optimum distribution coefficient for the
D ¼ 5∙1010 m2/s is obtained via fitting the surfactant between the two bulk phases.
Kinetics,
Fig. 14 Measured
dynamic surface tension of
SDS solutions at different
concentrations: 105 (●),
104 (▪), and 5∙104 (~)
mol/l in aqueous buffer of
pH 7 against hexane;
theoretical lines are
calculated for the Frumkin
(solid line) and Langmuir
(dotted line) adsorption
models (Javadi et al.
2010b)
the adsorption equilibrium is quickly attained

during a short adsorption time. However,
many of the applications, such as in emulsifica-
tion, dispersion, spraying, and extraction pro-
cesses, also involve rather fast processes like
creation of interface, deformations, break
off, and coalescence within short periods of
time, on the scale of milliseconds and even
shorter. Therefore, studies of fast dynamic
adsorption processes are an important request
as was discussed in the chapter on ▶ Interfacial
Dynamics Methods. Regarding the understand-
ing of such very fast dynamic adsorption
processes, reliable experimental data have to
be obtained, with emphasis on a minimum
influence by hydrodynamic effects. These
Surfactant Adsorption Kinetics, Fig. 15 Dynamic
hydrodynamic effects, which have to be
interfacial tension during adsorption with transfer of quantitatively separated from the interfacial
C13DMPO from an aqueous solution drop for different phenomena, refer to various sources, such
initial concentrations: 1, 2, 3, and 5 105 mol/l as liquid or gas flow inside the cell, the
(for curves a, b, c, and d) at a fresh water/hexane interface;
the solid curves are the theoretical calculations, according
capillary or the drop, and inertia of the
to Ravera et al. (2000) liquids.
Fast Dynamic Adsorption for Gas/Liquid Interfaces

Short-Time Dynamic Surface/Interfacial The problems of gas-liquid interfaces are less
Tension serious, and for the maximum bubble pressure
Often surfactants are used at rather high bulk tensiometry, the fastest dynamic tensiometry
concentrations. This leads to the fact that method, the most critical points are solved
Kinetics,
tensions of concentrated
TX45 solutions 50 mm/l ~
and 100 mmol/l ♦ measured
by the emerging bubble
(open symbols) and
maximum bubble pressure
(filled symbols) methods;
the solid curves are the
theoretical calculation for
a diffusion-controlled
mechanism and the
reorientation model
(Fainerman et al. 2009a)
(Kovalchuk and Dukhin 2001; Dukhin et al. also data from continuously growing drops
1998; Mishchuk et al. 2001). With this efficient (CGD) and the utilization of the maximum pres-
methodology, several experimental and theoreti- sure data, reached at the minimum drop radius,
cal studies have been published during the last can be used for studies of systems showing a fast
decade. In Fig. 16, the results of a recent study on adsorption kinetics. This experimental approach
the dynamic surface tensions of highly concen- is used for measuring the dynamic interfacial
trated TX45 solutions of 50 and 100 mmol/l tensions in Fig. 17 of a water drop growing in
are shown (Fainerman et al. 2009a). The solid a hexane solution containing 104 mol/l Span80
curves in this figure are the theoretical calcula- (Javadi et al. 2010a). It is shown that at least
tions considering diffusion-controlled mecha- a flow rate of about 8.5 mm3/s is necessary to
nism and the reorientation model. Very good reach the state of a fresh interface with interfacial
agreement with the experimental data for a wide tension close to that of the pure water/hexane
range of adsorption time from 0.001 to 1,000 s is interface. For the results shown in this figure,
observed. the hydrodynamic contributions were evaluated
experimentally and subtracted from the measured
Fast Dynamic Adsorption for Liquid/Liquid total pressure. This procedure is required to
Interfaces obtain real capillary pressure values which then S
The challenges of performing fast dynamic can be used for the calculation of interfacial ten-
experiments and their theoretical analysis are sion. The initial condition of the surface at the
much larger for liquid/liquid interfaces. The early stages of drop formation is a real challenge
main reasons are the more complicated hydrody- from a theoretical point of view. Figure 18 shows
namic effects of viscous flow inside the measur- a shift between the corresponding maximum cap-
ing cell, tubes, capillary tip, etc., and also the drag illary pressure and minimum drop radius at the
force of the motion of a drop (deformable inter- high flow rates (8.5 mm3/s). At the same time, for
face) inside the second immiscible liquid. low flow rates (0.85 mm3/s) as also presented in
For the stop growing drop (SGD) protocol in Fig. 18, such shift is negligible (Javadi et al.
capillary pressure tensiometry, the short-time 2010a).
dynamic interfacial tensions g(t) are presented Due to these problems in the CGD protocol,
in Fig. 14 for the range of 0.01–1,000 s. Similar pre-aged growing drop experiments (PGD) are
to the maximum bubble pressure tensiometry, also an interesting idea for studies of short-time
60
40
50
30 40
γ [mN/m]
A [mm2]
30
20
20
10
10
0 0
0.5 1 1.5 2
t [s]
Surfactant Adsorption Kinetics, Fig. 17 Dynamic hexane system and flow rates 0.85 ▪, 2 ♦, 4.8 ▼, and 8.5
interfacial tension (filled symbols) and drop surface area ~ mm3/s for water/solution system (104 mol/l Span80 in
A (open symbols) dependencies for growing water drop in hexane); the hydrodynamic effects are subtracted already
hexane using CGD protocol with different flow rates (real- (Javadi et al. 2010a)
time comparison), Q ¼ 0.85 mm3/s ● for pure water/
2 300 2 300
250 250
1.5 1.5
200 200
Pc [Pa]
Pc [Pa]
r [mm]
r [mm]
1 150 1 150
100 100
0.5 0.5
50 50
0 0 0 0
116 118 120 122 124 126 56.6 56.8 57 57.2 57.4
t [s] t [s]
Surfactant Adsorption Kinetics, Fig. 18 Capillary (left). The shift between corresponding maximum capil-
pressure values Pc (filled symbols) during growing drop lary pressure and minimum drop radius is clearly observed
of water (radius is shown with open symbols) in pure for high flow rates, which is insignificant for low flow
hexane (●) and 104 mol/l Span80 solution (▪) for high rates (Javadi et al. 2010a)
flow rate 8.5 mm3/s (right) and low flow rate 0.85 mm3/s
Surfactant Adsorption 3 60
Kinetics,
Fig. 19 Dynamic
interfacial tension (filled
symbols) and drop radius 50
(open symbols) 2.5
dependencies for growing
water drop in hexane using
40
PGD protocol with
different flow rates:
2
γ [mN/m]
0.85 mm3/s ● for pure
r [mm]
system and 0.42 ♦, 0.85 ▪, 30
2 ▼, 4.8 ◄, 8.5 ~, and
16.7 ► mm3/s for
surfactant solution
1.5
(105 mol/l Span80 in 20
hexane) (Javadi et al.
2010a)
10
1
0
0 0.2 0.4 0.6 0.8 1
t [s]
properties of liquid/liquid interfaces. Figure 19 Dilational Viscoelasticity of Surfactant

shows results of this approach for a water drop Solutions
growing in 105 mol/l Span80 hexane solution at
different flow rates (Javadi et al. 2010a). The Theoretical Basis
results for low and medium flow rates can be Adsorption kinetics of surfactants is closely
analyzed by a numerical simulations proposed related to another dynamic characteristic,
by MacLeod and Radke (1994) as long as the dilational viscoelasticity. Dilational deformation
initial conditions of the drop interface at equilib- of a fluid interface containing adsorbed surfactant
rium are known. However, the PGD protocol also molecules results in a change of interfacial ten-
shows limitation to reach a completely fresh sion depending on the deformation itself as well
interface at higher concentration (Javadi et al. as on the rate of deformation. Any relaxation
2010a). Currently, computational fluid dynamic process within the interfacial layer or adjacent
(CFD) simulations of the full set of Navier solution contributes to the dynamic response of S
Stokes equations without any simplifications the interface. If the deformations are sinusoidal
such as radial flow and homogeneous surface and small enough, the resulting change in surface
layers are under way. Such simulations, when tension can be represented as a sum of elastic and
successful, will allow a quantitative analysis of viscous terms, see Edwards et al. (1991), Krotov
data also at significantly higher liquid flow rates. (2009), and Ivanov et al. (2005):
However, the CFD packages for the simulation
of interfacial phenomena are still under devel-
Dg ¼ gðtÞ ge ¼ Eel a þ Za_ ¼ ðEel þ 2pinZÞa
opment, in particular regarding the interface
capturing and reconstruction techniques and (39)
also finding relevant numerical approaches for
the conservation of mass of the surface-active where Eel is the dilational surface elasticity,
components at the interface and between com- Z is the dilational surface viscosity,
Ðt
putational cells. a ¼ 0 a_ ðtÞdt ¼ lnð1 þ DA=A0 Þ ffi dA=A0 is the surface
deformation, A ¼ A0 þdA ¼ A0 þA1 expð2pintÞ is the where F is the Fourier transform operator.
area of surface at the time t (in complex representa- It should be stressed that the surface dilational
tion), dA is the change in the surface area, A1 is the viscoelastic modulus as a function of frequency,
amplitude of surface area oscillations, n is the E(in), is the inherent property of the system
frequency of oscillations, and i is the imaginary which does not depend on the type of excitation.
unit. Taking into account that deformation ampli- The only condition that has to be met is the small-
tude is very small and, therefore, A ffi A0 , ness of interfacial perturbations ensuring the linear
a ffi dðlnAÞ, Eq. 39 can be rewritten as: response. The experimental methods providing
the data on dilational surface elasticity are
EðinÞ E0 þ iE00 ¼ Eel þ 2pin discussed in details in the chapter ▶ Interfacial
Dynamics Methods.
dg dg dðln GÞ
¼ ¼ Except for the considerable practical impor-
dðln AÞ dðln GÞ dðln AÞ
tance, for example, for stability of foams and
dðln GÞ emulsions, the surface viscoelastic modulus pro-
E0 (40)
dðln AÞ vides a valuable information about time scales of
various relaxation processes occurring: diffu-
where E is the surface dilational viscoelasticity or sional exchange of surfactant between bulk and
surface dilational viscoelastic modulus, with its interface, reorientation and aggregation of mole-
real part E0 equal to the dilational surface elastic- cules at interface, etc. The theoretical description
ity and imaginary part E00 representing the contri- of relaxation due to diffusion-controlled
bution of the dilational surface viscosity, and E0 exchange of single surfactant molecules between
the limiting elasticity, that is, the high frequency interface and bulk was proposed by Lucassen and
limit corresponding to insoluble monolayer van den Tempel (1972):
behavior. Note that, although in this form of
presentation both the surface tension and the pffiffiffiffi
0 00 in
area of interface in Eq. 40 are functions of time, EðinÞ ¼ E þ iE ¼ E0 pffiffiffiffi pffiffiffiffiffiffiffi
in þ 2n0
the surface viscoelastic modulus depends on the pffiffiffiffiffiffiffiffiffiffi
frequency only. Using the area dependency on 1 þ n0 =n
¼ E0 pffiffiffiffiffiffiffiffiffiffi
time given above and representing the surface 1 þ 2 n0 =n þ 2n0 =n
tension as g ¼ ge þ g1 exp ið2pnt þ ’Þ, where pffiffiffiffiffiffiffiffiffiffi
n0 =n
g1 is the amplitude of surface tension oscillations þ iE0 p ffiffiffiffiffiffiffiffiffiffi (43)
and j is the phase shift between oscillations of 1 þ 2 n0 =n þ 2n0 =n
surface area and surface tension, the surface vis-
coelastic modulus can be represented as: or in terms of the viscoelasticity modulus |E|
and phase angle f between stress (dg) and strain
dg g ð nÞ
EðinÞ ¼ ¼ 1 expði’ðnÞÞ (41) (dA):
dðln AÞ A1
Equation 41 is a good basis for the treatment of 1
jEj ¼ E0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffiffiffiffi ;
experimental data based on sinusoidal oscilla- 1 þ 2 n0 =n þ 2n0 =n
tions of small amplitude. pffiffiffiffiffiffiffiffiffiffi (44)
If changes in surface area are still small, but n0 =n
f¼ pffiffiffiffiffiffiffiffiffiffi
not sinusoidal, the surface viscoelastic modulus 1 þ n0 =n
can be calculated using Fourier transformations
of measured time dependences, see Miller et al. with the characteristic frequency of diffusional
(1991) and Loglio et al. (2001): exchange given by:
FðDgðtÞÞ D
EðinÞ ¼ (42) n0 ¼ ðdc=dGÞ2 (45)
FðD ln AðtÞÞ 4p
The derivative dc/dG can be found from the take place was performed by Noskov and Loglio
corresponding adsorption isotherm. The latter (1998). The contribution of surface reactions into
are discussed in details in section “Adsorption surface rheology was considered thoroughly also
Isotherms for Surfactants” of this entry. The char- by Ivanov et al. (2005). The detailed discussion of
acteristic frequency can be easily found various relaxation processes (surface reactions,
from experimental data, as the maximum on reorientation, and aggregation of adsorbed mole-
the frequency dependence of imaginary part, cules at the interface) is presented in Ravera et al.
represented by the last term in the right-hand (2005), where the corresponding relationships for
side of Eq. 43, corresponds to frequency 2n0. calculation of viscoelastic modulus and charac-
If a liquid/liquid interface is considered teristic frequencies are derived.
and surfactant is soluble in both liquid phases, Surface elasticity of micellar solutions was con-
1 and 2, with partition coefficient K ¼ c2/c1, then sidered by Joos (1999), Noskov and Grigoriev
Eq. 46 takes the form: (2001), and Noskov (2002). The simplest equation
for complex viscoelastic modulus of micellar solu-
Deff tions in the limit when the characteristic time of
n0 ¼ ðdc1 =dGÞ2 (46)
4p micelles disintegration is much smaller than
the oscillation period as proposed by Joos (1999) is
with effective diffusion coefficient (Ravera et al.
1
2005) EðinÞ ¼ E0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
rffiffiffiffiffiffiffiffi!2 1 þ ð1 iÞ n ð1 þ bÞ 1 þ DDm b
n0
Dð2Þ
Deff ¼ D1 1þK (47) (49)
Dð1Þ
where Dm is the diffusion coefficient of micelle,
where D(i) is the diffusion coefficient in phase i. b ¼ (c CMC)/CMC, and CMC is the critical
The expression for the complex viscoelastic micelle concentration.
modulus for the case of mixture of two surfac- The presented above equations have been
tants was given in Aksenenko et al. (2006) and derived for a flat interface. Effect of the system
Aksenenko et al. (2007): geometry (flat surface, foam film, spherical drop,
rffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffi and liquid cylinder) is discussed in Joos (1999).
1 @P 2pin 2pin G2 For the spherical geometry, which is most impor-
E¼ a11 þ a12
B @ ln G1 G2 D1 D2 G1 tant from practical point of view and widely used

2pin in rheological measurements, for example, in
þ pffiffiffiffiffiffiffiffiffiffiffi ða11 a22 a21 a12 Þ profile analysis tensiometry and capillary pres-
D1 D2
rffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffi sure tensiometry (see chapter ▶ Interfacial
1 @P 2pin G1 2pin
þ a21 þ a22 Dynamics Methods), the equations for the case S
B @ ln G2 G1 D1 G2 D2
of diffusion-controlled adsorption kinetics are:
2pin
þ p ffiffiffiffiffiffiffiffiffiffiffi ða11 a22 a21 a12 Þ
D1 D2 1
EðinÞ ¼ E0 (50)
(48) 1 þ 2pinR
D
I dG ð1
dc
þ cRI Þ
where for diffusion from outside of a drop or a bubble and

rffiffiffiffiffiffiffiffiffi
2pin
B ¼1 þ a11
D 1
rffiffiffiffiffiffiffiffiffi1 EðinÞ ¼ E0
2pin 2pin 1þ 2pinRI dG ðcRI
D dc
cothðcRI Þ 1Þ
þ a22 þ pffiffiffiffiffiffiffiffiffiffiffi ða11 a22 a21 a12 Þ
D2 D1 D2 (51)
and aij ¼ ð@Gi =dcj Þck6¼j .
The general thermodynamic treatment of the for diffusion from inside of drop. Here, RI is the
paffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
system where different relaxation processes can drop/bubble radius, c ¼ 2pin=D.
Experimental dependency becomes considerable. Noteworthy

Experimental data on dilational surface rheology that critical micelle concentration for C11DMPO
provide valuable information about adsorption is about 1.3 mM and at concentrations close to
properties of surfactant solutions. First of all, critical micelle concentration processes of aggre-
they enable proving whether diffusion is gation/disaggregation in bulk phase can contrib-
a limiting step in adsorption kinetics. It is easy ute to the surface dilational viscoelasticity. Study
to obtain from Eqs. 43 and 45 that: performed for micellar solutions of C14EO8
(Fainerman et al. 2008b) has shown that rheolog-
ical behavior of this surfactant is described very
E0 2 þ E00 2
E0 ¼ (52) well by Eq. 49.
E0 E00 Equations 40, 43, and 45 show clearly that the
rffiffiffiffiffiffiffiffi dilational viscoelastic modulus contains informa-
dG E0 E00 D tion on equilibrium thermodynamic characteris-
¼ (53)
dc E00 4pn tics of the interfacial layer, that is, it is directly
related to the adsorption isotherm through the
The quantities in the left-hand side of Eqs. 52 quantities E0 and dc/dG. It is noteworthy that
and 53 are obviously frequency independent, the chosen thermodynamic model and its param-
whereas those in the right-hand side depend on eters have much more pronounced influence on
frequency. Therefore, if the calculations dilational viscoelasticity than on the shape of
according to Eqs. 52 and 53 using the experimen- surface tension isotherm. Therefore, analysis of
tal values for E0 and E00 provide frequency inde- experimental data on surface rheology provides
pendent values of E0 and dG/dc, the model of more precise knowledge of equilibrium adsorp-
Lucassen and van den Tempel is applicable and tion properties.
diffusion-controlled adsorption kinetics is valid. An impressive example of using the rheolog-
Corresponding analysis of experimental data for ical data to improve the knowledge on equilib-
alkyl dimethyl phosphine oxides (C1012DMPO) rium characteristics is the concept of intrinsic
presented by Wantke and Fruhner (2001) was compressibility of interfacial layer (see section
performed in Kovalchuk et al. (2004a). The “Adsorption Isotherms for Surfactants” of this
detailed experimental investigation and theoreti- entry). As discussed in details by Kovalchuk
cal analysis for oxyethylated alcohols C14EO8 et al. (2004b) and Kovalchuk et al. (2005), the
has been presented in Fainerman et al. (2008a). most commonly used thermodynamic models,
The diffusion-controlled adsorption kinetics was those of Langmuir and Frumkin, predict infinite
proved for these surfactants, at least for concen- growth of the limiting elasticity E0 at high sur-
trations well below the critical micelle concen- factant concentration, what contradicts the rheo-
tration. Experimental study on Brij 35 (C12EO23) logical experimental data. It was then assumed,
in the circular frequency range 0.3–10 s–1 taking into account experimental data provided
(Hansen 2008) has shown the only slow increase by grazing incidence X-ray diffraction, GIXD,
of E0 value with the frequency, and therefore it that the molar area of surfactant is not constant,
was concluded that diffusion relaxation is domi- but decreases with lateral pressure according to
nant for this solution at least for small frequen- Eqs. 19 and 20. In this case, the limiting (high-
cies. An example given in Fig. 20 shows the frequency) elasticity at higher surface coverages
results for undecyl dimethyl phosphine oxide is close to the reverse of 2-D compressibility e,
(C11DMPO). It is seen that results for three and rheological data enable rather precise esti-
smaller concentrations are described very well mating of this parameter. Figure 21 shows the
in assumption of diffusion kinetics of adsorption, experimental data on the limiting elasticity for
whereas at concentration 0.5 mM, there appears nonionic surfactant dodecyl dimethyl phosphine
already dependence on the frequency for both E0 oxide C12DMPO and theoretical curves based on
and dG/dc, and for concentration 1 mM, this original Frumkin model and modified Frumkin
Kinetics,
Fig. 20 Frequency
dependence of dG/dc and
E0 for C11DMPO solutions
with concentrations: ●
– 0.05 mM,▼ – 0.1 mM, ♦
– 0.2 mM, ~ – 0.5 mM, and
▪ – 1 mM, Kovalchuk et al.
(2004a)
model assuming intrinsic compressibility of solutions. Dodecanol is a nonionic surfactant

adsorbed layer. It is obvious that the model with which attracts special attention because it is
intrinsic compressibility describes experimental known that at temperatures below 20 C, the 2-D
data much more precisely. The similar results are aggregate formation occurs in interfacial layer if
reported for ionic surfactants alkyl trimethyl the surface coverage exceeds certain threshold
ammonium bromides in Stubenrauch et al. value (see Vollhardt et al. (2000)). Rheological
(2005). study (Liggieri et al. 2003) has shown that
The dilational rheology enables also studying according to the theoretical predictions (Ravera
other interfacial relaxation processes. One exam- et al. 2005), the elasticity modulus decreases
ple is the surface rheology of dodecanol sharply when aggregation process starts.
Kinetics, Fig. 21 Limiting
elasticity E0 versus
concentration for
C12DMPO. Experimental
data taken from Wantke
and Fruhner (2001);
Kovalchuk et al. (2004b),
theoretical curves: 1 –
original Frumkin model,
2 – Frumkin model with
intrinsic compressibility,
reproduced from
Kovalchuk et al. (2005)
Another example is the effect of molecular Surfactant Adsorption Kinetics, Table 3 Characteristic
reorientation relaxation on the rheological data. frequencies of relaxation processes for surfactants at the
water/hexane interface
Widely used in technical applications, Tritons are
described in the best way by the thermodynamic Diffusion Reorientation
Surfactant c, mol/cm3 n, Hz n, Hz
model accounting for reorientation. The detailed
C10E4 5∙109 3.9 246
rheological study of these surfactants based on
2∙108 75 547
reorientation model (see section “The
5∙108 217 881
Reorientation Model”) was performed by C10E5 1∙108 34 1,692
Fainerman et al. (2009b). It was predicted by C10E8 2∙108 2.2 25,000
theoretical analysis in the framework of C16E2 (Brij52) 1∙107 4.7 58.9
reorientation model, taking into account also the
effect of intrinsic compressibility of adsorbed
layer, that Tritons TX100, TX165, and TX405
should display a second maximum on the con- by Ravera et al. (2005). Diffusion adsorption
centration dependence of the elasticity modulus, kinetics was confirmed for this surfactant series
and those predictions were confirmed experimen- at both interfaces by the corresponding maximum
tally. The maximum becomes more pronouncing in the frequency dependency of imaginary part of
with the increase of the ethylene oxide chain complex viscoelastic modulus. The characteristic
length. The only one maximum was predicted frequency for reorientation was beyond the fre-
theoretically and observed experimentally for quency range used in Ravera et al. (2005); nev-
TX45 which has the shortest chain length ertheless, the experimental results enabled the
among the studied Tritons. Noteworthy that the estimation of this characteristic frequency. The
theoretical treatment was performed by taking data on the frequencies of relaxation processes
into account the real system geometry, namely, presented in Ravera et al. (2005) for water/hex-
Eq. 50 for measurements with buoyant bubble ane interface are listed in Table 3.
and Eq. 51 for measurements with pendant drop. The series of experiments was performed
The effect of reorientation relaxation on rheo- confirming validity of Eq. 48 for a mixture of
logical data for oxyethylated alcohols CnEOm at surfactants. For example, the results for the mix-
water/air and water/hexane interface was studied ture of C10DMPO and C14DMPO are reported by
Kinetics,
Fig. 22 Dependence of
viscoelasticity modulus
(left) and phase angle
(right) on concentration at
the oscillation frequency
f ¼ 0.01 Hz for individual
C14EO8 (red points, red
curve) and SDS (blue
points, blue curve)
solutions and SDS/ C14EO8
mixtures with molar ratio of
the components 5:1
(◊, solid curve), 30:1
(D, dotted curve), and 200:1
(□, dashed curve),
respectively; the points are
experimental data, and
lines are calculated
dependencies, according to
Fainerman et al. (2010d)
Aksenenko et al. (2007), where also the experi- results and corresponding theoretical curves cal-
mental results for the mixture of sodium dodecyl culated from Eq. 48 for the mixture of SDS and
sulfate (SDS) and dodecanol presented by C14EO8 are presented in Fig. 22 displaying
Wantke et al. (2003) are analyzed. For both sys- a good coincidence of experimental and theoret-
tems, good qualitative agreement between theory ical data.
and experiment has been obtained. Other exam-
ples of experimental studies on the surfactant
mixtures can be found in Fainerman et al. Summary and Conclusions
(2010d), Fainerman et al. (2010e) (SDS/
CmEOn), and Fainerman et al. (2010b) (C12EO5/ We can state that the adsorption kinetics of sur-
C14EO8 and TX45/TX165). The experimental factants at the water/air interface is essentially
understood. From literature data, we know (2007), as so far suitable experiments are yet
that so far all normal surfactants adsorb pending.
according to a diffusion mechanism. This is All the described recent achievements in the
true for ionic and nonionic surfactants; however, better understanding of liquid interfacial layers
the corresponding adsorption model has to be were made due to the great progress in the devel-
used as basis for the quantitative data analysis. opment of suitable experimental tools. Selected
We can also conclude that we can quantitatively methodologies and instruments are described in
describe the adsorption dynamics and the final the chapter ▶ Interfacial Dynamics Methods.
equilibrium state for mixed aqueous solution/air Together with advanced simulations, based, for
interfaces when we consider the correct adsorp- example, on CFD simulations or the free accessible
tion model for each of the compounds. OpenFOAM (Open Field Operation and Manipu-
Fainerman et al. (2010c) have shown that even lation) platform, we have now a situation where
ternary mixtures are perfectly described by the very complex systems can be quantitatively
available theory. described, even under rather dynamic conditions.
The situation is different when we consider
the interface between two immiscible liquids.
There is something known about the impact of List of Symbols
the type of oil; however, we are yet rather far
from a quantitative understanding. Depending A Surface area
on the aliphatic chain length of the surfactant A0 Surface area before expansion
A1 Amplitude of surface area oscillations
and the oil, different interfacial structures are
a Interaction constant (Frumkin model)
discussed, such as incorporation or squeezing
aij ð@Gi =dcj Þck6¼j , coefficients in Eq. 48
out of oil molecules. One of the peculiarities of
B Value in Eq. 48
the equilibrium behavior of a surfactant is that
b Adsorption equilibrium constant
remarkable adsorption effects are already mea-
c Surfactant bulk concentration
surable at bulk concentrations at which abso-
ci ith surface active component bulk
lutely no change in surface tension is observed concentration
at the solution/air surface. Recently, experi- c(k), k ¼ I, E Concentration of surfactant in kth region
ments were performed dealing with the surfac- cC Concentration of counterions (Frumkin
tant adsorption at the water/air interface in model for ionic surfactants)
presence of alkane vapor in the air. It could be cS/S Subsurface concentration of surfactant in
shown that this type of systems represents kth region
ðkÞ th
a bridge between the water/air and the water/ cS=S k ¼ I, E Subsurface concentration of i surface
active component
oil interface. More systematic investigations
ci,S/S Subsurface concentration of ith surface
are required to describe these bridging proper- active component
ties quantitatively. c* Mean ionic activity, Eq. 38
It has been discussed here also that relation D Diffusion coefficient of surfactant in
experiments with small perturbations are very solution
suitable to shine more light into the particular Di Diffusion coefficient of ith surface active
properties of interfacial layers. Relaxation data component in solution
D(k), k ¼ I, E Diffusion coefficient of surfactant in kth
obtained for a very broad frequency spectrum are
region
obviously most suitable to detect specific effects Deff Effective diffusion coefficient
of surfactants related to respective structures in Dm Diffusion coefficient of micelles
the molecule. dA Surface area change
Also yet quite unclear are the conditions E0 Limiting elasticity
at which the ionic charge plays a significant role E Surface dilational viscoelasticity
in the adsorption process, despite the very Eel Surface dilational elasticity
advanced recent analysis given by Ivanov et al. (continued)
E0 Real part of surface dilational
P Dimensionless surface pressure
viscoelasticity Pi0 Equilibrium surface pressure of
E00 Imaginary part of surface dilational component i
viscoelasticity B Constant in Eq. 16 (reorientation model)
F Fourier transform operator tD Diffusion relaxation time
f Function in Eq. 23 j Phase shift
g Function in Eq. 37 c Parameter in Eqs. 50 and 51
fi Activity coefficients of component i o Partial molar surface area
K Henry law adsorption constant o0 Partial molar surface area of solvent
KI,E distribution coefficient of surfactant molecule
between Ith and Eth region o1 Partial molar surface area for monomer
k12 Rate constant for the reorientation between (aggregation model)
states 1 and 2 in the surface on Partial molar surface area for n-mer
mi Number of moles of component i (aggregation model)
n Aggregation number (aggregation model) oi Partial molar surface area for the ith state of
P Pressure adsorbed molecule (reorientation model)
R Universal gas constant o10 Minimal partial molar area for the first state
RI Spherical interface radius of adsorbed molecule (reorientation model)
RE External radius of spherical cell oX partial molar surface area of the complex
r Radial coordinate
Acknowledgements The work was financially supported
T Absolute temperature by a project of the DFG SPP 1506 (Mi418/18-1), the
xi ¼ mi =Smi Molar fraction of component i European Space Agency (ESA FASES MAP AO-99-052
a Surface deformation (dilation) and PASTA), and the German Space Agency (DLR
a_ Rate of surface deformation (dilation) 50WM1129).
b Quantity introduced in Eq. 12
g Surface/interfacial tension
g0 Surface/interfacial tension of pure solvent Cross-References
g1 Amplitude of surface tension oscillations
ge Equilibrium value of surface/interfacial ▶ Interfacial Dynamics Methods
tension
G Adsorption
Gi Adsorption of component i References
G1 Partial adsorption of monomer
(aggregation model)
Aksenenko EV, Makievski AV, Miller R, Fainerman VB
GC Critical adsorption of aggregation (1998) Dynamic surface tension of aqueous alkyl
(aggregation model) dimethyl phosphine oxide solutions. Effect of the
G1, G2 Partial adsorptions (reorientation model) alkyl chain length. Coll Surf A 143:311–321
G1 Maximum adsorption Aksenenko EV, Kovalchuk VI, Fainerman VB, Miller R S
e Intrinsic compressibility coefficient (2006) Surface dilational rheology of mixed adsorp-
tion layers at liquid interfaces. Adv Coll Int Sci
Z Dilational surface viscosity
122:57–66
Y Dimensionless time in Eqs. 26, 27 Aksenenko EV, Kovalchuk VI, Fainerman VB, Miller
y Surface coverage R (2007) Surface dilational rheology of mixed surfac-
yX Surface coverage by surfactant ions and tants layers at liquid interfaces. J Phys Chem
counterions (Frumkin model for ionic C 111:14713–14719
surfactants) Aksenenko EV, Fainerman VB, Petkov JT, Miller R
m0s ðT; P; g Þ Standard chemical potential of component i (2010) Dynamic surface tension of mixed
i
oxyethylated surfactant solutions. Colloid Surf
n Frequency of oscillations
A 365:210–214
n0 Characteristic frequency of diffusional Alahverdjieva VS, Fainerman VB, Aksenenko EV, Leser
exchange ME, Miller R (2008) Adsorption of hen egg-white
x Quantity defined by Eq. 12 lysozyme at the air–water interface in presence of
P ¼ g0g Surface pressure sodium dodecyl sulphate. Colloids Surf
(continued) A 317:610–617
Aveyard R, Haydon DA (1965) Thermodynamic proper- Fainerman VB, Petkov JT, Miller R (2008b) Surface
ties of aliphatic hydrocarbon/water interfaces. dilational viscoelasticity of C14EO8 micellar solution
J Colloid Sci 20:2255–2261 studied by bubble profile analysis tensiometry. Lang-
Bergeron V (1997) Disjoining pressure film stability of muir 24:6447–6452
alkyltrimethylammonium bromide foam films. Lang- Fainerman VB, Lylyk SV, Aksenenko EV, Liggieri L,
muir 13:3474–3482 Makievski AV, Petkov JT, Yorke J, Miller R (2009a)
Butler JAV (1932) The thermodynamics of the surfaces of Adsorption layer characteristics of Triton surfactants.
solutions. Proc Roy Soc Ser A 135:348–375 Part 2. Dynamic surface tension and adsorption. Coll
Doss KSG (1938) Alterung der Oberfl€achen von Surf A 334:8–15
Lösungen; III. Aktivierte Anreicherung von gelösten Fainerman VB, Aksenenko EV, Lylyk SV, Makievski
Molek€ulen an der Oberfl€ache von Benzopurpurin- AV, Ravera F, Petkov JT, Yorke J, Miller R (2009b)
10B-Lösungen. Kolloid Z 84:138–143 Adsorption layer characteristics of Triton surfactants.
Dukhin SS, Kovalchuk VI, Fainerman VB, Miller 3. Dilational visco-elasticity. Coll Surf A 334:16–21
R (1998) Hydrodynamic processes in dynamic bubble Fainerman VB, Aksenenko EV, Lylyk SV, Petkov JT,
pressure experiments. Part 3. Oscillatory and aperiodic Yorke J, Miller R (2010a) Adsorption layer character-
modes of pressure variation in the capillary. Colloids istics of mixed sodium dodecyl sulfate/CnEOm solu-
Surf A 141:253–267 tions 1. Dynamic and equilibrium surface tension.
Edwards DA, Brenner H, Wasan DT (1991) Interfacial Langmuir 26:284–292
transport processes and rheology, Butterworth- Fainerman VB, Aksenenko EV, Petkov JT, Miller R
Heinemann series in chemical engineering. (2010b) Adsorption layer characteristics of mixed
Butterworth-Heinemann, Boston oxyethylated surfactant solutions. J Phys Chem
Fainerman VB, Lucassen-Reynders EH (2002) Adsorption B 114:4503–4508
of single and mixed ionic surfactants at fluid inter- Fainerman VB, Aksenenko EV, Petkov JT, Miller R
faces. Adv Coll Int Sci 96:295–323 (2010c) Adsorption layer characteristics of
Fainerman VB, Miller R (1996) Surface tension isotherms multicomponent surfactants solutions. Soft Matter
for surfactant adsorption layers including surface 6:4694–4700
aggregation. Langmuir 12:6011–6014 Fainerman VB, Aksenenko EV, Zholob SA, Petkov JT,
Fainerman VB, Miller R (2001) Simple method to esti- Yorke J, Miller R (2010d) Adsorption layer character-
mate surface tension of mixed surfactant solutions. istics of mixed SDS/CnEOm solutions. II. Dialtional
J Phys Chem B 105:11432–11438 viscoelasticity. Langmuir 26:1796–1801
Fainerman VB, Miller R (2009) Adsorption isotherms Fainerman VB, Aksenenko EV, Mys AV, Petkov JT,
at liquid interfaces. In: Somasundaran P, Hubbard Yorke J, Miller R (2010e) Adsorption layer character-
A (eds) Encyclopedia of surface and colloid science, istics of mixed SDS/CnEOm solutions. 3. Dynamic of
2nd edn, vol 1. pp 1–15. Taylor & Francis, London, UK adsorption and surface dilational rheology of micellar
Fainerman VB, Makievski AV, Miller R (1994) Analysis solutions. Langmuir 26:2424–2429
of dynamic surface tension of Na-alkyl sulphate solu- Ferri JK, Stebe KJ (1999) Soluble surfactants undergoing
tions, based on asymptotic solutions of adsorption surface phase transitions: a Maxwell construction
kinetics equations. Coll Surf A 87:61–75 and the dynamic surface tension. J Coll Int Sci
Fainerman VB, Miller R, W€ ustneck R (1997) Adsorption 209:1–9
isotherm and surface tension equation for a surfactant Frumkin A (1925) Die Kapillarkurve der höheren
with changing partial molar area, 2. Non-ideal surface Fetts€auren und die Zustandsgleichung der
layer. J Phys Chem 101:6479–6483 Oberfl€achenschicht. Z Phys Chem (Leipzig)
Fainerman VB, Lucassen-Reynders EH, Miller R (1998) 116:466–484
Adsorption of surfactants and proteins at liquid/fluid Gehlert U, Vollhardt D (2002) Molecular packing and
interfaces. Coll Surf A 143:141–165 textures of 1-Stearylamine-rac- glycerol monolayers.
Fainerman VB, Möbius D, Miller R (2001) Surfactants – Langmuir 18:688–693
chemistry, interfacial properties and application, vol 13, Gibbs JW (1928) The collected works, vol 1. Longmans/
Studies in interface science. Elsevier, Amsterdam Green, New York
Fainerman VB, Miller R, Kovalchuk VI (2002) Influence Hansen RS (1961) Diffusion and the kinetics of adsorption
of compressibility of adsorbed layer on the surface of aliphatic acids and alcohols at the water-air inter-
dilational elasticity. Langmuir 18:7748–7752 face. J Coll Sci 16:549–560
Fainerman VB, Kovalchuk VI, Aksenenko EV, Michel M, Hansen FK (2008) Surface dilational elasticity of poly
Leser ME, Miller R (2004) Models of two-dimensional (oxy ethylene)-based surfactants by oscillation and
solution assuming the internal compressibility of relaxation measurements of sessile bubbles. Langmuir
adsorbed molecules: a comparative analysis. J Phys 24:189–197
Chem 108:13700–13705 Henry W (1803) Experiments on the quantity of gases
Fainerman VB, Zholob SA, Petkov JT, Miller R (2008a) absorbed by water, at different temperatures, and
C14EO8 adsorption characteristics studied by drop and under different pressures. Phil Trans R Soc Lond
bubble profile tensiometry. Coll Surf A 323:56–62 93:29–42
Ivanov IB, Danov KD, Ananthapadmanabhan KP, Lips dodecanol monolayers at the water/air interface. Lang-
A (2005) Interfacial rheology of adsorbed layers with muir 19:10233–10240
surface reaction: on the origin of the dilatational vis- Loglio G, Pandolfini P, Miller R, Makievski AV, Ravera F,
cosity. Adv Coll Int Sci 114–115:61–92 Ferrari M, Liggieri L (2001) Drop and bubble shape
Ivanov IB, Martinova KG, Danov KD, Dimitrova D, analysis as a tool for dilational rheological studies of
Kavssery P, Ananthapadmanabhan KP, Lips interfacial layers. In: Möbius D, Miller R (eds) Novel
A (2007) Role of the counterions on the adsorption of methods to study interfacial layers, vol 11, Studies in
ionic surfactants. Adv Coll Int Sci 134–135:105–124 interface science. Elsevier, Amsterdam
Javadi A, Kr€agel J, Pandolfini P, Loglio G, Kovalchuk VI, Lu JR, Thomas RK, Aveyard R, Binks BP, Cooper P,
Aksenenko EV, Ravera F, Liggieri L, Miller R (2010a) Fletcher PDI, Sokolowski A, Penfold J (1992) Com-
Short time dynamic interfacial tension as studied by parison of neutron reflection and surface-tension mea-
the growing drop capillary pressure technique. Col- surements of the surface excess of
loids Surf A 365:62–69 tetradecyltrimethylammonium bromide layers at the
Javadi A, Mucic N, Vollhardt D, Fainerman VB, Miller air-water-interface. J Phys Chem 96:1383–1388
R (2010b) Adsorption characteristics of SDS at the Lu JR, Thomas RK, Binks BP, Fletcher PDI, Penfold
water/hexane interface. J Colloid Interf Sci J (1995) Structure and composition of dodecane layers
351:537–541 spread on aqueous-solutions of dodecyl- and
Joos P (1999) Dynamic surface phenomena. VSP, hexadecyltrimethylammonium bromides studied by
Dordrecht neutron reflection. J Phys Chem 99:4113–4123
Kotsmar CS, Pradines V, Alahverdjieva VS, Aksenenko Lu JR, Li ZX, Thomas RK, Binks BP, Crichton D,
EV, Fainerman VB, Kovalchuk VI, Kr€agel J, Leser Fletcher PDI, McNab JR, Penfold J (1998) The struc-
ME, Noskov BA, Miller R (2009) Thermodynamics, ture of monododecyl pentaethylene glycol monolayers
adsorption kinetics and rheology of mixed protein– with and without added dodecane at the air/solution
surfactant interfacial layers. Adv Colloid Interf Sci interface: a neutron reflection study. J Phys Chem
150:41–54 102:5785–5793
Kovalchuk VI, Dukhin SS (2001) Dynamic effects in Lucassen J, van den Tempel M (1972) Dynamic measure-
maximum bubble pressure experiments. Colloids ments of dilational properties of a liquid interface.
Surf A 192:131–155 Chem Eng Sci 27:1283–1291
Kovalchuk VI, Loglio G, Fainerman VB, Miller R (2004a) Lucassen-Reynders EH (1966) Surface equation of state
Interpretation of surface dilational elasticity data based for ionized surfactants. J Phys Chem 70:1777–1785
on an intrinsic two-dimensional interfacial compress- Lucassen-Reynders EH (1972) A surface equation of state
ibility model. J Coll Int Sci 270:475–482 for mixed surfactant monolayers. J Coll Interf Sci
Kovalchuk VI, Kr€agel J, Makievski AV, Ravera F, 41:156–167
Liggieri L, Loglio G, Fainerman VB, Miller Lunkenheimer K, Miller R (1979) On the purity of aque-
R (2004b) Rheological surface properties of ous surfactant solutions and the dynamic surface ten-
C12DMPO solution as obtained from amplitude- and sion behaviour. Tenside 16:312–316
phase-frequency characteristics of an oscillating bub- MacLeod CA, Radke CJ (1994) Surfactant exchange
ble system. J Coll Int Sci 280:498–505 kinetics at the air/water interface from the dynamic
Kovalchuk VI, Miller R, Fainerman VB, Loglio G (2005) tension of growing liquid drops. J Coll Interf Sci
Dilational rheology of adsorbed surfactant layers – 166:73–88
role of the intrinsic two-dimensional compressibility. Miller R (1990) Adsorption kinetics of surfactants at
Adv Coll Int Sci 114–115:303–312 fluid interfaces: experimental conditions and practice
Krotov VV (2009) Basics of interfacial rheology. In: of application of theoretical models. Coll Surf
Miller R, Liggieri L (eds) Interfacial rheology, pro- 46:75–83 S
gress in colloid and interface science, vol 1. Brill, Miller R, Loglio G, Tesei U, Schano KH (1991) Surface
Leiden, pp 1–37 relaxations as a tool for studying dynamic interfacial
Langmuir I (1917) The constitution and fundamental behaviour. Adv Coll Int Sci 37:73–96
properties of solids and liquids. Part ii. Liquids. J Am Miller R, Joos P, Fainerman VB (1994) Dynamic surface
Chem Soc 39:1848–1907 and interfacial tensions of surfactant and polymer solu-
Liggieri L, Ravera F, Ferrari M, Passerone A, Miller R tions. Adv Coll Interf Sci 49:249–302
(1997) Adsorption kinetics of alkyl phosphine oxides Miller R, Aksenenko EV, Liggieri L, Ravera F, Ferrari M,
at water/hexane interface, 2. of the adsorption with Fainerman VB (1999) Effect of the reorientation of
transport across the interface in finite systems theory. oxyethylated alcohol molecules within surface layer
J Colloid Interf Sci 186:46–52 on equilibrium and dynamic surface pressure. Lang-
Liggieri L, Ferrari M, Massa A, Ravera F (1999) Molec- muir 15:1328–1336
ular reorientation in the adsorption of some CiEj at the Miller R, Fainerman VB, Aksenenko EV, Leser ME,
water-air interface. Colloids Surf A 156:455–463 Michel M (2004a) Dynamic surface tension and
Liggieri L, Ravera F, Ferrari M (2003) Surface rheology adsorption kinetics of b-casein at the solution/air inter-
investigation of the 2-D phase transition in n- face. Langmuir 20:771–777
S 1126 Surfactant Adsorption Under Nonequilibrium
Miller R, Fainerman VB, Aksenenko EV (2004b) Wantke KD, Fruhner H (2001) Determination of surface
A simple method to estimate the dynamic surface dilational viscosity using the oscillating bubble
pressure of surfactant mixtures. Colloids Surf method. J Coll Int Sci 237:185–199
A 242:123–128 Wantke KD, Fruhner H, Örtegren J (2003) Surface dilata-
Milner SR (1907) On surface concentration; and the for- tional properties of mixed sodium dodecyl sulfate/
mation of liquid films. Phil Mag 13:96–101 dodecanol solutions. Coll Surf A 221:185–195
Mishchuk NA, Dukhin SS, Fainerman VB, Kovalchuk VI, Ward AFH, Tordai L (1946) Time-dependence of bound-
Miller R (2001) Hydrodynamic processes in dynamic ary tensions of solutions. I. The role of diffusion in
bubble pressure experiments. Part 5. The adsorption at time-effects. J Phys Chem 14:453–461
the surface of a growing bubble. Colloids Surf
A Physicochem Eng Aspects 192:157–175
Mucic N, Javadi A, Kovalchuk NM, Miller R (2011)
Dynamic of interfacial layers – experimental feasibil-
ities of adsorption kinetics and dilational rheology. Surfactant Adsorption Under
Adv Colloid Interf Sci 168:167–178 Nonequilibrium
Noskov BA (2002) Kinetics of adsorption from micellar
solutions. Adv Coll Int Sci 95:237–293
Noskov BA, Grigoriev DO (2001) Adsorption from micel- Nina M. Kovalchuk1,2 and Dieter Vollhardt1
1
lar solutions. In: Fainerman VB, Möbius D, Miller Max Planck Institute of Colloids and Interfaces,
R (eds) Surfactants – chemistry, interfacial properties, Potsdam/Golm, Germany
applications, vol 13, Studies in interface science. 2
Institute of Biocolloid Chemistry, Kiev, Ukraine
Elsevier, Amsterdam
Noskov BA, Loglio G (1998) Dynamic surface elasticity
of surfactant solutions. Coll Surf A 143:167–183
Peltonen LJ, Yliruusi J (2000) Surface pressure, hystere- Synonyms
sis, interfacial tension, and CMC of four sorbitan
monoesters at water-air, water-hexane, and hexane-
air interfaces. J Colloid Interf Sci 227:1–6 Chemical oscillation at water/oil interface;
Pradines V, Fainerman VB, Aksenenko EV, Kr€agel J, Electrical oscillation at water/oil interface;
Mucic N, Miller R (2010) Alkyltrimethylammonium Oscillatory Marangoni convection; Oscillatory
bromides adsorption at liquid/fluid interfaces in the
Marangoni instability; Periodic Marangoni
presence of neutral phosphate buffer. Colloids Surf
A 371:22–28 instability
Ravera F, Ferrari M, Liggieri L (2000) Adsorption and
partitioning of surfactants in liquid–liquid systems.
Adv Colloid Interf Sci 88:129–177
Ravera F, Ferrari M, Santini E, Liggieri L (2005) Influence
Keywords
of surface processes on the dilational visco-elasticity
of surfactant solutions. Adv Coll Int Sci 117:75–100 Compression of adsorption layer, liquid
Ross S (1945) The change of surface tension with time II. interface, marangoni instability, oscillation of
The surface tension of saponin solutions. J Phys Chem
interfacial tension, surfactant transfer, systems
49:377–386
Sharipova A, Aidarova S, Fainerman VB, Miller R (2011) far from equilibrium
Study of dynamic interfacial tension of alkyl sulphates
with different alkyl chain lengths at the water/hexane
interface. Col Surf A 382:181–185
Stubenrauch C, Fainerman VB, Aksenenko EV, Miller R
Definition
(2005) Adsorption behavior and dilational rheology of
the cationic alkyl trimethylammonium bromides at Auto-oscillation of interfacial tension appears in
water/air interface. J Phys Chem B 109:1505–1509 systems far from equilibrium where a surfactant
Sutherland KL (1952) The kinetic of adsorption at liquid
transfer occurs through or to a liquid interface.
surfaces. Aust J Sci Res A5:683–696
Vollhardt D, Fainerman VB (2010) Characterisation of Their origin is the Marangoni instability, that is,
phase transition in adsorbed monolayers at the air/ spontaneous development of convective motion
water interface. Adv Coll Interf Sci 154:1–19 due to gradients of interfacial tension supported
Vollhardt D, Fainerman VB, Emrich G (2000) Dynamic
by a surfactant transfer from the bulk to the inter-
and equilibrium surface pressure of adsorbed
dodecanol monolayers at the air/water interface. face. A specific common feature of the systems
J Phys Chem B 104:8536–8543 where auto-oscillations develop is limited area of
Surfactant Adsorption Under Nonequilibrium 1127 S
the interface. Compression of surfactant layer Z
near the walls causes termination of the
instability development resulting in an oscillatory C2
2
behavior instead of steady convection.
Interface
O
Overview X
1
Systems far from equilibrium demonstrate a rich C1
variety of self-organization phenomena. Exam-
ples can be found in any aspect of our life, in
nature, and society – in physics, chemistry, biol- Surfactant Adsorption Under Nonequilibrium,
ogy, economics, etc. (Schweizer 1997). One of Fig. 1 Mechanism of Marangoni instability due to sur-
the best known self-organization phenomena in factant transfer through the liquid interface
colloid and interface science is Marangoni insta-
bility arising in the systems with free liquid inter-
face, where surfactant or/and heat transfer takes changes of surface concentration gradients in
place (Colinet et al. 2001). Instability results in vicinity of point O is determined by the velocities
a spontaneous development of convection in the of convection in phases 1 and 2, which in turn
system, revealing itself either in formation of depend on the viscosities of these liquids, and by
steady convective cells, such as well-known concentration distributions in both x and
Benard cells, or in different oscillatory regimes z directions, depending on surfactant diffusion
with changes in the cell structure and wave coefficients in solvents 1 and 2. If the velocity
propagation. in phase 1 is larger than that in phase 2 (i.e.,
The first who considered theoretically the kinematic viscosity of phase 1 is larger than that
instability development due to a surfactant trans- of phase 2) and concentration distribution in both
fer were Sternling and Scriven (1959). It can be x and z directions is more sharp in phase 2 (i.e.,
shortly explained as follows. Let us consider two diffusion coefficient of surfactant is smaller in
immiscible and initially motionless liquid layers phase 2 than in phase 1), then interfacial concen-
brought into contact (Fig. 1). Liquids contain tration gradients and therefore interfacial tension
a surfactant (see also ▶ Surfactants) as a solute gradients in vicinity of point O increase. These
out of a partition equilibrium, namely, diffusion gradients cause the increase of convective veloc-
surfactant transfer that takes place from phase ity and, therefore, further increase of concentra-
1 to phase 2. There are always fluctuations of tion gradients close to point O. Thus, at certain
surfactant concentration over the system due to conditions indicated above, there is a feedback S
random thermal motion. Suppose that concentra- between interfacial concentration gradient and
tion in an arbitrary point O at the interface fluc- velocity of convective motion. This feedback
tuates to a larger value than in the neighborhood. enables the growth of concentration fluctuations
Then the interfacial tension in this point becomes and development of spontaneous convection in
smaller (see also ▶ Surfactant Adsorption), and the system. Such spontaneous transition from
liquids start to move out of this point because of initial state characterized by diffusion surfactant
the gradient of interfacial tension. The bulk transfer to convective one due to formation of
liquids are immediately involved in this convec- self-supporting interfacial tension gradients has
tive motion due to viscous forces and continuity been called Marangoni instability. Depending on
of liquid flow. the system properties, instability can reveal itself
Now, convection supplies to point O more as a quasi-steady convection or as oscillatory one.
concentrated solution from phase 1 and more If the changes in interfacial tension occur due to
diluted solution from phase 2. The direction of surfactant adsorption, one speaks about solutal
Marangoni instability. If heat instead of surfac- instability, which are related in particular to the
tant is an acting agent and interfacial tension properties of adsorbed monolayers, such as the
changes because of temperature, then thermal dilational surface viscoelasticity. This property
Marangoni instability takes place. Very thorough means that the contraction or expansion of
consideration of thermally driven instabilities is adsorbed layer gives rise to counteracting forces
given in Colinet et al. (2001) and Birikh et al. depending also on diffusion exchange of
(2003). The review of studies on solutal molecules with the adjacent bulk (see also
Marangoni instabilities presenting also criteria ▶ Adsorption). The viscoelastic response of
enabling prediction of instabilities is given in adsorbed layer is one of the key features in the
Birikh et al. (2003) and Kovalchuk and Vollhardt mechanism of repeated auto-oscillation of inter-
(2006a). Simultaneous action of buoyancy, facial tension considered below.
heat, and solute transfer is considered in The regular repeated auto-oscillations of
Nepomnyashchy et al. (2002). interfacial tension have been at first observed at
Sometimes, instability can be a desirable water/air interface by dissolution of a droplet of
event, as it can, for example, enhance consider- a surfactant in water bulk. Usually, during cer-
ably the mass transfer rate in extraction pro- tain induction period after the drop is formed at
cesses. Very often, however, it has to be the capillary tip, the surface tension remains
avoided because it can destabilize foams and constant and equal to that of pure water. After
emulsions or can lead to mistakes in measure- that, the oscillations of surface tension begin
ments of interfacial tension (see also ▶ Interfa- which are characterized by a sharp decrease of
cial Dynamics Methods). The instability and surface tension followed by its gradual increase.
related self-organization processes are the It is important to note that these oscillations
impressive examples of direct transformation of cannot be explained in the general frameworks
chemical energy in mechanical motion and, in of linear theory of Marangoni instability,
this way, can be used in microfluidic devices. because the criteria based on linear stability
Therefore, understanding the mechanisms of analysis predict only steady convection for
self-organization processes enabling the predic- transfer of surfactant from water bulk to water/
tion on how they will develop in real systems is air interface.
not only of fundamental but also of great practical The performed theoretical analysis based on
importance. direct numerical simulations allows to conclude
It should be stressed that the bulk concentra- that Marangoni instability in this system starts to
tion gradients of surfactant in perpendicular to develop as a steady convection, in full accor-
the interface direction are of key importance for dance to the criteria available, but terminates
the instability development because the surfac- later because of the compression of adsorbed
tant flux from the bulk and its adsorption at the layer near the vessel walls and the consequent
interface are crucial for formation of interfacial appearance of interfacial tension gradient
concentration gradients driving convection. suppressing instability. It was established later
Instability sets up only if the concentration that the same mechanism of periodic instability
gradient in the phase supplying the surfactant can be responsible for other oscillatory phenom-
exceeds a threshold value. ena, namely, the oscillations through a liquid
The above-mentioned criteria predicting the membrane (Arai and Kusu 2001) and the
instability onset are obtained as a rule by linear so-called chemical oscillations in a two-phase
stability analysis, that is, they relate to the initial system, where at least two solutes, one of them
stage of instability development. When the flow being a surfactant, are transferred in the opposite
velocity increases considerably, the nonlinear directions (Zieper and Oertel 1982). These phe-
effects become important. Then the development nomena, however, have been observed for ionic
of solutal Marangoni instability can reveal some surfactants interacting with other ionic solutes,
specific features in comparison to the thermal whereas experimental and theoretical studies on
auto-oscillation with dissolving drops were
performed mainly using nonionic surfactants.
The only several experimental investigations on
ionic surfactants are available. Therefore, as
a future work, the experimental and theoretical
study on auto-oscillation produced by ionic
surfactants and effect of other solutes on their
appearance and characteristic is desired.
The characteristics of auto-oscillation at air/
water or oil/water interface as well as oscillation
through liquid membrane depend essentially on
the solutes used. Therefore, it has been even
proposed to use the oscillation through the liquid
membrane for taste and drug sensors (Arai and
Kusu 2001). Obviously, to do that, comprehen-
sive understanding of oscillation mechanism and
possibilities to control them is necessary. Below,
the results of theoretical and experimental study Surfactant Adsorption Under Nonequilibrium,
of auto-oscillation are discussed in detail. Fig. 2 Geometry of the system used in mathematical
modeling on auto-oscillation: 1 – cylindrical vessel, 2,
3 – immiscible liquids, 4 – air, 5 – capillary filled with
surfactant, 6 – surfactant droplet, 7 – convective
Theory streamlines
The mathematical model used in numerical sim- rewritten in terms of vorticity and stream func-
ulations on auto-oscillation at air/water and oil/ tion. Dimensionless form of equations is
water interface reflects the main features of the
real experimental systems, where this phenome- @oi @ ðvri oi Þ @ ðvzi oi Þ Di
þ þ Sci
non has been observed, and is as follows (Fig. 2) @t @r @z D1
(Kovalchuk and Vollhardt 2005a). Two immisci- 2
@ oi @ 2 oi 1 @oi oi
ble, viscous, incompressible Newtonian liquids þ þ (1)
@r 2 @z2 r @r r2
are placed in a cylindrical vessel. A cylindrical
D2 c01 @ci
capillary is immersed into the liquids with its axis þ Rai Sci i2 ¼0
coinciding with the vessel axis, and spherical D1 c0i @r
droplet of a scarcely soluble surfactant is formed
at the tip of the capillary. All thermal effects are @ 2 Ci @ 2 Ci 1 @Ci
þ oi r ¼ 0 (2) S
neglected, and temperature is assumed to be @r 2 @z2 r @r
constant.
@ci @ ðvri ci Þ @ ðvzi ci Þ vri ci
The time evolution of the system under þ þ þ
@t @r @z r
consideration is described by Navier-Stokes, 2 (3)
continuity and convective diffusion equations. @ ci @ 2 ci 1 @ci
þ 2 þ ¼0
Boussinesq approximation is used, that is, @r 2 @z r @r
dependence of solution density on concentra-
tion of surfactant is taken into account only in where i ¼ 1 corresponds to the lower liquid
the buoyancy term. Independency of all vari- phase, where the droplet is situated; i ¼ 2
ables of angular coordinate is assumed due corresponds to another liquid phase;
@vri @vzi
to system symmetry and experimentally oi ¼ @z @r is the vorticity, scaled by D1/L2;
performed visualization of surface flow. In this t is the time, scaled by L2/D1; r is the radial
case, the system of governing equations can be coordinate and z is the normal to the interface
coordinate downward directed with z ¼ 0 on the surface and bulk (NE ! 0 corresponds to the
interface, both scaled by L; vri and vzi are the limiting case of insoluble monolayer when the
velocity components in radial and normal to the exchange with bulk is absent), KL and Gm are
interface directions respectively, both scaled by the parameters of the Langmuir isotherm related
D1/L; Di is the surfactant diffusion coefficient in to the liquid 1, DS is the surface diffusion coeffi-
phase i; L is the characteristic length, chosen here cient, and G is the Gibbs adsorption, scaled by
as the depth of the liquid layer 1; ci is the surfac- c01 KL Gm .
tant concentration in phase i scaled on surfactant The local equilibrium of the surfactant distri-
solubility in phase 1; Ci is the stream function, bution between the liquid phases at the interface
defined so that vri ¼ 1r @C 1 @Ci
@z , vzi ¼ r @r and
i
is assumed due to partition coefficient:
scaled by LD; and Sci ¼ Di is the Schmidt num-
ni
ber, Rai ¼ rgc0iniLDi @r

3
@ci is the Rayleigh number, ni is

i c2
0i ¼ kp at z ¼ 0 (6)
the kinematic viscosity of the i-th phase, c1

ri ¼ r0i cc0ii ðr0i rsi Þ is the solution
The first boundary condition for the vorticity
density, r0i is the density of pure i-th phase, rsi on liquid/liquid interface is obtained from the
is the density of the saturated surfactant solution tangential stress balance employing the
in the i-th phase, c0i is the surfactant solubility in Szyszkowski-Langmuir equation of state for
phase i, and g is the acceleration due to gravity. adsorbed surfactant:
Both liquids are supposed to be motionless in
the initial state, and the dimensionless surfactant r02 n2 1 @G
concentration is equal to unity at the droplet/ o1 ¼ o2 þ Ma at z ¼ 0
r01 n1 ð1 KL c0 GÞ @r
solvent interface and is equal to zero elsewhere.
(7)
No-slip boundary conditions are used on the
vessel walls, capillary, and the droplet surface. where Ma ¼ RTcr 01nK1LDG1m L is the Marangoni number,
The liquid/liquid interface is supposed 01
R is the universal gas constant, and T is the
nondeformable. The intrinsic surface viscosity absolute temperature. The second boundary
is neglected. The diffusion-controlled adsorption condition is the continuity of the tangential
kinetics is assumed with local equilibrium velocity component at the interface:
between the surface concentration and the
sublayer concentration described by the vr1 ¼ vr2 at z ¼ 0 (8)
Langmuir isotherm (see also ▶ Adsorption),
which in dimensionless form is The numerical simulations performed for var-
c1 ious sets of system parameters have shown that
G¼ (4) although surfactant transfer into the second liquid
1 þ KL c01 c1
phase and buoyancy force have a considerable
The equation for the surfactant mass balance impact on the oscillation appearance and charac-
on the interface is teristics, they do not change the general features
of oscillation mechanism, that is, theoretically

@G @ ðGvr Þ Gvr DS @ 2 G 1 @G oscillation can be observed in the systems where
þ þ þ
@t @r r D1 @r 2 r @r buoyancy is negligible and the surfactant is
insoluble in the second liquid phase. Therefore,
@c1 D2 @c2
NE for the sake of simplicity, the oscillation mecha-
@z D1 @z
nism will be discussed first for the case when
¼ 0 at z ¼ 0 (5) the droplet is dissolved under liquid/air interface,
and the density of solution is independent
Here, NE ¼ KLLGm is the exchange number of surfactant concentration as it was analyzed
related to surfactant exchange between the in Kovalchuk and Vollhardt (2002) and
Under Nonequilibrium,
Fig. 3 Development of
instability: time dependence
of the velocity growth rate.
1 – Ma ¼ 5.4 1010,
2 – Ma ¼ 5.4 109,
3 – Ma ¼ 5.4 108,
4 – Ma ¼ 5.4 107
(Taken from Kovalchuk and
Vollhardt (2004))
Kovalchuk and Vollhardt (2004). Then the effect increasing velocity, the surfactant transfer due
of buoyancy, partition, and the initial presence of to convection becomes comparable with that
other surfactants will be discussed. due to diffusion; the velocity growth rate begins
In the absence of buoyancy, initially, only to increase (Fig. 3), and after that, Marangoni
diffusion transfer of surfactant from the droplet instability develops very quickly. It is seen from
takes place in the system. After some noticeable Fig. 3 that instability development is determined
amount of surfactant reaches the surface, convec- by the Marangoni number. The larger the
tion begins to develop due to gradient of surface Marangoni number, the faster the instability
tension. It is obvious that such gradient should develops.
gradually appear because the concentration in the During the instability development, velocity
capillary region is higher than in the wall region increases by several orders of magnitude in a very
as the last is on a larger distance from the drop. It short time, much shorter than induction period
is clear from the streamlines, distribution shown and oscillation period. It brings a large amount
in Fig. 2, that there is a feedback between the of surfactant from the droplet to the surface and
velocity growth and the increase of surface spreads it over the surface. This results in an
tension gradient, which can result in instability abrupt decrease of surface tension. The stages of S
development: convective flux brings more con- surfactant transfer and spreading during the insta-
centrated solution to the capillary region and bility development are shown in Fig. 4. The
causes in this way the increase of surfactant con- typical time dependencies of surface tension and
centration gradient. Initially, before the surface velocity for two different capillary
instability onset, the convection is very slow, immersion depths are presented in Figs. 5 and 6
and its contribution into the surfactant transfer demonstrating that the abrupt decrease of surface
is negligible. At this period of time, the growth tension corresponds to the sharp increase of
rate of velocity is rather small and even decreases convective motion due to instability develop-
with time (Fig. 3), and the change of surface ment, whereas the gradual increase of surface
tension remains undetectable. This period of tension corresponds to rather slow stage, where
time with almost constant surface tension is the the velocity is close to zero.
induction period characterized by slow diffusion The comparison of Figs. 5 and 6 displays
surfactant transfer from the drop. However, with clearly that the change in the capillary immersion
Surfactant Adsorption Under Nonequilibrium, Fig. 4 Concentration distribution during the instability develop-
ment: (a) in the time moment t, (b) t + 10 s, (c) t + 20 s (Redrawn from Kovalchuk and Vollhardt 2002)
depth, while keeping all other parameters con- immersion depths, this decrease slows down, and
stant, results in the qualitative change of the sys- velocity is stabilized at certain relatively large
tem behavior. At small capillary immersion depth value, such that convective transfer of surfactant
4 mm, the system demonstrates an example of the remains to be dominating. At large immersion
regime of steady convection with monotonous depths, velocity falls to zero and even goes to
decrease of surface tension, predicted by linear small negative values; thereafter, the dominant
stability analysis. At larger capillary immersion mechanism of surfactant transfer becomes again
depth, the same system demonstrates oscillatory diffusion. Thus, in both cases, a force appears
behavior. The key point for understanding this counteracting the instability development, but in
difference is the fact that in both cases, the surface the case of small capillary immersion depths, it is
velocity begins to decrease after certain time, but it not powerful enough to return the system back to
decreases up to different values. At small capillary the stage of diffusion transfer.
Fig. 5 Results of
numerical simulation for
capillary immersion depth
8 mm: (a)surface tension,
(b) surface velocity
(Redrawn from Kovalchuk
and Vollhardt 2002)
S
To understand the nature of this force, it is the time when velocity is around its maximum. At
instructive to consider in more detail how the this time, practically the whole surface is already
surfactant spreads over the surface. The surface involved in convective motion. As the velocity on
velocity distribution in radial direction changes the capillary and vessel walls is equal to zero, large
with time. Initially, the maximum of surface velocity gradients appear, stretching the surface in
velocity is near the capillary, and velocity the capillary region and compressing it in the wall
decreases very quickly with distance from the region. As a result, the surface concentration
capillary. Later, during instability development, decreases in the capillary region (in comparison
more and more distant parts of the interface to that without stretching) and increases in the wall
become involved in convective motion. In other region. The latter causes appearance of the reverse
words, the convective wave spreading the gradient of surface concentration (adsorption) in
surfactant moves from the capillary to the wall. the wall region counteracting the surface convec-
Figure 7a shows the radial velocity distribution at tion as shown in Fig. 7b.
Fig. 6 Results of
numerical simulation for
capillary immersion depth
4 mm: (a) surface tension,
(b) surface velocity
(Redrawn from Kovalchuk
and Vollhardt 2002)
The further system behavior depends on reverse convective roll is shown in Fig. 8a. It
the balance between the surface velocity at should be stressed that magnitude of the velocity
the moment when the surface wave reaches the in this reverse roll is very low, that is, this part of
vessel wall and the strength of the formed reverse the surface can be rather considered as
concentration gradient. If the reverse concentra- a retardation zone. This retardation zone extends
tion gradient is not sufficient to suppress the then over almost the whole surface (Fig. 8b) and
velocity completely, then the situation presented breaks up the convective surfactant transfer from
in Fig. 6 occurs, characterized by regime of the droplet to the surface. The system returns to
steady convection. If reverse concentration gra- the slow stage with predominantly diffusion sur-
dient is large enough, then the velocity decreases, factant transfer. Then, formation of surfactant
first of all near the wall, and liquid near the wall concentration gradients occurs again as it was
even starts to move in the opposite direction – already described for the system evolution on its
from the wall to the capillary. Formation of the initial state. Gradually, as convection develops,
Fig. 7 To the mechanism
of instability termination:
(a) distribution of the
surface velocity: 1 – in the
time moment t, 2 – t + 1 s,
3 – t + 3 s, 4 – t + 4 s; (b)
distribution of surface
concentration: 2 – t + 1 s,
3–t+3s
S
the reverse convective roll shrinks (Fig. 8c) and capillary immersion depths to cell radius
eventually disappears. After that, convective (aspect ratio) corresponding to transition between
velocity grows very quickly again resulting in two regimes. The oscillation characteristics
an abrupt decrease of surface tension, that is, the depend on the geometry as well. The decrease
next oscillation starts. of the capillary immersion depth and/or the
The regime of instability development can be increase of the vessel radius leads to the decrease
controlled first of all by changing the system of oscillation period and amplitude. The droplet
geometry. The transition from steady convection size influences mainly the oscillation amplitude –
to oscillation occurs when the capillary immer- the amplitude increases with the increase of the
sion depth increases while the cell radius is kept droplet size.
constant or when the cell radius increases by By fixed system geometry, the surfactant
constant capillary immersion depth. In other properties can be used to control the oscillation
words, there is a threshold value of the ratio of characteristics (Kovalchuk and Vollhardt 2004).
Surfactant Adsorption Under Nonequilibrium, Fig. 8 Streamlines distribution in the liquid bulk during the
instability termination and diffusion stage (for details, see the text) (Taken from Kovalchuk and Vollhardt 2004)
The increase of the surfactant activity, into play the buoyancy convection (Kovalchuk
represented by the parameter a ¼ KLGm, results and Vollhardt 2005b). The buoyancy
in the increase of both oscillation period and contribution is described by the last term in
amplitude. The threshold value of the aspect Eq. 1. Buoyancy convection accelerates all pro-
ratio dividing oscillatory regime and regime of cesses in the system usually producing much
stable convection and oscillation amplitude higher contribution in the surfactant transfer
increases by the increase of surfactant solubility, than diffusion. Noteworthy that buoyancy takes
whereas period decreases. Increase of the part also in instability termination due to
diffusion coefficient causes decrease of both formation of reverse concentration gradient in
oscillation period and amplitude. the bulk phase.
Considerable effect on the phenomenon of If the density of the solution is smaller than the
oscillations has the dependency of the solution density of the solvent, then an increase in buoy-
density on surfactant concentration, as it brings ancy force results in a considerable decrease of
the induction period as well as oscillation period.
The presence of buoyancy force causes also an
increase of the critical ratio of the capillary
immersion depth to the vessel radius
corresponding to the transition from the regime
of steady convection to repeated oscillations.
When the surfactant not only adsorbs at the
interface but also dissolves into the second liquid
phase, then an additional acting parameter
appears – the partition coefficient, Eq. 6. An
increase in the partition coefficient results in
a decrease of the oscillation period and especially
in a decrease of the oscillation amplitude
Surfactant Adsorption Under Nonequilibrium,
(Kovalchuk and Vollhardt 2005a). At large
Fig. 9 Combined regime (Taken from Grigorieva et al.
enough partition coefficients, the oscillation and 2004)
development of instability itself become
nondetectable.
As according to the proposed mechanism, the The mechanism of auto-oscillations consid-
compression of adsorbed surfactant near the ves- ered above is applicable also to other similar
sel wall is the main reason of instability termina- phenomena which attract the scientific interest
tion; it can be expected that the initial presence of during the last several decades. Those phenom-
a surfactant in the system (besides the surfactant ena are oscillations in liquid membranes (Arai
dissolving from the drop) should be in favor of and Kusu 2001), and spontaneous chemically
oscillations instead of steady convection. Indeed, driven oscillations at water/oil interface first
the numerical simulations have shown that the investigated experimentally by Dupeyrat and
presence of some dissolved in bulk or spread on Nakache (Zieper and Oertel 1982). Such phe-
the surface surfactant results in a decrease of the nomena are often called chemical or electrical
threshold aspect ratio corresponding to the tran- oscillations. A review of experimental and theo-
sition from the regime of steady convection to retical studies on these systems is given in
repeated oscillations. By sufficiently high initial Kovalchuk and Vollhardt (2006a).
surfactant concentration, the force counteracting The liquid membrane oscillator (Fig. 10) con-
the surface motion becomes so large that it can sists of three liquid phases: (1) aqueous donor
prevent development of instability due to phase D containing an ionic surfactant and often
complete retardation of the surface. also a short-chain alcohol; (2) oil membrane
Numerical simulations have also predicted phase M (e.g., nitrobenzene, dichloromethane, S
that a combined regime shown in Fig. 9 is possi- nitromethane, octanol) containing a counter-
ble under certain conditions (Grigorieva at al. solute, a nonsurface-active ionic substance,
2003 and 2004). This regime can be explained which can react at the interface with the surfac-
as follows. When the initial surfactant concen- tant dissolved in donor phase; and (3) acceptor
tration is still smaller than that necessary to phase being pure water or aqueous solution of
support oscillatory regime at a given aspect nonsurface-active substances such as NaCl,
ratio, then steady convection begins first. It sup- HCl, and sucrose. Typical pairs of surfactant/
plies additional amount of surfactant to the sur- counter-solute are alkyltrimethylammonium
face, and, at a certain moment in time, the halogenide as a surfactant and picric acid as
amount of surfactant becomes already sufficient a counter-solute or sodium dodecyl sulfate/
to retard the surface and bring the system to the tetraethylammonium halogenide. The liquid
slow (diffusion) stage after which repeated membrane system can be placed in a U-shaped
oscillations begin. tube (Fig. 10a, b depending on the oil phase
Surfactant Adsorption a b c
Fig. 10 Schemes of M
various configurations of
D A 1 2 A D A
liquid membrane
oscillators 1 1
1 2 D A 2 2
M M
density), or it can be constructed from well as oscillation mechanism is not completely

a cylindrical open tube inserted in a cylindrical clear. It is however proved that the oscillations
vessel (Fig. 10c). are a result of periodically occurring very fast
It is important that, initially, this system is far spontaneous adsorption of the surfactant
out of partition equilibrium, that is, the surfactant followed by its gradual desorption or replacement
is well soluble in membrane phase, whereas by less surface-active substance. It has been
counter-solute is well soluble in water. Therefore, shown in several studies reviewed in Kovalchuk
the transfer of at least two species starts at once and Vollhardt (2006a) that electrical potential
through the donor/membrane interface after for- oscillations in this system are fully synchronized
mation of the liquid membrane system. More- with oscillations of interfacial tension.
over, a chemical reaction between surfactant The oscillation characteristics depend on all spe-
and counter-solute at the interface produces cies involved, including those initially dissolved
another surfactant with rather different surface in acceptor phase. That is why the liquid mem-
activity. The counter-solute is transferred also brane system has been considered as a good can-
through the membrane/acceptor interface to the didate for artificial sensor system which could
acceptor phase. Surfactant is transferred through distinguish between tastes or drugs (Arai and
the bulk membrane phase to the membrane/ Kusu 2001).
acceptor interface, adsorbs on this interface, and It is clear from Fig. 10 that from the geomet-
is transferred next into the aqueous acceptor rical point of view, the conditions at the mem-
phase. These equilibration processes result in brane/acceptor interface are rather similar to
the changes of adsorption and interfacial tension those at the air/water or oil/water interface
at both donor/membrane and membrane/acceptor where the auto-oscillations are produced due to
interfaces and, because the transferred species are diffusion from a point source, which were
ionic, electrical potentials through both discussed above. Indeed, in point 1 at the mem-
interfaces. brane/acceptor interface, more surfactant should
Usually, the measured quantities in liquid be adsorbed as the path from the donor/mem-
membrane oscillators are electric potential differ- brane interface to point 1 is shorter than to point
ences between the liquid phases: acceptor/donor, 2. Therefore, under the conditions when at the
acceptor/membrane, and membrane/donor. Mea- donor/membrane interface, no instability occurs,
surements start at once after three liquid phases and initially, the surfactant is transferred due to
are brought into contact. After a certain induction diffusion; the Marangoni convection should
period, spontaneous oscillations of electrical inevitably develop at the membrane/acceptor
potential start between donor and acceptor interface directed from point 1 to point 2. The
phase, similar in shape to those presented in convection brings to point 1 a solution more rich
Fig. 5. It is interesting that in the systems which on the surfactant; therefore, there is a feedback in
are chemically similar, oscillations can be the liquid membrane systems, very similar to that
detected either at donor/membrane or mem- considered above for the systems with a droplet
brane/acceptor interface. The reasons of that as dissolving under air/liquid or liquid/liquid
interface. Therefore, one can expect that the con- of numerical simulations have shown that oscil-
sidered earlier mechanism could be useful also lations according to the considered mechanism
for explanation of oscillations through the liquid appear in this system by such direction of the
membrane. surfactant transfer where steady convection is
The numerical simulations performed with expected according to the linear stability analysis
a mathematical model similar to described (Tadmouri et al. 2010). It should be noted that for
above but adapted to the geometry of the liquid this case, the oscillations obtained in numerical
membrane system, presented in Fig. 10c, have simulations are smoother than those observed
shown that indeed, if even only one surfactant is experimentally, probably because in this case,
transferred from the donor phase through the already three-dimensional flow patterns should
liquid membrane to the acceptor phase, the spon- be considered.
taneous oscillations can occur at the membrane/ Thus, the theoretical results on auto-
acceptor interface in full accordance to the oscillations can be summarized as follows. In
described above mechanism (Kovalchuk and the systems far from equilibrium, where surfac-
Vollhardt 2006b). tant transfer takes place to or through liquid inter-
The governing mechanism and the conditions face, self-organization processes often occur due
for oscillations at donor/membrane interface are to development of Marangoni instability. There
obviously the same as for chemically driven are criteria based on linear stability analysis
oscillations at water/oil interface mentioned enabling the prediction of the potential system
above (Zieper and Oertel 1982). Namely, the stability or instability and the regimes of instabil-
two-phase system where these oscillations have ity, steady convection, or oscillation. However,
been observed corresponds to donor/membrane the way of instability development depends on
part of a liquid membrane oscillator. It also con- nonlinear effects. Quick local contractions and
sists of an aqueous solution of ionic surfactant expansions of surfactant monolayers can result
and oil phase containing a counter-solute, both in considerable resistance forces terminating
solutes being out of partition equilibrium. Chem- instability development. Taking into account
ical reaction between the solutes near the inter- these forces allows to explain the development
face was considered to be playing a key role in of repeated spontaneous oscillations in several
establishing of these oscillations, although the systems where only steady convection is
direct mechanism remained unclear. The situa- predicted by linear stability analysis. Those are
tion became even more intriguing when it has such systems as air/liquid or liquid/liquid systems
been shown that the oscillations can be observed with a point source of surfactant under free inter-
in a very similar system, aqueous solution of face, producing auto-oscillations of interfacial
hexadecyl trimethyl ammonium bromide in tension, liquid membrane oscillators, and two-
contact with dichloromethane, but without any phase systems producing spontaneous chemically S
counter-solute and, therefore, without chemical driven oscillations at water/oil interface. Below,
reaction (Lavabre et al. 2005). the experimental data are presented proving the
It was supposed that the described above theoretical results.
mechanism could contribute also to the explana-
tion of these oscillations. To check that, the math-
ematical model was adapted to the geometry of Methods
two-phase system. As this geometry always gives
initially homogeneous distribution of a surfactant The auto-oscillation experiments in the system
over the interface, an artificial fluctuation of the with a surfactant droplet, considered here, have
surface concentration was introduced in the been performed in the geometry presented in
model, being many orders of magnitude smaller Fig. 2. The vessels used were from 40 to 75 mm
than the surface concentration itself, but still in diameter. The interfacial tension was measured
higher than the computational error. The results by electronic balance equipped with freshly
annealed platinum Wilhelmy plate or a cylinder about 500 times higher than that of dodecane.
made of high-density polyethylene. All surfac- After the droplet was formed, the buoyancy-
tants were of high purity and were used as pur- driven convection began in the system, transfer-
chased. Alkanes were checked for purity by ring the surfactant to the interface. The intensity
measuring alkane-water interfacial tension and, of surfactant transfer decreases in the series of
when necessary, were purified using freshly aliphatic alcohols from pentanol to nonanol
annealed powder of Al2O3. Ultrapure deionized rather noticeably. In the case of pentanol droplet,
water was produced by Purelab Plus. the transfer was so quick that oscillations begun
In another set of experiments, the solution of without any induction period, just during the
an ionic surfactant was gradually injected by droplet formation, whereas for octanol droplet
syringe in aqueous bulk at a prescribed distance under water/heptane interface, the velocity of
from the interface (Ikezoe et al. 2004). In these buoyancy-driven convection was comparable
experiments, the time-resolved quasi-elastic laser with the velocity of thermo convection, what
scattering (QELS) method was used to measure caused a poor reproducibility of the results at
interfacial tension. large capillary immersion depths (above 8 mm).
After the surfactant (colored with dye) reached
the interface, very rapid surface motion
Experimental Results proceeded, directed from the capillary to the ves-
sel walls. It was observed during about 1 s, and
Auto-oscillations generated by dissolution of after that, the motion has been terminated sud-
a surfactant droplet under liquid interface were denly. During the next period of time, there was
obtained with such surfactants as diethyl phthal- no detectable motion at all; even buoyancy-
ate, middle-chain (C5–C9) aliphatic alcohols, and driven convection was practically stopped
middle-chain fatty acids at water/air, water/ because of the mixing of solution in the bulk
heptane, and water/dodecane interfaces. These and concentration equilibration. Thereafter,
surfactants were chosen because they all have buoyancy surfactant transfer from the droplet
a limited solubility in water and form stable drop- started again beginning the next cycle of oscilla-
lets in aqueous phase which dissolve during tion. In their main features, these visualization
rather a long time. So, in oscillation experiments, experiments agree very well with the oscillation
a pentanol droplet with radius of about 1 mm was mechanism proposed above.
fully dissolved during about 2 h, whereas for An example of experimental data on auto-
octanol droplet, the full dissolution took more oscillation is presented in Fig. 11. Buoyancy
than 70 h. force and solubility are larger for hexanol in
To validate the mechanism of auto- comparison to octanol, whereas activity is
oscillations described above, the visualization smaller. All those factors according to the theory
experiments have been performed by using work in the same direction causing smaller oscil-
talcum powder added to the bulk of water as lation period for hexanol, what is confirmed
well as placed onto the surface or by adding experimentally in Fig. 11. Partition coefficient
a dye to the surfactant droplet. The visualization of hexanol between heptane and water is about
has shown that before the droplet was formed, 2, whereas partition coefficient for octanol is
weak thermal convection took place in the sys- about 40. Therefore, the theory predicts much
tem, despite of all precautions made to avoid it. more considerable decrease of oscillation ampli-
The intensity of convection depended on the liq- tude for octanol than for hexanol by transition
uid forming the upper layer being in contact with from water/air to water/heptane interface, what
air. Convection was negligible for the water/ is also confirmed experimentally (Fig. 11).
dodecane system (dodecane in contact with air) Another example showing the effect of surfac-
and was rather noticeable for water/heptane tant properties on oscillation characteristics is
system, as the vapor pressure of heptane is presented in Fig. 12. It is well known that the
Surfactant Adsorption Under Nonequilibrium, (b) octanol. Vessel diameter d ¼ 40 mm, capillary immer-
Fig. 11 Auto-oscillation of interfacial tension at water/ sion depth h ¼ 12 mm for hexanol, h ¼ 8 mm for octanol
air (1) and water/heptane (2) interface: (a) hexanol, (Taken from Kovalchuk and Vollhardt 2005a)
dissociated form of fatty acids is several orders of is a working parameter determining the transition
magnitude less surface active than from oscillations to steady convection, as it is
nondissociated form. The degree of dissociation seen from Fig. 13 displaying the proportionality
increases with the increase of pH. Therefore, by of the threshold value of capillary immersion
the increase of pH, the amount of active depth to the vessel diameter. Also, in full agree-
nondissociated acid dissolved from the droplet ment with the theory, it was found that the critical
decreases, what, according to the theory, should aspect ratio decreases with the increase of
result in an essential decrease of oscillation surfactant chain length. S
amplitude. The experimental confirmation of Experimental confirmation of theoretical pre-
this prediction was received recently in dictions was also obtained for the effect of initial
Kovalchuk and Vollhardt (2010) for octanoic presence of surfactant in the system for either
and nonanoic acids. The results for octanoic dissolved in bulk or spread on the interface sur-
acid are presented in Fig. 12. factant. Figure 14 presents results obtained by
Not only oscillatory regime but also regime of dissolution of heptanol droplet under water/air
steady convection with single oscillation interface in the presence of spread monolayer of
followed by continuous decrease of interfacial L-a-dimyristoylphosphatidylcholine (DMPC).
tension was experimentally detected for series These results demonstrate the possibility to con-
of aliphatic alcohols and fatty acids. Comprehen- trol the dynamic regimes accompanying the
sive experimental study on the effect of geometry transfer of surfactant from the bulk to the inter-
(Grigorieva et al. 2004) has shown that the ratio face by changing the initial surface coverage
of the capillary immersion depth to the cell radius y ¼ GGm (here, G stays for dimensional adsorption)
Fig. 12 Dependence of
oscillation characteristics
produced at water/air
interface by dissolution of
octanoic acid on pH:
1 – pH ¼ 5.15,
2 – pH ¼ 5.8, 3 – pH ¼ 6.2,
4 –pH ¼ 6.5 (Redrawn from
Kovalchuk and
Vollhardt 2010)
of spread monolayer. The results presented in

Fig. 14 are obtained for the vessel of diameter
75 mm. It follows from Fig. 13 that steady con-
vection is expected by dissolution of heptanol
droplet at immersion depths smaller than
13 mm. In the case considered here, the capillary
immersion depth is much smaller – 1 mm, and, as
expected, steady convection after the first oscil-
lation takes place in absence of spread surfactant
on the interface (Fig. 14, curve 1). Even small
amount of spread surfactant, y ¼ 0.038, causes
appearance of a combined regime (curve 2)
predicted theoretically (cf. Fig. 9). The repeated
Surfactant Adsorption Under Nonequilibrium,
oscillations appear in this case only after a very
Fig. 13 Dependence of the threshold value of capillary
long time, more than 30 h. By the increase of immersion depth corresponding to transition from the
surface coverage to y ¼ 0.118, oscillations appear regime of steady convection to oscillation on the vessel
already after 3 h (curve 3). The further increase of diameter. Heptanol droplet under water/air interface
(Redrawn from Grigorieva et al. 2004)
surface coverage results in transition from the
combined to pure oscillatory regime, as it was
predicted by the theory (curves 4 and 5). At least
by surface coverage more than 50 %, the interface from a small source situated in the liquid bulk
becomes fully retarded, and instability does not (see Ikezoe et al. 2004 and references herein). In
develop any more (curve 6). The decrease of the this case, the surfactant solution of desired con-
surface tension in this case is mainly due to centration is injected with a constant rate from the
diffusion transfer of the surfactant. tip of a syringe placed at certain distance from
There is another protocol to study auto- the interface. The advantage of this protocol is
oscillations produced by a surfactant transfer the possibility to study any surfactants, but not
Fig. 14 Surface tension
vs. time for heptanol
droplet dissolving under
water/air interface in
presence of spread DMPC
layer. Capillary immersion
depth h ¼ 1 mm, vessel
diameter 75 mm, surface
coverage 1 – y ¼ 0,
2 – y ¼ 0.038,
3 – y ¼ 0.118,
4 – y ¼ 0.313,
5 – y ¼ 0.382, 6 – y ¼ 0.522
(Redrawn from Grigorieva
et al. 2004)
only those enabling the droplet formation. In by convective motion, which was terminated
particular with this protocol, the oscillations gen- quickly and was not observed during the stage S
erated at water/nitrobenzene interface by injec- of gradual increase of interfacial tension.
tion of an anionic surfactant sodium dodecyl Theoretical analysis shows that the mecha-
sulfate (SDS) have been studied (Ikezoe et al. nism proposed for auto-oscillations in systems
2004). The results of this experimental study are with point source can be at least in a part actual
also in line with the theoretical results presented also for liquid membrane oscillators and two-
above. For example, the increase of the injection phase systems with chemical oscillations. To
rate, which can be considered as an increase of prove it, two model systems were investigated,
the surfactant solubility in the case of drop by where any chemical reactions and electrical
keeping unchanged its other properties, resulted charge effects are excluded as only one simple
(by the same position of syringe) in transition surfactant, being nonionic one, was used in these
from oscillatory regime to the regime of steady systems. The first model system was the two-
convection. In the oscillatory regime, each sharp phase system composed of an aqueous solution
decrease of interfacial tension was accompanied of oxyethylated alcohol, tetraethylene glycol
monooctyl ether (C8EO4), in contact with References

heptane. Oscillations were observed in a certain
range of concentrations of C8EO4 (Tadmouri Arai K, Kusu F (2001) Electrical potential oscillations
across a water-oil-water liquid membrane in the pres-
et al. 2010). The experimental results were
ence of drugs. In: Volkov AG (ed) Liquid interfaces in
discussed in the frameworks of the mechanism chemical, biological and pharmaceutical applications,
described above. vol 95, Surfactant science series. Marcel Dekker,
The second model system was the liquid mem- New York, pp 699–727
Birikh RV, Briskman VA, Velarge MG, Legros JC
brane system composed as follows (Kovalchuk
(2003) Liquid interfacial systems. Oscillations and
and Vollhardt 2006b). In the scheme presented in instability, vol 113, Surfactant science series. Marcel
Fig. 10c, pure water was used as a membrane Dekker, New York
phase, pure aliphatic alcohol (C5–C7) was used Colinet P, Legros JC, Velarde MG (2001) Nonlinear
dynamics of surface-tension-driven instabilities.
as a donor phase, and pure dodecane was used as
Willey-VCH, Berlin
an acceptor phase. The interfacial tension was Grigorieva OV, Kovalchuk NM, Grigoriev DO, Vollhatdt D
measured on the water/dodecane interface to (2003) Experimental studies on the geometrical charac-
detect oscillations. Oscillations were detected teristics determining the system behaviour of surface
tension autooscillations. J Coll Int Sci 261:490–497
for all alcohols used, and the oscillation period
Grigorieva OV, Kovalchuk NM, Grigoriev DO, Vollhatdt D
increased with the increase of the chain length as (2004) Spontaneous non-linear surface tension oscilla-
was expected. tions in the presence of a spread surfactant monolayer at
Thus, the results of experimental studies on the air/water interface. Coll Surf A 250:141–151
Ikezoe Y, Ishizaki S, Takahashi T, Yui H, Fujinami M,
auto-oscillations of interfacial tension are in good
Sawada T (2004) Hydrodynamically induced chemical
agreement with the predictions made in the oscillation at a water/nitrobenzene interface. J Colloid
frameworks of the theoretical approach based Interface Sci 275:298–304
on direct numerical simulations using equations Kovalchuk NM, Vollhardt D (2002) Theoretical descrip-
tion of repeated surface tension auto-oscillations. Phys
of physicochemical hydrodynamics. The model
Rev E 66:026302
experiments on other systems, where oscillation Kovalchuk NM, Vollhardt D (2004) Effect of substance
has been observed, demonstrated that the pro- properties on the appearance and characteristics of
posed mechanism can be useful also for explana- repeated surface tension auto-oscillations driven by
Marangoni force. Phys Rev E 69:016307
tion of oscillations through the liquid membranes
Kovalchuk NM, Vollhardt D (2005a) Nonlinear spontane-
and chemical oscillation in two-phase systems. ous oscillations at the liquid/liquid interface produced
Further work should be directed on the explana- by surfactant dissolution in the bulk phase. J Phys
tion of the role of other species involved in these Chem 109:22868–22875
Kovalchuk NM, Vollhardt D (2005b) Effect of buoyancy
phenomena.
on appearance and characteristics of surface tension
repeated auto-oscillations. J Phys Chem B 109:
Acknowledgment The German Research Council DFG, 15037–15047
project SPP 1506 (Mi418/18-1), is kindly acknowledged. Kovalchuk NM, Vollhardt D (2006a) Marangoni instabil-
Financial assistance by the Bundesministerium f€ ur ity and spontaneous non-linear oscillations produced
Bildung, Wissenschaft, Forschung und Technologie at liquid interfaces by surfactant transfer. Adv Colloid
(BMBF), and the Ukrainian Ministry of Education and Interface Sci 120:1–31
Science (common project UKR 10/039) are also gratefully Kovalchuk NM, Vollhardt D (2006b) Surfactant transfer
acknowledged. The work was supported by COST D43 through a liquid membrane: origin of spontaneous
Action. oscillations at the membrane acceptor phase interface.
J Phys Chem B 110:9774–9778
Kovalchuk NM, Vollhardt D (2010) Auto-oscillation of
surface tension: effect of pH on fatty acid systems.
Cross-References Langmuir 26:14624–14627
Lavabre D, Pradines V, Micheau JC, Pimienta V (2005)
▶ Adsorption Periodic Marangoni instability in surfactant (CTAB)
liquid/liquid mass transfer. J Phys Chem
▶ Interfacial Dynamics Methods
B 109:7582–7586
▶ Surfactant Adsorption Nepomnyashchy AA, Velarde MG, Colinet P (2002)
▶ Surfactants Interfacial phenomena and convection, vol 124,
Surfactant Mixtures, Polymer–Surfactant, and Surfactant Multilayer Structures 1145 S
Monographs and surveys in pure and applied adsorption properties. This is the basis of most
Mathematics. Chapman & Hall/CRC, Boca Raton surfactant-based formulations. Nonideal behav-
Schweizer F (ed) (1997) Self-organization of complex
structures. Gordon and Breach Science Publishers, ior is encountered and is determined by the
Amsterdam surfactant and polymer–surfactant interactions.
Sternling CV, Scriven LE (1959) Interfacial turbulence: Interplay between the surface and solution
hydrodynamic instability and the Marangoni effect. behavior plays an important role. The strong
AIChE J 5:514–523
Tadmouri R, Kovalchuk NM, Pimienta V, Vollhardt D, polyelectrolyte–surfactant complexation and
Micheau JC (2010) Transfer of oxyethylated alcohols the strong binding of multivalent counterions
through water/heptanes interface: transition from non- promote multilayer formation at interfaces. This
oscillatory to oscillatory behavior. Colloids Surf can be used to manipulate soft lubrication and to
A 354:134–142
Zieper J, Oertel H (eds) (1982) Convective transport and maximize surface delivery. Neutron reflectivity,
instability phenomena. Braun, Karlsruhe in combination with deuterium labeling, provides
a route to characterizing the fundamental proper-
ties of such mixtures.
Surfactant Aggregation Behavior

Overview
▶ Surfactant Self-Assembly
In the previous entry, the use of neutron reflec-
tivity to investigate surfactant adsorption at the
Surfactant Microheterogeneous air–solution and solid–solution interfaces has
Systems been described. The feature of the technique
which is of particular importance is the use of
▶ Surfactant Self-Assembly deuterium isotopic labeling to manipulate the
neutron refractive index. This can be used to
make the measurement surface specific and
to enable absolute-adsorbed amounts to be deter-
Surfactant Mixtures, Polymer- mined straightforwardly. Furthermore, it enables
Surfactant, and Surfactant access to the details of the structure of the surface
Multilayer Structures Studied by layer. So far, in the previous entry, only the
Neutron Reflectivity application of neutron reflectivity to simple
single-component systems which adsorb as
Jeffery Penfold monolayers has been described. However, the
ISIS Facility, Science & Technology Facilities technique can be applied to more complex
Council, Rutherford Appleton Laboratory, systems, and the ability to deuterium label differ- S
Chilton, Didcot, Oxon, UK ent components makes it an ideal probe of the
adsorption from multicomponent mixtures, such
as surfactant mixtures or polymer–surfactant
Keywords mixtures. It is also by no means limited to the
study of simple monolayer structures at the
Interfacial adsorption; Mixed surfactants; Multi- interface and is an ideal probe for the determina-
layers; Neutron reflectivity; Polymer-surfactant tion of surface self-assembled structures, such as
mixtures the surface multilayers that can occur in
surfactant-based systems. In this entry, the use of
Definition neutron reflectivity to probe the adsorption of sur-
factant mixtures and polymer–surfactant mixtures
Surfactant mixtures and polymer–surfactant and to determine more complex self-assembled
mixtures provide synergistic enhancements in multilayer structures at interfaces is described.
S 1146 Surfactant Mixtures, Polymer–Surfactant, and Surfactant Multilayer Structures
In order to develop a series of coherent themes, the potential applications. They have the ability to
description will be limited to the air–solution inter- provide extreme and persistent wetting and
face, which is the interface that has been most enhanced surface delivery to interfaces and are
extensively studied, but it can be equally well associated with many biological functions, such
applied to the solid–solution interface. as lung surfactants, and soft lubrication, such as
The widespread use of surfactants in the many conditioning. Neutron reflectivity has played
commercial, technological, and domestic appli- a key role in identifying and characterizing such
cations invariably involves mixtures of surfac- structures (Penfold et al. 2007b; Tucker et al.
tants (Scamehorn 1988; Abe and Scamehorn 2010) which can be promoted by the addition of
2005). In such formulations, different types of multivalent counterions, such as Al3+ or Ca2+,
surfactants are blended in order to optimize oligoamines, and polyelectrolytes.
different aspects of performance and to produce
greater flexibility in processing and formulation.
Furthermore, many of the raw surfactant ingredi- Neutron Reflectivity
ents are inherently mixtures of different isomers
and alkyl chain lengths. Understanding the nature The details of the neutron reflectivity method are
of such mixtures in their adsorption patterns and described in the previous entry, and the focus
in their self-assembly is hence of paramount here is only on the aspects relevant to the inves-
importance. Techniques such as surface tension tigation of mixed adsorption at the air–solution
and theoretical thermodynamic treatments such interface and of the treatment of multilayer
as the pseudo phase approximation (Holland and structures at the interface.
Rubnigh 1990) have provided much of the initial For an adsorbed monolayer of deuterium-
basis of understanding. However, in more recent labeled surfactant in null reflecting water, nrw
years, the use of neutron reflectivity, in combina- (a mixture of H2O and D2O with a refractive
tion with isotopic substitution, has transformed index of 1.0, the same as air), the reflected signal
our understanding of surfactant mixing and arises only from the adsorbed layer. Treating the
has challenged many aspects of their theoretical adsorbed layer as a single layer of uniform
treatment (Lu et al. 2000; Penfold and density and composition, as described in the
Thomas 2010). previous entry, the reflectivity is given by
Polymer–surfactant mixtures are being increas-
ingly used in many surfactant-based formulations k4 kt
(Goddard and Ananthapadmanabhan 1992) and 2
R ¼ 2r2 ð1 cos ktÞ ¼ 4b2 n2 sin (1)
16p 2
are used to manipulate adsorption and to modify
self-assembly and solution rheological properties. where k (Q is also sometimes used and is inter-
Polyelectrolyte–ionic surfactant mixtures, where changeable with k throughout this entry) is the
a strong surface interaction exists, give rise to wave vector transfer; t and r are thickness and
complex patterns of surface tension behavior scattering length density of the adsorbed layer,
which are difficult to interpret. Neutron reflection respectively; b is the sum of neutron scattering
(Taylor et al. 2007; Penfold et al. 2007a) has lengths of the adsorbed surfactant; and n is the
played a key role in developing an understanding refractive index (defined in the previous entry).
of how polyelectrolytes manipulate and enhance From this, the adsorbed amount, Г, can be calcu-
surfactant adsorption at interfaces and how struc- lated, where
tures more complex than a simple monolayer can
be found at the interface. rt 1
G¼ ¼ (2)
The formation of multilayer structures at inter- Na b Na A
faces, induced by the addition of polyions, is an
increasingly common and important structural where A is the area/molecule and Na is
signature of a wide range of phenomena and Avogadro’s number.
It is relatively straightforward to extend the is modified by an exponential decay constant,
approach encapsulated by Eqs. 1 and 2 to the such that
adsorption of a mixture of surfactants. For exam-
ple, in the simplest case of a binary mixture, rðiÞ ¼ rN Dr expðdi =Þ (5)
measurements can be made for either or both
components deuterium labeled and in nrw. and is a damping coefficient. This approach has
In which case, the reflected signal then arises proved to be a convenient way of parameterizing
from either component (the reflected signal will multilayer structures at the interface, as described
be dominated by the deuterium-labeled compo- by Penfold et al. (2007b) and Tucker et al. (2010).
nent) or from the total amount of surfactant at the
interface. Equation 2 is then extended to include
both components, such that Mixed Surfactants
b1 b2 Some typical neutron reflectivity data for a binary

rt ¼ þ (3)
A1 A2 and a ternary surfactant mixtures are shown in
Fig. 1a, b (Chen et al. 2010a, b).
where the subscripts refer to each of the surfac- The neutron reflectivity data in Fig. 1a are for
tant components at the interface, and it is readily a 60:40 mole ratio mixture of the mono-rhamnose
extended further to n components. and di-rhamnose rhamnolipids, R1 and R2;
Although in the previous entry the general glycolipids are described in more detail in the
formulisms associated with applying the optical previous entry. The solid lines are calculated
matrix methods to the calculation and analysis of curves for a single layer with t ¼ 20 Å, r ¼ 2.8
neutron reflectivity arising from structures more 106 Å2, and A ¼ 85 4 Å2 for d-R1/h-R2
complex than a single layer were described, they and with t ¼ 28 Å, r ¼ 0.7 106 Å2, and
are not necessarily the most convenient where A ¼ 300 10 Å2 for h-R1/d-R2, using Eq. 2.
there is a well-defined surfactant structure with This gives a surface mole fraction of R1 of 0.78
a repeat motif, such as surface multilayer struc- compared to a solution mole fraction of 0.6.
tures. There is an alternative approach based on The neutron reflectivity data in Fig. 1b are for
the kinematic approximation and developed by a 60:40 mole ratio mixture of rhamnolipid/LAS
Tidswell et al. (1990). The specular reflectivity (where LAS is the anionic surfactant sodium
for such a multilayer at the interface can be dodecylbenzene sulfonate) and with an equimo-
written as lar mixture of R1 and R2. The solid lines
are calculated curves for t ¼ 27 Å, r ¼ 1.1
16p2 X2N
106 Å2, and A ¼ 165 10 Å2 for d-R1/h-R2/h-
RðQÞ ¼ 2 ðr riþI Þ expðiQdi Þ
Q i¼0 i LAS; t ¼ 28 Å, r ¼ 0.8 106 Å2, and A ¼ S
(4) 270 10 Å2 for h-R1/d-R2/h-LAS; and t ¼ 24 Å,
2 2 2
Q s r ¼ 1.8 106 Å2, and A ¼ 82 4 Å2 for h-R1/
exp i 2
h-R2/d-LAS, using an extension of Eq. 2
for three components. This gives surface mole
where ri is the scattering length density of the ith fractions for R1, R2, and LAS of 0.28, 0.17, and
layer, i ¼ 0 represents the subphase, di is the 0.55, respectively, compared to the solution com-
distance of the interface between the ith and positions of 0.3, 0.3, and 0.4. In both cases, the
P
i + 1th layers from the subphase, di ¼ li , li is simultaneous equations established from modeling
i
the thickness of the ith layer, si is the roughness the reflectivity data for the different isotopic com-
between the ith and i + 1th layers, r(2 N + 1) is the binations, d-R1/h-R2 and h-R1/d-R2 for the R1/R2
upper bulk phase (air), and 2 N is the number of mixture and d-R1/h-R2/h-LAS, h-R1/d-R2/h-
layers (N is the number of bilayers). The defini- LAS, and h-R1/h-R2/d-LAS for the ternary
tion of the bilayer structure with increasing depth R1/R2/LAS mixture, were solved by least squares.
Surfactant Mixtures, a 1e−3

Polymer-Surfactant, and
Surfactant Multilayer
Structures Studied by 1mM dR1hR2 60:40
Neutron Reflectivity, 1mM hR1dR2 60:40
Fig. 1 Neutron reflectivity
data for R1/R2 and R1/R2/ 1e−4
LAS mixtures for (a) 1 mM
Reflectivity
60:40 mole ratio R1/R2 ( )
d-R1/h-R2/nrw, ( ) h-R1/
d-R2/nrw, and (b) 1 mM
60:40 mole ratio R1/R2
(1;1)/LAS, (red) d-R1/
h-R2/h-LAS, (blue) h-R1/ 1e−5
d-R2/h-LAS, (green) h-R1/
h-R2/d-LAS. The solid
lines are calculated curves
as described in the text
(# 2010 ACS (Reproduced
with permission of Chen 1e−6
et al. 2010a, b)) 0.1
Wavevector Transfer, Q (A−1)
b 1e−4
dR1 hR2(1:1)hLAS 60:40

hR1 dR2(1:1)hLAS 60:40
hR1 hR2(1:1)dLAS 60:40
Reflectivity
1e−5
0.1
Wavevector Transfer, Q (A−1)
The neutron reflectivity method, when used in cmc, there is no reliable method other than
a manner as described above, provides the oppor- neutron reflection for concentrations greater
tunity to not only determine absolutely the varia- than the cmc. Thermodynamical treatments,
tion in surface composition with solution such as regular solution theory, RST (Holland
composition at a fixed concentration, but impor- and Rubnigh 1990), where the departure from
tantly, the variation in surface composition with ideal mixing is described by a single interaction
solution concentration can be established. parameter, b, have been quite successful in
Although, in principle, this can be obtained predicting and quantifying cmc variations with
from methods like surface tension for concentra- solution compositions for a wide variety of
tions below the critical micellar concentration, mixed surfactants under different conditions.
evolution on surface composition with increasing
surface composition (mole fraction SDS (C12E6)
1.0 surfactant concentration above the cmc in

a binary surfactant mixture, and are an example
0.8 where the variation is well described by RST.
The data show that the surface immediately
0.6 above the cmc is initially rich in the more surface
active component, in this case the C12E6, and
0.4
evolves with increasing surfactant concentration
toward a surface composition close to the
0.2
solution composition. There is a shift in the
pattern of the adsorption of both components to
0.0
0.0001 0.001 0.01 lower concentration values with the addition of
solution concentration (M) the polyelectrolyte, poly-dmdaac. The solid lines
are model calculations from RST, and assuming
Surfactant Mixtures, Polymer-Surfactant, and
Surfactant Multilayer Structures Studied by Neutron an interaction parameter b 3.0, and in the
Reflectivity, Fig. 2 Variation of surface composition for absence of polyelectrolyte using measured cmc
70/30 SDS/C12E6/0.1 M NaCl with concentration at the values for the pure components and the mixture.
air–solution interface, for ( ) SDS, ( ) C12E6, (red) with-
The value for b is typical of that obtained from
out, and (blue) with 50 ppm poly-dmdaac. The solid lines
are RST calculations as described in the text (# 2005 cmc variations (Holland and Rubnigh 1990) for
ACS (Reproduced with permission of Penfold et al. ionic/nonionic surfactant mixtures in the pres-
2005a)) ence of electrolyte and is consistent with
a relatively modest departure from ideal mixing.
However, the variation in the surface composi- The addition of the polyelectrolyte can also be
tion, above the cmc, with solution composition accounted for by RST with the same parameters,
and concentration has not been rigorously inves- but with the additional assumption that the poly-
tigated. Neutron reflection measurements have, electrolyte effectively reduces the cmc of the
however, transformed our knowledge and under- ionic component, the SDS.
standing of surfactant mixing at interfaces and The adsorption data for the nonionic
especially at the air–solution interface. These surfactant mixture of triethylene monododecyl
studies have established the broad conditions ether, C12E3, and octaethylene monododecyl
where treatments such as RST are quite reliable ether, C12E8, derived from neutron reflectivity
and have highlighted areas and aspects where data, illustrate two important features that have
theory is deficient, and these aspects have been only been accessible using neutron reflection
critically reviewed recently by Penfold and (Penfold et al. 1995).
Thomas (2010). The purpose of this section of Figure 3a shows the total adsorption and the S
the entry is to show how neutron reflection adsorbed amounts of the individual components
contributes to this area rather than to provide an for an equimolar mixture of C12E3/C12E8 at con-
in-depth description of surfactant mixing. centrations from below to well in excess of the
The surface composition data shown in Fig. 2 mixed cmc, and the surface composition of each
for a 70:30 mole ratio mixture of the anionic component is shown in Fig. 3b. Although
surfactant sodium dodecyl sulfate, SDS and the predicted theoretically by Nikas et al. (1992),
nonionic surfactant hexaethylene monododecyl the data in Fig. 3 is the first experimental demon-
ether, C12E6 in 0.1 M NaCl, and with the addition stration of the abrupt change in the surface com-
of 50 ppm of the polyelectrolyte, poly(dimethyl- position at the mixed cmc. This change in
diallylammonium chloride), poly-dmdaac, at the monolayer composition can be explained in
air–water interface were obtained using neutron terms of changes in the distribution of two sur-
reflectivity (Penfold et al. 2005a). These data factant species between the bulk solution and the
encapsulate many of the features typical of the monolayer due to the onset of mixed micelle
Surfactant Mixtures, a 5
Structures Studied by
Adsorbed amount (×10−10 mol cm−1)

Neutron Reflectivity, 4
Fig. 3 (a) Variation in
surface excess, G,
(1010 mol cm2) with
solution concentration for 3
equimolar C12E3/C12E8
mixture, ( ) C12E3, ( )
C12E8, and ( ) total 2
adsorption. (b) variation in
surface composition ( )
mole fraction of C12E3 and
( )) mole fraction of C12E8 1
with surfactant
concentration (Reproduced
with permission of 0
Coll Surf)
1e−5 1e−4 1e−3 1e−2
Surfactant concentration (M)
b 1
0.8
Mole Fraction
0.6
0.4
0.2
0
10−5 10−4 10−3 0.01
Surfactant Conc. (M)
formation. That is, the amount of each surfactant greater than 100 times the mixed cmc, the surface
adsorbed is controlled by the free monomer con- is still dominated by the C12E3, and this is unex-
centration in solution, and this change at the cmc pected. In contrast, the micelle compositions are
where the free monomer concentration of the close to the solution composition. For ideal
more surface active component decreases due to mixing, and for that matter nonideal mixing, it
preferential uptake into mixed micelles. The is normally observed and theoretically predicted
other important feature of these data is that as that the surface and solution compositions will be
expected the measured cmc variation would close at solution concentrations well above the
imply that the two surfactants mix ideally or cmc. It is assumed that this is due to the unfavor-
close to ideally, yet at a solution concentration able packing constraints associated with the
much bulkier EO8 headgroup of the C12E8 com- 1.0
Surface composition (mole fraction R1)

ponent. This is partially accounted for in Hol-
land’s extension of RST to include a “force 0.8
field” term in the expression for the surface free

energy which is a combination of the surface 0.6
pressure and area/molecule and is associated

0.4
with packing constraints (Holland 1986). This
has implications for the structure of the mixed
0.2
monolayer compared to that of the individual
components, and this is discussed in more detail
0.0
later in the entry. 0.0 0.2 0.4 0.6 0.8 1.0
The reflectivity data shown in Fig. 1a for the Solution composition (mole fraction R1)
rhamnolipid mixture of R1 and R2 shows
a similar effect to the nonionic surfactant mixture Surfactant Multilayer Structures Studied by Neutron
described above. As shown by Chen et al. Reflectivity, Fig. 4 Surface composition (mole fraction
(2010a), the cmc values for R1 and R2 are fairly of R1) versus solution composition for 1 mM R1/R2
similar, and their variation with composition is (# 2010 ACS (Reproduced with permission of Chen
et al. 2010a))
well described by ideal mixing. However, the
data in Fig. 1a and its interpretation allude to
Surface composition (mole fraction R1, R2, LAS)
a rather different behavior at the interface.

1.0
This is shown more completely in Fig. 4, where
the surface composition at the air–solution inter- R1
0.8 R2
face for the R1/R2 mixture is plotted against the LAS
solution composition at a fixed surfactant concen-
0.6
tration of 1 mM.
The composition data in Fig. 4 show that the 0.4
surface is dominated by the R1 adsorption over
most of the solution composition range. As the 0.2
cmc values of R1 and R2 are similar and the cmc
variation with composition is ideal, the variation 0.0
in surface composition is not consistent with ideal 0.0 0.2 0.4 0.6 0.8 1.0
Solution composition (mole fraction R1, R2, LAS)
mixing or with RST. This is analogous to the
situation described above for the C12E3/C12E8 Surfactant Mixtures, Polymer-Surfactant, and
mixture, and the marked departure of the surface Surfactant Multilayer Structures Studied by Neutron
composition from the solution composition Reflectivity, Fig. 5 R1/R2/LAS surface composition
versus solution composition for 1 mM R1/R2/LAS and
S
in favor of R1 is attributed to the surface packing an R1/R2 mole ratio of 1:1. # 2010 ACS (Reproduced
constraints associated with the bulkier with permission of Chen et al. 2010b)
di-rhamnose headgroup of R2. This packing
constraint is also manifest in the ternary mixture
shown in Fig. 1b, where the binary rhamnolipid headgroup geometry have implications for the
mixture of R1 and R2 is blended with the anionic structure of the surface layer, and this was inves-
surfactant LAS. The analysis of the data from tigated using neutron reflectivity (Penfold et al.
Fig. 1b implies that the R2 competes less favor- 1998). A deuterium labeling scheme was used
ably than R1 and LAS for the surface, and this is where the ethylene oxide groups and the terminal
more clearly illustrated in Fig. 5, from Chen et al. hexyl groups of the alkyl chains were labeled.
(2010b). This provides maximum sensitivity to changes
As implied earlier, these packing constraints in the conformation of the ethylene oxide and
associated with the mismatch in surfactant alkyl chains upon mixing.
Surface composition (mole fraction non-ionic)

1.0
1 DHDAB C12E3
DHDAB C12E6
0.8 DHDAB C12E12
Volume Fraction
Solvent
Ideal mixing
0.6
0.5
C6 (E03)
E08 0.4
E03 C6 (E08)
0.2
0
−40 −20 0 20
0.0
Z (in ) 0.0 0.2 0.4 0.6 0.8 1.0
Solution composition (mole fraction non-ionic)
Surfactant Multilayer Structures Studied by Neutron Surfactant Mixtures, Polymer-Surfactant, and
Reflectivity, Fig. 6 Volume fraction profiles for Surfactant Multilayer Structures Studied by Neutron
5 105 M 30/70 C12EO3/C12EO8, showing the solvent Reflectivity, Fig. 7 Variation in surface composition
(---), alkyl chain (outer C6 only), and ethylene oxide (mole fraction nonionic surfactant) as a function of solu-
chains of C12EO3 and C12EO8 (Reproduced with permis- tion composition (mole fraction nonionic surfactant) for
sion of J Coll Int Sci) DHDAB/C12E3 (C12E6, C12E12) at a solution concentra-
tion of 1.5 mM. The solid lines join the data points as
a guide to the eye. The broken black line corresponds to
ideal mixing (# 2009 ACS (Reproduced with permission
The neutron reflection measurements were of Tucker et al. 2009))
made using 15 different isotopically labeled com-
binations and evaluated using the partial structure
factor approach described in the previous entry, associated with the onset of micelle formation,
Eq. 21. The resulting volume fraction distribu- as described earlier in more detail. Above the
tions for a 30/70 mole ratio mixture of C12EO3/ cmc, changes in micellar phase behavior, for
C12EO8 at a concentration of 5 105 M are example, growth from globular to worm-like
shown in Fig. 6. The results show that the frus- micelles, have not been associated with any sig-
tration caused by packing the EO8 group with the nificant changes in the surface adsorption prop-
EO3 group results in the EO8 group being less erties. In Fig. 7, the variation in surface
extended and more hydrated and the EO3 group composition with solution composition at
less hydrated than in the corresponding pure a fixed solution concentration for the dialkyl
monolayers, and this is consistent with the expec- chain cationic surfactant and dihexadecyldi-
tation from the theoretical work of Nikas et al. methylammonium bromide, DHDAB, with the
(1992). nonionic surfactants C12E3, C12E6, and C12E12
There are other circumstances for surfactant are shown.
mixtures where the surface adsorption and its For all three nonionic surfactants, there is an
departure from ideality are not well represented abrupt inset for the nonionic adsorption at
by current theoretical treatments, and the work of a particular solution composition. For C12E6 and
Tucker et al. (2009) on the surface adsorption of C12E12, the departure from ideal mixing is
dialkyl chain cationic/nonionic surfactant mix- dramatic. There is no nonionic adsorption below
tures is a notable example. There is an intrinsic a threshold composition 40 mol% nonionic sur-
interplay or interdependence between the surface factant in solution, and the dialkyl chain cationic
properties and the solution behavior, and this is surfactant dominates over much of the composi-
mediated by the free monomer concentration and tion range. For C12E3, the nonionic surfactant
composition in solution. This was illustrated in competes more effectively for the surface with
Fig. 4 for the C12E3/C12E8 mixture, where the the DHDAB above a threshold composition of
abrupt change in the surface adsorption is 20 mol% nonionic. Discontinuities in the cmc
variation with composition for all three nonionic critical aggregation concentration, the point at
surfactants are observed and occur at similar which SDS micelles start to form on the PEO
solution compositions to the threshold values chains in solution. The second breakpoint was
for the adsorption. Investigation of the solution identified as the cmc of the mixture, the point at
phase behavior by SANS showed that the transi- which free micelles start to form in solution.
tion in adsorption for DHDAB/C12E6 and Although surface tension was able to accurately
DHDAB/C12E12 coincided with a transition and correctly identify these transitions, there was
from micellar to micellar/lamellar coexistence. no reliable information about the adsorption
For DHDAB/C12E3, it was associated with behavior from such data. Cooke et al. (1998)
a transition from multi-lamellar to bi-lamellar used neutron reflection, in combination with
vesicle formation. Although the aggregate com- either d-PEO or d-SDS, to determine the relative
positions are close to the solution composition amounts of PEO and SDS adsorbed at the inter-
and change imperceptibly over the composition face. This is shown in Fig. 8 along with the
range measured, the free monomer concentra- surface tension response.
tions can be at least 1–3 orders smaller, and so, The neutron reflection data show that before the
quite subtle changes in aggregate composition cac, the PEO is progressively replaced by the SDS,
can result in relatively large changes in monomer and at the cmc, the PEO has been completely
concentration and composition. It is such changes removed from the surface. Hence, there is compet-
that are responsible for the features observed in itive adsorption between the PEO and SDS up to
the surface adsorption (Fig. 7), and Tucker et al. the point at which bulk PEO/SDSD micelles com-
(2009) made some relatively simple but plexes form, and at which point the PEO/SDS
non-rigorous estimates to demonstrate the solution complex is energetically more favorable
relationship between the adsorption and the cmc for the PEO than the surface.
variation. Purcell et al. (1998) demonstrated how impor-
tant insights can be gained from just
a combination of the surface tension response
Polymer–Surfactant Mixtures and measurements with deuterated surfactant,
and hence, only the quantification of the surfac-
Neutron reflection studies have substantially tant adsorption at the interface in SDS/poly
enhanced the understanding of the surface (vinylpyrrolidone), PVP, mixtures. The results
adsorption behavior of polymer/surfactant from the neutron reflectivity measurements are
mixtures, especially at the air–water interface, shown in Fig. 9 for SDS/PVP mixtures, for
and the recent developments have been exten- a 700 k MW PVP, and for two different PVP
sively reviewed (Taylor et al. 2007; Penfold concentrations (0.5 and 2 wt%). The surface
et al. 2007a). The purpose of this section is to tension behavior is broadly similar to that S
demonstrate and describe how neutron reflection described by Jones (1967) and observed for
has been used to quantify the adsorption in SDS/PEO by Cooke et al. (1998).
a variety of different polymer/surfactant mixtures Below the cac, as determined from surface
and how it provides an interpretation of some tension, there is a slight enhancement of the
quite complex surface tension data rather than SDS at the interface in the presence of PVP
present a comprehensive and in-depth discussion compared to SDS alone, indicative of
and description of the phenomena. a cooperative interaction at the interface.
In his pioneering study, Jones (1967) Between the cac and cmc, there is a depletion of
described the main features in the surface tension SDS from the surface due to bulk SDS/PVP com-
behavior of the weakly interacting polymer/sur- plex formation. Eventually, with increasing SDS
factant mixture of poly(ethylene oxide) and concentration, the amount of SDS at the interface
sodium dodecyl sulfate, SDS. The initial will start to approach that in the absence of
breakpoint (see Fig. 8) was ascribed to a cac, polymer.
55
CAC
12 + +
+ 51
Γ⫻1010/mol cm−2
+
+
γ/mNm−1
47
8 +
+
+
+
++
++
43
+ +
4 ++
++ 39
++
+ +
+ ++
++
+
0 35
−4.0 −3.5 −3.0 −2.5 −2.0 −1.5 −1.0 −0.5
Log10(CNaDS)
Surfactant Mixtures, Polymer-Surfactant, and segment/unit area, ( ) SDS adsorption, and (+) surface
Surfactant Multilayer Structures Studied by Neu- tension for SDS/0.1 wt% 25 k MW PEO. The solid lines
tron Reflectivity, Fig. 8 Adsorption isotherm and sur- are guides to the eye (# 1998 ACS (Reproduced with
face tension for PEO/SDS mixture, () PEO adsorption in permission of Cooke et al.))
Goddard (2001) discussed how interpreting

5
surface tension in the presence of a strong surface
surfactant–polyelectrolyte interaction is difficult 4
1010Γ/mol cm−2
and can be misleading. Neutron reflectivity has

played an important role in unraveling some of 3
the complexity in such systems by providing
direct information about the surface adsorption 2
from such mixtures. This is discussed in more
detail in recent reviews (Taylor et al. 2007; 1
Penfold et al. 2007a), and from a wide range of
0
studies on different system, the general patterns −5 −4 −3 −2 −1
of behavior can be rationalized in terms of the Log (CSDS)
competition between the formation of surface
polymer–surfactant complexes and solution Surfactant Mixtures, Polymer-Surfactant, and
Surfactant Multilayer Structures Studied by Neutron
complexes and their relative free energies of Reflectivity, Fig. 9 Variation in SDS adsorption versus
formation. As such, a complex pattern of surface SDS concentration for (o) SDS, ( ) SDS/0.5 wt% PVP,
tension behavior exists. and ( ) SDS/2 wt% PVP. The lines are guides to the eye
At one extreme of the surface tension behavior (# 1998 ACS (Reproduced with permission of Purcell
et al. 1998))
is the response of the mixture of SDS and
poly-dmdaac (Staples et al. 2002) in which
a pronounced peak in the surface tension is The initial sharp decrease in the surface
observed before the cmc, as shown in Fig. 10. tension at low SDS concentrations is an indica-
This is a system where the bulk complex and tion of a strong polymer/surfactant adsorption at
surface polymer/surfactant complex formation the interface due to surface complex formation.
have comparable free energy values. Penfold The peak in the surface tension at higher SDS
et al. (2007c) extended the initial neutron reflec- concentrations coincides with a region where the
tion study of Staples et al. (2002) to investigate solutions are cloudy and can show signs of phase
the impact of the surfactant alkyl chain length and separation. At even higher and at lower surfactant
make direct comparison with recent theoretical concentrations, the solutions are clear. The
developments. corresponding SDS adsorption was measured by
70
Surface tension (mN/m)
60
50
C10 alkyl sulfate / DMDAAC

40 C12 alkyl sulfate / DMDAAC
C14 alkyl sulfate / DMDAAC
C10 alkyl sulfate
C12 alkyl sulfate
C14 alkyl sulfate
30
1e−6 1e−5 1e−4 1e−3 1e−2 1e−1
Surfactant concentration (M)
Surfactant Mixtures, Polymer-Surfactant, and solid lines and (red) decyl, (green) dodecyl, and (blue)
Surfactant Multilayer Structures Studied by Neutron tetradecyl. The dashed lines are in the absence of polymer
Reflectivity, Fig. 10 Surface tension versus surfactant (# 2007 ACS (Published with permission of Penfold et al.
concentration for alkyl sulfates/20 ppm poly-dmdaac; the 2007c))
C10 alkyl sulfate

C12
Adsorbed amount (X10−10 mol cm−2)
5
C14
SDS alone
Surfactant Mixtures, 4
Structures Studied by 3
Neutron Reflectivity,
Fig. 11 Adsorbed amount
of surfactant for alkyl 2 S
sulfate/20 ppm poly-
dmdaac at the air–water
interface, (red) decyl, 1
(green) dodecyl, and (blue)
tetradecyl, (+) SDS in the
absence of polymer (#
0
2007 ACS (Reproduced
1e−6 1e−5 1e−4 1e−3 1e−2 1e−1
with permission of Penfold
et al. 2007c)) Surfactant concentration (M)
neutron reflection (using the combination d-SDS/ over the entire surfactant concentration range
poly-dmdaac/nrw) and is shown in Fig. 11. measured and for all three alkyl sulfates, is in
There are three key features that form the the form of a monolayer. Secondly, at the lower
neutron reflection data. Firstly, the adsorption, surfactant concentrations, consistent with the
Neutron Reflectivity, 70
Fig. 12 Surface tension
Surface tension (mN m−1)

for SDS/20 ppm
poly-dmdaac. The solid
line is a calculated curve 60
using Eq. 6 (# 2007c ACS
(Reproduced with
50
2007c))
40
30
−5 −4 −3 −2 −1
Log (SDS concentration, M)
surface tension data for sodium decyl and PSs, and a Langmuir isotherm, then the surface
dodecyl sulfate, there is marked enhancement in tension can be expressed using the Langmuir–
the surfactant adsorption compared to surfactant Szyszkowski equation, such that
in the absence of polymer due to a strong surface
polymer–surfactant interaction and surface g g0 ¼ RTG1 logð1 þ ks S þ kPSs PSs Þ (6)
polymer–surfactant complex formation. In the
region of the surface tension peak, the surfactant ks and kPSs are the surfactant and polymer/
is partially depleted from the surface, as polymer surfactant adsorption coefficients. The application
and surfactant are removed from the surface in of Eq. 6 to the surface tension data for SDS/poly-
favor of bulk complex formation (which may or dmdaac, shown in Fig. 10, is illustrated in Fig. 12.
may not phase separate). Complementary Similar neutron reflection studies have been
NR measurements for d-surfactant/h-polymer in made on SDS mixed with the pH-sensitive poly-
D2O (where only the polymer will be visible) electrolyte, poly(ethyleneimine), PEI (Penfold
show that at this point, the polymer is largely et al. 2005b). PEI is a strong, highly charged
removed from the surface. The subsequent polyelectrolyte at low pH and is essentially
decrease in surface tension is accompanied by a neutral polymer at high pH. It also exists in
increased SDS adsorption at the interface as the a linear form but more usually is in a highly
free monomer concentration of SDS in solution branched form. In its linear form, there is
increases. a strong surface interaction between the PEI and
Bell et al. (2007) used the insights from SDS, which is unexpectedly independent of pH,
neutron reflection measurements on a range and this aspect is discussed in more detail later.
of polymer–surfactant mixtures to develop However, this strong interaction gives rise to
a theoretical treatment for the surface behavior, a substantial enhancement of the SDS adsorption
using the law of mass action and based on the at very low SDS concentrations. The adsorption
competition between surface and solution data in Fig. 13 shows this effect and makes a -
complex formation. Assuming that the only comparison with the enhancement obtained
surface active components are the surfactant, S, from neutron reflectivity for SDS with
and the surface polymer–surfactant complexes, poly-dmdaac.
Adsorbed amount ( X 10−10 mol cm−1)

5
Fig. 13 Comparison of
SDS adsorption for ( ) 25 k
4
molecular weight linear
PEI/SDS pH 10, ( ) SDS,
( ) SDS/poly-dmdaac, and
( ) SDS/PEO (# 2005 3
ACS (Reproduced with
2005b)) 2
SDS
SDS/PEO
1 SDS / poly-dmdaac
SDS / PEI
0
1e−6 1e−5 1e−4 1e−3 1e−2
SDS concentration (M)
The other extreme of the surface tension The transition from a monolayer to the trilayer
behavior in the strongly interacting polyelectrolyte– structure is shown on Fig. 14 for 25 k MW
surfactant mixtures is epitomized in mixtures branched PEI/SDS at pH 10.
such as branched PEI/SDS and poly(styrene At the SDS concentration of 5 mM, the
sulfonate)/dodecyltrimethylammonium bromide, adsorbed layer is in the form of a monolayer
C12TAB. These mixtures represent situations 20 Å thick, whereas the data at 0.7 mM SDS
where the surface polymer/complex formation is is best described by two layers, a monolayer
energetically more favorable and dominate the 20 Å and a second more dilute layer 30 Å.
surface tension behavior. In such cases, the gen- The data for 0.3 mM SDS is distinctly different,
eral form of the surface tension response is and a pronounced interference fringe is visible.
different to that shown in Fig. 10, and the initial This is best described by three layers, an initial
decrease in surface tension at low surfactant con- layer 30 Å, followed by two layers 15 Å,
centrations is followed by a plateaux before where the first and third layers are relatively
a final decrease at the cmc. Neutron reflection dense containing deuterated alkyl chains with S
measurements have revealed that the structure a volume fraction 50 % and the middle layer
of the surface layer adsorbed at the interface, is predominantly solvent. Measurements made in
and its variation with concentration is more nrw and in D2O enabled the relative amounts of
complex than a simple monolayer. That is, it surfactant, solvent, and polymer to be estimated
evolves at some concentration along the surface for each of the layers. Under certain conditions,
tension plateaux from just a monolayer to there is a further transition from the trilayer
a structure which is described as three layers but structure to a multilayer structure at the interface.
represents the formation of a bilayer adjacent to For PEI, the trilayer structure and the transition to
the initial monolayer. In some circumstances, this multilayers are most prevalent for the lower MW
evolves to form a multilayer structure at the branched PEI (2 k MW) at the higher pH values
interface. Such surface structures are observed of 7–10 and with the addition of electrolyte.
for C12TAB/PSS (Taylor et al. 2002) and A typical reflectivity pattern that is observed
for SDS/branched PEI (Penfold et al. 2005b). when such multilayers form at the interface are
Surfactant Mixtures,
Structures Studied by 1e−3
Fig. 14 Neutron
reflectivity for SDS/25 k
molecular weight branched
Reflectivity
1e−4
PEI (20 ppm) at pH 10.0
and d-SDS/PEI/nrw,
( ) 5 103 M SDS,
( ) 7 104 M SDS, and
(D) 3 104 M SDS. The 1e−5
solid lines are model
calculations as described in
the text (#2005 ACS
(Reproduced with
permission of Penfold et al. 1e−6
2005b))
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
−1
Wavevector transfer, Q (Å )
shown in Fig. 15, for a 2 k branched PEI at pH 10, At pH 3, the electrostatic attraction between the
but where the PEI has been modified to include SDS and PEI dominates the interaction. At pH 10,
a single ethylene oxide group, PEI-EO1, as the effect is ascribed to a dipole–ion interaction
described by Zhang et al. (2011). between the amine nitrogen and the sulfate
The example presented in Fig. 15 was chosen headgroup which is augmented by a cooperative
as the measurements were made for three differ- hydrophobic interaction between the alkyl chains
ent isotopic combinations, d-SDS/PEI-EO1/nrw, of the attached surfactant molecules. By the time
h-SDS/PEI-dEO1/nrw, and d-SDS/PEI-EO1/ the MW of the amine has increased to pentaethy-
D2O. The solid lines in the figure are for the lenehexamine, the combined attraction at pH 10
same multilayer structure, with 20 bilayer repeats is comparable to that at low pH.
and a bilayer thickness of 38 Å, for each contrast The short-ranged nature of the interaction was
but where the changes in the scattering length demonstrated by the studies using different
density in the layers reflect the contrast. ethoxylated PEIs by Zhang et al. (2011) as
An unusual feature of the SDS/PEI adsorption described earlier. PEI-EO1 behaved very much
behavior is the pH dependence of the adsorption. It like PEI, but for EO3 and larger, the polymer–
would be expected that the strongest interaction surfactant interaction is similar to that observed
would exist at pH 3 where the polyelectrolyte is for SDS/PVP but even weaker. Furthermore, it
most highly charged. However, with the linear was shown that the pH dependence was more
PEI, the adsorption appeared to be independent consistent with the original expectation. That
of pH, and with the branched PEI, the multilayer is, the strongest interaction occurred at the
structures are most prevalent at higher pH. Penfold lowest pH.
et al. (2009) made a series of neutron reflection
measurements with SDS and a range of
oligoamines from ethylenediamine to pentaethyle- Role of Polyions and Multivalent Ions
nehexamine. In Fig. 16, the SDS adsorption for
104 M SDS and 5 104 M polyamine at pH 3 Surface multilayer formations at interfaces have
and 10 is plotted for increasing polyamine length. now been reported in a wide range of surfactant-
The SDS adsorption is more strongly depen- based systems but were initially observed in
dent on polyamine MW at pH 10 than at pH 3. relatively concentrated systems, such as the
Surfactant Mixtures, 1e-2
Surfactant Multilayer d-SD S/PEI-(hEO)1/nrw
Structures Studied by h-SD S/PEI-(dEO)1/nrw
Neutron Reflectivity, 1e-3
d-SD S/PEI-(hEO)1/D2O
Fig. 15 Neutron
reflectivity for 20 ppm
PEI-(EO)1/104 M SDS at
1e-4
Reflectivity
pH 10, for (red) d-SDS/
PEI-(hEO)1/nrw, (blue)
h-SDS/PEI-(dEO)1/nrw,
and (green) d-SDS/ 1e-5
PEI-(hEO)1/D2O. The solid
lines are model fits
assuming a multilayer
structure as described in the 1e-6
text (# 2011 ACS
(Reproduced with
permission of Zhang et al.
2011)) 1e-7
0.05 0.10 0.15 0.20 0.25 0.30
pH 10
pH 3
5
Adsorbed amount (⫻10−10 mol cm−2)
4
Surfactant Mixtures,
Surfactant Multilayer 3
Fig. 16 Variation in 2
adsorbed amount of SDS
(x 1010 mol cm2) with
polyamine molecular 1
weight for 104 M SDS,
5 104 M amine at pH 3 S
( ), and pH 10 ( ) (# 2009
0
ACS (Reproduced with
1 2 3 4 5 6 7
2009)) amine length
dialkyl chain cationic surfactants reported by formation is that the solutions exhibit extreme
McGillivray et al. (2003). More recently, as and persistent wetting properties on highly
described in the previous section, surface multi- hydrophobic surfaces, as reported by Penfold
layer structures have also been observed and et al. (2005a) for the SDS/PEI mixtures. Such
characterized by neutron reflection for relatively properties imply great potential for optimized
dilute systems in mixtures of ionic surfactants detergency and soft lubrication and the opportu-
and polyelectrolytes or oligoamines. One of the nity for enhanced delivery of active benefit
notable characteristics of such multilayer agents to surfaces and interfaces. At relatively
Surfactant Mixtures, 1e+0

Structures Studied by 1e−1
Fig. 17 Reflectivity for
1 mM 95/5 mole ratio 1e−2
SLES/C12E12/AlCl3/NaCl,
Reflectivity
( ) 0.0/3.0 mMAlCl3/
NaCl, ( ) 0.24/1.8 mM, ( ) 1e−3
0.36/1.2 mM, and ( ) 0.6/
0.0 mM. The different data
are shifted vertically for
1e−4
clarity. The solid lines are
model fits as described in
the text (# 2010 ACS
(Reproduced by permisison 1e−5
of Tucker et al. 2010))
1e−6
0.1
low surfactant concentrations, surface multilayer The measurements in Fig. 16 were made in
structures have been observed in a wider range of AlCl3/NaCl mixtures at constant ionic strength
systems and especially in systems where the but with increasing Al3+ concentration. In the
strong binding of multivalent counterions (Ca2+ absence of Al3+, the data are consistent with
or Al3+) induce such surface structures. Neutron a monolayer of surfactant adsorbed at the inter-
reflection measurements have played a central face with a thickness 18 Å and an area/molecule
role in identifying and characterizing the struc- 95 Å2. For 0.2 mM Al3+, a pronounced interfer-
tures from such systems. ence fringe is visible in the reflection data, and the
In their pioneering study, Alargova et al. data can be modeled as three layers of thicknesses
(2003) showed how the strong binding and 22, 15, and 24 Å, representative of an initial
complex formation of Al3+ to adjacent sodium monolayer and a single bilayer beneath. At
dodecyl dioxyethylene sulfate, SLES, molecules 0.36 mM Al3+, a double fringe pattern is
promoted substantial micellar growth and observed, and this was analyzed using the
a reduction in interfacial tension. They showed approach encapsulated in Eqs. 4 and 5.
that the onset of micellar growth coincided with It corresponds to three bilayers at the interface,
the counterion charge being in excess and that with a bilayer thickness 43 Å. At the highest Al3
+
coadsorption of a nonionic surfactant with concentration, 0.6 mM, the reflectivity pattern
a bulky headgroup, C12EO10, would disrupt the has evolved into the form of a single Bragg peak.
complexation and inhibit growth. Similar obser- This implies that there is now a much larger
vations were made for Ca2+, except that the number of bilayers in the multilayer structure at
effects were not so pronounced. Despite the ele- the interface, and the data are modeled as 20
gance of their work, they were unable to com- bilayers with a bilayer thickness 49 Å. Similar
ment on the surface properties in any detail. patterns of structure and evolution in the structure
Tucker et al. (2010) used neutron reflection to were reported as a function of composition, sur-
probe the surface structures induced by Al3+ in factant, and counterions concentration, such that
SLES/monododecyl dodecaethylene glycol, a surface phase diagram can be constructed. This
C12E12, mixtures. A typical evolution in reflec- is shown in Fig. 18a for SLES/C12E12.
tivity with increasing Al3+ concentration is A further feature illustrated in Fig. 17a is that
shown in Fig. 17. the onset of bilayer and multilayer formation is
SLES / C12E12 LAS / C12E8
0.7 2.5
multilayer
0.6
AlCl3 concentration (mM)
2.0
0.5
bilayer
CaCl2 conc (mM)

1.5 bilayer multillayer
0.4
0.3
1.0
0.2 monolayer
monolayer
0.5
0.1
0.0 0.0
0.75 0.80 0.85 0.90 0.95 1.00 1.05 75 80 85 90 95 100 105
Solution composition (mole fraction SLES) Mole % LAS
Surfactant Mixtures, Polymer-Surfactant, and (a) 1 mM SLES/C12E12/AlCl3 and (b) 2 mM LAS/C12E8

Surfactant Multilayer Structures Studied by Neutron (# 2010 ACS (Published with permission of Tucker
Reflectivity, Fig. 18 Surface phase behavior for et al. 2010))
shifted to progressively higher Al3+ concentra- To further investigate the impact of the sur-
tions as the solution composition becomes richer factant architecture on the surface multilayer for-
in the nonionic cosurfactant. This is mation, Tucker et al. (2011) investigated the role
a manifestation of the same observation made of different LAS isomers on surface multilayer
for micellar growth, where the EO12 headgroup formation. LAS-6 is an essentially symmetrical
inhibits or partially inhibits the strong SLES/Al3+ dialkyl chain surfactant with the phenyl ring
complexation, which drives the structural joined to the middle of the alkyl chain (the “C6”
changes both in solution and at the interface. position) and LAS-4 is an asymmetric isomer
Although the addition of Ca2+ was shown to functionalized at the “C4” position. Mixtures of
promote micellar growth (Alargova et al. 2003; LAS-6 and LAS-4 were studied in the presence of
Tucker et al. 2010), the complexation between CaCl2. The reflectivity profiles in Fig. 19a, b are
SLES and Ca2+ is not sufficiently strong to drive for a 2-mM mixture of LAS-6 and LAS-4 in
multilayer formation at the interface. However, 1 mM CaCl2, for two different isotopic combina-
for the dialkyl chain anionic surfactant LAS, Ca2+ tions, d-LAS-6/h-LAS-4 and h-LAS-6/d-LAS-4,
promotes multilayer formation, and Penfold et al. and for a range of LAS-6/LAS-4 compositions.
(2007) used neutron reflectivity to investigate its The reflectivity data illustrate some interesting S
surface phase behavior in combination with the and important features. The single Bragg peak
nonionic cosurfactant C12E8, and the resulting provides a clear signature of multilayer formation
phase diagram is shown in Fig. 18b. The compar- at the interface, but the detailed information is
ison between SLES/C12E12 and LAS/C12E8 limited as discussed earlier. It provides primarily
shown in Fig. 18 shows that relatively higher a quantification of the bilayer thickness. The
counterion concentrations are required to width of the Bragg peak is also related to 1/N,
promote multilayer formation with LAS/Ca2+ where N is the number of bilayers, and is addi-
than for SLES/Al3+, and this is a reflection of tionally broadened by the instrumental resolu-
the relative strengths of the binding of Ca2+ and tion, DQ (Dk). Hence, from data such as shown
Al3+. Furthermore, the bilayer region extends in Fig. 16 for 1 mM 95/5 mole ratio SLES/
over a broader range of counterion concentrations C12E12/0.6 mM AlCl3, it is often only possible
for SLES/Al3+ than for LAS/Ca2+, although the to put a limit on the minimum value for N.
reason for this is not yet entirely clear. However, the data in Fig. 19a, b show some
Surfactant Mixtures, a 1e−3

Surfactant Multilayer 80/20
Structures Studied by 60/40
40/60
20/80
Fig. 19 Neutron
reflectivity data for 2-mM
mixture of LAS-6 and 1e−4
LAS-4 in 1 mM CaCl2 in
Reflectivity
nrw and compositions from
20/80 to 80/20 mole
fraction (see details in
figure legend),
(a) d-LAS-6/h-LAS-4 and
(b) h-LAS-6/d-LAS-4. The 1e−5
solid lines are calculated
curves as described in the
text (# 2011 ACS
(Published with permission
of Tucker et al. 2011))
0.1
b 1e−3
20/80
40/60
50/50
60/40
1e−4
Reflectivity
1e−5
0.1
Wave vector transfer, Q(A−1)
additional features which enable the value for where the data for 0.36 mM AlCl3 shows the
N to be determined more precisely. At lower interplay between the interference fringes
k values, below the Bragg peak, the interference from the total thickness and the Bragg peak, but
fringes visible arise from the total thickness of the for a structure with only three bilayers.
bilayer structure at the interface. Hence, ensuring For N much larger than as illustrated in Fig. 19,
that the modeling encompasses the Bragg peak this as not generally possible, and the instrumen-
position and the total thickness interference tal resolution damps the interference fringes
fringes N can be determined relatively accurately. sufficiently that they are no longer visible. The
This is also well illustrated in the data in Fig. 16, key model parameters for the d-LAS-6/h-LAS4
and h-LAS-6/d-LAS-4 data are slightly different. Finally, the potential to use such surface struc-
For d-LAS-6/h-LAS4, the bilayer thickness is tures and mixtures as vehicles to optimally
31 to 32 Å and N ranges from 11 to 17, and deliver active benefit agents to surface and inter-
for h-LAS-6/d-LAS-4, the bilayer thickness is faces has been relatively unexplored, and neutron
33 to 35 Å and N is 19. Importantly, the data reflection studies could provide the key insights
show that both LAS isomers contribute to the for such systems.
multilayer structure. The slight difference in the
bilayer thickness reflects the different alkyl chain
asymmetries and lengths of the alkyl chains in the Cross-References
two LAS isomers (where the difference in the mean
length is estimated to be 2.5 Å) and illustrates that ▶ Adsorption Isotherm
the packing differences are accommodated by ▶ Adsorption of Surfactants
changes within the alkyl chain region of the bilayer. ▶ Critical Micelle Concentration
Furthermore, the multilayer structure where the ▶ Detergency
LAS-4 is visible extends further from the surface ▶ Detergents
and is systematically thicker, and this implies ▶ Gibbs Adsorption Isotherm
a variation in the tendency to form the surface ▶ Interaction Parameter
multilayer structures in favor of the LAS-4 isomer. ▶ Interface, Air-Liquid
▶ Mixed Film Theory
▶ Polyelectrolyte-Surfactant Interaction
Summary and Future Prospect ▶ Specific Surface Interactions
▶ Surfactant Adsorption
It has been demonstrated in this entry how neutron ▶ Surfactant Monolayers
reflectivity can be used to probe the adsorption of ▶ Surfactant Self-assembly
mixed multicomponent systems and reveal and ▶ Surfactants
characterize surface structures that are more com- ▶ Wetting
plex than simple monolayers. The power of the
technique when combined with deuterium labeling
offers great potential for the study of such systems
References
in some detail. The approach has already
transformed our understanding of mixed surfactant Abe M, Scamehorn JF (2005) Mixed surfactant systems,
adsorption and of the adsorption of polymer/ vol 124, Surfactant science series. Marcel Dekker,
surfactant mixtures and has highlighted aspects New York
Alargova RG, Petkov JT, Petsev DN (2003) Micellisation
where current theoretical treatments are deficient.
and interfacial properties of alkyl oxyethylene
There remain many more interesting systems and sulphate surfactants in the presence of multivalent S
aspects to pursue. counterions. J Colloid Interface Sci 261:1
In the more complex surface structures, the Bell CG, Breward CJW, Howell PD, Penfold J, Thomas
RK (2007) Mathematical modelling of polymer-
lateral order and organization and the associated
surfactant systems using the law of mass action.
off-specular scattering still remain a largely Langmuir 23:6042
unexplored and unquantified area. New instru- Chen ML, Penfold J, Thomas RK, Smyth TJP, Perfumo A,
mentation and the combination with other com- Marchant R, Banat IM, Stevenson P, Parry A, Tucker I,
Grillo I (2010a) Solution self-assembly and adsorption
plementary techniques herald the potential for at the air-water interface of the monorhamnose and
much progress in these areas. Furthermore, dirhamnose rhamnolipids and their mixtures.
developments in sources and instrumentation Langmuir 26:18281
offer the potential to probe not only equilibrium Chen ML, Penfold J, Thomas RK, Smythe TJP, Perfumo A,
Marchant R, Banat IM, Stevenson P, Parry A, Tucker I,
properties but also the kinetics of formation and
Grillo I (2010b) Mixing behaviour of the biosurfactant,
disassociation of such structures, and this is of rhamnolipd, with a conventional anionic surfactant,
critical importance in the context of applications. sodium dodecylbenzene sulfonate. Langmuir 26:17958
S 1164 Surfactant Molecule
Cooke DJ, Lu JR, Thomas RK, Staples EA, Penfold Purcell IP, Lu JR, Thomas RK, Am H, Penfold J (1998)
J (1998) The interaction between PEO and SDS stud- The adsorption of SDS at the surface of an aqueous
ied by neutron reflectivity. J Phys Chem B 102:4912 solution of PVP studied by neutron reflectivity.
Goddard ED (2001) Polymer/surfactant mixtures: interfa- Langmuir 14:1637
cial aspects. J Colloid Interface Sci 256:228 Scamehorn JF (1988) Mixed surfactant systems, vol 311,
Goddard ED, Ananthapadmanabhan KP (1992) Interaction ACS symposium series. ACS, Washington, DC
of surfactants with polymers and proteins. CRC Press, Staples E, Tucker I, Penfold J, Warren N, Thomas RK,
Boca Raton Taylor DJF (2002) Organisation of polymer-surfactant
Holland PH (1986) Nonideality at the interface in mixed mixtures at the air-water interface: SDS and
micellar systems. Colloid Surf 19:171 poly-dmdaac. Langmuir 18:5147
Holland PM, Rubnigh DN (1990) Cationic surfactants, Taylor DJF, Thomas RK, Hines JD, Humphreys K,
vol 37, Surfactant science series. Marcel Dekker, Penfold J (2002) The adsorption of oppositely charged
New York polyelectrolyte/surfactant mixtures at the air-water
Jones MN (1967) The interaction of SDS with PEO. interface; neutron reflection from dodecyltrimethy-
J Colloid Interface Sci 23:36 lammonium bromide/sodium poly(styrene sulfonate)
Lu JR, Thomas RK, Penfold J (2000) Surfactant layers at and SDS/poly(vinyl pyridinium chloride). Langmuir
the air-water interface, structure and composition. Adv 18:9787
Colloid Interface Sci 84:143 Taylor DJF, Thomas RK, Penfold J (2007) Adsorption of
McGillivray DJ, Thomas RK, Rennie AR, Penfold J, Sivia polymer-surfactant at the air-water interface. Adv
DS (2003) Ordered structures of di-chain cationic Colloid Interface Sci 132:69
surfactants at interfaces. Langmuir 19:7719 Tidswell IM, Ocko BM, Pershan PS, Wasserman SR,
Nikas YF, Puuvada S, Blankschtein D (1992) Surface Whitesides GM, Axe JD (1990) X-ray specular reflec-
tension in aqueous nonionic surfactant mixtures. tion studies of silicon coated by organic monolayers.
Langmuir 8:2680 Phys Rev B 41:1111
Penfold J, Thomas RK (2010) Mixed surfactant adsorp- Tucker I, Penfold J, Thomas RK, Tildesley DT (2009)
tion at the air-water interface. RSC Ann Rep Prog Interplay between the surface adsorption and solution
Chem Sect C 106:14 phase behaviour in di-alkyl chain cationic – nonionic
Penfold J, Staples E, Tucker I, Thomas RK (1998) The surfactant mixtures. Langmuir 25:3924
structure of the mixed nonionic surfactant monolayer Tucker I, Petkov JT, Penfold J, Thomas RK, Petsev DN,
of C12E3 and C12E8 at the air-water interface. J Colloid Dong CC, Golding S, Gibson C, Grillo I (2010) The
Interface Sci 201:223 impact of multivalent counterions, Al3+, on the surface
Penfold J, Staples E, Thompson LJ, Tucker I (1995) The adsorption and self-assembly of the anionic surfactant
composition of nonionic surfactant mixtures at the alkyloxyethylene sulphate and anionic – nonionic
air-water interface as determined by neutron reflectiv- surfactant mixtures. Langmuir 26:16699
ity. Colloid Surf A 102:127 Tucker I, Penfold J, Thomas RK, Dong CC, Golding S,
Penfold J, Tucker I, Thomas RK (2005a) Polyelectrolyte Gibson C, Grillo I (2011) The variation in equilibrium
modified solid surfaces: the consequences for ionic surface adsorption behaviour in alkyl benzene sulphate
and mixed ionic/nonionic surfactant adsorption. solutions using varying electrolyte concentrations.
Langmuir 21:11757 Langmuir 27:6674
Penfold J, Tucker I, Thomas RK, Zhang J (2005b) Adsorp- Zhang XL, Taylor DJF, Thomas RK, Penfold J (2011)
tion of polymer/surfactant mixtures at the air/water Adsorption of polyelectrolyte/surfactant mixtures at
interface: PEI/SDS. Langmuir 21:10061 the air-water interface: modified PEI/SDS. Langmuir
Penfold J, Thomas RK, Taylor DJF (2007a) Polyelectro- 27:2601
lyte-surfactant mixtures at the air-water interface.
Curr Opin Colloid Interface Sci 12:337
Penfold J, Thomas RK, Dong CC, Tucker I, Metcalfe K,
Golding S (2007b) Equilibrium surface adsorption
behaviour in complex anionic – nonionic surfactant
mixtures. Langmuir 23:10140 Surfactant Molecule
Penfold J, Tucker I, Thomas RK, Taylor DJF, Zhang XL,
Bell CG, Breward CJW, Howell PD (2007c) The inter-
Tharwat Tadros
action between sodium alkyl sulphate surfactants and
the oppositely charged polyelectrolyte, poly-dmdaac, Wokingham, Berkshire, UK
at the air-water interface: the role of alkyl chain length
and electrolyte and comparison with theoretical
predictions. Langmuir 23:3128
Penfold J, Thomas RK, Zhang XL, Taylor DJF
Synonyms
(2009) Nature of amine-surfactant interactions at the
air-solution interface. Langmuir 25:3972 Surfactants
Surfactant Monolayers 1165 S
Definition equilibrium, and the rate of exchange between
a surfactant molecule and the micelle may vary
Surface-active agents (usually referred to as sur- by orders of magnitude, depending on the
factants) are amphipathic molecules consisting of structure of the surfactant molecule.
a nonpolar hydrophobic portion, usually
a straight or branched hydrocarbon or fluorocar-
bon chain containing 8–18 carbon atoms, which Cross-References
is attached to a polar or ionic portion (hydro-
philic). The hydrophilic portion can, therefore, ▶ Interfacial Tension
be nonionic, ionic, or zwitterionic, accompanied ▶ Micellization
by counterions in the last two cases. The hydro- ▶ Surface Tension
carbon chain interacts weakly with the water ▶ Surfactants
molecules in an aqueous environment, whereas
the polar or ionic head group interacts strongly
with water molecules via dipole or ion-dipole
interactions. It is this strong interaction with the Surfactant Monolayers
water molecules which renders the surfactant sol-
uble in water. However, the cooperative action of Dieter Vollhardt
dispersion and hydrogen bonding between the Max Planck Institute of Colloids and Interfaces,
water molecules tends to squeeze the hydrocar- Potsdam/Golm, Germany
bon chain out of the water, and hence these chains
are referred to as hydrophobic. The balance
between hydrophobic and hydrophilic part of Synonyms
the molecule gives these systems their special
properties, for example, accumulation at various Monolayer phases; Phase characterization of
interfaces and association in solution (to form amphiphilic monolayers; Phase diagram of Lang-
micelles). The driving force for surfactant muir monolayers; Phase features of insoluble
adsorption is the lowering of the free energy of monolayers; Phase properties of insoluble
the phase boundary. The interfacial free energy monolayers
per unit area is the amount of work required to
expand the interface. This interfacial free energy,
referred to as surface or interfacial tension, g, is Keywords
given in mJm2 or mNm1. Adsorption of sur-
factant molecules at the interface lowers g, and Brewster angle microscopy (BAM); chiral
the higher the surfactant adsorption (i.e., the more discrimination; chiralility; equation of state; S
dense the layer is), the larger the reduction in g. equilibrium; Grazing incidence X-ray diffraction
The degree of surfactant adsorption at the inter- (GIXD); langmuir monolayers; lattice structure;
face depends on surfactant structure and the mesoscopic morphology; molecular recognition;
nature of the two phases that meet the interface. non-equilibrium; phase behaviour; surface
Surface-active agents also aggregate in solution pressure isotherms; thermodynamics
forming micelles. The driving force micelle for-
mation (or micellization) is the reduction of con-
tact between the hydrocarbon chain and water, Definition
thereby reducing the free energy of the system. In
the micelle, the surfactant hydrophobic groups Rapid progress in understanding the molecular
are directed toward the interior of the aggregate, organization of condensed monolayer phases
and the polar head groups are directed toward the is demonstrated. The combination of P-A
solvent. These micelles are in dynamic isotherms with BAM imaging and GIXD
S 1166 Surfactant Monolayers
measurements is optimal for the experimental are decisive for the features of the monolayer.
phase characterization of amphiphilic mono- The fluid/condensed transitions were treated in
layers. The development of new equations of the framework of a quasi-chemical approach,
state allows the theoretical description of the where the mass action law was employed for the
fluid-/condensed-phase transition region decisive description of the monomer/aggregate
for the features of the monolayer. New equilibrium.
mesoscopic length-scale characteristics of mono- The quasi-chemical monomer/aggregate equi-
layer material are available, such as morphology librium model based on a combination of the
and textural features of the domains, equilibrium generalized Volmer’s equation and the equation
and nonequilibrium states, chiral discrimination of state for the case of monomers and large aggre-
effects, and molecular recognition of non-sur- gates (area A < Ac) was developed (Fainerman
face-active species dissolved in the aqueous and Vollhardt 1999) and frequently used to
subphase. Information about 2D lattice structure describe the P-A isotherms of various amphi-
and phase behavior is based on GIXD results. philic monolayers. As shown in numerous
papers, the application of these equations of
state is in good (very often in excellent) agree-
Overview ment with the experimental results obtained for
monolayers of very different amphiphiles (see,
In the two recent decades, rapid progress in the e.g., Vollhardt and Fainerman (2006)).
understanding of the molecular organization of This equation of state for monolayers derived
condensed monolayer phases has been made. in the framework of a quasi-chemical approach
Over many decades in the time before, surface was generalized by addition of a term which takes
tension measurements have been the main source into account the entropy nonideality of mixing of
of information about insoluble monolayers (see monolayer components with different sizes,
also ▶ Interface, Air-liquid). The existence of namely, monomers and clusters of amphiphilic
different condensed monolayer phases could be molecules. The results calculated with this equa-
concluded from the surface pressure–molecular tion of state have shown good agreement with the
area (P-A) isotherms. However, the classical experimental P-A isotherms obtained for various
equations of state allowed only the characteriza- types of amphiphilic monolayers.
tion of the fluid (gaseous, liquid-expanded) state The studies of the phase and structural features
of Langmuir monolayers. The fundamental pro- of different amphiphilic acid amide monolayers
gress in the experimental determination of the demonstrate impressively the existence of
main characteristics of Langmuir monolayers in two condensed phases indicated by the existence
mesoscopic and molecular scale required also of two plateau regions in the P-A isotherms
improved thermodynamic possibilities to at low temperatures. Correspondingly, the
describe the surface pressure–area (P-A) iso- equation of state for monolayers with a fluid
therms. To fill this gap, thermodynamically (G, LE)-/condensed (LC)-phase transition was
based equations of state have been developed further generalized to take into account further
which consider also the aggregation of the mono- (two or more) phase transitions between con-
layer material to the condensed phase. Under densed phases provided that the two-dimensional
optimal conditions, the P-A isotherms show compressibility undergoes a jump at the phase
a sharp break, where the main phase transition transition, whereas within any condensed phase
from a fluid (gaseous (G) or liquid-expanded of the monolayer, this compressibility is propor-
(LE)) to the liquid-condensed (LC) states takes tional to the surface pressure (Fainerman and
place. For the first time, those parts of the P-A Vollhardt 2009). This model used in combination
isotherm can be characterized which represent with the general stability principles allows
the transition region to the condensed phase and successful description of the experimental
surface pressure isotherms with an additional discrimination effects. The chiral nature of the
condensed/condensed-phase transition. enantiomeric and racemic amphiphiles is apparent
It is important to note that all theoretical by characteristic differences in the domain shape
models do not involve any arbitrary parameters. and the sense of curvatures in the textures. Both
All necessary model parameters are either calcu- homo- and heterochiral preferences for chiral
lated from the fluid (LE, G) monolayer branch of interaction have been observed depending on the
the P-A isotherm or evaluated from independent intermolecular interaction of the amphiphilic
GIXD experiments. molecules (Nandi and Vollhardt 2003). In the
The combination of P-A isotherms with Brew- case of strong homochiral preference, spontane-
ster angle microscopy (BAM) imaging and grazing ous separation of racemic mixtures is expected.
incidence X-ray diffraction (GIXD) measurements New information about the properties of
is optimal for the experimental phase characteriza- mesoscopic length-scale aggregates was obtained
tion of amphiphilic monolayers. In the two recent about chiral discrimination effects.
decades, a number of experimental observations A large number of studies of interfacial molec-
have confirmed that the morphology of the ular recognition made self-assembly at the inter-
condensed-phase domains which are formed in faces under formation of supramolecular
the fluid-/condensed-phase transition region of structures to a separate research field (Vollhardt
the P-A isotherm provides mesoscopic length- et al. 2011). The topic of molecular recognition is
scale information of the monolayer material. The as broad as its abundance in nature. A variety of
advantages of utilizing Langmuir monolayers as artificial host–guest assemblies has been planned
models result mainly from the easiness with which to understand molecular recognition systems
experimental variables can be manipulated. with new properties and functions. The air/water
An unsuspected morphological variety of interface represents a unique hetero-dielectric
condensed-phase domains has been found that medium where two phases which have
were formed in Langmuir monolayers (Vollhardt completely different dielectric constants are in
1996) and in adequate molecular recognition contact. In comparison to the aqueous bulk solu-
systems (Vollhardt 2005). In all these systems, tion, the efficiency of molecular recognition is
shape and size of the monolayer domains as well essentially enhanced at the air/water interface.
as their textural features depend sensitively on Molecular recognition at the air–water interface
the chemical structure of the amphiphiles by assembly between monolayer components and
and the monolayer state determined mainly by guest materials dissolved in the aqueous subphase
surface pressure and temperature. The inner is one of the most important research topics in this
textures of domains can be explained by field. At interfacial molecular recognition, the
consideration of their geometry and lattice in Langmuir monolayer formed by receptor mole-
dependence of the tilt angle of the alkyl chains cules binds non-surface-active target molecules S
and can be described by a geometric concept on dissolved in the aqueous subphase by complemen-
the basis of the molecular packing. tary hydrogen bonding and/or acid–base interac-
The growth rate of the condensed-phase tion (Vollhardt 2005). A selected amphiphilic
domains can be so rapid that metastable melamine type represents a good candidate as
nonequilibrium structures of fractal-like or host monolayers for molecular recognition of
dendritic character are formed. In particular, in dissolved components on the basis of complemen-
the presence of strong hydrogen-bonding tary hydrogen bonding. A recognition system on
systems, the nonequilibrium state can be the basis of acid–base interaction is represented
preserved in a frozen state. The high sensitivity by cationic heptadecyl-benzamidinium chloride
of the morphological features of the condensed- monolayer as host component and sodium
phase domains to the specific stereochemistry has phenylacetate dissolved in the aqueous subphase
been visualized by impressive chiral as guest component.
The introduction of synchrotron X-ray the fluid (gaseous, liquid-expanded) state. To fill
diffraction at grazing incidence (GIXD) has made this gap, thermodynamically based equations of
it feasible to determine the two-dimensional lattice state were developed which consider also the
structure and phase behavior of amphiphilic aggregation of the monolayer material to
monolayers. It has been demonstrated that not condensed phase. Thus, for the first time, those
only in Langmuir monolayers but even in parts of the P-A isotherm can be characterized
the condensed phase of an adsorbed monolayer, which represent the transition region to the
the mesoscopic textural features can be related to condensed phase and are decisive for the features
the two-dimensional lattice structure. of the monolayer. As shown in numerous papers,
To date, numerous new interesting in situ the application of these equations of state are in
results about morphological differences and good (very often in excellent) agreement with the
two-dimensional lattice structures of condensed- experimental results obtained for monolayers of
phase domains of Langmuir monolayers cannot very different amphiphiles.
be explained by macroscopic theories and needed Under optimal experimental conditions, the
new information about the detailed orientation P-A isotherms of Langmuir monolayers show
and distance dependence of the intermolecular a sharp break, where the main phase transition
interaction between molecules within the aggre- from a fluid (gaseous (G) or liquid-expanded
gates. First new bottom-up studies have been (LE)) to the liquid-condensed (LC) states occurs.
directed toward understanding the driving forces Several attempts were made to give an explana-
for the aggregation (condensation) processes of tion for the nonhorizontal shape of P-A
amphiphilic monolayers. A first step to gain isotherms in 2D transition region. In these stud-
insight into the role of the interactions involved ies, fluid/condensed transitions were treated in
is to correlate the main characteristics of the framework of a quasi-chemical approach,
mesoscopic length-scale aggregates of mimetic where the mass action law was employed for the
systems with the molecular structure and polar- description of the monomer/aggregate equilib-
ity. Different types of interactions including rium. The theoretical combination on the
hydrogen bonding have been successfully con- basis of a quasi-chemical approach resulted in
sidered using semiempirical quantum chemical rather simple equations which contain a few
methods. parameters only.
The equation of state for monolayers in the
fluid (G or LE) state is represented by the
Theory Volmer-type equation:
Equations of State for P-A Isotherms with kT

P¼ Pcoh ; (1)
One Condensed Phase mðA oÞ
Since the early periods of research on Langmuir
monolayers over many decades, surface-tension where P is the surface pressure; k is the
measurements have been the main source of Boltzmann constant; T is temperature; o is the
information about insoluble monolayers. The partial molecular area for monomers (or the lim-
existence of different condensed monolayer iting area of molecule in the gaseous state); A is
phases could be concluded from the P-A the area per molecule; Pcoh is the cohesion pres-
isotherms. In the past two decades, the fundamen- sure, which accounts for the intermolecular inter-
tal progress in the experimental determination of action under the condition of a constant value of
the main characteristics of Langmuir monolayers the enthalpy activity coefficient in the mono-
in mesoscopic and molecular scale required also layer; and the m value allows for the association
improved thermodynamic possibilities to (m > 1) or dissociation (m < 1) degree of the
describe the P-A isotherms, as the classical equa- amphiphilic molecules in the monolayer, or the
tions of state allowed only the characterization of size of these molecules.
Furthermore, the generalized Volmer’s equa- the limiting area per molecule in the gaseous state.
tion which describes the multicomponent insolu- This fact is accounted for by the parameter e.
ble monolayer of monomer or aggregates was However, to reproduce the experimental behavior
derived: of ocl vs P for a number of amphiphiles, the e
P value should consist of two terms, e ¼ e0 þ P:
Gi
P ¼ RT P
i
B; (2) ocl ¼ oð1 eÞ ¼ oð1 e0 PÞ; (6)
1 Gi oi
i
where e0 is the relative jump of the area per mole-
where R is the gas law constant, Gi is the adsorption cule during phase transition and is a
(surface concentration) value of the ith component two-dimensional compressibility (or intrinsic com-
or state, and B is the integral Volmer constant. pressibility) coefficient of the condensed mono-
For the case of monomers and large aggre- layer. For example, the intrinsic compressibility
gates (area A < Ac), the equation of state was can be physically understood by the change of the
obtained using a combination of the generalized tilt angle of molecules upon surface-layer
Volmer’s equation (Eq. 2) and the quasi-chemical compression, accompanied by a corresponding
monomers/aggregates equilibrium model. It was increase in the surface-layer thickness.
assumed that the area per one molecule in the The quasi-chemical monomers/aggregates
aggregate can differ from the area per free mono- equilibrium model based on a combination of
mer molecule. The condition which describes the the generalized Volmer’s equation (Eq. 2) and
equilibrium between the monomers and aggre- the equation of state for the case of monomers
gates was formulated considering the free surface and large aggregates (area A < Ac) (Eq. 3) has
existing in the monolayer. The following equa- been frequently used to describe the surface pres-
tion of state derived in the framework of a quasi- sure–area (P-A) isotherms of various amphi-
chemical approach for bimodal distribution (large philic monolayers. For example, for the
clusters and monomers) was obtained (Fainerman thermodynamic characterization of the mono-
and Vollhardt 1999): layers, the P-A isotherms of 2C11H23-melamine,
which behaves thermodynamically as a usual
kTab amphiphilic monolayer, have been studied for
P¼ Pcoh : (3)
mðA o½1 þ eðab 1ÞÞ different temperatures in the accessible range
(Vollhardt et al. 2005). Similarly, as shown in
In this equation, the parameter a expresses the Fig. 1, for taking into account the contribution
dependence of the aggregation constant on the of the entropy nonideality caused by the mixing
surface pressure: of monomers and clusters, the quasi-chemical
monomers/aggregates equilibrium model S
A P Pc (Fainerman and Vollhardt 2008) describes per-
a¼ exp e o ; (4)
Ac kT fectly the experimental P-A isotherms.
For the calculation of the enthalpy of two-
and the parameter b is the fraction of the mono- dimensional phase transition DH0, a rigorous
layer free from aggregates: equation was derived and applied to the experi-
mental results (Vollhardt and Fainerman 2006):
b ¼ 1 þ oð1 eÞða 1Þ=A; (5)
RT 2 dPc =dT RT 2 dx
where e ¼ 1– ocl/o, ocl is the area per monomer DH 0 ¼ ¼ ; (7)
in a cluster and Ac the molecular area which x dPc =dx x dT
corresponds to the onset of the phase transition
(i.e., at P ¼ Pc). If Ac < 2o, then b ¼ a. The area where x ¼ ocl/Ac and R is the gas constant. The
per molecule in the cluster can be different from values of x ¼ ocl/Ac can be calculated using the
Surfactant Monolayers, 40
Fig. 1 Experimental
(points) and theoretical
(lines) P-A isotherms of
2C11H23-melamine 29.2, 25.4, 23.1, 20.0, 15.8, 12.5 °C
monolayers spread on 30
Surface Pressure, mN/m

water measured in the
temperature range between
12.5 C and 29.2 C. The
values of the model
20
parameters of Eq. 12 are
listed in Table 1
10
0
0.35 0.4 0.45 0.5 0.55 0.6 0.65
Area, nm2
25
20
15
πc, mN/m
10 C10
Surfactant Monolayers, C11
Fig. 2 Temperature
5 C12
dependence of the phase
transition pressure Pc of
the monolayers of 2C10H21-
0
melamine, 2C11H23-
0 10 20 30 40 50
melamine, and 2C12H25-
melamine spread on water T, °C
experimental values of Ac and o ¼ 0.36 nm2 for calculation of DH0 (see Table 1), a constant
2C11H23-melamine and the different tempera- dx/dT value for each melamine homologue can
tures. x depends linearly from T. This can be used.
be clearly seen by the presentation of the temper- In the equation of state (Eq. 3), the contribu-
ature dependence of the phase transition point Pc tion of the entropy nonideality, caused by the
(the break point in the P-A isotherm at onset of mixing of monomers and clusters, to the surface
the phase transition) in Fig. 2. For these homolo- pressure can be taken into account. Assuming
gous melamine amphiphiles which have that the surface layer consists of the mixture
similar dx/dT values, the fluid-/condensed-phase of small (molecular area o1) and large (molecular
coexistence region is in the measurable tempera- area o2, o2 > o1) molecules, the contribution
ture range (see Fig. 2). Therefore, for the arisen from the nonideal entropy of mixing of
Surfactant Monolayers, Table 1 Parameters of Eqs. 1, 12, and 3 calculated for 2C11H23-melamine monolayers at
different temperatures
Temperature, C 12.5 15.8 20.0 23.1 25.4 29.2
o, nm2 0.40/0.36 0.39/0.36 0.4/0.36 0.39/0.36 0.38/0.36 0.38/0.36
Ac, nm2 0.58 0.55 0.52 0.51/0.5 0.5 0.48
Pcoh, mN/m 9.6/16.2 12.9/15.7 12.2/16.6 12.9/16.6 14.0/16.6 12.8/15.6
e0 0.19 0.19/0.18 0.22 0.23/0.24 0.22 0.21
m 2.0/1.0 1.4/1.02 1.53/0.97 1.33/0.92 1.13/0.91 1.15/0.91
DH0, kJ/mol 8.9 8.8 8.5 8.4 8.2 8.1
molecules with different sizes on the basis of the aggregates is equal to b ¼ A/Ac, for the 2D
Flory–Huggins theory was obtained as phase transition region (A < Ac), one obtains the
expression for the negative nonideal entropy con-
tribution to surface pressure:
kT o1
DP ¼ y2 1 ; (8)
o1 o2 kT A2
DP ¼ 1 2 : (10)
A Ac
where y2 ¼ o2/A is the monolayer coverage by
large molecules. This results in a surface pressure The theoretical dependences Eq. 9 correspond
decrease of the monolayer. For a molecular mix- to the experimental data. It was shown that an
ture with equal sizes, that means for monolayers increase of Ac (decrease of yc) results in
which are ideal with respect to entropy, Eq. 8 a decrease of the isotherm slope in the 2D transi-
becomes zero. tion region. Taking into account Eqs. 4-6,
The nonideal entropy contribution in Eq. 7 can one obtains the most detailed expression for
be extended to the case of phase transition from DP if e > 0:
the fluid (G, LE) to the condensed (LC) state of
monolayers if one takes into account that the area kT
of clusters nocl (n is the aggregation number of DP ¼ ð1 abÞ: (11)
A
a cluster) essentially exceeds the area of the
monomer, nocl > > o. Then, it follows from Introducing Eq. 11 into Eq. 3, the equation of
Eq. 8 that state for the surface layer in the fluid (G, LE)/
condensed (LC) transition region which takes
into account the entropy nonideality of mixing
kT kT A
DP ¼ ½y yc ¼ 1 ; (9) of monomers and clusters takes the form S
o A Ac
(Fainerman and Vollhardt 2008):
where yc ¼ o/Ac and y ¼ o/A are the monolayer

1 kTab kT
coverages in the phase transition point which P¼ ð1 abÞ Pcoh :
m A o½1 þ eðab 1Þ A
corresponds to the initial point of the fluid/con-
densed transition and at A < Ac, respectively. In (12)
the fluid/condensed transition region, the fraction
of the surface covered by monomers does The advantages of the proposed model
not remain constant. Rather, it depends on the according to Eq. 12 can be demonstrated using the
fraction of the surface covered by aggregates experimental results for 2,4-di(n-undecylamino)-
(cf. Eq. 5). Therefore, Eq. 9 should be modified 6-amino-1,3,5-triazine (2C11H23-melamine)
accordingly. Taking into account that, for e ¼ 0, (Vollhardt et al. 2005). In Fig. 1, six experimental
the fraction of the monolayer free from P-A isotherms of 2C11H23-melamine monolayers
Surfactant Monolayers,
Fig. 3 P-A isotherms of
TDHPA monolayers
measured in the
5 C and 25 C. The arrow
indicates the transition
point between two
condensed phases. The
formula of the TDHPA
molecule (R ¼ C14H29) is
inserted
measured between 12.5 C and 29.2 C are amide (TDHPA) monolayers between 5 C and
presented by symbols. It can be seen that over the 20 C with the inserted chemical formula of the
whole temperature range studied, 2C11H23- TDHPA amphiphile. Two pronounced phase
melamine forms a two-phase coexistence region, transitions which are indicated by two plateau
whose extent decreases clearly as the temperature regions at temperatures of 15 C can be clearly
increases. The lines represent the theoretical curves seen. At large areas per molecule, the transition
obtained on the basis of the equations of state of from the low-density fluid phase to a condensed
Eq. 1 (A > Ac) and Eq. 12 (A < Ac), respectively. phase is indicated by a break in the slope of the
The corresponding parameters of the isotherms are isotherm followed by the large plateau region. At
listed in Table 1. The values in the numerator small areas per molecule and a surface pressure
correspond to the value calculated from Eqs. 1 of 16 mN/m, the isotherms show, after a second
and 12. break in the slope of the isotherm, a second small
The calculated values show that the aggrega- plateau with a width of 0.02 nm2. The phase
tion number for the small aggregates m which transition between the two condensed phases is
exist in the fluid state at A > Ac decreases as the also of first order as a break in the slope of the
temperature increases. The values calculated on isotherm followed by a plateau region indicates
this basis confirm a good agreement with the a first-order phase transition.
experimental data. The equation of state valid for bimodal distri-
Theoretical description of P-A isotherms bution (aggregated molecules/monomers) (Eq. 3)
in the case of more than one condensed phase applies to the fluid-/condensed-phase coexis-
Equation 3 for monolayers with a fluid (G, LE)-/ tence. It is shown in Eq. 6 that the e value should
condensed (LC)-phase transition was further consist of two terms, e ¼ e0 + P, where e0 is the
generalized to take into account further phase relative jump of the area per molecule during the
transitions between condensed phases and phase transition and is a relative two-
successfully used to describe the experimental dimensional compressibility of the condensed
surface pressure isotherms with an additional con- monolayer. The application of this equation of
densed/condensed-phase transition (Fainerman state is capable of providing a thermodynamic
and Vollhardt 2003). description of P-A isotherms indicating two or
The studies of the phase and structural features more phase transitions (Fainerman and Vollhardt
of different amphiphilic acid amide monolayers 2009). The two-dimensional compressibility
demonstrate impressively the existence of two undergoes a jump at the second-phase transition,
condensed phases (Vollhardt and Fainerman whereas within one of the two condensed phases,
2006). For example, Fig. 3 shows the P-A iso- this compressibility is proportional to the surface
therms of N-tetradecyl-b-hydroxy-propionic acid pressure.
40 Surfactant Monolayers, Table 2 Characteristics of the
surface tension data TDHPA monolayer at 10 C
Surface Pressure [mN/m]
GIXD-data
30 Curve 1 2
o, nm2 0.247 0.247
II
20 Ac, nm2 0.45 0.37
I , m/mN 0.004 0.001
10
e0 0.06 0.2
m 1.4 1.4
0
0.2 0.3 0.4 0.5 0.6 Pcoh 8.06 8.06
Area [nm2/molecule]
0.26
0.25
Axy [nm2/molecule]
0.24
1 molecular area (Axy) decreases also in the higher
0.23 surface pressure range but with a slope 4 times
0.22 lower than in the region of the first condensed
0.21
2 phase. At a surface pressure value of 16 mN/m,
which corresponds to the second-phase transition
0.2
0 10 20 30 40 point in Fig. 4, above, the area per TDHPA mol-
Π [mN/m] ecule obtained from the lattice structure
(intermolecular distances, polar tilt angle)
Surfactant Monolayers, Fig. 4 Experimental P-A iso- changes drastically, shown by the arrow in
therm of the TDHPA monolayer at T ¼ 10 C (cf. Fig. 3).
Dark points correspond to the values of pressure and area Fig. 4, bottom. It follows from the data of the
per molecule for which the GIXD experiments were THPAA monolayer at 10 C (see Fig. 4, bottom)
performed. The arrows indicate the two-phase transitions that o ¼ 0.247 nm2 (intersection point of the
(above). Dependence of the TDHPA molecular area par- straight line 1 with the ordinate), 1 ¼ 0.004 m/
allel to the surface on the surface pressure obtained from
the GIXD data (bottom) mN, 2 ¼ 0.001 m/mN (slopes of the straight
lines 1 and 2, respectively), and the variation of
e0 at the second-phase transition of 16 mN/m
(jump between lines 1 and 2) De0 ¼ 0.14. The
To demonstrate the procedure of the thermo- values of m and Pcoh, that enter Eqs. 3 and 5,
dynamic analysis for the case of two-phase tran- were calculated from the fitting of the isotherm
sitions, the P-A isotherm of the TDHPA branch at A > Ac (Fig. 4, above). The parameter
monolayer measured at 10 C was selected in values that describe the P-A isotherm both in the
Fig. 4, above, where the positions of the gaseous monolayer region and in the phase coex- S
two-phase transitions are marked by arrows. istence region at A values after the first-phase
The dark points correspond to the values of the transition point are listed in the column for
surface pressure and area per molecule, for curve 1 of Table 2, and the corresponding theo-
which GIXD experiments were performed. retical curve 1 is presented in Fig. 5.
The dependence of the Axy values of the projected The analysis of the theoretical curves shown in
area of the TDHPA molecule parallel to the sur- Fig. 5 is based on the general principle that the
face on the surface pressure is shown in Fig. 4, behavior of any thermodynamic system is deter-
bottom. In the surface pressure range between the mined by the stability condition for spread or
two-phase transitions shown in Fig. 4, bottom adsorbed monolayers: dP/dA 0. Consequently,
(5.5–16 mN/m), the area per molecule in the the dependence of surface pressure on the area
condensed state (Axy) decreases linearly with the per molecule should possess a minimum slope to
increase in the surface pressure. Above the sec- be thermodynamically stable. The thermodynam-
ond-phase transition (16 mN/m and higher), the ically stable course of the TDHPA isotherm is
Surfactant Monolayers, 30
Fig. 5 Theoretical P-A
isotherms 1 and 2 for the
TDHPA monolayer

calculated from Eqs. 1, 3,
and 4 with the parameters 20
listed in Table 2. For C
detailed discussion, see the 2 B
text
10
1 A
0
0.2 0.3 0.4 0.5 0.6
Area [nm2/molecule]
25
20
II
15
Fig. 6 Comparison 10
between the experimental I
(points) and theoretical
(line) P-A isotherms of the
5
TDHPA monolayer using
the theoretical model
for two-phase transitions.
0
I – first-phase transition
0.1 0.3 0.5 0.7
point, II – second-phase
transition point Area [nm2/molecule]
shown by the arrows in Fig. 5. The arrow on the Then, the system returns to curve 1. In point C,
right-hand side of point A corresponds to the fluid the system proceeds also along the branch
(LE) state of the monolayer. Starting in point A, corresponding to the condensed state with the
generally two routes are possible at further com- larger density (e0 ¼ 0.2) which possesses the
pression. According to the stability condition, the lower slope. However, at very rapid compression,
system follows the route characterized by the the route along the more steep line 1 can be also
lower slope of the isotherm along curve 1 across realized by instable monolayers (monolayers
the fluid-/condensed-phase coexistence region. In oversaturated with the condensed phase with the
the case that the compression of the monolayer is lower density).
very rapid, a thermodynamically unstable state In Fig. 6, the thermodynamically stable P-A
can arise by oversaturation. Then, when reaching isotherm (Fig. 5) obtained by the proposed model
point A, the system continues to proceed along in combination with the general stability princi-
curve 2 (the state of the monolayer oversaturated ple is compared with the experimental TDHPA
with monomers) up to point B which occurs due isotherm presented in Fig. 4 and shows a satisfac-
to the formation of the second condensed phase. tory agreement with the experiment.
Fig. 7 Formation of
Langmuir monolayers by
spreading
It is important to note that the theoretical incidence X-ray diffraction (GIXD) measure-
models do not involve any arbitrary parameters. ments is optimal for the experimental phase char-
All necessary model parameters are either calcu- acterization of amphiphilic monolayers. In the
lated from the fluid (LE, G) monolayer branch of two recent decades, a number of experimental
the P-A isotherm or evaluated from independent observations have confirmed that the morphology
GIXD experiments. The generalization of the of the condensed-phase domains which are
equation of states presents a first possibility to formed in the fluid-/condensed-phase transition
thermodynamically describe P-A isotherms for region of the P-A isotherm provides mesoscopic
the case that two or more phase transitions in the length-scale information of the monolayer
monolayer take place, provided that the two- material. The advantages of utilizing Langmuir
dimensional compressibility undergoes a jump monolayers as models result mainly from the
at the phase transition, whereas within any con- easiness with which experimental variables can
densed phase of the monolayer, the compressibil- be manipulated. Previously, the shape of the P-A
ity is proportional to the surface pressure. isotherms was used to obtain information about
the phase behavior of the monolayers. However,
measurements of P-A isotherms provide thermo-
Methods dynamic results, and hence, the manifestation of
structural effects in terms of interaction at
A simple and fundamental possibility to form mesoscopic and molecular level is limited on
Langmuir monolayers at the air/water interface this basis.
is represented by a well-known procedure (see
also ▶ Interface, Air-Liquid). In this case, long- Surface Pressure Techniques S
chain amphiphiles, insoluble in water, are Thermodynamic conclusions on the phase behav-
dissolved in an organic solvent and are directly ior of amphiphilic monolayers can be drawn con-
spread at the aqueous surface. Figure 7 shows sidering the P-A isotherms at different
schematically the formation of Langmuir mono- temperatures. The P-A isotherms are usually
layers by spreading. The air/water interface rep- recorded using a computer-interfaced film bal-
resents a unique hetero-dielectric medium where ance with a Wilhelmy-type pressure measuring
two phases which have completely different system. According to the Wilhelmy method, the
dielectric constants are in contact. This interfacial surface pressure can be measured with
distance possesses a molecularly flat plane a roughened glass plate or a filter paper plate
although in a dynamic state. with an accuracy of 0.1 mN·m1 and the area
The combination of surface pressure–molecu- per molecule with 0.5 Å2.
lar area (P-A) isotherms with Brewster angle A very generalized P-A isotherm (Fig. 8) is
microscopy (BAM) imaging and grazing helpful to understand the phase behavior and the
Fig. 8 Generalized P-A
isotherm
two-dimensional morphology of amphiphilic isotherm type T1. In this case, the P-A isotherm
monolayers (Vollhardt 1996). Three characteris- does not allow the exact determination of the
tic types of generalized P-A isotherms represen- first-order phase transition point. Already after
tative for a certain temperature region (T1 < T2 < spreading, the monolayer is in the state of fluid-/
T3) can be measured. Only isotherm type T2 is condensed-phase coexistence so that regularly
suited for the determination of the mesoscopic shaped domains cannot be formed by nucleation
domain characteristics of the monolayer. and growth of the condensed phase.
Neglecting details, generally two single phases On the other hand, at temperatures high enough,
of Langmuir monolayers can be distinguished, a continuous increase of the surface pressure with
namely, the fluid (gaseous, liquid-expanded decreasing area per molecule suggests that for this
(LE)) and condensed (liquid-condensed (LC), isotherm type T3, phase transition to the condensed
solid (S)) phases. The experimental P-A iso- phases does not occur over the whole area range.
therms can indicate further condensed/con- The dashed border line with an apex for the fluid/
densed-phase transition by a small plateau condensed coexistence region indicates that with
region (first order) or a kink in the steep part of decreasing temperature, the surface pressure of the
the isotherm. fluid/condensed coexistence also decreases, but
The most interesting type T2 of P-A isotherm the area range for the coexisting phases becomes
shows a nonhorizontal plateau region after more extended. This third isotherm type T3 is not
a break point in the isotherm characteristic for of interest for studies of the condensed-phase char-
the main phase transition of first order from the acteristics of monolayer.
fluid to the condensed phase. The plateau repre-
sents the main fluid-/condensed-phase transition Optical Methods: Brewster Angle Microscopy
region. The lower the temperature, the lower the (BAM) and Fluorescence Microscopy
surface pressure of the plateau and the more Knowledge of the morphology of condensed-
extended is the molecular area of the plateau. phase domains formed after the main phase tran-
At temperatures low enough and with increas- sition in the two-phase coexistence region of
ing alkyl chain length of the amphiphiles, the Langmuir monolayers progressed rapidly with
two-phase coexistence region exists already at the development of two highly sensitive imaging
a surface pressure P 0, that is, this is the general techniques, namely, fluorescence microscopy
Fig. 9 Physical principle
of Brewster angle
microscopy
Fig. 10 Schematic
diagram of the
experimental setup for
combination of surface
pressure measurements and
Brewster angle microscopy S
and BAM. Fluorescence microscopy, developed the shape of the condensed-phase domains and
some years before BAM, yielded first images their inner texture can be visualized. In a typical
which have unequivocally shown the coexistence experimental setup, a Brewster angle microscope
of two phases in the “plateau” region of the P-A is mounted to a computer-interfaced film balance
isotherm. (Fig. 10). The microscope is equipped with a spe-
The principle of BAM is based on the fact that cial scanning technique for providing sharp
at the Brewster angle of incidence, a parallel (p) images. To obtain BAM images real in scale
polarized laser beam has zero reflectance, and the and angle, the CCD sensor of the camera is tilted
presence of a condensed monolayer phase leads according to the Scheimpflug condition. The
to a change in the refractive index and thus to application of a green laser allows a resolution
a measurable change in reflectivity (Fig. 9). Both of the BAM images of 4 mm.
BAM studies of Langmuir and also Gibbs

monolayers revealed an enormous variety of
domains in the condensed phase and allow direct
visualization of the domain shape and inner tex-
ture (Vollhardt 2006). The structures of the
corresponding molecules have profound influ-
ence in driving these domain shapes as well as
the molecular orientation within a particular
domain. The texture gives a direct indication
about the orientation of the average molecular
orientation in the monolayer. However,
a characterization of the orientational order in
a condensed phase requires that the observed
Surfactant Monolayers, Fig. 11 Principles of GIXD
textures occur under well-defined conditions.
experiments
BAM can also distinguish between the jumps of
orientations and a continuous change in orienta-
tion. Careful observation must be made whether
the lines that separate the dark and bright areas The more ordered the monolayer phase, the lower
remain at a fixed position or are shifted, respec- the solubility. Consequently, the contrast arises
tively, when the analyzer is rotated. Often, this from the different solubility of the probe in the
can be better achieved observing a sequence of two phases. In polarized fluorescence micros-
images using a video than observing single copy, the arising light is polarized. Then, the
images. It may be noted that external probes are fluorescence emission depends on the orientation
not used in BAM technique, and it is therefore of the probe so that images of the molecular tilt
advantageous compared to the fluorescence tech- orientation in monolayers can be obtained. This
nique, which requires dye molecules as probe to technique has two drawbacks. First, in some
investigate the monolayer morphology. Thus, cases, the addition of trace components to the
careful investigation using BAM can reveal system, that is, also fluorescent probes, can con-
many microscopic details of the monolayer with- siderably modify its properties. Secondly, crys-
out any external perturbation (Vollhardt 1996). talline and highly ordered phases reject the
In fluorescence microscopy, an insoluble fluo- fluorescent molecules nearly completely, and
rescent amphiphilic marker must be added to the consequently, their texture is difficult to observe.
monolayer material which is excited using
a high-pressure mercury lamp. Then, the mono- Grazing Incidence X-ray Diffraction (GIXD)
layer state is observed with a microscope GIXD has been established as an effective
mounted with a camera. A variety of probes method for studying the lattice structures in con-
have been used. Few commonly used probes densed monolayers of amphiphiles at the air–
are, for example, 4-(hexadecylamino)-7- water interface. The GIXD experiments were
nitrobenz-oxa-1,3-diazole; L-a-phosphatidylcho- performed at the BW1 beamline, HASYLAB,
line-b-(NBD-aminohexyl)-g-palmitoyl; sn-1,2 DESY (Hamburg, Germany).
dipalmitoyl-3-glycerophosphatidyl ethanolamine In the experimental setup at BW1, a Langmuir
(DPPE) sulforhodamine; dipalmitoyl nitroben- film balance equipped with a single movable
zoxadiazole phosphatidylethanolamine barrier and a Wilhelmy plate for monitoring the
(DP-NBD-PE); and 1-palmitoyl-2-{2-[(7-nitro- lateral pressure were placed in a hermetically
2-1,3-benzoxadiazole-4-yl) amino] dodecanoyl} closed container filled with helium (Fig. 11).
phosphatidylcholine (NBD-PC). When mono- A monochromatic X-ray beam (l ¼ 1.304 Å)
layer phases coexist, the solubility of the fluores- struck the water surface at a grazing incidence
cent probe depends on the ordering of the phases. angle ai ¼ 0.85ac (ac ¼ 0.13 ) and illuminated
roughly 100 mm2 of the monolayer surface (with
2-mm width and 50-mm length of the footprint
area). During the course of an experiment
through a slow lateral movement of the trough,
the X-ray footprint moved across the monolayer
to avoid sample damage by the strong X-ray
beam. A linear position-sensitive detector (PSD,
OEM-100-M, Braun, Garching, Germany) mea-
sured the diffracted signal and was rotated to scan
the in-plane Qxy component values of the scatter-
ing vector. The vertical channels of the PSD
measured the out-of-plane Qz component of the
scattering vector between 0.0 and 0.85 Å1. The
diffraction data consisted of Bragg peaks at diag-
nostic Qxy values. These peaks were obtained by
summing the diffracted intensity at each in-plane
Qxy value over a defined vertical angle or Qz
window. The in-plane lattice repeat distances
d of the ordered structures in the monolayer
were calculated from the Bragg peak positions: Surfactant Monolayers, Fig. 12 Domain growth
d ¼ 2p/Qxy. To access the extent of the crystalline along the compression isotherm of a 1-monopalmitoyl-
order in the monolayer, the in-plane coherence rac-glycerol monolayer at 23 C
length, Lxy, was approximated from the full width
at half maximum (fwhm) of the Bragg peaks region of the P-A isotherm. The observed mor-
using Lxy 0.9 (2p)/fwhm (Qxy). The diffracted phological variety results from the complicated
intensity normal to the interface was integrated interplay between microscopic interfacial
over the Qxy window containing the diffraction dynamics and external driving forces.
peak to calculate the corresponding Bragg rod. On the other hand, under the conditions where
The thickness of the monolayer was estimated isotherms of the low-temperature (T1) type exist,
from the fwhm of the Bragg rod using 0.9 (2p)/ irregularly shaped condensed-phase domains are
fwhm (Qz). More experimental details are formed depending on the conditions of spreading
described in the literature. and monolayer preparation. Regular textures can-
not be generated from the irregular textures by
simple compression; rather, they must be
Characterization transformed in the fluid state by increasing the S
temperature. Then, well-defined condensed-
Effect of Monolayer Conditions on the phase domains are reformed by subsequent
Domain Morphology cooling.
The best developed condensed-phase domains Under the conditions of compression or
are formed after the main phase transition point decompression in the region A Ac of the P-A
within the two-phase coexistence region under isotherm of T2 type, the domain shape changes
the conditions of P-A isotherms of the T2 type. with progressive compression. This is demon-
At compression to smaller area values then that of strated by the morphological changes of an alkyl
the main phase transition point Ac, first small monoglycerolester monolayer (Fig. 12). The P-A
condensed-phase domains are formed following isotherm of 1-monopalmitoyl-rac-glycerol
the mechanisms of nucleation and growth. The monolayers at 23 C having a well-developed
main mesoscopic domain characteristics are “plateau” region shows after the break point the
developed in the fluid/condensed transition first condensed-phase domains which grow up to
a size between 0.5 and 1.0 mm at slow compres- new circular condensed-phase domains are
sion rates. Number and size of the domains formed within the holes (Fig. 13, bottom).
increase with further compression so that they
touch each other at the end of the plateau region. Nonequilibrium Textures
Then, accompanied by the further increase of Recent BAM studies of various chemical struc-
the surface pressure, the domains impinge ture amphiphilic monolayers have shown that
increasingly on each other. This leads to a more different nonequilibrium structures of fractal-
dense packing and deformation of the domains, like or dendritic character can be visualized
and the gaps between the domains are filled. (Vollhardt 1996). Metastable microstructures
Finally, the condensed-phase domains are can be formed far from equilibrium if the growth
packed so tightly that a complete state is formed kinetics is so rapid that the stable equilibrium
by the condensed phase without any gaps. phase has not enough time to reach its lowest
However, the pronounced inner texture of the energy state on microscopic level (see also
monoglycerolester domains becomes preserved ▶ Nucleation). The relaxation times of the
essentially up to the monolayer collapse. As nonequilibrium textures of the various types of
inner textures indicate molecular tilt, it can be amphiphiles are very different. In particular, the
concluded that in the present case of presence of strong hydrogen-bonding systems
monoglycerolester monolayers, the molecules preserves the nonequilibrium state in a frozen
are tilted up to the collapse. The morphology of state. Whereas fatty acid ethyl esters relax within
the monolayer at decompression depends largely a few minutes or seconds, a number of phospho-
on the state of the compressed monolayer before lipids need more than 1 h or most of the long-
starting the decompression. If the elements of chain amides stay in the dendritic state. Chiral
the original ordering are preserved, such discrimination effects in the nonequilibrium con-
as shown for 1-monopalmitoyl-rac-glycerol densed-phase structures of phospholipid mono-
monolayers, decompression of the monolayer layers are discussed in the section Chiral
leads again to the preceding domain structure. Discrimination Effects.
In other amphiphilic monolayers, like those of Different stages of supersaturation produced by
long-chain alcohols (C14, C16), the tilted alkyl high or nearly jump-like compression rates can
chains of the 2D condensed phase are erected initiate the formation of fractal-like patterns of
into a position perpendicular to the surface at dense branching far from equilibrium. For exam-
compression to high surface pressures. Then, ple, 1-monopalmitoyl-rac-glycerol monolayers
a homogeneous solid state without microscopic develop a broad spectrum of nonequilibrium
defects can be formed. Tetradecanol monolayers states: fractal-like domain patterns at very high
develop circular domains in the main phase compression rates, dendritic-shaped domains
transition region at compression (Fig. 13, top). at medium compression rates, and circular equi-
At compression to high surface pressures librium shapes at very low compression rates.
(>25 mN/m), the domains coalesce to a A sequence of relaxing condensed-phase patterns
homogeneously reflecting condensed monolayer of 1-monopalmitoyl-rac-glycerol for the transition
in perpendicular orientation to the surface. At of shape and inner texture from nonequilibrium to
decompression of the monolayer in this state, equilibrium forms is demonstrated in Fig. 14.
small circular holes of the fluid phase are formed After compression stop, the domain shape relaxes
within the homogeneous condensed phase, which from dense-branched dendritic nonequilibrium
at further decompression are growing to very patterns (1) over less-branched patterns (2) to
large holes separated by thin condensed-phase the circular equilibrium textures (3) within
lamella so that such systems can be considered approximately 5 min. In all cases, the dendrites
to be a two-dimensional foam (Fig. 13, middle). It meet in a center and develop the number of
is interesting to note that in the case of renewed dendritic arms into the number of segments of
compression during the decompression process, the circular domains, that is, in the dendritic
Fig. 13 Morphological
variety at compression and
decompression of myristyl
alcohol monolayers,
T ¼ 25 C
Formation of circular condensed-phase domäns during compression at

A=0.35 nm2/molecule (left) and A=0.32 nm2/molecule (right).
Metastable holes in a continuous condensed phase formed in the plateau region of a

myrstoyl alcohol monolayer upon decompression starting at π = 30 mNm–1. (left)
A = 35 nm2/molecule (right) two-dimensional foam at A = 37 nm2/molecole.
Circular condensed-phase domains within the holes of the foam-like structures formed S
by a renewed step-like compression of a metastable myristyl alcohol monolayer within the
plateau region.
growth regime, the growth directions are corre- depends primarily on the chemical structure of
lated to crystallographic directions of the lattice. the amphiphilic monolayer, but also on other
system parameters, such as the surface pressure
Morphological Variety Primarily Caused by and pH of the subphase. The features of the
the Chemical Structure of the Amphiphilic domain morphology are affected both by the
Monolayer number and length of the alkyl chains and by
Regular condensed-phase domains are formed the chemical structure of the polar head groups.
during the main phase transition in the “plateau” In many cases, alkyl chain lengths between C14–
region of the P-A isotherms of Langmuir mono- C16 are preferred to provide the best precondi-
layers. The observed morphological variety tions for the formation of well-developed domain
1.0 observed. Then, the segments can meet at points

situated at the edge or special points inside of the
0.8 domains. Finally, continuous changes in the ori-
entation of molecules have been observed in
roundness r r
more or less circular domains.

0.6
A representative example for a regular domain
is shown in Fig. 15. Most of the 1-monoglyceride
0.4 domains are subdivided into seven segments of
different uniform brightness that meet at a center,
0.2 at which the alkyl chain tilt azimuth jumps by
a defined angle of 51 . The chains have differ-
0.0 ent azimuthal orientation in each segment. Based
0 50 100 150 200 250 300 350 on the geometric analysis of this domain texture,
time t (s) it was concluded that the chains are tilted along
the bisector of each segment and that the segment
Surfactant Monolayers, Fig. 14 Domain roundness–
time (r-t) diagram of nonequilibrium domains in boundaries are along next-nearest-neighbor
1-monopalmitoyl-rac-glycerol monolayers. Three relaxing directions.
domain shapes after compression stop in the plateau region As seen in Fig. 15, the domain textures show
of the P-A isotherm (T ¼ 23 C) pronounced defect lines where jump-like changes
of the molecular orientation occur between
segments of homogeneous orientation. This is of
textures. In many cases, the condensed-phase special interest for the explanation of the domain
domains show a pronounced substructure which textures by consideration of the domain geometry
can be most effectively studied by BAM if an and its lattice in dependence of the tilt angle of
analyzer is positioned in the reflected beam path. the alkyl chains. The inner structure can be
The substructure inside the domains is correlated described by a geometric concept on the basis of
to the molecular packing in the amphiphilic the molecular packing. The defect lines run along
monolayers which contain the aliphatic chains dense lattice rows of the two-dimensional lattice.
as the structure-determining unit. If the cross- Therefore, a well-defined change in molecular
sectional area of the head group is larger than orientation occurs at the boundary between two
that of the alkyl chain(s), the tightest alkyl chain’s segments. The segment boundaries of circular
packing can only be realized when they are tilted domains represent mirror lines as well as contact
to the aqueous interface. This also holds if two or lines of segments rotated by 360 /n around the
more alkyl chains of medium length are present meeting point of segments. On this basis, numer-
in the molecule. ous domain types can be derived geometrically.
Numerous various amphiphiles, such as In reality, however, only two types have been
tetradecanol, fatty acids, fatty acid ethyl or observed (i) with chains tilted outward, and
methyl esters (C16, C18), and 1-monoglycerides (ii) with rotation geometry. Both types can be
(C16, C18), but also complicated molecular rec- distinguished easily from each other by BAM if
ognition systems form circular domains which the two polarizers of the BAM are in parallel
can have well-defined textures. In regular tex- position. The distortion of the lattice due to the
tures, these domains are subdivided in up to tilt, t, of the alkyl chains causes a deviation of the
seven segments of different reflectivity by sharp segment angles from 60 . The segment angle can
straight lines at which the chain tilt azimuth be calculated as function of the polar tilt angle.
jumps by a defined angle. Usually, the segments For example, type (i) with the chain tilt t330
meet in the center. Apart from the regular tex- provides segment angles of 51 , as expected
tures, also those with irregular or special arrange- for sevenfold domains, whereas for both types
ment of segment of uniform orientation can be with t 15 , the deviations from 60 are <2
Fig. 15 BAM image of
a round domain of
1-monopalmitoyl-rac-
glycerol monolayers and
schematic presentation of
the tilt direction of the alkyl
chains
Fig. 16 Left and right
ethyl palmitate domains
subdivided into 6 segments
by defect lines and
schematic presentation of
S
the tilt direction of the alkyl
chains perpendicular to the
bisectors of domains
suggesting consequently sixfold domains. The 51 , whereas sixfold domains of type (ii) were
segment angles experimentally observed in observed for the expected right and left forms of
monolayers of 1-monopalmitoyl glycerol and ethyl palmitate monolayers for which a segment
ethyl palmitate agreed with those calculated angle of 61 was calculated (Fig. 16). Both in the
from 2D lattice data obtained by GIXD measure- sevenfold texture of 1-monopalmitoyl glycerol
ments. 1-Monopalmitoyl glycerol monolayers domains and in the sixfold texture of
develop sevenfold domains of the (i) type in the ethyl palmitate domains, the defect lines
agreement with a calculated segment angle of are along two next-nearest-neighbor (NNN)
Fig. 17 BAM image of a
cardioid domain of
1-monopalmitoyl-
rac-glycerol monolayers
and schematic presentation
of the tilt direction of the
alkyl chains
directions, but in the case of the 1-monopalmitoyl perpendicular to the boundary in the whole
glycerol domains, the chains are tilted domain (Fig. 16, right). The centered domains
toward nearest-neighbor (NN) direction have lower total line energy than the cardioid-
(Fig. 15, right), whereas in the case of the shaped domains, but the total length of segment
ethyl palmitate domains, they are tilted in the boundaries is higher. However, neither the
NNN direction (Fig. 16, right). Consequently, molecular origin of the anisotropic line tension
the alkyl chains of the 1-monopalmitoyl glycerol nor the origin of the widespread positions of the
domains are tilted toward the bisectors of the segment center is understood. Possibly, the posi-
segments, and those of the ethyl palmitate tion of the segment center is determined by the
domains are tilted perpendicular to the bisectors. directional dependence of the growth rate of the
In some cases, cardioid-shaped domains with domains.
segment contact at the edges are formed (Fig. 17). Small changes in the polarity by subtle varia-
The outer segment shapes are distorted because tion in the head-group structure can dramatically
of anisotropic line tension. The line energy for affect shape and organization of the condensed-
a domain is the line tension integrated over the phase domains. Four types (amide, ether, ester,
whole domain boundary. Minimum line energy and amine abbreviated as ADD, ETD, ESD,
can be realized for a domain shape for which the and AMD, respectively) of racemic amphiphilic
boundary area is not the minimum at the constant monoglycerols have been compared to investigate
domain area when the line tension is how systematic alteration of the head-group struc-
anisotropic. The boundary length of a cardioid- ture can affect the domain shape (Fig. 18). The
shaped domain is only 4 % larger than that of the crystalline nature of the domains increases from
circular domain of the same area. The minimum amine (lowest) to amide (highest). The
line tension is realized when the domain bound- monoglycerol amine domains are of fractal nature
ary is perpendicular to the tilt azimuth of the alkyl with certain fluidity. The amide domains are
tails. Thus, the part of the domain boundary, thin and brittle and break up when interdomain
where the tilt azimuth of alkyl tails is nearly collision occurs. Amide domains have the highest
perpendicular, is shorter in circular domain com- orientational correlation within the domain prop-
pared to that in cardioid-shaped domains if the agated up to a few millimeters. Correspondingly,
contact point of segments is situated at the the orientational correlation in monoglycerol
domain boundary. Therefore, the domains amide domains is several orders of magnitude
become cardioid shaped if the contact point of higher than that of the monoglycerol amine.
the segments is situated at the domain boundary. The monoglycerol ester and ether domains have
When the segment boundaries meet in the center, crystallinity and orientational correlation of
the molecules are observed to be tilted nearly the structure between the monoglycerol amide
a R R R O R O
CH2 CH2 C C
HN O O HN
CH2 CH2 CH2 CH2
C C C C
HO CH2OH HO CH2OH HO CH2OH HO CH2OH
H H H H
Amine diol Ether diol Ester diol Amide diol
100 μm
Surfactant Monolayers, Fig. 18 (a) Comparison of the chemical structure of the four monoglycerols. (b) Represen-
tative BAM images of four diols
and amine. The molecular azimuthal projection shape-dependent dipolar repulsion, which favors
changes within the ester and ether domains in elongated shapes, and the line tension between
monoglycerol monolayers. However, the the condensed-phase domain and the surrounding
domains of the monoglycerol ester are of closed fluid phase, which favors compactness of the
and round shape, whereas the domains of domain. According to the usual temperature
the monoglycerol ether are branched and with dependence of the domain forms, a compact
elongated arms. shape is favored with decreasing temperature,
Line tension and electrostatic repulsion are the whereas elongated shapes are preferred with
main factors controlling the domain shape. increasing temperature. This is in agreement
A comparison of the dipolar interaction can pro- with the expectation that the electrostatic repul-
vide insight into the role of the dipolar repulsion. sion between head groups is reduced by the
The dipole moments of the four monoglycerols increase in the static dielectric constant of the
calculated by semiempirical quantum mechanical subphase and the line tension increases
technique follow the sequence AMD < ETD < with decreasing temperature. However, the S
ESD < ADD. The dipolar interaction energies condensed-phase domains of 1-hexadecyl-rac-
calculated by ab initio calculations for glycerol monolayers show anomalous tempera-
monoglycerol head-group segments with differ- ture dependence as compact domain shapes are
ent possible mutual orientations between the formed at higher temperature and elongated
dipoles follow also the same trend and agree arms or even long stripes are developed as the
qualitatively with the observed changes in the temperature is lowered (Fig. 19). Such behavior
domain shape of the monoglycerols obtained by indicates the increasing importance of the dipolar
BAM and GIXD techniques. These theoretical interaction at lower temperature.
calculations have conclusively shown that the Quantum mechanical calculations of molecu-
electrostatic interaction plays an important role lar dipole moments and the analysis on the basis
in the mesoscopic domain architecture. of GIXD data reveal that the azimuthal orienta-
It is largely accepted that the shape transition tion of the molecule enhances the regular
is governed by the competition between the arrangement of in-plane dipole moment at lower
a b c
d R
CH2
CH2
C
HO CH2OH
H
Surfactant Monolayers, Fig. 19 Domain shapes of 1- 308.15 K, p ¼ 18 mN/m. (d) The chemical structure of the
O-hexadecyl-rac-glycerol monolayers at (a) 278.15 K, amphiphile
p ¼ 0.1 mN/m, (b) 288.15 K, p ¼ 2.2 mN/m, and (c)
temperatures, which dictates the anomalous depends crucially on their chiral structure. Chiral
shape transition. Higher in-plane dipole moments objects are abundant in nature and are operative
at lower temperatures favor elongated shapes in many biological systems at both microscopic
along the azimuthal tilt direction. The ordering as well as macroscopic levels. Ordinary chiral
of amphiphiles will be more at lower tempera- examples are proteins and their constituent
ture, which is also expected to favor the align- amino acids, nucleic acids and their constituent
ment of dipoles toward the azimuthal direction sugars, as well as membranes and their constitu-
and thereby increase the repulsive interaction in ent lipids and membrane proteins. In most cases,
that direction. The result has significance in con- only one enantiomer is biologically active, and
trolling the shapes of mesoscopic aggregates by the biological units lose their functionality
tuning temperature and molecular structure. when their chiral structures are altered (Nandi
and Vollhardt 2009). A further challenge is to
Chiral Discrimination Effects understand the well-known problem of
Chirality is a fundamentally important property homochiral evolution in nature.
of biological molecules so that there is a. contin- A chiral molecule has at least one asymmetric
ued interest in studying the effect of chirality on carbon atom connected to four different chemical
the monolayer morphology (Nandi and Vollhardt groups so that permutation of any pair of these
2003). The functionality of these molecules groups leads to the mirror image of the original
molecule. Amphiphilic monolayers with one or mesoscopic length scales of chiral monolayer
more chiral centers represent unique systems to material. Previously, chiral discrimination has
study chirality-dependent interactions under been concluded from differences in the shape of
defined conditions (Nandi and Vollhardt 2007). the P-A isotherms. However, measurements of
Understanding of chiral interactions in the two- P-A isotherms provide thermodynamic informa-
dimensional monolayers is relatively easier com- tion, and hence, the manifestation of the chirality
pared to that in three-dimensional systems due to effect in terms of interaction at the molecular
reduction of one degree of freedom. level is limited.
Three categories of enantiomer pair mixtures Both homo- and heterochiral preferences for
are distinguished: (1) homochiral discrimination chiral interaction have been observed depending
with favored interaction between the same enan- on the intermolecular interaction of the amphi-
tiomers (ED-D or EL-L > ED-L), (2) heterochiral philic molecules. Correspondingly, the nature of
interaction with favored interaction between the the chiral discrimination is determined by the
different enantiomers (ED-D or EL-L < ED-L) molecular structure and the respective two-
where for strong interactions racemic compounds dimensional lattice structure of the monolayer
can be formed, and (3) textural differences forming amphiphiles. In the case of strong
between the different chiral forms can also homochiral preference, spontaneous separation
occur in case of very small energetic differences of racemic mixtures is expected. For monolayers
in the interaction between the same and different of various chiral amphiphiles, typical examples
enantiomers. for handedness of the two enantiomeric forms in
With the development of Brewster angle opposite directions have been demonstrated by
microscopy (BAM) and fluorescence micros- a mirror-image relation of the respective con-
copy, two optical techniques have been available densed-phase domains.
to allow the direct observation of the effect of the A variety of chiral discrimination has been
molecular chiral structure of the enantiomeric evident. This concerns not only general differ-
forms on the morphology of the condensed- ences in shape and inner texture of the con-
phase domains. The combination of surface pres- densed-phase domains between the enantiomeric
sure studies with Brewster angle microscopy forms and the racemic mixtures but also between
(BAM) imaging and grazing incidence X-ray the two enantiomeric forms. The two enantio-
diffraction (GIXD) measurements is optimal for meric forms can show a striking stereometric
the experimental characterization of chiral mirror-image relation in the domain shape as
discrimination in amphiphilic monolayers. well as clockwise or counterclockwise curvatures
Particularly, the development of the condensed- of the domain arms, the defect lines, or regions of
phase domains formed in the fluid/condensed the same azimuthal orientation depending on the
transition region can be directly visualized. chemical structure of the amphiphile. S
The shape and inner texture provide valuable In the following, the effect of chirality on the
information about the long-range ordering of morphology of the condensed-phase domains for
the domains. The enormous variety of domain amphiphilic monolayers of various chemical
morphology found for the various monolayer structures is discussed.
amphiphiles has evidenced that their chemical The chiral structure-dependent interaction in
structure has a profound influence in driving the amino acid amphiphiles is fascinating. Recent
domain shapes as well as the molecular orienta- comprehensive studies of the condensed-phase
tion within a particular domain. In the two recent domains of various amino acid amphiphiles,
decades, a number of experimental observations such as N-palmitoyl aspartic acid, N-palmitoyl
have confirmed that the morphology of the or N-stearoyl serine methyl ester, and
condensed-phase domains which are formed in N-palmitoyl threonine and its diastereomer
the fluid-/condensed-phase transition region of N-palmitoyl allo-threonine and their methyl
the P-A isotherm expresses chirality at esters, N-palmitoyl or N-myristoyl alanine, have
Surfactant Monolayers, Fig. 20 Chiral discrimination of the condensed-phase domains of N-stearoylserine methyl
ester: (left) D-enantiomer; (middle) L-enantiomer; (right) 1:1 DL-racemate. Image size 60 60 mm
Surfactant Monolayers, Fig. 21 Chiral discrimination of the condensed-phase domains of N-a-palmitoyl threonine:
(left) D-enantiomer; (middle) L-enantiomer; (right) 1:1 DL-racemate. Image size 300 300 mm
revealed a large morphological variety. The two periphery or opposite growth directions in newly
enantiomeric forms of numerous amino acid formed domains can be observed (Fig. 20, right).
amphiphiles show curvatures in opposite direc- In the case of N-palmitoyl-threonine mono-
tions. Usually, the domain shape of the 1:1 race- layers, the condensed-phase domains of the enan-
mic mixtures is different to that of the tiomeric monolayers are curved at one end of the
enantiomeric forms, but very often, oppositely domain in sense of the respective chiral form,
curved texture elements can be observed within again in opposite direction to each other
a single domain suggesting at least elements of (Fig. 21, left and middle). The domains of the
homochiral preference. 1:1 racemic monolayers grow from a center with
For example, the domains of the two enantio- curvature in opposite direction (Fig. 21, right),
meric forms of N-stearoyl serine methyl ester again suggesting to chiral segregation within
monolayers are spiral-shaped curved in opposite a domain.
direction (Fig. 20, left). The 1:1 racemic mono- As demonstrated by the mesoscopic domain
layer tends to develop compact domains in equi- shapes of amphiphilic monoglycerols, slight var-
librium, but both of the opposite curvatures iation in the dipolar head-group structure of the
within a domain in form of hooks at the domain amphiphile and the related change in the dipolar
Fig. 22 Chiral
discrimination in
1-O-hexadecylglycerol
domains, T ¼ 23 C,
A 0.3 nm/molecule:
(right) racemic mixture
with curvatures of the
spirals in two opposite
directions; (left)
S-enantiomer having only
clockwise curved spirals
Surfactant Monolayers, Fig. 23 Chiral discrimination in 1-stearylamine-glycerol (SAG) monolayers.

(a) S-Enantiomer, (b) R-enantiomer, (c) 1:1 racemate. Bar length: 100 mm
interaction can also influence the aggregate in 1-stearylamine-glycerol monolayers. Filigree S

shape. Moreover, 1-O-hexadecylglycerol mono- domains are curved in clockwise direction
layers reveal interesting chiral discrimination for the R (+) enantiomer and counterclockwise
effects (Fig. 22). Whereas the spirals formed at for the S () enantiomer, whereas the racemic
the domain periphery of S-enantiomeric 1-O- mixture evolves fractal-like domains without any
hexadecylglycerol are only clockwise curved specific sense of direction (Fig. 23).
according to its handedness, those of the 1:1 Numerous studies of biomimetic monolayers
racemic mixture are both clockwise and counter- are based on phospholipids. Although nearly no
clockwise curved. The opposite curvatures of differences exist between the P-A isotherms of
these spiral structures reflect obviously segrega- enantiomeric and racemic forms, striking chiral
tion of the two enantiomeric forms. discrimination effects have been observed. Sys-
Striking chiral discrimination of the tematic studies of the domain morphology of
condensed-phase domain’s effects was also phospholipid monolayers have shown how chiral-
observed by fluorescence microscopic studies ity is basically evident in the domain morphology.
Surfactant Monolayers, Fig. 24 Representative equi- monolayers with curved defect lines. (right) L-DPPE
librium domains of phospholipids. (left) L-DPPC mono- monolayers with curvature in the brightness change within
layers with curved domain arms. (middle) L-DPP(Me)2E the domains. Image size 500 500 mm
Surfactant Monolayers, Fig. 25 Representative chiral influence. (middle) Curved axes and arms for enan-
nonequilibrium domains of phospholipids. (left) L-DPPC tiomeric monolayers (L-DPPE). (right) Straight axes and
monolayer far from equilibrium with fractal-like (disor- arms for racemic monolayers (DL-DPPE). Image size 500
dered) growth without chiral influence. Monolayer not so 500 mm
far from equilibrium with dendritic (ordered) growth with
In equilibrium, phospholipid monolayers form conditions (Fig. 25). In the case that the
compact domains of different shapes and inner nonequilibrium state of the monolayer is far
textures (Fig. 24). The chirality of the enantio- from equilibrium, non-ordered fractal-like struc-
meric forms of the compact domains can be vis- tures are formed so that chiral effects cannot be
ible in clockwise or counterclockwise curvatures observed (Fig. 25, left). In the corresponding
of the defect lines; domain arms, for example, racemic 1:1 mixtures of different phospholipids,
triskelions; or regions of the same azimuthal ori- curvatures are not formed, but rather non-curved
entation. The curved defect lines can meet within compact domains or straight dendrite arms are
the domain or at their edge. In all cases, the two developed (Fig. 25, right).
enantiomers show curvatures of opposite sense Chiral discrimination in the equilibrium
both in equilibrium and nonequilibrium struc- domain shapes can be represented by the model
tures. As the relaxation time for reaching equilib- phospholipid DPPC (Fig. 26). Triskelions are the
rium is rather high, nonequilibrium structures in characteristic equilibrium shapes of the two enan-
form of dendritic or fractal-like structures are tiomeric forms which are oppositely curved. The
observed under the corresponding experimental lines running through the arms of the enantiomeric
Surfactant Monolayers, Fig. 26 Chiral discrimination arms of the racemic DL-DPPC (right) are straight. Image
in DPPC monolayers. The two enantiomeric triskelions size 100 100 mm
(left and middle) are curved in opposite direction. The
domains indicate a strong continuous change in the monolayer components and guest materials
orientation. The racemic DPPC monolayers form dissolved in the aqueous subphase is one of the
non-curved compact domain structures. most important research topics in this field
The effect of very small changes in the chemi- (Vollhardt 2008). The air/water interface repre-
cal structure on the domain morphology can be sents a unique hetero-dielectric medium, where
demonstrated by the phospholipids dipalmitoyl- two phases which have completely different
phosphatidylethanolamine (DPPE),dipalmitoyl- dielectric constants are in contact. In comparison
phosphatidyl-N-monomethylethanolamine [DPP- to the aqueous bulk solution, the efficiency of
(Me)E], dipalmitoylphosphatidyl-N-dimethyl- molecular recognition is essentially enhanced at
ethanolamine [DPP(Me)2E], and dipalmitoylpho- the air/water interface.
sphatidylcholine (DPPC), which have the same Despite the impressive progress in conceiving
alkyl chain length and are different only with a large variety of novel supramolecular systems,
respect to the number of methyl groups at the comprehensive studies of the main characteristics
nitrogen of the head group (Fig. 27). The BAM of the supramolecular entities, particularly the
images of Fig. 27 show the different domain effect of the non-surface-active component
shapes in equilibrium and that chirality affects dissolved in the aqueous subphase on structure
their texture always by curvatures corresponding and phase properties of the amphiphilic host
to the enantiomeric forms. monolayer, are limited. The first contributions
of our recent work for filling this gap are summa-
Interfacial Molecular Recognition rized in this section (Vollhardt 2005). S
A large number of studies of interfacial molecu- At interfacial molecular recognition, the sys-
lar recognition made self-assembly at the inter- tems have been constructed by a Langmuir mono-
faces under formation of supramolecular layer formed by receptor molecules which bind
structures to a separate research field. The topic the non-surface-active target molecules dissolved
of molecular recognition is as broad as its abun- in the aqueous subphase by complementary
dance in nature. A variety of artificial host–guest hydrogen bonding and/or acid–base interaction.
assemblies has been planned to understand A selected amphiphilic melamine type has
molecular recognition systems with new proper- been discovered to be a good candidate as host
ties and functions. Highly ordered molecular monolayers for molecular recognition of
architectures have been formed, and numerous dissolved components on the basis of comple-
well-designed interfacial recognition systems mentary hydrogen bonding because a
have been reported. Molecular recognition at pronounced two-phase coexistence region over
the air–water interface by assembly between the entire accessible temperature range provides
o o o o
o o o o o o o o
c o c o c o c o
c c c c
c c c c
o o o o
o P o− o P o− o P o− o P o
−
o o o o
+ − + −
N N N N
H H H H CH3 CH3 CH3 CH3
H CH3 H CH3
DPPE DPP(Me)E DPP(Me)2E DPPC
DPPE DPP(Me)E DPP(Me)2E DPPC
Surfactant Monolayers, Fig. 27 Effect of systematic N-monomethylethanolamine; DPP-(Me)2E, dipalmitoyl-

head-group variation of chiral phospholipids on the texture phosphatidyl-N,N-dimethylethanolamine; and DPPC,
of the equilibrium domains: DPPE, dipalmitoylphosphati- dipalmitoylphosphatidylcholine
dylethanolamine; DPP(Me)E, dipalmitoylphosphatidyl-
good conditions for the study of the main charac- at 25 C (Fig. 28). 2C11H23-melamine mono-
teristics. Recognition studies of selected non-sur- layers spread on pure water show the well-
face-active pyrimidine derivatives are developed plateau region at a surface pressure
demonstrated with 2,4-di(n-undecylamino)-6- of 17–18 mN/m. The presence of non-surface-
amino-1,3,5-triazine (2C11H23-melamine) as active pyrimidine derivatives in the aqueous
host monolayer. The molecular recognition of subphase leads to the disappearance of the pla-
all non-surface-active pyrimidine derivatives teau region and the onset of a steep pressure
dissolved in the aqueous subphase changes dras- increase at essentially smaller area values per
tically and in specific way the characteristic fea- molecule. However, whereas the equilibrium
tures (P-A isotherms, morphology of the P-A isotherms of the 2C11H23-melamine mono-
condensed-phase domains) of the 2C11H23- layer on thymine or uracil are identical, they
melamine monolayers. are different to that obtained at barbituric acid
At first, consider the equilibrium P-A recognition where the shift of the steep pressure
isotherms of the 2C11H23-melamine monolayer increase to lower areas indicates a more
Surfactant Monolayers, Fig. 28 Molecular recognition of thymine, uracil, or barbituric acid by a 2C11H23-melamine
monolayer
condensed state (Fig. 28). The strong change in with an inner alkyl chain texture essentially ori-
the shape of the isotherms implies molecular ented parallel to the periphery.
recognition of the non-surface-active pyrimidine The 2C11H23-melamine–thymine domains are
derivative by the 2C11H23-melamine monolayer dumbbell shaped (Fig. 28, bottom left), not yet
under formation of supramolecular assemblies. observed in other Langmuir monolayers so far.
This is strongly confirmed by the results of The 2C11H23-melamine-uracil domains are S
Brewster angle microscopy (BAM) imaging. nearly circular with well-developed inner texture.
Figure 28 demonstrates impressively the highly The domain growth shows a striking tendency to
specific domain morphology displayed by the two-dimensional Ostwald ripening (Fig. 28, bot-
supramolecular entities after molecular recogni- tom middle) (see also ▶ Ostwald Ripening).
tion of the various pyrimidine derivatives. The morphology of the 2C11H23-melamine–
Pure 2C11H23-melamine monolayers develop barbituric acid monolayers is completely differ-
only small compact, non-textured condensed- ent to those of thymine and uracil. This is in the
phase domains (Fig. 28, above left). On the agreement with the shift of the isotherms to lower
other hand, before the steep pressure increase at area values in comparison to the thymine/uracil
zero pressure, the monolayers of the supramolec- assemblies. Large homogeneous areas of con-
ular 2C11H23-melamine assemblies with thymine densed phase are formed (Fig. 28, bottom right)
or uracil form completely different well-shaped and transfer at smaller areas per molecule to
and highly specific condensed-phase domains a homogeneous condensed monolayer.
Fig. 29 Chemical
structure of interfacial
supramolecular assemblies
between 2C11H23-
melamine and pyrimidine
derivatives dissolved in the
aqueous subphase with
thymine and uracil (left),
with barbituric acid (right) N N N N N N N
H H H H H H
O O O O
N N N N N N
H H H H H
(H3C) N N (CH3) N N N N
N N N
H H H H H H
N O O N O O O O
Surfactant Monolayers, Fig. 30 Molecular recognition monolayers: (left) thymine, (middle) uracil, and (right)
of pyrimidine derivatives dissolved in aqueous subphase barbituric acid. Image size: 350 350 mm
by the modified (CH3(CH2)11O(CH2)3)2-melamine-type
To understand the striking differences in the The high specificity of the domain morphol-
main characteristics between the supramolecular ogy is demonstrated by modification of the mel-
assemblies of the three pyrimidine derivatives, amine-type monolayer and molecular recognition
observed by the isotherms and BAM, their differ- of the same pyrimidine derivatives thymine, ura-
ent chemical structures should be looked at. The cil, and barbituric acid. The amphiphilic alkyl
head group of 2C11H23-melamine has two bind- chains were modified to (CH3(CH2)11O
ing faces on both sides. The binding faces of the (CH2)3)2-melamine, and Fig. 30 shows the spe-
studied pyrimidine derivatives are different. Thy- cific domain morphology of the supramolecular
mine and uracil have only one binding face so assemblies after molecular recognition of
that at interfacial molecular recognition, comple- thymine, uracil, and barbituric acid. The conspic-
mentary hydrogen bonding of two thymine or uous differences to the unmodified 2C11H23-
uracil molecules by one 2C11H23-melamine mol- melamine are obvious (compare Figs. 28 and 30).
ecule should be concluded. On the other hand, Alternatively to complementary hydrogen
barbituric acid has two binding faces on the same bonding, interfacial molecular recognition sys-
plane as 2C11H23-melamine, so that a linearly tems can be based on acid–base interaction.
extended hydrogen-bonding network is formed Here a supramolecular system is demonstrated
at the air–water interface. The chemical structure consisting of the cationic heptadecyl-
of the two different types of supramolecular enti- benzamidinium chloride monolayer as host com-
ties is shown in Fig. 29. ponent and sodium phenylacetate dissolved in the
aqueous subphase as guest component. The substructure within the condensed-phase domains,
experimental results obtained by molecular rec- consisting of long filigree strings and the favored
ognition illustrate clearly the drastic changes in formation of bilayers overgrowing the strings,
the surface features characterized by coupling the indicates a linearly extended amidinium–carbox-
results of the P-A isotherms, BAM, and AFM ylate interfacial structure of the base and acid
studies (Fig. 31). The absence of a plateau region component in alternating sequence.
in the isotherm of heptadecyl-benzamidinium
chloride monolayer spread on water indicates Lattice Structure of the Condensed Phases
that these monolayers are already after spreading The GIXD measurements contribute to the char-
in the state of two-phase coexistence of irregu- acterization of the condensed phase formed after
larly shaped condensed phase with the surround- the phase transition point in the molecular scale.
ing fluid (gaseous) phase. At all surface pressures All Langmuir monolayers studied to date by
with P > 0 mN/m, a homogeneously reflecting GIXD consist of “powders” of randomly ordered
condensed phase is formed. The P-A isotherms 2D crystallites on the water surface.
of the heptadecyl-benzamidinium chloride Usually only a few reflexes of the tightest
monolayers spread on aqueous solutions of the lattice rows of the freely rotating alkyl chain are
guest components sodium phenylacetate or observed which determine the unit cell dimen-
sodium benzoate are completely different. This sions and molecular orientation. They pack in
concerns mainly the pronounced plateau formed three different cell types: hexagonal (a ¼ b,
at different surface pressures (10 mN/m for g ¼ 120 ); distorted hexagonal (a ¼ b, g 6¼
phenylacetate and 24.5 mN/m for benzoate) 120 ), which may be more properly described as
for the specific carboxylate anions. Different to centered rectangular; and oblique. The molecules
the behavior of usual Langmuir monolayers, the in a hexagonal cell are aligned vertically relative
increase of the surface pressure is approximately to the plane of the monolayer, while in the
finished at the end of the plateau region at area per distorted hexagonal cell, they are tilted along
molecule values corresponding tightest mono- a symmetry direction and in the oblique cell
layer coverage. The completely different shape along an arbitrary direction.
of the P-A isotherms reveals that with the molec- The hexagonal cell has three equivalent lattice
ular recognition of the counter ions, also the phase spacings. The contour plot from these three
behavior of the monolayer is totally changed. reflections appears as a single reflex at qz ¼ 0 Å1
Unusual condensed-phase domains of (Fig. 32, top left).
a special texture and topography are formed in The corresponding contour plots for the
the plateau region of the P-A isotherm of the distorted hexagonal cell comprise two peaks. It
heptadecyl-benzamidinium/phenylacetate com- results a centered rectangular lattice (i) with tilt of
plex. The round domains consist of numerous the alkyl chains in nearest-neighbor (NN) direc- S
filigree branches having small gaps between tion with one reflex at qz ¼ 0 Å1 and the other
them. The domains grow from a center in all one at qz > 0 Å1 (Fig. 32, top right) and (ii) with
directions to a large size, and finally, they contact tilt of the alkyl chains in next-nearest-neighbor
to each other. The AFM image of Fig. 31, bottom, (NNN) direction with both reflexes at qz > 0 Å1
reveals that the strings of molecular thickness are (Fig. 32, bottom left).
partly overgrown by a second molecular layer In the oblique cell, the degree to which the
although some free gaps remain visible between three reflections are resolved depends upon the
them. The continuous transition to a bilayer sub- deviation from centered rectangular (i.e.,
structure of the condensed-phase domains devel- distorted hexagonal) cell symmetry. That means
oping within the plateau of the P-A isotherm is the an oblique lattice is realized if the tilt direction is
reason that a further increase of surface pressure no lattice row. The distances of all lattice rows
does not take place at compression to low molec- are different so that three reflexes at qz > 0 Å1
ular area values. The development of a are observed (Fig. 32, bottom right).
Fig. 31 Heptadecyl-
benzamidinium chloride
monolayer. (above) P-A
isotherms; (middle) BAM
image after spreading on
a 1-mM aqueous of sodium
phenylacetate solution
obtained in the plateau
region of the P-A isotherm;
(bottom) AFM image after
transfer onto a silicon wafer
reveals partial overgrowth
of first molecular layers
with second layers
Surfactant Monolayers, Qxy Qxy
Fig. 32 Schematic
presentation of the contour
plots of the corrected
diffraction intensities as
a function of the in-plane
Qxy and out-of-plane Qz
components of the
scattering vector
Qz Qz
rectangular centred cell
hexagonal cell
with NN chain tilt
Qxy Qxy
Qz Qz
rectangular centred cell oblique cell
with NNN chain tilt
Now, with the acquired information from sur- The oblique intermediate (I) phase character-
face-sensitive GIXD technique, the packing of istic for a chiral lattice occurs both in the enan-
such Langmuir monolayers can be determined tiomeric E-Gl and in the racemic R-Gl monolayer
almost at atomic level. Although in the meantime with an alkyl chain lattice and a chain tilt in
GIXD measurements have been carried out on a direction between the NN and NNN phases of
several Langmuir monolayer systems, the mono- the rectangular centered lattice. The NN ! NNN
layer structures corresponding to the phase dia- phase transition splits in two successive NN ! I
grams should be demonstrated by GIXD only for and I ! NNN transitions. The NN ! I transition
the chiral discrimination of monostearoyl- occurs abruptly, whereas the I ! NNN transition
glycerol monolayers. takes place continuously. In both the racemic S
To understand in more detail the reason for the R-Gl and the enantiomeric E-Gl, the NN ! I
temperature effect on the chiral discrimination of transition begins at 5 C at zero pressure and
monostearoylglycerol monolayers, a systematic changes linearly to higher pressures with increas-
comparison of the phase and lattice features of ing temperature but with a stronger slope for
the racemic 1-stearoyl-rac-glycerol (R-Gl) and R-Gl. The proportion of the I phase is very
enantiomeric 3-stearoyl-sn-glycerol (E-Gl) was different for the enantiomeric and racemic
performed. The surface pressure–temperature dia- forms. Symmetry splitting in the racemic R-Gl
grams demonstrate that three phases occur both in monolayer occurs only at low temperatures in
the racemic R-Gl monolayers (Fig. 33) and in the a very small pressure region (Fig. 33), so that
enantiomeric E-Gl (Fig. 34) but with completely the I phase was overseen in previous works. At
different proportion of the single phases and phase 20 C already, the I phase disappears in the
transition pressures. Consequently, chiral discrim- racemic monolayer, and a direct NN ! NNN
ination can be clearly seen. transition takes place. On the other hand, in the
Fig. 33 Surface pressure– 50
NNN
temperature diagram of the
racemic 1-stearoyl-rac- I
45
glycerol monolayer NN
40
35 NNN
30
π [mN/m]
25
20
15 NN
10
5 I
0
0 5 10 15 20 25
T[°C]
enantiomeric E-Gl monolayers, the I phase exists monolayers measured between 10 C and 3 C.
over a broad pressure range at all temperatures, The area change after the second inflection point
but at low temperatures, it is so dominant that at is very small compared with the main phase tran-
5 C it was found over the whole accessible sition. The entropy change of this second transi-
pressure range (Fig. 34). tion is small with only 10 % of that observed for
The properties and transition of monolayer the main phase transitions at the same tempera-
phases can be fundamentally characterized on ture. It is interesting to note that this second phase
the basis of the surface pressure and GIXD transition between two condensed phases of
results. This is demonstrated by the monolayers TDAHA monolayers is existent mainly at such
of a highly purified N-acylated ethanolamine low temperatures (10 C) where the fluid-/con-
(C13H27-CO-NH-C2H4OH, TDAHA). densed-phase transition approaches to zero
At temperatures measured between 10 C and pressure.
36 C, the characteristics of the P-A isotherms, Diffraction studies were performed on water
which indicate the main first-order phase transi- at 5 C and at different lateral pressures.
tion from the fluid phase to the condensed phase, Figure 35, right, shows selected contour plots of
are similar to those of usual amphiphiles. How- the corrected X-ray diffraction intensities as
ever, at temperatures of 10 C, a second inflec- a function of the in-plane scattering vector com-
tion point whose transition pressure increases ponent Qxy and the out-of-plane scattering vector
strongly with increasing temperature proves the component Qz obtained at different lateral pres-
existence of a second first-order phase transition sures. The three Bragg peaks observed at
between two condensed phases. Figure 35, all pressures between the first and the second
left, shows three P-A isotherms of TDAHA phase transition show that the structure of the
Fig. 34 Surface pressure– 50
temperature diagram of the
enantiomeric 3-stearoyl-sn-
45 NNN
glycerol monolayer
40
35
30
π [mN/m]
25 I
20
15
10 NN
0
0 5 10 15 20 25
T [°C]
Surfactant Monolayers, 1.0

Fig. 35 (left) P-A 40 10°C
5°C 0.8
isotherms of TDAHA
3°C 0.6
monolayers spread on
0.4
water and measured in the 30
0.2
Qx, Å–1

π [mN/m]
3 C and 10 C. (right) 0
GIXD contour plots of the 20 1.0
corrected diffraction 0.8
intensities as a function of 0.6
the in-plane Qxy and out-of- 10 S
0.4
plane Qz components of the
0.2
scattering vector for
0
TDAHA monolayers at 0
5 C and different pressures 1.4 1.5 1.6
10 20 30 40 50 60
(bottom, 10 mN/m; top, 30 2 Qxy, Å–1
A,Å /molecule
mN/m)
condensed monolayer phase in this region orthorhombic structure with NNN (next nearest
is oblique (Fig. 35, right, bottom). The alkyl neighbors) tilted alkyl chains (Fig. 35, right, top).
chains are strongly tilted in a nonsymmetry The corresponding lattice parameters of the
direction. Above the second transition pressure, TDAHA monolayer measured at 5 C are listed
the condensed monolayer phase changes to an in Table 3.
Surfactant Monolayers, Table 3 Lattice parameters a, mimic biological phenomena, such as chiral dis-
b, c and a, b, g of the unit cell, lattice distortion, chain tilt crimination effects and molecular recognition
t from the surface normal, in-plane area Axy per chain,
cross-sectional area A0 of TDAHA monolayers at 5 C, (Vollhardt et al. 2011). Mimetic systems have
and different lateral pressures P been used as simple model systems in which
´ chiral effects and molecular recognition are
a (Å) a ( )
´ well-known but can be studied much more easily
P b (Å) b ( ) t Axy A0 compared to their biological counterparts. The
´ ´ ´
(mN/m) c (Å) g ( ) Distortion ( ) (Å2) (Å2) advantages of utilizing Langmuir monolayers as
4 4.885 123.1 0.06825 29.3 22.0 19.2 models result mainly from the easiness with
5.081 119.4 which experimental variables can be manipulated.
5.179 117.4
Understanding the reason for chiral prefer-
10 4.891 122.4 0.05310 27.8 21.8 19.3
ence in nature is also of practical importance.
5.037 119.6
At present, single-enantiomer drugs constitute
5.120 117.9
more than 30 % of the total therapeutic drugs
20 4.883 122.3 0.04910 26.3 21.6 19.4
5.021 119.6
and constitute more than 50 % of total sales
5.093 118.1
in particular classes, such as cardiovascular,
30 4.887 115.5 0.08839 7.4 18.9 18.8 antibiotic and antifungal, hormone, cancer,
4.580 122.2 and hematology-related drugs. The annual
4.580 122.2 sales of single-enantiomer drugs exceed
US $100 billion.
Chiral interfaces and molecular recognition
phenomena are of special interest not only for
Calculating the lattice parameters from the the understanding of biological recognition pro-
X-ray data for these two phases shows that cesses but also for the potential application in
the cross-sectional area, A0, amounts to 19.3 Å2 material science, such as applications ranging
in the oblique phase and decreases to 18.8 Å2 in from enantioselective catalysis, nonlinear optical
the orthorhombic phase. The tilt is in NNN direc- devices, sensors, and smart coatings to chiral
tion and the distortion in NN (nearest neighbor) separations. The topics of chirality and molecular
direction; therefore, this orthorhombic phase can recognition are as broad as their abundance
be designated as L2’. The dimensions of the in nature. Chirality is operative in countless
projected rectangular unit cell containing two biological systems at both microscopic and
molecules are 4.887 7.683 Å2 and indicate macroscopic levels. Despite the fact that the
a dense molecular packing of the herringbone structure–function relationship plays an impor-
(HB) arrangement, usually found for the packing tant role in the biological system, the molecular
of hydrocarbon chains in bulk crystals. The tilt understanding of this correlation for many bio-
angle with respect to the surface normal, t, logical molecules is yet to be understood in detail.
decreases only marginally with increasing Important information about the nature’s princi-
pressure (0.18 per mN/m) in the oblique phase. ples in building the molecular machinery which
The transition into the orthorhombic L2’ phase are perfected along the evolutionary pathway can
leads to an abrupt decrease of the tilt angle by be gained through such studies. One can gain
almost 20 and a decrease in the cross-sectional control over the functionality of the biological
area by 2.5 %. molecule by tuning its structure provided its cor-
relation between chiral structure and/or molecu-
lar recognition process and function is known to
Application their molecular detail.
Another interesting approach is to use two-
Langmuir monolayers at the air–water interface dimensional arrangements at the air–solution
have successfully been used as simple model to interface to regulate molecular recognition
processes because in comparison to the bulk References
aqueous solution, the efficiency of molecular rec-
ognition is essentially enhanced at the air/water Ariga K, Hill JP, Lee MV, Vinu A, Chravet R (2008)
Challenges and breakthroughs in recent research
interface. The advantage of the principle of the
on self-assembly. Sci Technol Adv Mater 9(014109):
molecular recognition is that highly ordered 1–96
molecular architectures with new properties and Fainerman VB, Vollhardt D (1999) Equation of state for
functions can be successfully designed by spe- Langmuir monolayers with two-dimensional phase
transition. J Phys Chem B 103:145–150
cific composition of two or more various molec-
Fainerman VB, Vollhardt D (2003) Equation of state for
ular components. Therefore, molecular monolayers under consideration of two-dimensional
recognition at the air–water interface by assem- compressibility in the condensed state. J Phys Chem
bly between monolayer components and guest B 107:3098–3100
Fainerman VB, Vollhardt D (2008) Equation of state for
materials dissolved in the aqueous subphase has
the phase coexistence region of insoluble monolayers
been developed to an important research topic. under consideration of the entropy nonideality. J Phys
Recognition of various aqueous guest mole- Chem B 112:1477–1481
cules, such as nucleic acids, their bases, nucleo- Fainerman VB, Vollhardt D (2009) Equation of state for
monolayers with additional phase transition between
sides, and nucleotides; barbituric acid and related
condensed phases of different compressibility. J Phys
derivatives; amino acids; and saccharides, as well Chem B 113:6311–6313
as chiral recognition within membrane compo- Nandi N, Vollhardt D (2003) Effect of molecular chirality
nents and/or between amphiphile and aqueous on the morphology of biomimetic Langmuir mono-
layers. Chem Rev 103:4033–4075
guests at the air/water interface has been studied
Nandi N, Vollhardt D (2007) Molecular interactions in
(Ariga et al. 2008). amphiphilic assemblies: theoretical perspective. Acc
Recent advances in the exploration of chiral Chem Res 40:351–360
discrimination effects and of the highly specific Nandi N, Vollhardt D (2009) Chirality and Molecular
Recognition in Biomimetic Organized Films.
structural changes of the host monolayer by rec-
In: Ariga K, Nalwa HS (eds) Bottom-up
ognition of non-surface-active guest components Nanofabrication: Supramolecules, Self-Assemblies,
have been initiated. The development of first and Organized Films, vol 5. American Scientific
thermodynamic and kinetic approaches of inter- Publishers (ASP), Stevenson Ranch, CA,
pp 131–159, Chapter 4
facial molecular recognition models is of current
Vollhardt D (1996) Morphology and phase behaviour
interest to mimic functions in naturally occurring of monolayers. Adv Colloid Interface Sci 64:
systems. Significant progress is made toward 143–171
understanding the model systems. However, Vollhardt D (2005) Effect of interfacial molecular recog-
nition of Non-surface-active species on the main char-
these model studies are still limited considering
acteristics of monolayers. Adv Colloid Interface Sci
the diversity of natural and synthetic chiral mol- 116:63–80
ecules as well as molecular recognition pro- Vollhardt D (2006) Monolayers at air-water interfaces,
cesses. The evidence of highly specific features Morphology of. In: Somasundaran P (ed) Encyclope-
dia of surface and colloid science, vol 5, 5th edn.
S
of chiral and interfacial supramolecular model
Taylor and Francis, New York, pp 4104–4118
systems and understanding the interaction at Vollhardt D (2008) Interfacial molecular recognition of
molecular level may be the first step toward the non-surface-active species at Langmuir monolayers.
bottom-up approach toward the development of Curr Opin Colloid Interface Sci 13:31–39
Vollhardt D, Fainerman VB (2006) Progress in character-
novel functional materials.
ization of Langmuir monolayers by consideration
of compressibility. Adv Colloid Interface Sci
127:83–97
Cross-References Vollhardt D, Fainerman VB, Liu F (2005) Thermody-
namic and structural characterization of amphiphilic
▶ Interface, Air-Liquid melamine-type monolayers. J Phys Chem
B 109:11706–11711
▶ Nucleation
Vollhardt D, Nandi N, Dutta Banik S (2011)
▶ Ostwald Ripening Nanoaggregate shapes at the air/water interface. Phys
▶ Surfactants Chem Chem Phys 13:4812–4829
S 1202 Surfactant Self-Assembly
broadly divided into monolayer-based films or

Surfactant Self-Assembly bilayer-based films. Qualitatively, one can ratio-
nalize the preferential aggregates formed and the
Eduardo F. Marques1 and Bruno F. B. Silva2 sequence of phases that occurs when a given
1
Department of Chemistry and Biochemistry, variable is changed, in terms of the surfactant
Faculty of Science, University of Porto, Porto, packing parameter or the spontaneous curvature
Portugal of the surfactant film.
2
Physical Chemistry, Centre for Chemistry and
Chemical Engineering, Lund University, Lund,
Sweden Overview
The presence of amphiphilic molecules is ubiq-

Synonyms uitous in nature, technology, and everyday life.
Both naturally occurring and synthetic surfac-
Self-organization of amphiphiles; Surfactant tants, in aqueous and nonaqueous systems, find
aggregation behavior; Surfactant microheter- widespread functions and uses. Phospholipids
ogeneous systems are the major components of cell membranes,
and bile salts have a vital role in the emulsifica-
tion of dietary fat so that the body can absorb it.
Keywords The vital lung surfactant is a mixed lipid–pro-
tein complex, and without it, the lung alveoli
Bilayer; cubic phase; emulsion; Fontell scheme; would not be able to expand and contract. Syn-
hexagonal phase; hydrophobic/hydrophilic; inter- thetic surfactants are employed as cleaning
mediate phase; lamellar phase; liquid crystal; agents, as wetting agents to disperse dry pow-
micelle; micelle growth; microemulsion; normal ders, as emulsifiers for oil-containing formula-
phase; phase behavior; phase diagram; polar/non- tions, as germicides, to impart stability against
polar; reverse phase; self-assembly; solubility; aggregation and as agents to provide rheology
sponge phase; spontaneous curvature; surfactant; control, among many other applications.
surfactant packing parameter; swelling; unimer Amphiphiles are therefore such a relevant class
of chemical compounds that it is essential to
have a good understanding of their solution
Definition properties.
The fundamental ability of surfactants to self-
Surfactant self-assembly is the property that associate in water (and a few other polar solvents)
surfactant molecules have of being able to into a wealth of colloidal structures is generally
self-associate in a given solvent, thus forming designated as surfactant self-assembly. The struc-
different types of aggregates and structures of ture and dynamics of the aggregates formed by
colloidal dimensions. Surfactant self-assembly surfactants vary widely, according not only to
is, like adsorption, largely a consequence of the chemical structure ab initio but also to system
hydrophobic effect, but is influenced by many variables, such as composition and temperature.
other factors, such as the repulsive interactions Micelles and bilayers constitute the basic build-
between polar headgroups. Self-assembled struc- ing blocks of surfactant phases, occurring either
tures can be of finite size – which is the case of in the form of discrete aggregates or virtually
micelles, liposomes, and microemulsion unlimited assemblies.
nanodroplets – or unlimitedly large, with connec- It is important to recognize that the macro-
tivity in one, two, or three dimensions – such as in scopic phase behavior and bulk properties of
liquid crystals. The basic building blocks of all surfactant systems are often determined by the
these discrete and infinite structures can be structure at nanoscopic level and by
Surfactant Self-Assembly 1203 S
interaggregate interactions. Hence, a grasp of the dispersed in water, in order to avoid unfavorable
fine details of surfactant self-assembly is key to contact between the hydrophobic tails and the
an understanding of many colloidal and solvent, the surfactant molecules tend to adsorb
biological phenomena, and directly relevant also at interfaces such as water–air, and water–oil
for many practical uses of these compounds. (when oil is also present), lowering drastically
This entry is organized as follows. First, we the interfacial tension. This property makes sur-
will address the molecular structure of surfactants factants, for instance, excellent foam stabilizers
and its consequences for the physicochemical and emulsifiers (Fig. 2), two of the reasons why
properties and solution behavior of this class of they are widely used as detergents and food
molecules. In order to gain physical insight and additives.
a unifying picture, the main rationalizing models An alternative way for the hydrophobic chains
for surfactant self-assembly will be presented and to avoid contact with water is through the forma-
compared. After an overview of the types of tion of aggregates, like micelles, in which the
aggregates and phases that surfactants can build, chains are packed together in the aggregate inte-
we will deal with the structural and dynamic rior, with the polar heads facing the solvent. This
properties of the most common ones, namely, happens when the unimer concentration in water
micelles, bilayer structures, liquid crystals, and reaches a critical value (more exactly, a narrow
micro- and macroemulsions. In the last section, concentration range), defined as the critical
the essential features of the most commonly used micelle concentration, cmc. Beyond this point,
experimental methods for characterizing micelles coexist with the unimer. As the concen-
self-assembled structures will be covered. tration of surfactant increases, long-range interac-
As a complement to this entry, and for a broad tions between aggregates and packing constraints
coverage of the phase behavior of different types induce ordering, and therefore, lyotropic liquid-
of surfactants and surfactant-containing systems, crystalline phases emerge naturally. These topics
readers are suggested to consult the entry will be dealt with more fully in the next sections.
“▶ Surfactants, Phase Behavior.” For now, we will focus on the basic properties of
the most representative classes of surfactants.
Basic Properties of Surfactants Classes of Surfactants

The schematic representation in Fig. 1 corre-
Amphiphilic Structure and Its Consequences sponds to a single-chained surfactant. In fact,
Surfactants are molecules made of two well- there are several different types of synthetic and
separated parts (Fig. 1), headgroup and tail, with natural amphiphiles, and the need for novel struc-
very different affinities toward water (or oil). The tures for different applications keeps the devel-
headgroup is polar and has affinity toward water opment of new surfactants an extremely active S
(hydrophilic). The tail, normally a hydrocarbon field. With such a variety of surfactant types
chain, is nonpolar and avoids contact with water (Jönsson et al. 1998; Tadros 2005), their classifi-
(hydrophobic), preferring oily environments cation becomes somewhat intricate. However,
(lipophilic). The existence of two parts with they can be broadly classified in two convenient
very different affinities toward a given solvent, ways: (1) according to headgroup charge – and
water or oil, makes surfactants amphiphilic in that thus as anionic, cationic, nonionic, zwitterionic,
solvent. If the solvent is equally good or bad for and catanionic and (2) according to the number of
head and tail, then the amphiphilic character is chains – as single-, double-, triple-chained,
lost. For a general solvent, one can accordingly etc. Single- and double-chained surfactants are
use the terms solvophilic and solvophobic. also often referred to as micellizing and swelling
It is this amphiphilic character that lies behind surfactants, respectively.
the hydrophobic effect, the key thermodynamic Within these classes, many subclasses can be
driving force for surfactant self-assembly. When also attributed considering the chemical nature of
Surfactant
Self-Assembly,
Fig. 1 General structure of
a surfactant molecule
Surfactant
Self-Assembly,
Fig. 2 Schematic
various forms of surfactant
adsorption and
self-association in aqueous
solution
the headgroups (carboxylates, sugar-based, promote water solubility, calcium and magne-
amino acid-based, etc.), and some also fall out sium ions (divalent) promote solubility in oil.
of this classification, such as bolaamphiphiles These divalent ions are also present in higher
and gemini surfactants. The most important concentration in hard water, which affects the
classes will be briefly described below and are detergency properties of surfactants. The perfor-
listed in Table 1. mance of surfactants in hard water, in terms of
sensitivity, follows the order: carboxylate
Anionic Surfactants < phosphate < sulfate sulfonate.
These are the most common synthetic surfac-
tants, widely used as emulsifiers and in detergent Cationic Surfactants
formulations (Jönsson et al. 1998). The most fre- Cationic surfactants are less commonly used in
quent are single-chained amphiphiles with general than anionics, in part due to the fact that
carboxylate, sulfate, sulfonate, and phosphate they are less environment-friendly. Nevertheless,
headgroups, with alkyl moieties ranging from they are still pivotal for many technological
12 to 16 carbons. The most frequently found applications, like surface enhancement. Since
counterions are sodium, potassium, ammonium, the majority of surfaces, such as metals, plastics,
calcium, and different protonated alkyl amines. fibers, and cell membranes, are negatively
While sodium and potassium ions (monovalent) charged, they have a tendency to adsorb onto
Surfactant Self-Assembly, Table 1 Examples of surfactants and their classification according to the headgroup
charge
Type Name Structure General class
Anionic Sodium dodecylsulfate O Na+ Alkylsulfates
−
O S O
Sodium octylsulfonate O Na+ Alkylsulfonates

−
S O
Cationic Hexadecyltrimethyl- Br− Alkyltrimethyl-

ammonium bromide N+ ammonium halides
Didodecyldimethylammonium Dialkyldimethyl-
N+ Br−
bromide ammonium halides
Nonionic Dodecylether O Poly(ethylene oxide)

pentaethyleneglicol OH O O
surfactants
O O
Octylglucoside CH2OH Alkylglucosides

O
O OH
OH
OH
Zwitterionic Dimyristoyl- O Dialkyl-

phosphatidylcholine C O CH2
phosphatidylcholines
O
C O CH
O
N+ CH2
O P O
O−
Catanionic Dodecyltrimethylammonium Catanionic surfactants

N+
dodecanoate
O−
S
them, modifying the surface properties (Jönsson Nonionic Surfactants
et al. 1998). They are usually found with quater- Nonionic surfactants have polar but uncharged
nary ammonium or amine headgroups. headgroups. They are the second largest class of
A major advantage of the quaternary ammonium surfactants, after anionics, and are widely used as
is the ease with which the number of chains and detergents and emulsifiers, among other technical
size of the headgroup can be changed, affecting applications. A major advantage of this type of
drastically the self-assembly structure of surfactants is their insensitivity to hard water and
the formed aggregates. For instance, replacing a a large compatibility with all other classes of
methyl group from DTAB (dodecyltrimethy- surfactants (Jönsson et al. 1998).
lammonium bromide) by another dodecyl alkyl The most common subclass of nonionics is the
chain, yielding DDAB (didodecyldimethy- poly(ethylene oxide) type, also known simply as
lammonium bromide), changes the preferred CmEn, where m stands for the number of carbons
aggregate structure from micelles to bilayers. in the hydrophobic chain, and n stands for the
number of ethylene oxide (OC2H4) units. surfactants, zwitterionics are also compatible
The higher the number of En units, the higher with most of the other surfactant classes and are
the polarity of the headgroup. An important fea- rather insensitive to hard water. Nevertheless,
ture of these systems is that with increasing tem- a disadvantage is their high cost.
perature, the headgroups become less polar due to
dehydration of the En units. This leads to a rich Catanionic Surfactants
self-assembly behavior and a typical aggregate Catanionic surfactants are made of paired anionic
structure evolution from micelles to bilayers to and cationic amphiphilic ions, with alkyl chains
reverse micelles as temperature is increased. of more than six carbon atoms. Hence, they can
Another relevant subclass of nonionics is be viewed as pseudo-zwitterionic amphiphiles,
sugar-derived surfactants (e.g., alkylglucosides, with the important difference that the two
c.f. Table 1). For this group, temperature plays charged headgroups are not covalently linked.
a much weaker role, mainly due to the strong Catanionic surfactants have aggregation proper-
hydrogen bonds between water and the sugar ties in water that can be very different from the
headgroup. Hence, self-assembled structures are original ionic surfactants they derived from,
much more stable with temperature compared to depending on the length and number of chains
ethoxylate-based surfactants. In addition, the of each ionic part (Khan and Marques 1998;
various possible configurations of the sugar Zemb and Dubois 2003).
headgroup configuration, as well as the different
possibilities of linkage between the hydrophobic Gemini Surfactants and Bolaamphiphiles
chains and headgroup, can give rise to complex Gemini or dimeric surfactants are made of two
self-assembly (Stubenrauch 2001). surfactants linked by a covalent spacer, most
commonly at the headgroup level. Compared
Zwitterionic Surfactants with the analogous monomeric counterparts,
Zwitterionic surfactants have both a positive and gemini have lower cmc values, higher surface
negative charge in the headgroup and thus are activity, and different aggregation behavior
globally neutral. Amphoteric surfactants are (Menger and Keiper 2000; Zana 2002). The
those that change from cationic via zwitterionic length and flexibility of the spacer chain dramat-
to anionic on increasing pH, meaning that they ically influence their phase behavior.
are only zwitterionic in a certain range of pH. Bolaamphiphiles are surfactants which bear
The solution properties of amphoteric surfactants a headgroup at each end of a hydrocarbon chain
are thus strongly dependent on pH. Amphoteric (Jönsson et al. 1998). Compared with the analo-
surfactants with sulfate and sulfonate groups in gous single-chained surfactants, the addition of the
the anionic site remain zwitterionic down to low second headgroup increases the unimer solubility,
pH (Jönsson et al. 1998). increases the cmc, and decreases the aggregation
Important subclasses of zwitterionic surfac- number. The aggregate types are also changed in
tants are phospholipids, which are the main accordance with the new molecular structure.
constituents of the cell membranes, and some
amino acid-derived surfactants. Amino acids are
chiral molecules, and when amino acid-based Models for Surfactant Self-assembly
surfactants are enantiopure, tilting of the chiral
molecules with respect to the local bilayer normal Packing Parameter
vector may twist bilayers into ribbon or microtu- The solubilization of hydrocarbon chains by
bular structures (Shimizu et al. 2005). water implies the formation of highly ordered
Most types of zwitterionic surfactants have cages of water molecules surrounding the chains
very low eye and skin irritation, which makes (so-called clathrates). This effect has a high
them well suited for personal and health care entropic cost and lies on the basis of the
formulations. In common with nonionic hydrophobic effect (Tanford 1980). As a means
of reducing the unfavorable contact of chains
with water, surfactant molecules will form aggre-
gates in solution, typically micelles, where the
chains are enclosed in the aggregate core
(cf. micelle in Fig. 2). The underlying reason
that surfactants do not phase-separate like an
ordinary oil–water mixture is the unfavorable
steric and electrostatic interactions (the latter for cone cylinder wedge
ionic surfactants) between the hydrated Ps = 1/3 Ps = 1 Ps > 1
headgroups. Hence, while the hydrophobic effect
defines the onset of aggregation (micellization), Surfactant Self-Assembly, Fig. 3 Illustration of the Ps
the repulsions between the headgroups define its concept and its relation to geometrical shapes
end, and as a consequence, the formation of
aggregates (micelles) with well-defined sizes is ahglhc. In this case, Ps > 1, and the molecules
favored. are more similar to inverted cones or wedges,
If the surfactant has two chains instead of one, forming reverse micelles of cylindrical or spher-
the same principles apply, but due to geometrical ical type. These considerations are illustrated in
reasons, the higher volume of the hydrophobic Fig. 3 and summarized in Table 2.
part relative to the headgroup area will not favor The volume, Vhc, and length, lhc, of a fully
the packing into spherical micelles, and extended saturated hydrocarbon chain can be
bilayer formation is preferred. Hence, while the calculated according to the expressions:
amphiphilic character of surfactants favors
self-assembly into aggregates, it is the volume Vhc =nm3 ¼ 0:0274 þ 0:0269 nC (2)
of the hydrophobic chain relative to that defined
by the optimal headgroup area that will mainly
lhc =nm ¼ 0:154 þ 0:127 nC (3)
determine which aggregate will be formed. This
ratio is known as the surfactant packing parame-
ter (Evans and Wennerström 1999; Israelachvili where nc is the number of carbon atoms in the
1985), designated by Ps, and calculated as chain. The term 0.0274 nm3 in Eq. (2) comes
from the fact that the volume of a methyl group
Vhc is about twice as large as that of a methylene
Ps ¼ (1)
ahg lhc group. The value 0.127 nm in Eq. (3) is the C-C
bond length (0.154 nm) projected onto the fully
where Vhc and lhc are the volume and length of the stretched chain axis, while 0.154 nm is the van
fully extended hydrocarbon chain, and ahg is the der Waals radius of the terminal methyl group S
effective (hydrated) headgroup area. In general (0.210 nm) minus half the bond length of the first
terms, Ps defines a cylinder of area ahg and length C atom not in the hydrocarbon core (0.056 nm).
lhc as the reference structure forming a bilayer of The value of ahg is somewhat more difficult to
zero curvature. If the volume of the alkyl chain is estimate since it varies strongly with the medium
equal to the cylinder volume, Ps ¼ 1, and conditions. For instance, in ionic surfactants, ahg
a bilayer is favored. If the alkyl chain volume is is higher than the physical cross section of the
smaller, the overall shape of the surfactant is headgroup due to repulsive interactions between
more similar to a cone or a truncated cone, and the headgroups. Adding salt decreases ahg
Ps is much lower than one. In this case, the most because the electrostatic charge is more screened.
favorable packing will result in curved aggre- Hence, Ps increases, which in turn may lead to
gates like spherical micelles (Ps ¼ 0.33) or cylin- growth from spherical to elongated micelles.
drical micelles (Ps ¼ 0.330.50). The same Finally, it should be stressed that these
applies to hydrocarbon volumes higher than arguments are only valid when interactions
Surfactant Self-Assembly, Table 2 Aggregate structures and their relation with Ps and Ho
Aggregate Ps Ho Aggregate structure
Spherical micelle 1/3 1/R
Cylindrical micelle 1/2 1/(2R)
Bilayer 1 0
Bicontinuous bilayer 1 0
Reverse cylindrical micelle >1 1/R
Reverse spherical
micelle
R1
R2 R2 = ∞
R1
R1
R2
Surfactant Self-Assembly, Fig. 4 Main radii of curvature in a sphere (left, both radii are positive), cylinder (center, R1
is positive and R2 is infinite), and saddle point (R1 is negative and R2 is positive)
between aggregates can be neglected. At high The bending modulus k defines how difficult it is
concentrations, interaggregate interactions can to bend the surfactant film, and typically, its
also induce shape changes (c.f. sections “Addi- magnitude is within 1–20 kBT units. The saddle-
tional Parameters Affecting Self-assembly” and splay modulus k defines the topology of the sur-
“Lyotropic Liquid Crystals”). factant film.
Qualitatively, the concept of Ho is analogous
Spontaneous Curvature to the Ps since the preferred curvature will
As an alternative to the concept of surfactant also depend on the relative proportion of polar
packing parameter, one can rationalize aggrega- and nonpolar volumes. However, the theory
tion on the basis of the so-called spontaneous behind it is based on a consideration of the
curvature, Ho, of the surfactant film, that is, the mechanical properties of the film as a whole.
preferred mean curvature adopted by the film, in In general terms, Ps is related with individual
the absence of mechanical constraints. Ho is molecules, while H is related with a continuum
defined as with global physical properties. Table 4 lists the
relation between Ps and Ho for a series of aggre-
1 1 1 gate geometries.
Ho ¼ þ (4)
2 R1 R2 The spontaneous curvature and gc concepts
are particularly useful in describing different
where R1 and R2 are the radii of curvature in two morphologies in bilayers (planes, vesicles, and
perpendicular directions (Fig. 4). For spheres, R1 sponge phases), since the Gaussian curvature
and R2 have the same value. For cylinders, one of and saddle-splay moduli are explicitly accounted
the R is finite, while the other is infinite. For for.
planes, both R are infinite, and for a saddle
shape, R1 and R2 are finite but have opposite Additional Parameters Affecting S
signs. Self-assembly
Bending the amphiphilic surface away from As seen in the previous two sections, the packing
its preferred curvature to a different one, H, has parameter and spontaneous curvature are primar-
an energy cost, given by ily determined by the surfactant structure. In
addition, external parameters such as tempera-
K
gc ¼ 2k ðH Ho Þ2 þ k (5) ture, salt, cosolutes, and surfactant concentration
strongly influence Ps and Ho and thus the type of
where k is the mean curvature (or normal bend- self-assembled structures formed (Evans and
ing) modulus, k is the Gaussian curvature (or Wennerström 1999; Lindman and Wennerström
saddle-splay) modulus, and K is the Gaussian 1980). In this section, the most important param-
curvature, defined as eters are summarized.
• Salt. The effective headgroup area of surfac-
K ¼ ðR1 R2 Þ1 (6) tants is larger than their actual physical
size due to electrostatic or steric repulsions. In Ps (decrease in Ho). Cosurfactants are espe-
ionic surfactants, if salt is added, the ionic cially useful to tune the spontaneous curvature
charge at the headgroup is further screened, in microemulsions, where maximum
leading to a reduction of the effective water–oil emulsification is reached when
headgroup area ahg. This results in a decrease Ho 0. If a cosurfactant is added to
of the packing parameter Ps and increase of a bilayer, the membrane bending rigidity can
the spontaneous curvature Ho. also be greatly reduced.
On a larger length scale, without affecting • Ionic Surfactant of Opposite Charge. Adding
Ps to a great extent, salt screening reduces the a surfactant of opposite charge to a system
electrostatic repulsions between particles, composed of ionic surfactant will have an
which can induce flocculation of aggregates effect similar to adding a cosurfactant but
(e.g., vesicle aggregation). much more dramatic, since each added mol-
• Temperature. For ionic surfactants, tempera- ecule will neutralize one charge of the initial
ture has a modest effect on Ps and Ho. On one system. A micellar system can be easily
hand, the gauche conformations of the hydro- changed to bilayers or very thermally stable
carbons are increased (reducing Ho), but on crystalline precipitates. The surfactant mix-
the other hand, the electrostatic Gibbs energy tures are called catanionic mixtures and
that is dominantly entropic is also increased will be discussed more fully in section
(increasing Ho). These two effects combined “Vesicles.”
lead to a slight increase of Ho with T in ionic • Type of Oil. When oil is present (such as in
systems. microemulsions), in the form of short-chain
For nonionic surfactants of the poly(ethyl- alkanes like hexane or octane, it can pene-
ene oxide) type, CmEn, temperature has strong trate to some extent into the hydrophobic
effects and opposite of what is found for region of surfactant monolayers, increasing
ionics. An increase in temperature makes the the effective volume of this region and
headgroups increasingly less hydrophilic, increasing Ps (decreasing Ho). The shorter
which in turn leads to a reduction in headgroup the oil chain is, the higher this effect
area. This favors a continuous increase of Ps becomes.
and decrease of Ho. Depending on the non- • Surfactant Concentration. This parameter is
ionic surfactant in question, simply by increas- more difficult to predict, as it affects Ps and
ing the temperature, the system can change Ho in an intricate way. By increasing surfac-
from normal micelles, through bilayers, to tant concentration, the available volume for
reverse structures. aggregates becomes smaller; hence,
• Cosurfactant. Long-chain alcohols interaggregate interactions become stronger.
(e.g., octanol), among other weak amphiphiles When the aggregates become overcrowded,
of similar type, do not aggregate spontane- they may rearrange their structure to more
ously by themselves but significantly alter optimally fit the available volume. In addition
the self-assembling of surfactants when they to these effects, and for ionic surfactants,
are present. For this reason, they are called a decrease in the available water effectively
cosurfactants. Due to their weakly polar increases the counterion concentration,
headgroups, these molecules interact favor- which will reduce the headgroup area. This
ably with surfactant molecules. The slight will favor a progressive transition from posi-
increase in headgroup area weakens electro- tive to negative spontaneous curvatures. This
static and/or steric repulsions between the sur- interplay of effects is captured in the Fontell
factant headgroups, whereas the hydrophobic scheme (Fig. 5), which idealizes a natural
chain increases significantly the hydrophobic sequence of self-assembled structures and
volume of the resulting surfactant–co- phases as a function of surfactant
surfactant “pair.” This leads to an increase of concentration.
Surfactant Self-Assembly, Fig. 5 The Fontell scheme, idealizing a logic sequence of amphiphilic mesophases as
a function of surfactant concentration. Gray areas represent regions where cubic phases can appear
Self-assembled Structures and Phases Surfactant Self-Assembly, Table 3 Self-assembled

structures formed by surfactants (Jönsson et al. 1998)
Bulk Self-assembly Homogeneous (one-phase) systems
The variety of amphiphilic molecular structures, Isotropic solution phases
Micellar phases (dilute and concentrated; normal and
together with the number of variables that affect
reverse)
the packing parameter Ps and spontaneous curva- Microemulsions (discrete droplets and bicontinuous
ture Ho, lead to a wealth of self-organized struc- structures)
tures in solution (Evans and Wennerström 1999; Vesicle solutions
Tiddy 1980). Before going further to the detailed Liquid-crystalline phases
Lamellar
characterization of these phases, it is instructive Hexagonal (normal and reverse)
to classify them from the point of view of phase Cubic: several structures known, including water-
science (Table 3). continuous, oil-continuous, and bicontinuous
Surfactant systems are thus better described if Intermediate and deformed phases, including nematic
lyotropic
divided into homogeneous, or single-phase, sys-
Solid phases
tems and heterogeneous, or multiphasic, systems. Several different structures
The homogeneous systems may in turn be Heterogeneous (two or more phases) systems
divided as: (1) solutions (e.g., micellar phase); Emulsions
(2) liquid-crystalline phases (e.g., lamellar Suspensions
phase); and (3) crystalline phases. Vesicles/liposomes
Foams
Crystalline phases are characterized by having
Adsorbed surfactant layers and films
both long- and short-range order. Solutions, on Gels S
the other hand, are intrinsically disordered at
short- and long-range scales, even though the
very existence of micellar aggregates presup-
poses a certain degree of molecular organization Heterogeneous systems, owing to a higher
in the bulk. There is also a residual liquid struc- diversity, are somewhat more difficult to subdi-
ture owing to spatial correlations between the vide, but essentially, they encompass emulsions,
aggregates – see, for example, (Hansen and suspensions, foams, gels, and adsorbed films.
McDonald 1986) for a discussion on this In some cases, the classification can be subtle.
topic. Liquid-crystalline phases present disorder For instance, vesicles are considered a true
at short scales but some type of orientational thermodynamic phase in some systems, but
order, and distinct translational order over large more often, they occur in others when a lamellar
distances, in one, two, or three dimensions phase is dispersed in excess solvent by
(Demus et al. 1998; Tiddy 1980). shear forces. In the former case, they are
considered as a homogeneous system (solution), Micelles and Bilayers

in the later as belonging to a heterogeneous
(biphasic) one. Micelles and bilayers are the basic building
blocks of surfactant self-assemblies. For
Self-assembly at Interfaces instance, discrete cubic phases are composed of
Surfactant self-assembly occurs not only in the discrete micelles arranged in a cubic lattice, and
bulk but also at interfaces, like the gas/liquid and bicontinuous cubic phases can be either based on
solid/liquid interfaces. Many water-insoluble infinite branched rodlike micelles or in saddle-
compounds, such as long-chain fatty acids and like curved bilayers.
surfactants, can be spread from an organic Micelles are either discrete or continuous
solvent on an aqueous solution to form monomo- aggregates with overall circular cross section.
lecular films adsorbed at the gas–liquid interface, Discrete micelles include spherical, prolate-
designated as monolayers (Adamson and shaped, and disk-shaped aggregates, whereas
Gast 1997). The polar groups of the amphiphiles continuous micelles can include long
in monolayers point into the aqueous environ- unconnected rods with variable rigidity or
ment, whereas the hydrophobic tails point toward branched rods. These aggregates may occur in
air. These adsorbed films have characteristic water-rich or oil-rich solutions. For normal
physicochemical properties, the study of which micelles, the surfactant film is curved toward oil
is of great interest in colloid and interface or upon itself (Ps < 1, Ho > 0), whereas in reverse
science. micelles the film curves in the opposite direction
The molecules in monolayers can be (avoiding water) and thus reversing its curvature
self-organized in different ways, especially (Ps > 1, Ho < 0).
when they are tightly packed, depending on the Although different nomenclature is used in the
lateral forces acting upon them. The formation of literature, in this entry, we will designate
different 2-D condensed phases depends on a normal micelle solution phase by L1 and
molecular structure, temperature, and surface a reverse micelle solution phase by L2. The sub-
pressure. These condensed phases exhibit scripts 1 and 2 refer in general to normal phases
a range of structures that are 2D liquid-crystalline and reverse curvature phases, respectively.
or crystalline phases, with different molecular Throughout this entry, the focus will be on
packings and molecular tilts with respect to the normal micelles, as they are by far more common
planar interface. Although amphiphiles form in binary surfactant–water systems. Reverse
lyotropic liquid crystal phases in the bulk, these micelles are formed in some nonpolar solvents
2D phases are reminiscent of surfactant layers in and normally require also the presence of water in
thermotropic smectic liquid crystals (which form their core (to adjust the micelle curvature to the
when solid phases are heated). preferred curvature of the surfactant film). They
It is also possible to construct multilayers by will be discussed further in the section
successive deposition of monolayer films onto “Microemulsions and Emulsions.”
a solid substrate, in a process called the Lang- Bilayers are composed of two films of amphi-
muir–Blodgett technique. The deposited films are phile (usually a surfactant or lipid) in a solvent, in
called Langmuir–Blodgett (or LB) films, and which the solvophobic region of each film faces
they are of great relevance for a number of tech- the other and the solvophilic part is facing the
nical applications (Roberts 1985), such as solvent. Bilayers normally have zero spontaneous
nonlinear optics, gas sensors, biosensors, and curvature and therefore are symmetric toward
insulating films for semiconductor devices water and oil (i.e., have no normal or reverse
(diodes and transistors). For more on interfacial counterparts). Depending on a range of
thins films, readers are referred to specialized intermolecular forces (such as hydration, van
reviews (Roberts 1985). der Waals, electrostatic, and undulation forces),
bilayers can either swell in water, oil, or both.
Owing to different morphologies and different
states of the hydrophobic chains, these aggre-
gates have different designations, as will be
seen in more detail in the section “Bilayers.”
Spherical Micelles
Micelles can be regarded as liquid hydrocarbon
nanodroplets, with the surfactant headgroups at
the interface (Jönsson et al. 1998). They form as
a means of minimizing the contact of hydrocar-
bon chains with water, while at the same time,
keeping the headgroups separated from each Surfactant Self-Assembly, Fig. 6 Simplified model of
other as much as possible due to steric or electro- a spherical micelle, highlighting (1) the preferred surfac-
static repulsions. This compromise is reflected on tant conic shape (Ps 0.33), (2) the micellar radius that is
approximately equal to the extended chain length of the
the values of the packing parameter and sponta- surfactant molecule, (3) the exposure of a significant frac-
neous curvature, which indicate which structure tion of hydrocarbon surface to water, and (4) the counter-
is optimal to pack surfactants with a certain ion diffuse cloud
molecular structure. Micelles are normally spher-
ical up to moderate concentrations when the close vicinity to the interface, there are about
packing parameter Ps is ca. 1/3, that is, when 50–80 % of the surfactant counterions
the surfactant structure approaches a cone shape (depending on the nature of the counterion). If
favoring the packing into spheres (Fig. 6). the counterions are inorganic – the most common
The micellar radius is approximately equal to situation – they are very mobile and loosely asso-
the length of the fully extended hydrocarbon ciated with the micelle, so that the micelle can be
chain lhc of the surfactant (c.f. Eq. 3), and the viewed as a macro-ion with a diffuse neutralizing
aggregation number N can be obtained by the cloud surrounding it.
ratio of micelle volume to hydrocarbon chain
volume Vhc (Eq. 7): Micellization Process
Despite their amphiphilic character, surfactants
Vmic 4pl3hc have some solubility in water in the form of
N¼ ¼ (7)
Vhc 3Vhc unimers. For micelle formation to occur, the
unimer concentration has to exceed a critical
This aggregation number is usually very well value, designated as critical micelle concentra-
defined for long-chain surfactants (e.g., for SDS, tion (cmc). Hence, micellization only occurs if S
N is about 60 at 25 C), but for smaller chain the unimer solubility exceeds the cmc. This
surfactants, the distribution of N values starts to occurs above the Krafft temperature, which is
be wider (more polydisperse system). the temperature at which the unimer solubility
Even though the micellar radius is approxi- equals the cmc (Fig. 7a). Above the cmc, micelles
mately equal to the length of the fully extended coexist with unimers, the concentration of which
surfactant chain, the micellar core is known to remains approximately constant and equal to the
be liquid, with most of the chains in a disordered cmc.
state with several gauche conformations. Because micelles have distinct properties as
The headgroups do not fully cover the compared to the free unimers, at the cmc, many
water/hydrocarbon interface, so there is always physical properties of the aqueous solutions show
a fraction of hydrocarbon that is exposed to water inflection points in their variation when surfac-
in this region. In the case of ionic surfactants, in tant concentration is increased (Fig. 7b).
a b
cmc curve osmotic
Magnitude of physical property

pressure
unimers unimers + micelles conductivity

turbidity
T / ºC
Krafft surfactant
solubility curve
temperature self-diffusion
surface tension
hydrated crystals
cmc
surfactant concentration cmc surfactant concentration
Surfactant Self-Assembly, Fig. 7 (a) Phase diagram solutions. The dashed line corresponds to the cmc
for a binary surfactant–water system nearby the cmc. (b) (Adapted from Jönsson et al. (1998))
Variation of some physical properties in surfactant
As previously mentioned, micellization is considered to be an infinitely diluted solution, in

driven by two opposing factors: the hydrophobic Eq. 8, the cmc should be expressed in mole fraction
effect that promotes association and the headgroup units. Because in this model, micelles are consid-
repulsions that define the mean aggregation num- ered a different (condensed) phase, the surfactant
ber. A more quantitative picture can be obtained by chemical potential in the micelles is equal to its
considering several thermodynamic models. Here, standard chemical potential in the micelles:
we will present the most commonly considered
approaches: the pseudo phase-separation model
mmic ¼ momic (9)
and the mass action (or equilibrium) model
(Evans and Wennerström 1999; Hunter 2001).
At phase equilibrium:
Pseudo Phase-Separation Model
There is some analogy between the onset of mS ¼ mmic (10)
micelle formation and a phase-separation pro-
cess, in particular, in what regards the high Combining Eqs. 8–10 yields
cooperativity of micellization. Adding another
molecule to a group of N self-assembled mole- momic ¼ moS þ RT ln cmc (11)
cules becomes more and more favorable as N
increases, just like in a phase-separation process. The standard molar Gibbs energy of
When aggregation occurs, micelles coexist with micellization, DmicGm , which represents the dif-
unimers, with the unimer concentration ference between the surfactant chemical potential
remaining approximately constant and equal to in the micelle and its standard chemical potential
the cmc. Considering the chemical potential of at infinite dilution, per mole, can be written as
surfactant in the unimer state, mS, one has (assum-
ing unitary activity coefficient for the unimer) Dmic Gom ¼ momic moS ¼ RT ln cmc (12)
mS ¼ moS þ RT ln cmc (8) Although Eq. 12 provides a useful approxima-

tion for obtaining DmicGm , and the phase-
where moS is the unimer standard chemical poten- separation model captures several features of
tial, R is the gas constant, and T the absolute micelle formation, it fails to describe the
temperature. Since the standard state is reasonably start–stop mechanism of this process. In fact,
Surfactant
Self-Assembly,
Fig. 8 Fraction of added
surfactant that
assembles into the micelle
(∂(N[SN])/∂([S]T) versus
total surfactant
concentration ([S]T). As N
increases, the micellization
process becomes more
abrupt (Adapted from
Evans and Wennerström
(1999))
micelles are not unlimited assemblies but Considering that for most surfactants N > 50
aggregates with a finite aggregation number. and that near the cmc |SN|<<|S|, the first term in
Eq. 16 can be ignored, and this equation reduces
Mass Action (or Equilibrium) Model to (since |S| ¼ cmc)
A somewhat more advanced way of modeling the
micellization process considers that a specific Dmic Gom ¼ RT ln cmc (17)
aggregation number N dominates over all the
others. In this case, N surfactant unimers (S) which is the same result as Eq. 12.
assemble into a SN aggregate, as described by From Eqs. 13–15, a solution for the derivative
∂(N[SN])/∂([S]T), which describes what fraction
NS $ SN (13) of added surfactant enters the micelle, can be
obtained straightforwardly. This is shown graph-
jSN j ically in Fig. 8. As can be seen, when N is
KN ¼ (14)
jSjN increased, ∂(N[SN])/∂([S]T) changes more
abruptly from zero to one. When N increases to
With only unimers and N-aggregates, the total infinity, the result of the phase-separation model
surfactant concentration expressed in terms of is obtained, with a discontinuity in the derivative S
moles of unimer, |S|T, is given by at the cmc. For finite values of N, the aggregation
process is somewhat gradual, and by definition,
jSjT ¼ NjSN j þ jSj ¼ NKN jSjN þ jSj (15) the cmc is the point at which ∂(N[SN])/
∂([S]T) ¼ 0.5, that is, the point at which an
From the thermodynamic relation between the added unimer has the same probability of enter-
standard Gibbs energy and the equilibrium con- ing the micelle or remaining in solution.
stant, one can obtain the molar Gibbs energy of The case of nonionic surfactants was consid-
micellization through ered above, for simplicity. For ionic surfactants,
however, the equilibrium model also provides an
Dmic Go expression for DmicGm . Let us assume the
Dmic Gom ¼
N surfactant is anionic and designated by SC+. In
RT the micellization process, N free unimers, S,
¼ ln jSN j þ RT ln jSj (16)
N associate with P counterions, C+, to yield
a micelle of net negative charge N-P. This pro- spherical aggregates because this is the geometry
cess can then be described by with highest mean curvature, which maximizes
the surface area per headgroup ahg. If salt is
ðNPÞ
NS þ PCþ $ SN (18) added, the headgroup charges are more effec-
tively screened, leading to a decrease in the unfa-
with vorable repulsions and a decrease in ahg. Hence,
the packing parameter Ps is increased and the
ðNPÞ spontaneous curvature Ho decreased. This might
jSN j
KN ¼ (19) lead to a sphere-to-rod transition, as is the case
jS jN jCþ jP with cetyltrimethylammonium bromide (CTAB)
upon the addition of salt.
The degree of micelle ionization, a, is defined An alternative way of reducing headgroup
as repulsive interactions and inducing a sphere-
to-rod transition is by replacing the inorganic
NP P counterions by organic ones, like for instance
a¼ ¼1 (20)
N N salicylate. The higher hydrophobicity of organic
ions will lead to their incorporation into the
The standard molar Gibbs energy of micelle, while the opposite charge effectively
micellization can be derived in a similar fashion reduces the repulsive interactions between the
as above: headgroups at the interface. Hence, a sphere-to-
rod transition is also favored. The same effect can
Dmic Go also be achieved by the addition of cosurfactants
Dmic Gom ¼
N (such as long-chain alcohols), nonionic surfac-
RT ðNPÞ tants, or surfactants of opposite charge. In these
¼ lnSN þ RT lnjS j
N cases, however, a significant change in hydropho-
þ ð1 aÞRT lnjCþ j (21) bic volume might also take place, which can also
induce a sphere-to-bilayer transition.
When N is in the 50–100 range, as is the case Conversely, in nonionics, there are no charges
for most spherical micelles, the lnjSN(N-P)-j/N present. In surfactants of the type CmEn, Ho is
term becomes again negligible. Furthermore, if adjusted by changing temperature. Starting from
no salt is added, both |S-| and |C+| can be replaced spherical micelles, if T is further increased, Ho
with the cmc, and Eq. 21 can be simplified to will decrease, and as a result, a sphere-to-rod
transition is also observed.
Dmic Gom ¼ ð2 aÞRT ln cmc (22)
End-cap Energy, Temperature, and Size
The factor (2-a) is often called the Gibbs pre- Predictions
factor. It should be noted that Eq. 22 is valid for When a rod of finite size forms, in order to avoid
monomeric surfactants with monovalent counter- exposure of the hydrophobic core to water, two
ions, in the absence of salt. In the presence of salt, end-caps of spherical shape form at the edges of
other expressions can be derived. Besides, further the rods. Since the local curvature of these
expressions are available for other types of ionic end-caps is different from the rest of the rod, an
surfactants, for example, with divalent counter- additional energetic penalty is created. This is,
ions, bolaforms, gemini, etc. (Zana 1996). however, attenuated when two or more rods
merge, since at each merging point, two end-caps
Micelle Growth are eliminated (Fig. 9). It is also a highly cooper-
Driving Forces ative process, since the larger the number of
Due to unfavorable headgroup repulsions of like- merging rods, the larger the reduction in
charges, ionic surfactants usually assemble in end-caps. For n merging aggregates, the number
Rec considered, due to the interaction of charges
along the micelle backbone that favor shorter
R micelles. Equation 23 then takes the form
< L > F1=2 exp½ðEc Eel Þ=2kB T (25)
The electrostatic contribution Eel is given by
Eel ¼ kB TlB Rv2 F1=2 (26)

Surfactant Self-Assembly, Fig. 9 Schematic represen-
tation of two rodlike micelles, showing the end-caps of
different curvature. If both micelles merge, two end-caps where lB is the Bjerrum length, R the micelle
vanish. R and Rec are the local radii of the micelle and cross-section radius, and n is the effective charge
end-cap, respectively. In this case, Rec ¼ R, and per unit length.
Ho,mic ¼ 1/2R and Ho,ec ¼ 1/R If the radius of the micelle R is similar to the
end-cap radius Rec, as in Fig. 9, the end-cap
of end-caps decreases by 2n2. Hence, when energy Ec can be estimated through
a transition from spherical to rodlike micelles
occurs, these have a tendency to grow to very 3
Ec ¼ 8pk RH0 þ 4p
k (27)
large sizes, in some cases to the micron scale 4
(Dreiss 2007).
When two rods merge, there is also a decrease which requires knowledge of the spontaneous
in translational entropy since the overall number curvature Ho and moduli k and k . Alternatively,
of aggregates is reduced. Considering these two it can be determined experimentally through the
opposing factors toward micelle growth, and zero shear viscosity, by the relation
through a mean-field treatment, the average
1=2
micelle size (or contour length < L>) can be 0 k1 F7=2 exp½ðEc Eel Þ=2kB T (28)
obtained by
where k1 is a temperature dependent rate-constant
< L > F1=2 expðEc =2kB TÞ (23) (per unic arc length and per unit time), related
with the break time of an elongated micelle.
where F is the surfactant volume fraction, Ec the In addition to curvature changes, micelle growth
end-cap energy required to form two hemispher- can also be induced by concentration increase, like
ical end-caps as a result of rod scission, kB is the in the CTAB-water system (no added salt), and
Boltzmann constant, and T the absolute tempera- some nonionics. In this case, Eq. 23 should
ture. The size distribution is broad, and the num- have the factor F1/2 replaced by Fa, in which a is S
ber density of micelles of size L is given by a value still not precisely known, varying signifi-
cantly in the literature (Dreiss 2007).
nðLÞ expðL= < L >Þ (24)
Living Polymers and Their Structure
From Eq. 23, it is clear that a temperature Micelle growth as dictated by Eq. 23 generally
increase will oppose micelle growth, due to favors very long and flexible aggregates, so-called
a drop in the number of rods that causes the wormlike micelles, where L can easily reach the
translational entropy to decrease. As mentioned micrometer length scale. These aggregates possess
before, an exception comes from nonionic sur- very similar properties with polymers in solution,
factants, where an increase in temperature favors for which they are also commonly referred to as
micelle growth, due to a decrease in Ho. living polymers. The term “living” comes from the
For charged micelles in the absence of excess fact that, unlike the covalently bound polymer
salt, an additional electrostatic term Eel has to be chains (that can only relax through reptation),
Surfactant Self-Assembly, Fig. 10 Schematic repre- overall radius of gyration Rg, the persistence length lp, and
sentation of a wormlike micelle, showing characteristic the cross-section radius R. For clarity, the magnification is
length-scales: the contour length L (see dashed line), the not drawn to scale (Adapted from Dreiss (2007))
wormlike micelles can break and rejoin, being Equation 29 is not very useful from a practical
more dynamic systems. Nevertheless, structurally, point of view. Alternatively, and in analogy with
both systems are very similar. polymers, if the wormlike micelle is viewed as
In addition to the overall contour length < L > a random walk with step length 2lp (or b), for
and R, elongated micelles have additional struc- a sufficiently high number of Nr (or high L, since
tural parameters (Dreiss 2007; Evans and L / Nr ) we have
Wennerström 1999). Bearing in mind that the
local structure of the elongated micelle is cylindri- D E l L
p
R2g ¼ (30)
cal, the length in which this cylinder keeps its 3
orientational correlation (i.e. remains roughly
straight) is known as the persistence length, Due to excluded volume effects, the wormlike
lp (Fig. 10). This parameter is of central impor- micelle cannot be properly modeled by a random
tance since it gives a measure of the rigidity of the walk, and Rg as given by Eq. 30 is slightly
micelles, which strongly affect the overall system underestimated. Still, this equation does capture
properties. A high lp /< L > ratio means that the the essential qualitative features, the most impor-
micelle is very rigid, while the opposite is true for tant being the increase of the overall coil size Rg
flexible systems. Normally, charged micelles are for fixed L just by increasing lp. This is expected
more rigid than uncharged ones. An alternative to since a more rigid internal structure forces the
the persistence length is the Kuhn length b, which coil to expand.
is simply twice the value of lp (b ¼ 2lp) and can As with polymers, when the surfactant volume
also be used. fraction increases, at a certain threshold value
Another important parameter is the radius of (the overlap concentration, F*), the micellar
gyration, Rg. It gives a measure of the coil size of coils start to overlap, while the micelle ends
the elongated micelle. Formally, it is a measure of start to entangle with neighboring coils (Fig. 11).
the average distance of the micelle segments rel- This entanglement increases drastically the
ative to its center of mass (Fig. 10). If the worm- viscosity of the solution, while concomitantly,
like chain is considered to be an ensemble of Nr a network with solid-like behavior is formed.
connected rigid rods with length lp and distance rj Hence, above F*, a gel is formed, with the
to the center of mass rCM, Rg can be obtained by micelles entangled in a network displaying
solid-like behavior and the water diffusing within
D E 1 X Nr D
2 E the lattice like a normal liquid. An additional
R2g ¼ rj rCM (29) structural parameter here is the mesh size, z, of
N j¼1
the network.
Surfactant
Self-Assembly,
Fig. 11 Schematic
representation of the state
of the wormlike micelle
solution with increasing
surfactant volume fraction.
z is a hydrodynamic
correlation length
(typically identified with
mesh size of the network)
Branched Micelles
After rods are formed, if Ho continues to decrease H2
and before a lamellar phase is formed, the
micelles may become branched. This happens
because a branching point has local curvature
H2 smaller than in the rest of the cylindrical
micelle, H1 (Fig. 12). Hence, by adjusting the
number density of branches f, the micelle can H1
reach the overall spontaneous curvature Ho ¼ f
H2 + (1f) H1 of the surfactant. In nonionics, the
branch point density is coupled with surfactant
volume fraction F, increasing when F increases.
Hence, when branches are formed, liquid–liquid Surfactant Self-Assembly, Fig. 12 A branch point in
an elongated micelle. The branch point has lower curva-
phase separation is also observed (cloud point in ture than the rodlike sections (H2 < H1)
nonionics).
Compared with entangled wormlike micelles,
branched micelle systems have a much lower
viscosity, since branches can slide along the surfactant mixtures, however, the existence of
cylindrical backbone reducing significantly the more than one compound makes it possible for
relaxation time. a micellizing surfactant of higher spontaneous
curvature to accumulate at these edges, stabiliz- S
Disklike Micelles ing the disks. This should still have some ener-
In addition to cylindrical micelles, bilayer getic cost due to the entropic loss in segregating
disklike micelles can also occur in surfactant the surfactants in specific locations, unless they
systems, though they are much rarer than the are immiscible.
former (Almgren 2000). Indeed, in most of the Generally, oil, cosurfactant, and other compo-
cases where they are observed, they occur as nents can also be incorporated in the bilayer,
intermediate structures in the micelle–vesicle affecting its physical properties.
transition. Compared to cylindrical micelles, in
the disk-shaped geometry, the end-cap extends Bilayers
over the entire perimeter of the disk (Fig. 13). Bilayers are thin membranes made of two layers
This has a high energetic cost, and consequently, of amphiphilic molecules, occurring both natu-
disklike micelles are almost always disfavored rally and artificially. Bilayers can be found in
with respect to lamellae or vesicles. In some three different morphological forms: closed
Surfactant Self-Assembly, Fig. 13 A disklike micelle composed of two partially immiscible surfactants. The dark
gray has higher Ho and prefers to accumulate on the disk rim. Rdm is the disk radius and d is the bilayer thickness
Surfactant
Self-Assembly,
Fig. 14 Schematic
drawing of (a) local bilayer
structure and (b) type of
bilayer phases formed:
vesicles, lamellar (La), and
bicontinuous (or sponge)
L3 phase. d is the bilayer
thickness, R is the vesicle
radius, d the lamellar repeat
distance (or lamellar
spacing), and R1 and R2 are
the principal radii of
curvature of the L3 phase
(vesicles), planar, and bicontinuous (Fig. 14). parameters regulating the function of membrane
The origins of these different forms, as well as proteins and other relevant physiological proper-
their physical properties such as elasticity/rigid- ties (Evans and Wennerström 1999).
ity of the membrane, can to a great extent be Bilayers can also be produced artificially by
understood using the concepts of curvature elas- dispersing appropriate amphiphiles in solvents.
tic theory, previously introduced (Eq. 5 and They are very appealing colloidal aggregates for
Table 4). fundamental studies in colloid science and serve
Natural bilayers are found in living organisms as cell models for biochemical and biophysical
mostly in the form of cell membranes, which are investigations (Lasic 1993). They also find
composed mainly of phospholipids but also important uses in technical applications, for
proteins, cholesterol, and glycolipids, among instance in health-care products, drug delivery,
other biomolecules (Lipowsky 1995). Cell mem- and microreactor chemistry, to cite but a few
branes act as diffusion barriers and generate an applications.
intracellular environment that serves to organize Bilayers are commonly formed by surfactants
the multitude of metabolic processes that occur in of the swelling type, which have a solubility in
the living cell. Additionally, the lipid bilayer acts water that is much lower than that of typical
also as a two-dimensional solvent for membrane micellizing surfactants. Hence, the unimer con-
proteins. Many of the physical properties such as centration in bilayer systems is much lower and
fluidity/rigidity are expected to be key dynamic processes that depend on unimer
Surfactant Self-Assembly, Table 4 Mean curvature H, Gaussian curvature K, local curvature energy gc, and relative
stability of different bilayer morphologies for one-component bilayers (Ho ¼ 0)
Morphology H K gc Stability
Cylinder 1/2R 0 gc ¼ 2Rk 2 Unstable compared to planar bilayers (gc is
always positive)
Lamellar sheet 0 0 0 Most stable morphology, unless k is positive

< 2k
(favoring a sponge phase) or k
(favoring vesicles)
Vesicle 1/R 1/R2 1 < 2k

Stable if k
gc ¼ Þ
ð2k þ k
R2
Sponge 0 1/R2 gc ¼ Rk2 >0

Stable if k
exchange are much slower. Nonetheless, upon Mismatches between the cross-sectional areas of S
carefully tuning of the overall curvature by chains and headgroups can lead to tilted or inter-
adding salt, cosurfactant, or increasing the con- digitated chains, or rippled surfaces (Fig. 15).
centration, bilayers can also be found in Above the gel-to-liquid crystal transition temper-
micellizing surfactant systems. Another possibil- ature, the chains gain conformational disorder
ity is by tuning the curvature of nonionic surfac- (fluidity) and hence the bilayer becomes thinner,
tants through temperature increase. somewhat less than twice the surfactant molecu-
Depending on temperature, bilayers can be lar length. The surfactant molecules have now
found in two different states. Below a certain enhanced mobility along the bilayer film, which
temperature (gel-to-liquid crystal transition behaves like a two-dimensional liquid and is in
temperature), the alkyl chains are crystallized fact in a liquid-crystalline state. Even though
and molecular motions severely restricted, with lateral diffusion of surfactant/lipid molecules
the bilayer behaving essentially like a hydrated along the bilayer plane is fast, their exchange
two-dimensional solid, designated as a gel phase. from one layer to the other through the center
Surfactant
Self-Assembly,
Fig. 15 Bilayer structures
in the gel different phases:
upper left, Lb; lower left,
Pb; upper right, LbIΝ; and
lower left, Lb’
(Adapted from Evans and
Wennerström (1999))
(flip-flop motion) is very slow. This is mainly due case is thus a result of a balance between the disk
to the unfavorable interactions of the headgroups edge unfavorable energy and the bending energy
with the oily bilayer core. In fact, the existence of required to close the disks and form the vesicles.
these two domains of very different polarities Vesicles produced in this way, by breaking up
within bilayers makes them effective diffusional swollen lamellae using mechanical energy, are
barriers for very polar or very nonpolar metastable. With time, they flocculate and fuse,
molecules. reforming the equilibrium lamellar phase in
excess solvent they were derived from. This is
Vesicles mainly due to the fact that the bilayers generally
Vesicles or liposomes are bilayer shells ranging have a spontaneous curvature of zero value, and
from a few nanometers to some micrometers, bending them requires extra energy (Lasic et al.
enclosing a pool of solvent in their interior. 2001). This is the general case in most of the
They are important particles in cell biology, par- vesicles, for example, those formed in binary
ticipating in many cell mechanisms. Besides, surfactant–water and lipid–water systems,
their unique properties of encapsulation and com- reported in the literature. However, in some
partmentalization can be exploited toward prac- systems (typically bilayers with more than one
tical applications, such as drug delivery and component), vesicles are thought to exist in ther-
microreactor chemistry. A good understanding modynamic equilibrium, rather than only kineti-
of their properties and formation mechanisms is cally stabilized. Comprehensive reviews are
thus of the utmost importance (Antonietti and available in the literature covering both experi-
Förster 2003; Lasic 1993; Marques 2010). mental investigations on this topic and theoretical
Artificial vesicles are normally produced by models behind them (Marques 2010; Šegota and
dispersing a lamellar liquid-crystalline phase in Težak 2006).
excess solvent and adding external mechanical Different explanations and mechanisms are
energy, for instance by sonication or extrusion. put forth for the thermodynamic stability of
The lamellar phase is broken into small bilayer vesicle aggregates. One of these cases is thought
disks that have edges with exposed hydrophobic to occur when k < 2k (Table 4). This situation
alkyl chains in contact with the solvent. As is, however, barely observed, since the bending
a response to this destabilization of the lamellar modulus can take values from a few to tenths of
phase, at a critical disk radius, the disks will close kBT units, whereas k is (in modulus) in the order
and form vesicles. The vesicle formation in this of a few kBT. A deeper discussion on these
such as stimuli–responsive formulations. Hence,
significant effort has been put on the understand-
ing of the nature of this structural change.
The vesicle-to-micelle transition, as well as the
reverse micelle-to-vesicle transition, may be
attained by a number of different ways. One of
the most common ones is to induce the change of
vesicles to micelles via the addition of detergent
(single-chained surfactant) to liposomes, normally
constituted by phospholipids or other double-
chained amphiphiles (Majhi and Blume 2002).
This will have the effect of decreasing the packing
Surfactant Self-Assembly, Fig. 16 A mixed bilayer parameter (increasing Ho),which after a certain
containing a double-chained and a single-chained amphi-
critical value will dictate the formation of micelles
phile: (a) with equal partition of the two surfactants, the
bilayer spontaneous curvature is zero; (b) for the case instead of vesicles. This transition is normally
where an asymmetric partitioning of the two surfactants continuous, and vesicles, disks, and micelles are
occurs, a spontaneous curvature different from zero is often found to coexist within the same macro-
attained. Legend: Hout curvature of the outer monolayer;
scopic phase over a narrow composition range.
Hin curvature of the inner monolayer; and Ho spontaneous
curvature The reverse micelle-to-vesicle transition can
be attained by diluting a mixed micelle solution,
as is the case for lecithin–bile salt mixtures. Since
moduli and on the parameters affecting them can the micelle-forming surfactant has a much higher
be found in the literature (Safran 1999). solubility than the swelling surfactant or lipid, it
An alternative way of having equilibrium ves- will migrate from the micelles to bulk in order to
icles is by a spontaneous curvature mechanism, keep the unimer concentration constant, as
using more than one surfactant type in the dilution progresses. Hence, the mixed micelles
bilayers (Safran et al. 1991). This is notably become increasingly enriched in the bilayer-
found in some mixtures of cationic and anionic forming surfactant, and a transition to bilayer
surfactants (cf. section “Basic Properties of Sur- disks is eventually induced (Fig. 17). These
factants”). This extra compositional degree of disks will have some single-chained surfactant
freedom enables the two bilayer leaflets to have stabilizing the rims, but they will still be very
distinct compositions, which in turn can result in unstable due to the unfavorable end-cap energy.
different spontaneous curvatures. Under some Hence, they will grow by colliding with each
conditions, especially when the inner layer has other. Since the single-chained surfactant is nor-
a curvature similar in magnitude to the outer layer mally ionic, this growth is controlled by electro- S
but opposite in sign (Hout –Hin), the overall statics, and the kinetics is usually slow (hours).
membrane curvature may be different from Small disks cannot close to form vesicles because
zero and favor vesicles instead of a planar the vesicle radius would be prohibitively small.
bilayer (Fig. 16). However, as soon as disks grow above a certain
critical radius (rcr), vesicles are not forbidden by
Vesicle/Micelle Transitions size any longer and disks rapidly close into
The transition between micelles and vesicles is vesicles. Vesicles formed in this way – by dilu-
a challenging problem in surfactant tion of mixed micelles – are normally not ther-
self-assembly studies (Evans and Wennerström modynamically stable, but due to the fact that
1999; Marques 2010). It is also relevant for many they still have some residual charge, they can be
biological processes, like membrane protein kinetically stable over long periods of time, with
reconstitution and fat emulsification by bile their size distribution remaining unchanged
salts, and in biotechnology-related applications, (Leng et al. 2002).
disk disk disk

formation growth closure
fast slow fast
rcr
(<seconds) (≈hours) (<sec)
Rves
Surfactant Self-Assembly, Fig. 17 Schematic representation of the mechanism of micelle-to-vesicle transition

involving disklike aggregates as intermediate structures in the lecithin-bile salt system (Adapted from Leng et al. (2002))
Conversely, the vesicle-to-micelle transition Lyotropic Liquid Crystals

can also be achieved in a similar way in
catanionic mixtures, that is, by mixing surfactants Definition and Classification of Liquid
of opposite charge (Marques 2010). By progres- Crystals
sively adding a micellizing ionic surfactant to Liquid crystals are a state of matter in which the
micelles of opposite charge, these aggregates degree of molecular ordering is intermediate
will start to grow into wormlike micelles, and at between that of the three-dimensional long-range
a certain critical mixing ratio, vesicles will start positional and orientational order found in
to form. The driving force for this process is the crystals, in one extreme, and the absence of
continuous change in Ps and Ho as both ionic short- and long-range order of liquids and solu-
surfactants associate (Schmölzer et al. 2002). tions, in the other extreme. The intermediate
Another classical transition studied is the order of liquid crystals implies some type of
vesicle formation when micelles are heated up long-range orientational order and the existence
in nonionic surfactants. As seen in section of periodic lengths, while the positional order
“Basic Properties of Surfactants,” these surfac- may or may not occur in some dimensions
tants have a spontaneous curvature that depends (Demus et al. 1998; Hamley 2007).
strongly on temperature. At lower temperatures, Liquid crystals, also referred to by
the mean curvature is positive, favoring micelles. mesophases, are fundamentally divided between
However, if temperature is increased to a value thermotropic and lyotropic phases. Thermotropic
where the spontaneous curvature is approxi- phases are formed by a broad range of molecules
mately zero, a lamellar phase is the preferred (designated by mesogens) by action of tempera-
structure. If this is done in a region of the phase ture, in the absence of any solvent. In contrast,
diagram where the lamellar phase coexists with lyotropic mesophases are formed in the presence
excess solvent, vesicles will be formed. Also in of solvents, thus requiring at least two compo-
this case, vesicle formation has been found to be nents in the system (amphiphile and solvent).
intermediated by disk formation (Bryskhe Even though thermotropic liquid crystals can
et al. 2005). be viewed to a certain extent as self-assembled
In summary, even though each system structures (and many amphiphiles also form
involves the change of different variables and them), this entry is centered in solvent-driven
different pathways, ultimately, the driving force self-assembly, and therefore, only lyotropic
for the transitions between vesicles and micelles mesophases will be considered. For an in-depth
is always a tunable change of Ps (and Ho), as overview of liquid crystals of both types, several
could be anticipated from the notions discussed references can be consulted (Collings and Hird
in section “Models for Surfactant 1997; Demus et al. 1998; Hamley 2007; Tiddy
Self-assembly.” 1980).
Surfactant Self-Assembly, Table 5 Common liquid- One of these factors arises from the fact that
crystalline phases and respective dimensionality (Adapted distances between aggregates are shorter, which
from Seddon and Templer (1995))
makes interaggregate interactions become stron-
Phase Description Dimensionality ger. Often, at such proximity, the net interaction
La Lamellar 1-D between aggregates is repulsive, due to
H Hexagonal 2-D short-range entropic (or hydration) forces
R Rectangular 2-D
(Israelachvili and Wennerström 1996), and there-
M Oblique 2-D
fore, the particles will try to order in a way that
I Discrete cubic 3-D
minimizes the overall Gibbs energy. The aggre-
Q Bicontinuous cubic 3-D
gates may also become deformed to minimize the
T Tetragonal 3-D
Rh Rhombohedral 3-D
unfavorable interactions and accommodate better
in the available volume (e.g., deformation of
spherical micelles to prolates).
Another factor, which is connected with the
Lyotropic liquid crystals display a rich poly- previous ones, is the physical impossibility of
morphism, showing phases made of aggregates packing more aggregates within a given lattice
that are either discrete or infinite, and which after a certain point in concentration is reached,
arrange in space – when there is some degree of for example, when the particles become
positional order – in one, two, or three dimen- overcrowded and there is no more volume left
sions. The most commonly used nomenclature for additional particles. Hence, as the concentra-
for these lyotropic systems is that proposed by tion keeps increasing, a transition to a structure
Luzzati (1968; Seddon and Templer 1995), that allows a larger amount of surfactant per unit
which will be adopted here, with the exception cell (concomitantly, a smaller amount of water) is
of the discrete cubic phase, which we will distin- thus necessary to avoid overcrowding.
guish from bicontinuous (Table 5). The lattice If only geometric arguments are taken into
type is denoted by a capital letter, for example, account, the optimal way of arranging surfactant
H for Hexagonal, whereas subscripts 1 and 2, as and water as the surfactant concentration is
mentioned before, are used to denote normal increased is L1 ! I1 ! H1 ! Q1 ! La !
(oil-in-water) and reverse (water-in-oil) phases, Q2 ! H2 ! I2 ! L2 (for sake of simplicity,
respectively. When applicable, for instance in the intermediate phases were excluded). As can be
lamellar phase, a Greek subscript is also used to noticed, this phase sequence is arranged in
denote the alkyl chain conformation: b for ordered decreasing order of surfactant film curvature. The
solid-like phases and a for liquid-like phases. rationale for this is obvious: an increase in surfac-
tant concentration implies a reduction on the
Formation of Lyotropic Liquid Crystals amount of water. This fact forces the aggregates S
The formation of lyotropic liquid crystals is very to curve into water, enclosing it and promoting
common in surfactant systems. These structures reverse phases. In surfactant–oil systems, the
are normally found at high surfactant concentra- sequence would be reversed if surfactant concen-
tion, or when the surfactant is poorly soluble in tration was to be increased at the expense of oil.
a solvent (usually water), forming a mesophase The phase sequence described above is often
with a limited capacity of incorporating solvent called the Fontell scheme, briefly mentioned
in its structure. As with solution phases, the pack- before (Fig. 5). It is only an idealized sequence,
ing parameter (Ps) and spontaneous curvature since hardly any surfactant system will show all
(Ho) continue to be central parameters to describe of these phases. It also assumes for simplicity that
and predict the liquid-crystalline structure. How- the starting aggregate is discrete spherical
ever, additional factors have now to be taken into micelles. However, for instance, in the case of
account when predicting the overall structures a surfactant that forms rodlike micelles,
formed. a hexagonal phase rather than a discrete cubic
would likely be the first lyotropic phase to form

when the rods become densely packed. Besides,
swelling surfactants (Ps 1) almost invariably
form bilayer phases as the first aggregation state.
We will now focus on a detailed view of the
structure of the different mesophases and some of
their most characteristic physical properties.
Surfactant Self-Assembly, Fig. 18 Left: A body-
Discrete Cubic Phases centered cubic phase made of discrete prolate-shaped
Normal discrete cubic phases usually occur micelles (Adapted from Fontell (1990)). Right: A reverse
between micellar solution phases and hexagonal cubic phase with Fd3m lattice, evidencing the two types of
phases (c.f. Fig. 5). These phases are optically micelles (Adapted from Seddon and Templer (1995))
isotropic and normally very viscous. The struc-
ture is composed of small discrete micelles
arranged in a three-dimensional cubic lattice In the H1 phase, the surfactants aggregate in
with typical dimensions in the 2–25 nm range. circular cylindrical micelles that pack into
The most common lattice type is the a hexagonal lattice with a continuous water
body-centered cubic (BCC) with space group region filling the space between the micellar
Pm3n. The micellar discrete units are similar to rods. The micelle radius is close to that of the
those found in the solution phases (c.f. section surfactant molecular length. In the inverse H2
“Spherical Micelles”) but commonly distorted to phase, the structure is composed of reverse micel-
a prolate shape (axial ratio 1–2), which allows lar cylinders with water cores surrounded by the
a more efficient packing (Fontell 1990; Jönsson surfactant headgroups, with the remaining
et al. 1998). In Fig. 18a, one of the proposed volume filled by the fluid hydrocarbon chains at
models for the structure of this phase is shown. essentially uniform liquid hydrocarbon density
Reverse discrete cubic phases can also form. (Seddon and Templer 1995). The water core
Usually, for geometric reasons and due to repul- radius typically varies in the range 2–20 nm. In
sive interactions, the headgroups are not able to the hydrophobic region, additional oil can
pack by themselves, and some water is required normally be also incorporated. For some systems,
to fill the void and allow the aggregates to adjust the shape of the cylinders may be deformed,
to the preferred curvature. The radius of the water leading to a 2-D lattice of lower symmetry, such
droplets in these reverse micelles is typically in as rectangular or oblique (intermediate phases).
the range 2–20 nm. The first well established Due to this structural anisotropy in two dimen-
structure to be composed of reverse micelles is sions, hexagonal phases are very viscous and
in fact composed of two types of micelles with optically birefringent.
very different radii arranged on a Fd3m lattice
(eight large and sixteen small micelles per unit Lamellar Phase
cell, Fig. 18b). This phase usually implies the Lamellar phases are stacks of bilayers in the
presence of at least two amphiphilic components liquid-crystalline state separated by layers of
in order to partially segregate into the two types solvent. They are characterized by a repeat
of micelles (Seddon and Templer 1995). distance d ¼ l + d, where d is the bilayer thickness
(somewhat less than twice the surfactant mole-
Hexagonal Phase cule length) and l is the solvent layer thickness.
This phase consists of a hexagonal arrangement Bilayers can swell either in water or oil, in
of unlimitedly long cylindrical micelles (Fig. 19). a process that leads to an increase of d. This
As with cubic phases, these micelles can be repeat distance can vary over wide ranges and is
normal or reverse, yielding normal (H1) and governed by a balance of attractive and repulsive
reverse (H2) hexagonal phases. forces. Compared to most other liquid-crystalline
Surfactant Self-Assembly, Fig. 19 Left: Illustration of reverse hexagonal phase (H2). As with H1, the distance
the normal hexagonal phase (H1). The distance between between each rod is constant, and the water channels’
each rod is constant, and the rod radius is similar to the radius is typically 2–20 nm
surfactant extended length (lhc). Right: Illustration of the
phases, the La phase is not necessarily too

concentrated in surfactant.
Bilayers in lamellar phases are in a 2-D
liquid-like state (Fig. 14), and may have struc-
tural defects that cause them to close up on them-
selves, forming multilayer tubes or onion-like
shells (Mortensen 2001). The structural anisot-
ropy of this phase implies different refractive
Surfactant Self-Assembly, Fig. 20 Cubic phases of
indices between the different spatial directions, Ia3d (left) and Pn3m (right) reverse structures (Adapted
yielding characteristic birefringent features under from Seddon (1990))
polarized light.
One of the best examples for a natural lamellar Reverse bicontinuous phases normally occur
phase is the outermost layer of the skin, the between lamellar and reverse hexagonal phases.
stratum corneum (Bouwstra et al. 2003). Typically, they are based on the Ia3d, Pn3m, and
Although this layer has a complex structure, one Im3m space groups, with the surfactant arranged
of its essential elements is the existence of in a periodic minimal surface, separating two
a lamellar liquid crystal composed of stacked independent water networks. In the Ia3d group,
bilayers that act as an effective barrier for the surfactant is arranged in a gyroid-minimal
strongly polar and strongly nonpolar molecules. surface and the water channels are connected
three by three (Fig. 20, left). In the Pn3m, the
Bicontinuous and Sponge Phases surfactant is arranged on a F-minimal surface, S
As with discrete cubic phases, bicontinuous ones with the two water networks displaying
may be either normal (oil-in-water) or reverse a tetrahedral structure (connected four by four)
(water-in-oil). However, now, they can be formed arranged in a double-diamond lattice (Fig. 20,
by either interconnected micelles or bilayers with right). In the Im3m case, the surfactant is
negative Gaussian curvature (c.f. Eq. 6 and Fig. 4). arranged in a Schwartz P-minimal surface, with
Normal bicontinuous phases are typically the two water networks displaying an orthogonal
based on the Ia3d space group, consisting of two structure with the water channels connected six
interwoven (but not connected) networks of sur- by six (Fig. 21).
factant micelles connected three by three at bifur- The sponge phase is a disordered version of
cation points, separated by the G-minimal surface the cubic bicontinuous, and as such, it is
(Seddon and Templer 1995). These phases typi- a solution phase. It occurs mainly in
cally occur between normal hexagonal and microemulsion systems, which will be discussed
lamellar phases. further in this entry.
Dispersions of two immiscible liquids (like oil

and water), with an emulsifier agent present, that
are thermodynamically unstable are not
microemulsions, but rather classified as emul-
sions. Besides this fundamental thermodynamic
difference, microemulsions and emulsions also
differ considerably in their structure and dynam-
ics (Adamson and Gast 1997; Evans and
Wennerström 1999; Hunter 2001). Also, equilib-
rium systems composed of oil, water, and surfac-
Surfactant Self-Assembly, Fig. 21 Im3m tant, but with a liquid-crystalline structure
(bilayer-based) bicontinuous cubic phase with the two (e.g., discrete and bicontinuous cubic), are not
distinct solvent networks highlighted in light and dark classified as microemulsions either.
gray. This structure corresponds to a P-minimal surface We will now focus on the most important
self-assembly features of microemulsion phases,
and leave further distinctive features between
Intermediate Phases microemulsions and emulsions for later in this
In addition to the more commonly found liquid- section.
crystalline phases, there is a range of additional
mesophases observed in some systems between Microemulsions: Structure and Behavior
the hexagonal and lamellar phases, different from Usually one classifies microemulsions as:
cubic bicontinuous ones. Those mesophases are (1) oil-in-water (o/w), if the system is composed
normally anisotropic in structure, birefringent, of discrete oil droplets dispersed in the aqueous
and more fluid than cubic ones. They can be medium; (2) water-in-oil (w/o), when water
divided in three groups according to symmetry: droplets are dispersed in the oil medium; and
(1) rectangular or ribbon structures; (2) layered (3) bicontinuous, if both mediums are continu-
mesh structures; and (3) bicontinuous structures ously dispersed in one another, without the
(Fig. 22). For a deeper view of intermediate formation of discrete droplets (Fig. 23).
phases the reader is suggested to consult the spe- Bicontinuous microemulsions can also be
cialized literature (Holmes 1998; Kékicheff and named as sponge phases, and can be viewed as
Cabane 1988; Seddon and Templer 1995). a disordered version of the bicontinuous
cubic phase.
This rich phase behavior makes
Microemulsions and Emulsions microemulsions challenging structures from a
fundamental point of view, as well as important
Microemulsions are thermodynamically stable in a broad range of technical applications, such as
solutions containing a strongly nonpolar compo- emulsion technology, detergency, food engineer-
nent (usually oil) and a strongly polar component ing, drug delivery, and nano-template chemistry
(usually water), stabilized at the interface by an (Schwuger et al. 1995).
amphiphile, typically a surfactant. They are The complex phase behavior exhibited by
homogeneous, macroscopically clear and opti- microemulsions can be understood by the relative
cally isotropic, but present a rich polymorphism amounts of the three components and by the
at the nanoscopic scale (Danielsson and Lindman mechanical properties of the amphiphilic film.
1981). This is due to the different ways of com- The mechanical properties of the surfactant film
partmentalization of the two bulk media (oil and can be thermodynamically described by the flex-
water) in the available volume, with the surfac- ible surface model introduced before, namely,
tant that is situated at the oil–water interface Eq. 5. Because now only one monolayer is pre-
playing a central role. sent (when compared with bilayers), Ho assumes
Surfactant
Self-Assembly,
Fig. 22 Examples of
intermediate phase
structures. Left: central
rectangular ribbon phase;
Right: rhombohedral mesh
phase (evidencing the
characteristic ABC
stacking)
Surfactant
Self-Assembly,
Fig. 23 Schematic
representation of the three
types of microemulsions.
Within these three major
types of microemulsions,
different structures may
exist. The bicontinuous
microemulsion figure on
top was reprinted with
permission from Arleth
et al. (2001)
a more central role. For instance, if Ho is positive, For this discussion, it is easier if one considers
the surfactant film has a tendency to curve toward two types of radii. One is the optimal radius
oil (positive curvature), and form o/w dictated by the spontaneous curvature of the sur-
microemulsion droplets, whereas if Ho is nega- factant film (Ro ¼ 1/Ho). The other one is the
tive, the film prefers to curve toward water optimal radius dictated by composition (Rc). To
(negative curvature), and w/o droplets are estimate Rc one assumes that all the amphiphile is
favored. If Ho is zero, the film has no preference located and homogeneously distributed in the S
for water or oil and typically a bicontinuous oil–water interface of the droplet. Hence, for the
phase is formed. case of spherical droplets, there is a constraint
between total droplet volume (the volume frac-
Self-assembly Control Through Composition tion of the discontinuous phase, plus a fraction
Having the mechanical properties of the surfac- a of the surfactant film), and total droplet surface
tant film established by Eq. 5, one can predict in (obtained from the interfacial area of the film).
a simple way the global self-assembling behavior Thus, in the case of o/w microemulsions, Rc is
of microemulsions, only by knowing the compo- given by (Evans and Wennerström 1999):
sition of the system, that is, the relative amounts
of water, oil, and surfactant. Let us consider the Rc ¼ 3ðFo =Fs þ aÞðvs =ao Þ (31)
formation of discrete droplets at relative high
dilution, so that interaggregate interactions can where Fo and Fs are the volume fractions of oil
be neglected. and surfactant, respectively, vs is the molecular
a Fixed Ho
fdp /fs decreases
b Ho decreases
Fixed composition
Surfactant Self-Assembly, Fig. 24 Schematic repre- formed by decreasing the Fdp/Fs ratio, Ho is still nonzero
sentation of the droplet change in structure with variation and the bicontinuous network is swollen in the continuous
of (a) the relative volumes of the dispersed phase Fdp and phase. If a bicontinuous network is formed by changing
surfactant Fs, and (b) spontaneous curvature Ho. Situa- Ho to zero, the surfactant film is balanced and prefers
tions (a) and (b) are different. If a bicontinuous network is equal amounts of water and oil
volume of the surfactant, and a0 its interfacial deviance from the spontaneous curvature. In this
area in the film. In the w/o microemulsion case, case, however, rather than phase-separating
one only has to replace the volume fraction of oil excess surfactant in order to reach the ideal Ro,
by the volume fraction of water Fw. The actual it is energetically more favorable for the system
radius and shape that the microemulsion droplet to relax by deforming the droplets to non-
will adopt depends on these two radii, but typi- spherical shapes, like prolates, wormlike
cally for spherical droplets the mean radius is in micelles, or branched micelles, where the aver-
the range of 5–20 nm. age curvature is lower than the equivalent spher-
If Rc is larger than Ro, the ideal droplets dic- ical shape (Fig. 24).
tated by composition are bigger than the optimal In summary, by manipulating the relative
value dictated by the spontaneous curvature, and amounts of water and oil for a fixed Ho,
the surfactant film would have a higher energy the shape of microemulsion droplets can shift
due to its deviance from the curvature Ho. In this from spherical to prolate or even elongated
case, the system can lower the total energy by micelle shapes.
phase-separating the excess oil, and achieving the
optimal curvature Ho. Due to the fact that there is Self-assembly Control Through Spontaneous
a certain amount of oil left to emulsify, this Curvature
attained boundary is called the emulsification The spontaneous curvature Ho of microemulsions
failure boundary. The droplets here incorporate alone can also be tuned in various ways, as briefly
the maximum quantity of oil to achieve the min- mentioned in section “Additional Parameters
imum curvature energy. Hence, if the entropy of Affecting Self-assembly.” The same general
mixing can be neglected, droplets are expected to principles as in binary micellizing surfactant
be spherical. systems apply for microemulsions.
For the opposite case of Rc being smaller than Thus, the addition of salt to ionic surfactant-
Ro (e.g., a situation where the amount of oil is based microemulsions screens the charges, lead-
smaller, and not enough for the droplets to incor- ing to a decrease in Ho. A cosurfactant will lead to
porate, reaching the ideal Ro), the surfactant an even more significant decrease in Ho due to an
would have a higher energy too, due to its increase in the hydrophobic volume of the
Surfactant
Self-Assembly,
Fig. 25 The different
droplet size and curvature
of an emulsion (left) and
a microemulsion (right)
surfactant film, in addition to a slight decrease in Emulsions

headgroup repulsions. A third method of lower- Emulsions are composed by two immiscible
ing the spontaneous curvature is through hydro- liquids, typically water and oil, and emulsifier.
carbons (oils) that are also able to penetrate Under intense agitation, fine droplets of one
slightly on the oily part of the surfactant film. phase dispersed in the other (either oil-in-water
The penetration is higher the shorter the chain or water-in-oil) may form, stabilized by the emul-
length of the oil is, making the curvature decrease sifier. Typical emulsifiers can be surfactants,
stronger. water-soluble polymers, and hydrophobic parti-
Also in close similarity with binary surfactant- cles. Even with a proper stabilizing agent, these
water systems, raising temperature (T) in non- systems are not thermodynamically stable, unlike
ionic surfactants of the CmEn type will lower microemulsions; hence, after some time they
Ho, leading to a very rich phase behavior in breakdown and undergo macroscopic phase
microemulsions. For ionic surfactants, the influ- separation. The breakdown process may result
ence of T is the opposite. On one hand, the gauche from different mechanisms, such as sedimenta-
conformations of the hydrocarbons are increased tion/creaming, flocculation, coalescence, and
(reducing Ho), but on the other hand, the electro- Ostwald ripening (Jönsson et al. 1998;
static Gibbs energy that is dominantly entropic is Tadros 2005).
also increased (increasing Ho). These two effects Emulsions and microemulsions are fundamen-
combined lead to a slight increase of Ho with T; tally different. The most distinctive difference is
nevertheless, for ionic surfactants temperature the thermodynamic stability mentioned before,
effects are weaker than in nonionics. but other characteristics such as droplet size,
The decrease of Ho, while keeping the system interfacial area, and monolayer curvature also
composition constant, could lead to a growth of differ (Adamson and Gast 1997; Evans and S
the spherical droplets, in the same fashion as Wennerström 1999; Hunter 2001). While
decreasing the amount of oil in o/w droplets microemulsion droplets are typically small
(Fig. 24). As Ho becomes smaller, bigger droplets (5–20 nm), emulsion droplets have sizes typically
would be favored. However, because the compo- between 1 and 10 mm (about three orders of
sition is fixed, this is forbidden and the droplets magnitude larger). Due to the large size of the
distort instead to prolates as a means of decreas- droplets the interfacial area between oil and water
ing the average curvature. If Ho approaches zero, is also much smaller when compared to
a bicontinuous phase is formed. This microemulsions, and the interfacial curvature is
bicontinuous structure is somewhat different also much smaller (Fig. 25).
from the ones found for nonzero Ho. Here, In addition, for emulsions, in contrast with
because Ho is zero, the surfactant film is microemulsions, factors such as agitation time
balanced, and thus it prefers to accommodate and agitation strength, temperature of mixing
equal amounts of oil and water. and storage, among other parameters related to
the process of droplet formation, affect strongly (i.e., there is no transfer of energy between the
the final droplet size and stability. For this reason, sample and the beam). In most scattering exper-
many of the concepts of self-assembly based on iments, a well-collimated beam of radiation
surfactant film spontaneous curvature and pack- (light, neutrons, or X-rays) with wavelength l is
ing parameter (Ho and Ps) so useful in micelles, incident on a sample. Most of the incident radia-
bilayers, and microemulsions, are not directly tion will be transmitted, while some fraction can
applicable in emulsions. In fact, many of the be absorbed and some can be scattered at an angle
concepts used in the prediction of emulsion y > 0. The angle-dependent scattering intensity is
behavior and stability are more complicated related with the density of scattering material of
than the simple concepts introduced in this entry the sample by a Fourier transformation, and
since a number of additional parameters (often of therefore, contains the complete information on
dynamic nature) have to be included (Jönsson the structure of the sample in the available angu-
et al. 1998). lar range. Because the wavelengths of X-rays and
neutrons are usually much smaller than the size of
the particles under investigation, the scattered
Characterization Methods fields deviate from zero essentially only at small
scattering angles. For this reason, the terms
Several experimental methods can be used to small-angle X-ray scattering (SAXS) and
characterize surfactant aggregates and phases. small-angle neutron scattering (SANS) are used.
Owing to the richness in structure and dynamics Rather than angles (y), a more convenient
of these phases, covering different characteristic parameter to use in small-angle scattering is the
time scales, spatial lengths, thermodynamic scattering vector q, defined as:
states, and media (e.g., in heterogeneous sys-
tems), a single characterization method is usually 4p y
q¼ sin (32)
insufficient to obtain the needed information. l 2
Therefore, a combination of methods that com-
plement each other is typically necessary for The parameter q is also related with real space
a good picture of the systems at microscopic distances r through
level. Moreover, each technique presents both
strengths and limitations in the characterization 2p
r¼ (33)
of surfactant self-assemblies. A concise overview q
of the most informative methods (scattering,
microscopy, and self-diffusion NMR) will be The theory underlying scattering from X-rays,
given in this entry, with reference to other neutrons and light (in this last case, only when the
available ones. scattering particles are small or have a very low
refractive index ratio to the solvent) is similar,
Radiation Scattering and what is different is the nature of the interac-
Scattering methods can be used to investigate the tion between these different radiations and the
structure, the organization, and the dynamics of scatterers. X-rays interact with the electrons,
self-assembly systems with radiation such as neu- and the contrast between scatterers and solvent
trons, X-rays, and light (Glatter and Kratky 1982; is given by the electron density difference
Lindner and Zemb 2002). The structural informa- between both. When it comes to neutrons, the
tion that can be obtained can span spatial lengths contrast is given by the difference of scattering
from one to thousands of nanometers, and time lengths, and with light, the contrast is propor-
lengths of seconds to picoseconds. tional to the refractive index increment.
In this section we will only consider elastic One major advantage of neutron scattering is
scattering, where the energy of the incident beam the very large scattering length difference
is practically conserved in the scattering event between protium (1H) and deuterium (2H).
Surfactant Self-Assembly, Fig. 26 Different ways of between the shell and the matrix is slightly lower.
changing the contrast between core-shell micelles and (d) The scattering length density of the micelle shell is
their matrix. (a) The scattering length densities from sol- matched to the solvent, hence, only the micelle core is
vent, shell, and micelle core are different. (b) The scatter- visible to neutrons. (e) The scattering length density of
ing length density of the micelle core is matched to the the solvent is matched to the micelle shell, thus, only the
solvent, hence, only the micelle shell is visible to neu- micelle core is visible to neutrons. Compared to (d), the
trons. (c) The scattering length density of the solvent is contrast between the matrix and the micelle core is higher.
matched to the micelle core, hence, only the micelle shell (f) The three scattering lengths are equal, and the system is
is visible to neutrons. Compared to (b), the contrast said to be contrast matched, resulting in no scattering
This enables the researcher to hide or highlight fitted with the form factor of polydisperse vesi-
certain sections of aggregates through change of cles, and through the fit the average radius, the
the relative amounts of protium and deuterium size polydispersity, and the bilayer thickness can
in the solvent and aggregates. A good example be obtained. If a good initial guess is not available
can be extracted from microemulsion droplets for the particle structure, this method is not
(Fig. 26), where by partial deuteration of either efficient.
the solvent, the oil, or the surfactant, one can Alternatively, the scattering data may allow
observe either the oily core, the surfactant mono- the extraction of additional structural parameters
layer, or both (Bagger-Jörgensen et al. 1997). that can either be used as important as-is infor-
mation, or used to refine a model for subsequent
Solution Scattering form factor fittings. If the scattering data extends
As mentioned before, the q-dependent scattering to sufficiently low q values (low angles), the
intensity is related with the density of scattering radius of gyration can be obtained through the S
material of the sample by a Fourier transforma- Guinier approximation:
tion. If the individual particles are uncorrelated
from each other (i.e., sufficiently high dilutions),
IðqÞ ¼ Ið0Þeq Rg =3
2 2
(34)
the scattering curve I(q) versus q contains infor-
mation on the orientational average of the particle
in the so-called form factor. To obtain the struc- In addition to Rg, and depending on the system,
ture of the particle under investigation, one the scattering data can be described by power law
method is to fit the scattering curve with known decays of the type I(q) / qD , where D is related
form factor expressions. This is a good approach with the dimensionality of the aggregate. D ¼ 0 for
if one has already some knowledge on the system globular micelles, D ¼ 1 for straight cylinders, and
under investigation. For instance, if from D ¼ 2 for bilayers. Usually flexible cylinder or
cryo-TEM it is known that the sample is com- branched micelles will have a D value between 1
posed of unilamellar vesicles, the data can be and 2, with the latter limit obtained for Gaussian
Surfactant Self-Assembly, Fig. 27 Left: hypothetical R1 and Rg1 < q < lp1. Right: pair-distance distribution
scattering curve for a wormlike chain (showed in the function p(r) for a homogeneous sphere (straight) and
insert), showing the scattering regimes of lp1 < q < homogeneous prolate (dashed line)
chains. Using log-log plots of I(q) vs. q these the tail to higher r. The IFT method is thus a good
asymptotic behaviors can be easily identified as way of gaining model-independent information
regions where the dependence is linear, with about the particle structure. If the particles have
a slope of –D. In some cases, an aggregate can inhomogeneous scattering length densities
display different regimes with different D in (e.g., a vesicle), the p(r) function may have
a single scattering curve. This is the case of some minima or even negative values at some r.
elongated micelles, where D is approximately 1 at
lp1 < q < R1 and ca. 1.7 at z1 (or Rg1) < q < Scattering From Liquid-Crystalline Phases
lp1 reflecting the different lengths present for this In liquid-crystalline phases, the periodic arrange-
type of aggregate (Fig. 27, left). ment of surfactant structures in planes of symme-
Alternatively, an Indirect Fourier Transform try gives rise to a diffraction pattern when
(IFT) method can be employed on the scattering exposed to X-rays or neutrons. This diffraction
data – c.f. Chap. 5 in (Lindner and Zemb pattern is characterized by more or less
2002) – yielding the pair-distribution distance well-defined peaks on the scattering intensity
function p(r). This function describes the number when the Bragg condition is met, that is, when
of ways in which one can choose a vector of a given q value matches the interplanar spacing d
length r that connects scattering material within according to (note the similarity with Eq. 33)
the particle. This contains valuable information
on the particle structure, since the height of p(r) is 2p
dhkl ¼ (35)
proportional to the number of distances that can qhkl
be found inside the particle within the interval
r and r + dr. where h, k, and l are the Miller indexes, used to
As can be seen in Fig. 27 (right), for a sphere define the planes that will give rise to diffraction
the distribution of p(r) is symmetric, whereas for peaks (crystallographic planes).
an elongated particle (such as a prolate), the peak For a lamellar phase, the periodicity is only in
of p(r) moves to lower r, while the tail moves to one dimension requiring only one index (and only
higher r. This is because the number of small one repeat distance), and dh ¼ 2ph/qh. Hence the
transverse distances perpendicular to the particle diffraction pattern of a lamellar phase is
long axis increases (hence the larger peak at qh/q1 ¼ 1, 2, 3, etc. (for h ¼ 1, 2, 3, etc.), with
lower r). Conversely, there are less (but larger) the lamellar spacing d given by the index h ¼ 1
distances parallel to the long axis, which move (i.e., d ¼ 2p/q).
a dilution series (where the lattice spacing a con-
tinuously increases), one can identify if the phase
is normal or reverse, and through a suitable
model, determine the micellar radius. Diffraction
methods, in particular X-ray diffraction or SAXS
are therefore the method of choice to identify
liquid-crystalline phases, since along with phase
assignment, they also provide structural
parameters.
For phases with lower symmetry (like inter-
mediate phases) the unit cell is defined by more
than one size (e.g., a, b, and c, rather than just a as
Surfactant Self-Assembly, Fig. 28 Hexagonal lattice. in the previous lattices). This changes the order-
The rods (either normal or reverse) have radius R. The
ing of reflections, which prevents the phase
distance between rods is a and the first interplanar distance
to originate a Bragg peak is d10 assignment in a straightforward way as Table 6
would predict. A more elaborate analysis is there-
fore required when studying intermediate phases.
For the hexagonal phase, the periodicity
now is two-dimensional, and so two Miller Dynamic Light Scattering
indexes (h and k) are required (Kittel 2005), In dynamic light scattering (DLS) what is mea-
according to: sured is the fluctuation of scattered intensity at
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi fixed angles. Since this fluctuation results from
4ðh2 þ hk þ k2 Þ the motion of particles, it is correlated with
qhk ¼ 2p (36)
3a2 the particle diffusion coefficient. The auto-
correlation function can be calculated from
where a defines the lattice size (Fig. 28). The first these time-dependent fluctuations of the inten-
four diffraction peaks of a hexagonal lattice come sity, and the collective diffusion coefficient Dc
from the planes defined by hk ¼ 10, 11, 20, and can be extracted from this auto-correlation func-
21. Through Eq. 36 these peaks are expected to tion. From the diffusion coefficient, the hydrody-
occur at a qhk/q10 ratio of 1, √3, 2, and √7. When namic radius RH can be determined according to
a diffraction pattern like this is observed, there is the Stokes–Einstein equation:
a strong indication that a hexagonal phase
is present. kB T
RH ¼ (38)
For the cubic phases, the periodicities occur in 6pDc
three dimensions that require the three indexes S
hkl. Here the interplanar distances are obtained where kB is the Boltzmann constant, T is the
according to absolute temperature and is the solvent viscos-
ity. DLS can be used straightforwardly to deter-
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
h2 þ k 2 þ l 2 mine the size of particles that are relatively
qhkl ¼ 2p (37) monodisperse, and whose size is well below the
a
radiation wavelength. If the sample is polydis-
The different cubic lattices give rise to differ- perse, then more advanced analysis methods
ent diffraction planes with different Miller dis- have to be employed (c.f. Chap. 9 in (Lindner
tances. Table 6 shows the expected ratios and Zemb 2002)) Some limitations of DLS
between q values for the lamellar, hexagonal, include also the need to have samples that are
and some cubic phases that can aid in the identi- dust-free, since a minute amount of dust may
fication of liquid-crystalline phases. Knowing the dominate the sample scattering and hide the
system composition, and often with the aid of signal from the particles of interest. Most of
Surfactant Self-Assembly, Table 6 Sequence of normalized qhkl values for different liquid-crystalline phases
Lamellar Hexagonal Cubic Pn3m Cubic Fd3m Cubic Im3m Cubic Ia3d
√1 √1 √2 √3 √2 √6
√4 √3 √3 √8 √4 √8
√9 √4 √4 √11 √6 √14
√16 √7 √6 √12 √8 √16
√25 √9 √8 √16 √10 √20
√36 √12 √9 √19 √12 √22
√49 √13 √10 √24 √14 √24
√64 √16 √11 √27 √16 √26
the analysis methods also require the sample to be with a platinum–carbon (or alternatively
very dilute, in order to avoid multiple scattering. a tungsten–tantalum) alloy. To improve the sam-
ple mechanical properties, pure carbon is also
Microscopy projected on the sample at an angle of 90 . The
Electron Microscopy Methods replica is then cleaned in a proper solvent and
In electron microscopy, the electron beam is then introduced into a transmission electron
usually generated either by thermionic effect or microscope. In the cryo-SEM technique, the vit-
by field emission under conditions of high rified samples have to be cryo-fractured and par-
vacuum, with pressures below 105 to 106 Pa. tially sublimated (in order to reveal the internal
Due to the high vacuum conditions inside the structure), followed by sputter coating with gold
microscope chamber, samples where solvent is or gold/palladium alloy, before observation in the
present need to be either dried or frozen. To enable scanning electron microscope.
a good focus, the aggregates have to be fixed as Because of its relevance and widespread use,
well, which is only possible in the absence of we will consider below cryo-TEM in a little more
solvent or in vitrified samples. For surfactant- detail, whereas for the other electron microscopy
containing aqueous samples carefully vitrified techniques readers are referred to in-depth
specimens are prepared for imaging, since dehy- reviews available in the literature (Goldstein
dration is detrimental to self-assembled structures. et al. 2003; Mondain-Monval 2005; van Zanten
Both transmission electron microscopy and Zasadzinski 2005).
(TEM) and scanning electron microscopy
(SEM) are employed for the imaging of surfac- Cryogenic Transmission Electron Microscopy
tant systems. Cryo-TEM is the most useful tech- Cryogenic transmission electron microscopy
nique for aqueous dilute solutions of low (cryo-TEM) is a well established technique and
viscosity, for instance those containing large constitutes an excellent tool for visualization of
micelles, vesicles and other bilayer structures, colloidal aggregates in liquid samples (Almgren
cubosomes, etc. Freeze-fracture TEM et al. 2000; Talmon 1996) provided that they are
(Mondain-Monval 2005; van Zanten and properly vitrified, that is, frozen at very low tem-
Zasadzinski 2005) and cryogenic SEM peratures and at such a speed that crystalline ice
(Goldstein et al. 2003) are typically employed (which would destroy the aggregates) has no time
for more surfactant-concentrated, viscous sam- to form. With this technique one can observe
ples, for instance containing large multilamellar aggregate structure in real space, with
vesicles (onionsomes) and lyotropic liquid crys- a resolution down to the nanometer scale. One
tals (in most cases, lamellar phases). advantage of the cryo-TEM imaging is that the
In the FF-TEM method, the sample is first recorded images, after calibration, report a good
vitrified at low temperature, fractured in estimation of the size of the imaged objects since
a vacuum chamber and shadowed in the surface they are a simple projection of their shape.
Cryo-TEM can be very useful to character- polarized light and two Nomarski prisms, can be
ize bilayers and micelle systems, such as vesi- used (Kachar et al. 1984; Miller et al. 1987).
cles, wormlike and branched micelles A particularly important use of polarized light
(Almgren et al. 2000; Talmon 1996). However, microscopy is on the identification of liquid-
when analyzing cryo-TEM images, care should crystalline phases (Dierking 2003; Rosevear
also be taken in order not to confuse possible 1954, 1968). In this case, when polarized light
artifacts that might have formed during the propagates through liquid-crystalline structures
blotting, vitrification, or transfer process, with with defects, placed between slide and coverslip,
the real aggregates in the sample. The most characteristic birefringent textures are formed
common contaminants are frost and crystalline that often allow for the phase assignment
ice. For instance, the blotting process can (i.e., lamellar and hexagonal phases).
induce aggregate shape changes, and large
aggregates can also be excluded from the poly- Nuclear Magnetic Resonance
mer holes, due to size limitations. Objects A number of NMR tools can be used to charac-
larger than 500 nm are usually not observed in terize surfactant-self-assembled systems,
cryo-TEM studies. Another limitation is the use namely, the measurement of chemical shifts,
of under-focus, used to enhance contrast, since relaxation times, and self-diffusion coefficients.
this procedure overestimates the size of objects In the case of chemical shifts, they are useful for
smaller than 4 nm. obtaining information on cmc, organic counter-
Another drawback is that as with any other ion binding, alkyl chain conformation, and solu-
microscopy technique, as one increases the mag- bilization, among other things. Longitudinal and
nification to highlight fine structural features, transverse relaxation measurements provide
the sampling of the system is compromised. In essentially information on micellar dynamics. It
order to overcome this difficulty, a very exhaus- is self-diffusion NMR by far the most common
tive scan of the whole system should be and powerful NMR method for the investigation
employed, which is tedious and time consuming. of surfactant systems. In this entry, we will focus
This is especially critical in polydisperse shortly on this technique and the reader is
systems, and therefore, cryo-TEM should in referred to specialized reviews for more in-depth
those cases, be combined with a complementary coverage of NMR methods applied to surfactant
technique, such as SANS or self-diffusion systems (Furo 2005; Söderman and Stilbs 1994;
NMR that give a better ensemble average of Söderman et al. 2004).
aggregate size.
Pulsed Field Gradient Self-diffusion NMR
Light Microscopy Methods The Pulsed Field Gradient (PFG) NMR technique
Light microscopy resolution is much lower than for self-diffusion, also known as Diffusion S
electron microscopy (typically the lower limit is Ordered Spectroscopy (DOSY), is based on
around 200 nm), and therefore aggregates with a combination of a sequence of radiofrequency
small cross-section-like micelles and small (rf) pulses leading to the formation of spin
vesicles cannot be visualized. Nevertheless, this echoes, and magnetic field gradient pulses
technique is still useful to visualize large which “label” the spins magnetically (Stilbs
unilamellar vesicles or multilamellar vesicles (the 1987). Compared to other diffusion methods,
dimensions of which can reach several microns), NMR offers several important advantages:
emulsion droplets, and other supramolecular speed, chemical selectivity, broad range of acces-
assemblies like fibers and microtubules (Kachar sible diffusion coefficients, and non-invasiveness
et al. 1984; Miller et al. 1987). To enhance the (Furo 2005).
contrast between these aggregates and the medium In this method, one monitors the attenuation
where they are suspended, the technique of differ- of a spin–echo resulting from the dephasing of
ential interference contrast (DIC), which employs the nuclear spins due to the combined effect of
translational motion of the spins and the impo- molecules or the degree of micelle counterion
sition of the gradient pulses. The two field dissociation. Further uses of this method include
gradient pulses applied have a time separation the determination of the composition of mixed
D, which is the experimental time scale (typi- micelles, size estimation of small vesicles, and
cally on the order of 1–1,000 ms). The self- investigation of micelle–vesicle transition.
diffusion coefficient of the diffusing spins is In microemulsion systems, one can measure
directly extracted from the echo attenuation. almost independently the aliphatic protons from
The application of Fourier transformation of oil and surfactant, the surfactant headgroup
the time-domain echo signifies that individual protons, and the water protons. For o/w droplets,
diffusion coefficients of different components the diffusion of oil and surfactant should be very
in multicomponent systems can be determined similar, reflecting the diffusion coefficient of the
simultaneously, as long as they have resolved droplets. The diffusion of water should be faster
NMR signals in the frequency domain. (since the system is water-continuous) but
In fact, NMR diffusometry is especially useful slightly attenuated by obstruction from the o/w
as one can track the diffusivity of molecules in droplets. With the same rationale, if the system is
different sites within the sample. Because the bicontinuous, oil would diffuse much faster than
chemical shift of protons is very sensitive to the when confined in droplets, and with a different
chemical surroundings, one can often track dif- coefficient than the surfactant. A great usefulness
ferent sections of molecules or aggregates. If the of self-diffusion NMR is, therefore, the ability to
investigated molecules are present in sites with distinguish between discrete and connected/bi-
different diffusion coefficients and the molecular continuous phases.
exchange between sites is fast on the time scale of
D, the experimental diffusion coefficient is the Further Methods
population average of the specific diffusion A range of further experimental methods can be
coefficients. Thus, for instance, in a system of employed for direct or indirect microstructural
surfactant micelles with fast exchange between characterization of surfactant systems.
micelles and unimers, the observed surfactant A comprehensive reference to all of them is out-
diffusion coefficient, Dobs, is the following pop- side the scope here, but we will briefly mention
ulation-weighted average: some of the more common ones, with examples
of their usefulness.
Dobs ¼ Pmic Dmic þ ð1 Pmic ÞDunim (39) Within the panoply of methods for cmc
determination (Evans and Wennerström
where Dmic and Dunim are the self-diffusion coef- 1999), surface tension measurements and
ficients of micelle-bound surfactant and free conductimetry stand out as the standard ones,
unimer surfactant, respectively, and Pmic is the with the former allowing also for the determina-
mole fraction of micelle-bound surfactant. tion of molecular areas at the air–water interface
Hence, the concentration dependence of Dobs and the latter yielding the degree of micelle
can be used to determine the surfactant cmc. On ionization.
the other hand, fast unimer-micelle exchange With respect to calorimetric methods, differ-
implies that Dobs 6¼ Dmic and, therefore, Dobs ential scanning calorimetry allows for the assess-
can be used for measuring micellar size and ment of the thermodynamic parameters involved
shape only if DmicPmic > (1Pmic)Dunim. An in phase transitions, with a particular usefulness
alternative is to take advantage of the diffusivity for instance in the characterization of the gel-to-
of very hydrophobic molecules that reside liquid crystal transition of bilayer structures of
entirely in the micellar core. Expressions analo- myriad compositions (Biltonen and Lichtenberg
gous to Eq. 39 can also be used to estimate the 1993; Chernik 1995). Isothermal titration
degree of micellar solubilization of other calorimetry is often used for the determination
of cmc and micellization enthalpies of surfac- (onionsomes), and these phenomena on their
tants, as well as the assessment of partition own are important for a wide range of applica-
equilibria in bilayer systems and interaction tions (Butler 1999). For more on the rheology of
enthalpies in systems with strongly interacting surfactant solutions, dispersions, and
cosolutes (Blandamer et al. 1998; Heerklotz and mesophases, several advanced textbooks can be
Seelig 2000). consulted (Bergström 1994; Larson 1999).
Besides NMR, other molecular spectroscopy
techniques can be used for probing structure and
(more often) dynamics in surfactant structures. Conclusion
Fluorescence spectroscopy methods, both
steady-state and time-resolved, rely on fluores- Throughout this entry, we have seen that
cent probing molecules and quenchers and are surfactants, because of their simple solvophilic–
particularly useful for determining micelle aggre- solvophobic character, and despite the decep-
gation numbers and micelle fluidities (Almgren tively simple molecular structure that lies behind
and Lofroth 1981). In the case of fluorescence it, can form a fascinating range of self-organized
anisotropy, the probes yield information on structures of varying complexity. Between
amphiphile dynamics at various depths of bilayer unimer solutions and the highly ordered crystal-
membranes, in both gel and liquid-crystalline line phases, as extremes, lie a variety of phases,
states (Lentz 1993). the structures of which depend on the chemical
Finally, we refer to rheological methods structure of the surfactant, surfactant concentra-
applied to surfactant systems, which constitute tion, and other parameters such as temperature,
by themselves a complex and diverse area salt, oil, cosolutes, etc. Micelles, vesicles, bila-
(Jönsson et al. 1998). Different types of structures yers, microemulsions, and liquid-crystalline
exhibit different rheological properties, so it is phases are among the ways surfactant molecules
possible in some cases (although in general find to pack in solvents. These different phases
a difficult task) to correlate the structure of sam- play a relevant role in many fundamental aspects
ple with flow behavior and vice versa. of colloid and interface science, helping us to
For instance, micellar growth into cylindrical understand colloidal interactions. They are also
and threadlike micelles is usually accompanied pivotal in myriad applications of colloid science
by viscosity increases, while changes in in the modern world, including detergency, emul-
microemulsion structure can also be probed by sion technology, oil recovery, foodstuffs,
viscosity changes. Typically, dilute solutions cosmetics, medicine, and pharmaceuticals. Last
show Newtonian behavior, while concentrated but not least, they lie on the basis of many impor-
ones can be Newtonian or viscoelastic, tant biological phenomena, further helping us to
depending on structure. gain insight into them. S
The viscosity is rather high for cubic and
hexagonal liquid crystals but much lower for
lamellar systems. Cubic phases have very high Cross-References
viscosities and may even possess a yield value,
showing non-Newtonian plastic behavior; hexag- ▶ Electrical Double Layers
onal phases are pseudoplastic. When subject to ▶ Emulsion Formation
shear, liquid crystals generally exhibit complex ▶ Emulsion Stabilization
and distinctive rheological responses because ▶ Microemulsions
shear itself can change the structure of these ▶ Nanoemulsions
phases. For instance, lamellar phases show ▶ Phase Diagram of Langmuir Monolayers
a complex interplay between planar sheets, ▶ Polyelectrolyte Dynamics
bilayer fragments, and multilamellar vesicles ▶ Polymeric Surfactant
▶ Rheological Techniques Danielsson I, Lindman B (1981) The definition of

▶ Rheology microemulsion. Colloids Surf 3:391–392
Demus D, Goodby J, Gray GW, Spiess H-W, Vill
▶ Self-Assembly of Surfactants V (1998) Handbook of liquid crystals. Wiley-VCH
▶ Surfactant Adsorption Verlag GmbH, Weinheim
▶ Surfactants Dierking I (2003) Textures of liquid crystals. Wiley-VCH
▶ Surfactants, Phase Behavior Verlag GmbH & Co KGaA, Weinheim
Dreiss CA (2007) Wormlike micelles: where do we stand?
Recent developments, linear rheology and scattering
techniques. Soft Matter 3:956–970
References Evans DF, Wennerström H (1999) The colloidal domain:
where physics, chemistry, biology and technology
Adamson AW, Gast AP (1997) Physical chemistry of meet, 2nd edn. Wiley, New York
surfaces, 6th edn. Wiley, New York Fontell K (1990) Cubic phases in surfactant and surfac-
Almgren M (2000) Mixed micelles and other structures in tant-like lipid systems. Colloid Polym Sci
the solubilization of bilayer lipid membranes by sur- 268:264–285
factants. Biochim Biophys Acta Biomembranes Furo I (2005) NMR spectroscopy of micelles and related
1508:146–163 systems. J Mol Liq 117:117–137
Almgren M, Edwards K, Karlsson G (2000) Cryo transmis- Glatter O, Kratky O (1982) Small angle X-ray scattering.
sion electron microscopy of liposomes and related struc- Academic, London
tures. Colloids Surf A Physicochem Eng Asp 174:3–21 Goldstein J, Newbury DE, Joy DC, Lyman CE, Echlin P,
Almgren M, Lofroth JE (1981) Determination of micelle Lifshin E, Sawyer LC, Michael JR (2003) Scanning
aggregation numbers and micelle fluidities from time- electron microscopy and x-ray microanalysis, 3rd edn.
resolved fluorescent quenching studies. J Colloid Kluwer Academic/Plenum Publishers, New York
Interface Sci 81:486–499 Hamley IW (2007) Introduction to soft matter – revised
Antonietti M, Förster S (2003) Vesicles and Liposomes: edition. Wiley, Chichester
a self-assembly principle beyond lipids. Adv Mater Hansen JP, McDonald IR (1986) Theory of simple liquids.
15:1323–1333 Academic, San Diego
Arleth L, Marčelja S, Zemb T (2001) Gaussian random Heerklotz H, Seelig J (2000) Titration calorimetry of
fields with two level-cuts – Model for asymmetric surfactant-membrane partitioning and membrane sol-
microemulsions with nonzero spontaneous curvature? ubilization. Biochim Biophys Acta Biomembranes
J Chem Phys 115:3923–3936 1508:69–85
Bagger-Jörgensen H, Olsson U, Mortensen K (1997) Micro- Holmes MC (1998) Intermediate phases of surfactant-
structure in a ternary microemulsion studied by small water mixtures. Curr Opin Colloid Interf Sci
angle neutron scattering. Langmuir 13:1413–1421 3:485–492
Bergström L (1994) Rheology of concentrated disper- Hunter RJ (2001) Foundations of colloid science, 2nd edn.
sions. In: Pugh RJ, Bergström L (eds) Surface ad Oxford University Press, Oxford
colloid chemistry in advanced ceramic processing. Israelachvili J (1985) Intermolecular and surface forces
Marcel Dekker, New York with application to colloidal and biological systems.
Biltonen RL, Lichtenberg D (1993) The use of differential Academic, London
scanning calorimetry as a tool to characterize liposome Israelachvili J, Wennerström H (1996) Role of hydration
preparations. Chem Phys Lipids 64:129–142 and water structure in biological and colloidal interac-
Blandamer MJ, Cullis PM, Engberts J (1998) Titration tions. Nature 379:219–225
microcalorimetry. J Chem Soc Faraday Trans Jönsson B, Lindman B, Holmberg K, Kronberg B (1998)
94:2261–2267 Surfactants and polymers in aqueous solution, 1st edn.
Bouwstra JA, Honeywell-Nguyen PL, Gooris GS, Ponec Wiley, Chichester
M (2003) Structure of the skin barrier and its modula- Kachar B, Evans DF, Ninham BW (1984) Video enhanced
tion by vesicular formulations. Prog Lipid Res 42:1–36 differential interference contrast microscopy: a new
Bryskhe K, Bulut S, Olsson U (2005) Vesicle formation tool for the study of association colloids and prebiotic
from temperature jumps in a nonionic surfactant sys- assemblies. J Colloid Interf Sci 100:287
tem. J Phys Chem B 109:9265–9275 Kékicheff P, Cabane B (1988) Crystallography of systems
Butler P (1999) Shear induced structures and transforma- with long periods: a neutron-scattering study of
tions in complex fluids. Curr Opin Colloid Interf Sci sodium dodeeyl sulfate/water mesophases. Acta
4:214–221 Crystallogr Sect B Struct Sci 44:395–406
Chernik GG (1995) Phase-equilibria in phospholipid Khan A, Marques E (1998) Catanionic surfactants. In:
water-systems. Adv Colloid Interf Sci 61:65–129 Robb ID (ed) Specialist surfactants. Blackie Academic
Collings PJ, Hird M (1997) Introduction to liquid & Professional, London, pp 37–76
crystals – chemistry and physics, 1st edn. Taylor & Kittel C (2005) Introduction to solid state physics, 8th edn.
Francis, London Wiley, Hoboken
Surfactant Solution 1241 S
Larson RG (1999) The structure and rheology of complex angle X-Ray scattering sudies of Micellar transforma-
fluids. Oxford University Press, New York tions. Phys Rev Lett 88:258301
Lasic DD (1993) Liposomes: from physics to applications. Schwuger MJ, Stickdorn K, Schomacker R (1995)
Elsevier, Amsterdam Microemulsions in technical processes. Chem Rev
Lasic DD, Joannic R, Keller BC, Frederik PM, Auvray 95:849–864
L (2001) Spontaneous vesiculation. Adv Colloid Interf Seddon JM (1990) Structure of the inverted hexagonal
Sci 89–90:337–349 (HII) phase, and non-lamellar transitions of lipids.
Leng J, Egelhaaf SU, Cates ME (2002) Kinetic pathway of Biochim Biophys Acta 1031:1–69
spontaneous vesicle formation. Europhys Lett Seddon JM, Templer RH (1995) Polymorphism of lipid-
59:311–317 water systems. In: Lipowsky R, Sackmann E (eds)
Lentz BR (1993) Use of fluorescent probes to monitor Handbook of biological physics. Elsevier Science B.
molecular order and motions within liposome bilayers. V., Amsterdam
Chem Phys Lipids 64:99–116 Šegota S, Težak Ð (2006) Spontaneous formation of ves-
Lindman B, Wennerström H (1980) Micelles: topics in icles. Adv Colloid Interf Sci 121:51–75
current chemistry. Springer, Berlin Shimizu T, Masuda M, Minamikawa H (2005) Supramo-
Lindner P, Zemb T (2002) Neutrons, X-rays and light: lecular nanotube architectures based on amphiphilic
scattering methods applied to soft condensed matter, molecules. Chem Rev 105:1401–1444
1st edn. Elsevier Science B.V, Amsterdam Söderman O, Stilbs P (1994) NMR studies of complex
Lipowsky R (1995) The morphology of lipid membranes. surfactant systems. Prog Nucl Magn Reson Spectrosc
Curr Opin Struct Biol 5:531–540 26:445–482
Luzzati V (1968) X-ray diffraction studies of lipid-water Söderman O, Stilbs P, Price WS (2004) NMR studies
systems. In: Chapman D (ed) Biological membranes. of surfactants. Concepts Magn Reson A 23A:
Academic, London, pp 71–123 121–135
Majhi PR, Blume A (2002) Temperature induced micelle- Stilbs P (1987) Fourier transform pulsed-gradient spin-
vesicle transition in DMPC-SDS and DMPC-DTAB echo studies of molecular diffusion. Prog Nucl Magn
mixtures by calorimetry and dynamic light scattering. Reson Spectrosc 19:1–45
J Phys Chem B 106:10753–10763 Stubenrauch C (2001) Sugar surfactants – aggregation,
Marques EF (2010) Surfactant vesicles: formation, prop- interfacial, and adsorption phenomena. Curr Opin
erties, and stability. In: Somasundaran P, Hubbard Colloid Interf Sci 6:160–170
A (eds) Encyclopedia of surface and colloid science, Tadros TF (2005) Applied Surfactants. Wiley-VCH
2nd edn. Taylor and Francis, New York Verlag GmbH & Co, Weinheim
Menger FM, Keiper JS (2000) Gemini surfactants. Angew Talmon Y (1996) Transmission electron microscopy of
Chem Int Ed 39:1906–1920 complex fluids: the state of the art. Ber Bunsen Ges
Miller DD, Bellare JR, Evans DF, Talmon Y, Ninham BW Phys Chem 3:364–372
(1987) Meaning and structure of amphiphilic phases: Tanford C (1980) The hydrophobic effect: formation of
inferences from video-enhanced microscopy and micelles and biological membranes. Wiley, New York
cryotransmission electron microscopy. J Phys Chem Tiddy GJT (1980) Surfactant-water liquid-crystal phases.
91:674–685 Phys Rep Rev Sect Phys Lett 57:1–46
Mondain-Monval O (2005) Freeze fracture TEM investi- van Zanten R, Zasadzinski JA (2005) Using cryo-electron
gations in liquid crystals. Curr Opin Colloid Interf Sci microscopy to determine thermodynamic and elastic
10:250–255 properties of membranes. Curr Opin Colloid Interf Sci
Mortensen K (2001) Structural studies of lamellar surfac- 10:261–268
tant systems under shear. Curr Opin Colloid Interf Sci Zana R (1996) Critical micellization concentration of
6:140–145 surfactants in aqueous solution and free energy of S
Roberts GG (1985) An applied science perspective of micellization. Langmuir 12:1208–1211
Langmuir-Blodgett films. Adv Phys 34:475–512 Zana R (2002) Dimeric and oligomeric surfactants.
Rosevear FB (1954) The microscopy of the liquid crystal- Behavior at interfaces and in aqueous solution:
line neat and middle phases of soaps and synthetic a review. Adv Colloid Interf Sci 97:205–253
detergents. J Am Oil Chem Soc 31:628–639 Zemb T, Dubois M (2003) Catanionic microcrystals:
Rosevear FB (1968) Liquid crystals – mesomorphic organic platelets, gigadalton “molecules”, or ionic
phases of surfactants compositions. J Soc Cosm solids? Aust J Chem 56:971–979
Chem 19:581–594
Safran SA (1999) Curvature elasticity of thin films. Adv
Phys 48:395–448
Safran SA, Pincus PA, Andelman D, MacKintosh FC
(1991) Stability and phase behavior of mixed surfac-
tant vesicles. Phys Rev A 43:1071–1078
Surfactant Solution
Schmölzer S, Gr€abner D, Gradzielski M, Narayanan
T (2002) Millisecond-range time-resolved small- ▶ Surface Tension
S 1242 Surfactants
BAn (or ABn), where B is the “anchor” chain

Surfactants and A is the stabilizing chain(s), are important
for stabilization of various disperse systems.
Tharwat Tadros The solution properties of surfactants show
Wokingham, Berkshire, UK abrupt change at a critical concentration that
is consistent with the fact that above this con-
centration surfactant molecules or ions associ-
Synonyms ate to form micelles. This concentration is
defined as the critical micelle concentration,
Surface-active agents cmc, whose value depends on the structure of
the surfactant molecule. Ionic surfactants show
a rapid increase in their solubility above
Keywords a certain temperature (the Krafft temperature)
at which the solubility is equal to the
Amphiphiles; lyophobic; lyophilic; surface- cmc. Nonionic surfactants of the ethoxylate
active agents; surfactants; anionic; cationic; type show cloudiness above a certain tempera-
amphoteric; nonionic; alkyl polyglucosides; ture defined as the cloud point (CP) of the
sucrose esters; gemini surfactants; surfactant solution. The cloud point depends
biosurfactants; lipids; polymeric; krafft tempera- on the balance between hydrophilic (polyethyl-
ture; critical micelle concentration; cmc; surface ene oxide) and hydrophobic (alkyl chain)
tension; micellization; dynamics; equilibrium groups. At high concentrations, surfactant solu-
aspects; mixtures; surfactant–polymer interactions form high-order assembly structures
tion; adsorption; air/liquid; liquid/liquid; solid/ of the hexagonal, cubic, and lamellar phases.
liquid Micellization is a dynamic process
characterized by two relaxation processes. The
free energy of micelle formation (equilibrium
Definition aspects) can be calculated from the
cmc. The enthalpy of micellization is small
Surface-active agents (abbreviated surfactants) and positive (endothermic), whereas the
are amphiphilic molecules consisting of a entropy of micellization is large and positive.
lyophobic (hydrophobic when in water) and The driving force of micelle formation is deter-
lyophilic (hydrophilic when in water) portions. mined by the “hydrophobic effect” which is
They are classified according to the nature of entropy driven. Mixtures of surfactants give
the head group: anionic, cationic, amphoteric mixed micelles, and the cmc of the mixed
(zwitterionic), and nonionic. Speciality surfac- micelle depends on the net interaction between
tants of fluorocarbons and silicones can lower the surfactant molecules. Surfactants interact
the surface tension of water to values lower with water-soluble molecules forming associa-
than conventional surfactants. Several surfac- tion structures. The adsorption of surfactants at
tants are based on mono- and oligosaccharides the air/liquid and liquid/liquid interface can be
(alkyl polyglucoside, sucrose esters). Some measured using surface and interfacial tension
molecules contain two head groups and an measurements and application of the Gibbs
alky chain attached to each of the head group, adsorption isotherm. The adsorption of surfac-
referred to as gemini surfactants. Naturally tants on solid surfaces depends on the nature of
occurring surfactants such as lipids are essen- surfactant and substrate. The adsorption of
tial for applications in pharmacy and food ionic and nonionic surfactants on hydrophobic
products. Biosurfactants are produced by enzy- surfaces is determined by hydrophobic interac-
matic reactions which can offer selectivity. tion between the alkyl chain and the surface.
Polymeric surfactants of the A-B, A-B-A, and Ionic and nonionic surfactants interact with
Surfactants 1243 S
hydrophilic surfaces with the polar head the interface. This interfacial free energy,
groups. Various structures can be identified referred to as surface or interfacial tension, g,
such hemi-micelles, bilayers, etc. is given in mJm2 or mNm1. Adsorption
of surfactant molecules at the interface lowers
g, and the higher the surfactant adsorption (i.e.,
Overview the more dense the layer is), the larger the
reduction in g. The degree of surfactant adsorp-
Surface-active agents (usually referred to as tion at the interface depends on surfactant
surfactants) are sometimes referred to amphi- structure and the nature of the two phases that
philic molecules, implying that they consist of meet the interface (Tadros 1984; Holmberg
at least two parts, one which is soluble in et al. 2003a). Surface-active agents also aggre-
a specific liquid (the lyophilic part) and one gate in solution, forming micelles. The driving
which is insoluble (the lyophobic part). If the force micelle formation (or micellization) is the
fluid is water, one refers to the hydrophilic and reduction of contact between the hydrocarbon
hydrophobic parts, respectively. In this case, chain and water, thereby reducing the free
the molecule consists of a nonpolar hydropho- energy of the system. In the micelle, the sur-
bic portion, usually a straight or branched factant hydrophobic groups are directed toward
hydrocarbon or fluorocarbon chain containing the interior of the aggregate, and the polar head
8–18 carbon atoms, which is attached to a polar groups are directed toward the solvent. These
or ionic portion (hydrophilic). The hydrophilic micelles are in dynamic equilibrium, and the
portion can, therefore, be nonionic, ionic, or rate of exchange between a surfactant molecule
zwitterionic, accompanied by counterions in and the micelle may vary by orders of magni-
the last two cases. The hydrocarbon chain tude, depending on the structure of the surfac-
interacts weakly with the water molecules in tant molecule.
an aqueous environment, whereas the polar or Surfactants find application in almost
ionic head group interacts strongly with water every chemical industry of which the following
molecules via dipole or ion–dipole interactions. may be worth mentioning: detergents, paints,
It is this strong interaction with the water mol- dyestuffs, cosmetics, pharmaceuticals, agro-
ecules which renders the surfactant soluble in chemicals, fibers, plastics, etc. Moreover, sur-
water. However, the cooperative action of dis- factants play a major role in the oil industry,
persion and hydrogen bonding between the for example, in enhanced and tertiary oil recov-
water molecules tends to squeeze the hydrocar- ery. They are also occasionally used for environ-
bon chain out of the water, and hence these mental protection, for example, in oil slick
chains are referred to as hydrophobic. As we dispersants. Therefore, a fundamental under-
will see later, the balance between hydrophobic standing of the physical chemistry of surface- S
and hydrophilic part of the molecule gives active agents, their unusual properties, and their
these systems their special properties, for phase behavior is essential for most industrial
example, accumulation at various interfaces chemists. In addition, an understanding of the
and association in solution (to form micelles). basic phenomena involved in the application of
In addition to the name surface-active agents, surfactants, such as in the preparation of emul-
these molecules are often called by other names sions and suspensions and their subsequent sta-
which include surfactants, association colloids, bilization, in microemulsions, in wetting
colloidal electrolytes, amphipathic compounds, spreading and adhesion, etc., is of vital impor-
tensides, etc. tance in arriving at the right composition and
The driving force for surfactant adsorption control of the system involved (Tadros 1984;
is the lowering of the free energy of the phase Holmberg et al. 2003a). This is particularly the
boundary. The interfacial free energy per unit case with many formulations in the chemical
area is the amount of work required to expand industry mentioned above.
S 1244 Surfactants
It should be stated that commercially produced Anionic Surfactants

surfactants are not pure chemicals, and within each These are the most widely used class of surfac-
chemical type, there can be tremendous variation. tants in industrial applications (Porter 1994;
This can be understood, since surfactants are pre- Linfield 1967; Lucasssen-Reynders 1981). This
pared from various feedstocks, namely, petro- is due to their relatively low cost of manufacture,
chemicals, natural vegetable oils, and natural and they are practically used in every type of
animal fats. It is important to realize that in every detergent. For optimum detergency, the hydro-
case the hydrophobic group exists as a mixture of phobic chain is a linear alkyl group with a chain
chains of different lengths. The same applied to the length in the region of 12–16 C atoms, and the
polar head group, for example, in the case of polar head group should be at the end of the
polyethylene oxide (the major component of non- chain. Linear chains are preferred since they are
ionic surfactants) which consists of a distribution more effective and more degradable than the
of ethylene oxide units. Hence, products that may branched chains. The most commonly used
be given the same generic name could vary a great hydrophilic groups are carboxylates, sulfates,
deal in their properties, and the formulation chem- sulfonates, and phosphates. A general formula
ist should bear this in mind when choosing may be ascribed to anionic surfactants as follows:
a surfactant from a particular manufacturer. It is Carboxylates: CnH2n+1 COO X+
advisable to obtain as much information as possi- Sulfates: CnH2n+1 OSO3 X+
ble from the manufacturer about the properties of Sulfonates: CnH2n+1 SO3 X+
the surfactant chosen such as its suitability for the Phosphates: CnH2n+1 OPO(OH)O X+
job, its batch to batch variation, its toxicity, with n being the range 8–16 atoms and the
etc. The manufacturer usually has more informa- counterion X+ is usually Na+.
tion on the surfactant than that printed in the data Several other anionic surfactants are commer-
sheet, and in most cases such information is given cially available such as sulfosuccinates,
on request. isethionates (esters of isothionic acid with the
general formula RCOOCH2–CH2–SO3Na) and
taurates (derivatives of methyl taurine with the
General Classification of Surface-Active general formula RCON(R0 ) CH2–CH2–SO3Na),
Agents and sarcosinates (with the general formula
RCON(R0 )COONa), and these are sometimes
A simple classification of surfactants based on the used for special applications. Below, a brief
nature of the hydrophilic group is commonly description of the above anionic classes is given
used. Four main classes may be distinguished, with some of their applications.
namely, anionic, cationic, amphoteric, and
nonionic. A useful technical reference is Carboxylates
McCutcheon (McCutcheon), which is produced These are perhaps the earliest known
annually to update the list of available surfac- surfactants since they constitute the earliest
tants. A recent text by van Os et al. (1993) listing soaps, for example, sodium or potassium stearate,
the physicochemical properties of selected C17H35COONa, sodium myristate,
anionic, cationic, and nonionic surfactants has C14H29COONa. The alkyl group may contain
been published by Elsevier. Another useful text unsaturated portions, for example, sodium oleate,
is the Handbook of Surfactants by Porter (1994). which contains one double bond in the C17 alkyl
It should be mentioned also that a fifth class of chain. Most commercial soaps will be a mixture
surfactants, usually referred to as polymeric of fatty acids obtained from tallow, coconut oil,
surfactants, has been used for many years for palm oil, etc. The main attraction of these simple
preparation of emulsions and suspensions and soaps is their low cost, their ready biodegradabil-
their stabilization. ity, and low toxicity. Their main disadvantage is
Surfactants 1245 S
their ready precipitation in water containing biva- dodecyl sulfate (abbreviated as SDS and some-
lent ions such as Ca2+ and Mg2+. To avoid their times referred to as sodium lauryl sulfate) which
precipitation in hard water, the carboxylates are is extensively used both for fundamental studies
modified by introducing some hydrophilic as well as in many applications in industry. At
chains, for example, ethoxy carboxylates with room temperature (25 C), this surfactant is
the general structure RO(CH2CH2O)nCH2COO, quite soluble, and 30 % aqueous solutions are
ester carboxylates containing hydroxyl or multi fairly fluid (low viscosity). However, below 25
COOH groups, and sarcosinates which contain an C, the surfactant may separate out as a soft paste
amide group with the general structure RCON as the temperature falls below its Krafft point
(R0 )COO. The addition of the ethoxylated (the temperature above which the surfactant
groups results in increased water solubility and shows a rapid increase in solubility with further
enhanced chemical stability (no hydrolysis). The increase of temperature). The latter depends on
modified ether carboxylates are also more the distribution of chain lengths in the alkyl
compatible with electrolytes. They are also com- chain; the wider the distribution, the lower the
patible with other nonionic, amphoteric, and Krafft temperature. Thus, by controlling this
sometimes even cationic surfactants. The ester distribution, one may achieve a Krafft tempera-
carboxylates are very soluble in water, but they ture of 10 C. As the surfactant concentration
suffer from the problem of hydrolysis. The is increased to 30–40 % (depending on the dis-
sarcosinates are not very soluble in acid or neutral tribution of chain length in the alkyl group), the
solutions, but they are quite soluble in alkaline viscosity of the solution increases very rapidly
media. They are compatible with other anionics, and may produce a gel but then falls at about
nonionics, and cationics. The phosphate esters 60–70 % to give a pourable liquid, after which it
have very interesting properties being intermedi- increases again to a gel. The concentration at
ate between ethoxylated nonionics and sulfated which the minimum occurs varies according to
derivatives. They have good compatibility with the alcohol sulfate used and also the presence of
inorganic builders, and they can be good emulsi- impurities such as unsaturated alcohol. The vis-
fiers. A specific salt of a fatty acid is lithium cosity of the aqueous solutions can be reduced
12-hydroxystearic acid that forms the major con- by addition of short-chain alcohols and glycols.
stituent of greases. The critical micelle concentration (cmc) of SDS
(the concentration above which the properties of
Sulfates the solution show abrupt changes) is 8
These are the largest and most important class of 103 mol dm3 (0.24 %). The alkyl sulfates
synthetic surfactants, which were produced by give good foaming properties with an optimum
reaction of an alcohol with sulfuric acid, that is, at C12–C14. As with the carboxylates, the sulfate
they are esters of sulfuric acid. In practice, sul- surfactants are also chemically modified to S
furic acid is seldom used, and chlorosulfonic or change their properties. The most common mod-
sulfur dioxide/air mixtures are the most common ification is to introduce some ethylene oxide
methods of sulfating the alcohol. However, due units in the chain, usually referred to as alcohol
to their chemical instability (hydrolyzing to the ether sulfates. These are made by sulfation of
alcohol, particularly in acid solutions), they are ethoxylated alcohols. For example, sodium
now overtaken by the sulfonates which are dodecyl 3-mole ether sulfate which is essen-
chemically stable. The properties of sulfate sur- tially dodecyl alcohol reacted with 3 moles
factants depend on the nature of the alkyl chain EO, then sulfated and neutralized by NaOH.
and the sulfate group. The alkali metal salts The presence of PEO confers improved solubil-
show good solubility in water, but they tend to ity when compared with the straight alcohol
be affected by the presence of electrolytes. The sulfates. In addition, the surfactant becomes
most common sulfate surfactant is sodium more compatible with electrolytes in aqueous
S 1246 Surfactants
solution. The ether sulfates are also more chem- sulfonate is 5 103 mol dm3 (0.18 %). The
ically stable than the alcohol sulfates. The cmc main disadvantage of LABS is their effect on the
of the ether sulfates is also lower than the skin, and hence they cannot be used in personal
corresponding surfactant without the EO units. care formulations.
The viscosity behavior of aqueous solutions is Another class of sulfonates is the a-olefin
similar to alcohol sulfates, giving gels in the sulfonates which are prepared by reacting linear
range 30–60 %. The ether sulfates show a pro- a-olefin with sulfur trioxide, typically yielding
nounced salt effect, with significant increase of a mixture of alkene sulfonates (60–70 %),
the viscosity of a dilute solution on addition of 3- and 4-hydroxyalkane sulfonates (30 %),
electrolytes such as NaCl. The ether sulfates are and some disulfonates and other species. The
commonly used in hand dishwashing and in two main a-olefin fractions used as starting mate-
shampoos, in combination with amphoteric rial are c12-C16 and C16-C18. Fatty acid and ester
surfactants. sulfonates are produced by sulfonation of unsat-
urated fatty acids or esters. A good example is
sulfonated oleic acid:
Sulfonates
With sulfonates, the sulfur atom is directly CH3(CH2)7CH(CH2)8COOH
attached to the carbon atom of the alkyl group, SO3H
and this gives the molecule stability against
hydrolysis, when compared with the sulfates A special class of sulfonates are
(whereby the sulfur atom is indirectly linked to sulfosuccinates which are esters of sulfosuccinic
the carbon of the hydrophobe via an oxygen acid:
atom). The alkyl aryl sulfonates are the most
CH2COOH
common type of these surfactants (e.g., sodium
alkyl benzene sulfonate), and these are usually HSO3 CH COOH
prepared by reaction of sulfuric acid with alkyl
aryl hydrocarbons, for example, dodecyl ben- Both mono- and diesters are produced.
zene. A special class of sulfonate surfactants are A widely used diester in many formulations is
the naphthalene and alkyl naphthalene sulfonates sodium di(2-ethylhexyl)sulfosuccinate (that is
which are commonly used as dispersants. As with sold commercially under the trade name Aerosol
the sulfates, some chemical modification is used OT). The cmc of the diesters is very low, in the
by introducing ethylene oxide units, for example, region of 0.06 % for C6–C8 sodium salts, and they
sodium nonyl phenol 2-mole ethoxylate ethane give a minimum in the surface tension of
sulfonate C9H19C6H4(OCH2CH2)2SO3Na+. 26 mNm1 for the C8 diester. Thus, these mole-
The paraffin sulfonates are produced by cules are excellent wetting agents. The diesters
sulfo-oxidation of normal linear paraffins with are soluble both in water and in many organic
sulfur dioxide and oxygen and catalyzed with solvents. They are particularly useful for prepa-
ultraviolet or gamma radiation. The resulting ration of water-in-oil (W/O) microemulsions.
alkane sulfonic acid is neutralized with NaOH.
These surfactants have excellent water solubility
and biodegradability. They are also compatible Phosphate-Containing Anionic Surfactants
with many aqueous ions. The linear alkyl ben- Both alkyl phosphates and alkyl ether phosphates
zene sulfonates (LABS) are manufactured from are made by treating the fatty alcohol or alcohol
alkyl benzene, and the alkyl chain length can vary ethoxylates with a phosphorylating agent, usually
from C8 to C15, and their properties are mainly phosphorous pentoxide, P4O10. The reaction
influenced by the average molecular weight and yields a mixture of mono- and diesters of phos-
the spread of carbon number of the alkyl side phoric acid. The ratio of the two esters is deter-
chain. The cmc of sodium dodecyl benzene mined by the ratio of the reactants and the amount
Surfactants 1247 S
of water present in the reaction mixture. Cationic surfactants can also be modified
The physicochemical properties of the alkyl by incorporating polyethylene oxide chains, for
phosphate surfactants depend on the ratio of the example, dodecyl methyl polyethylene oxide
esters. Phosphate surfactants are used in the metal ammonium chloride having the structure
working industry due to their anticorrosive
properties. C12H25 (CH2CH2O)nH
+
N Cl−
Cationic Surfactants
The most common cationic surfactants are the CH3 (CH2CH2O)nH
quaternary ammonium compounds (Jungermana
1970; Rubingh and Holland 1991) with the gen- Cationic surfactants are generally water solu-
eral formula R0 R00 R000 R0000 N+X, where X is usu- ble when there is only one long alkyl group.
ally chloride ion and R represents alkyl groups. When there are two or more long-chain
These quaternaries are made by reacting an hydrophobes, the product becomes dispersible
appropriate tertiary amine with an organic halide in water and soluble in organic solvents. They
or organic sulfate. A common class of cationics are generally compatible with most inorganic
is the alkyl trimethyl ammonium chloride, ions and hard water, but they are incompatible
where R contains 8–18 C atoms, for example, with metasilicates and highly condensed phos-
dodecyl trimethyl ammonium chloride, phates. They are also incompatible with protein-
C12H25(CH3)3NCl. Another widely used cationic like materials. Cationics are generally stable to
surfactant class is that containing two long-chain pH changes, both acid and alkaline. They are
alkyl group, that is, dialkyl dimethyl ammonium incompatible with most anionic surfactants, but
chloride, with the alkyl groups having a chain they are compatible with nonionics. These cat-
length of 8–18 C atoms. These dialkyl surfactants ionic surfactants are insoluble in hydrocarbon
are less soluble in water than the monoalkyl qua- oils. In contrast, cationics with two or more long
ternary compounds, but they are commonly used alkyl chains are soluble in hydrocarbon solvents,
in detergents as fabric softeners. A widely used but they become only dispersible in water (some-
cationic surfactant is alkyl dimethyl benzyl times forming bilayer vesicle-type structures).
ammonium chloride (sometimes referred to as They are generally chemically stable and can
benzalkonium chloride and widely used as bac- tolerate electrolytes. The cmc of cationic surfac-
tericide), having the structure, tants is close to that of anionics with the same
C12H25 CH3 alkyl chain length. For example, the cmc of
+ benzalkonium chloride is 0.17 %. The prime use
N Cl– of cationic surfactants is their tendency to adsorb
CH2 CH3 at negatively charged surfaces, for example, S
anticorrosive agents for steel, flotation collectors
for mineral ores, dispersants for inorganic pig-
Imidazolines can also form quaternaries, the ments, antistatic agents for plastics, antistatic
most common product being the ditallow deriva- agents and fabric softeners, hair conditioners,
tive quaternized with dimethyl sulfate: and anticaking agent for fertilizers and as
bactericides.
CH3
Amphoteric (Zwitterionic) Surfactants
[C17H35 C--N--CH2-CH2-NH-CO-C17H35]+
These are surfactants containing both cationic
and anionic groups (Buestein and Hiliton 1982).
N CH CH3SO4−
The most common amphoterics are the N-alkyl
C betaines which are derivatives of trimethyl
H glycine (CH3)3NCH2COOH (that was described
S 1248 Surfactants
as betaine). An example of betaine surfactant those of cationic and anionic surfactants, respec-
is lauryl amido propyl dimethyl betaine tively. Zwitterionic surfactants have excellent
C12H25CON(CH3)2CH2COOH. These alkyl beta- dermatological properties. They also exhibit
ines are sometimes described as alkyl dimethyl low eye irritation, and they are frequently used
glycinates. The main characteristic of amphoteric in shampoos and other personal care products
surfactants is their dependence on the pH of the (cosmetics).
solution in which they are dissolved. In acid pH
solutions, the molecule acquires a positive Nonionic Surfactants
charge, and it behaves like a cationic, whereas The most common nonionic surfactants are those
in alkaline pH solutions, they become negatively based on ethylene oxide, referred to as
charged and behave like an anionic. A specific pH ethoxylated surfactants (Schick 1966, 1987;
can be defined at which both ionic groups show Schonfeldt 1970). Several classes can be
equal ionization (the isoelectric point of the mol- distinguished: alcohol ethoxylates, alkyl
ecule). This can be described by the following phenol ethoxylates, fatty acid ethoxylates,
scheme: monoalkaolamide ethoxylates, sorbitan ester
ethoxylates, fatty amine ethoxylates, and ethyl-
N+....COOH ↔ N+...COO– ↔ NH...COO– ene oxide-propylene oxide copolymers (some-
acid pH < 3 isoelectric pH > 6 alkaline times referred to as polymeric surfactants).
Another important class of nonionics is the
Amphoteric surfactants are sometimes multihydroxy products such as glycol esters,
referred to as zwitterionic molecules. They glycerol (and polyglycerol) esters, glucosides
are soluble in water, but the solubility shows (and polyglucosides), and sucrose esters. Amine
a minimum at the isoelectric point. Ampho- oxides and sulfinyl surfactants represent non-
terics show excellent compatibility with other ionics with a small head group.
surfactants, forming mixed micelles. They
are chemically stable both in acids and in Alcohol Ethoxylates
alkalis. The surface activity of amphoterics These are generally produced by ethoxylation of
varies widely, and it depends on the distance a fatty chain alcohol such as dodecanol. Several
between the charged groups, and they generic names are given to this class of
show a maximum in surface activity at the surfactants such as ethoxylated fatty alcohols,
isoelectric point. alkyl polyoxyethylene glycol, monoalkyl
Another class of amphoteric is the N-alkyl polyethylene oxide glycol ethers, etc. A typical
amino propionates having the structure example is dodecyl hexaoxyethylene glycol
R-NHCH2CH2COOH. The NH group is reactive monoether with the chemical formula
and can react with another acid molecule (e.g., C12H25(OCH2CH2O)6OH (sometimes abbrevi-
acrylic) to form an amino dipropionate R-N ated as C12E6). In practice, the starting alcohol
(CH2CH2COOH)2. Alkyl imidazoline-based will have a distribution of alkyl chain lengths, and
product can also be produced by reacting alkyl the resulting ethoxylate will have a distribution of
imidazoline with a chloro acid. However, the ethylene oxide (EO) chain length. Thus, the
imidazoline ring breaks down during the forma- numbers listed in the literature refer to average
tion of the amphoteric. numbers.
The change in charge with pH of amphoteric The cmc of nonionic surfactants is about two
surfactants affects their properties, such as wet- orders of magnitude lower than the
ting, detergency, foaming, etc. At the isoelectric corresponding anionics with the same alkyl
point (i.e.p), the properties of amphoterics chain length. At a given alkyl chain length, the
resemble those of nonionics very closely. cmc decreases with decrease in the number of
Below and above the properties shift toward EO units. The solubility of the alcohol
Surfactants 1249 S
ethoxylates depends both on the alkyl chain The most common surfactants of this type are
length and on the number of ethylene oxide those based on nonyl phenol. These surfactants
units in the molecule. Molecules with an aver- are cheap to produce, but they suffer from
age alkyl chain length of 12 C atoms and the problem of biodegradability and potential
containing more than 5 EO units are usually toxicity (the by-product of degradation is
soluble in water at room temperature. However, nonyl phenol which has considerable toxicity
as the temperature of the solution is gradually for fish and mammals). In spite of these
raised, the solution becomes cloudy (as a result problems, nonyl phenol ethoxylates are still
of dehydration of the PEO chain and the change used in many industrial properties due to their
in the conformation of the PEO chain), and the advantageous properties, such as their solubil-
temperature at which this occurs is referred to as ity both in aqueous and nonaqueous media,
the cloud point (CP) of the surfactant. At a given their good emulsification, and dispersion
alkyl chain length, the CP increases with properties, etc.
increase in the EO chain of the molecule. The
CP changes with change of concentration of the Fatty Acid Ethoxylates
surfactant solution, and the trade literature usu- These are produced by reaction of ethylene
ally quotes the CP of a 1 % solution. The CP is oxide with a fatty acid or a polyglycol, and
also affected by the presence of electrolyte in they have the general formula RCOO-
the aqueous solution. Most electrolytes lower (CH2CH2O)nH. When a polyglycol is used,
the CP of a nonionic surfactant solution. Non- a mixture of mono- and diester (RCOO-
ionics tend to have maximum surface activity (CH2CH2O)n-OCOR) is produced. These sur-
near to the cloud point. The CP of most non- factants are generally soluble in water, provided
ionics increases markedly on the addition of there is enough EO units and the alkyl chain
small quantities of anionic surfactants. The sur- length of the acid is not too long. The monoesters
face tension of alcohol ethoxylate solutions are much more soluble in water than the diesters.
decreases with increase in its concentration, In the latter case, a longer EO chain is required to
till it reaches its cmc, after which it remains render the molecule soluble. The surfactants are
constant with further increase in its concentra- compatible with aqueous ions, provided there is
tion. The minimum surface tension reached at not much unreacted acid. However, these surfac-
and above the cmc decreases with decrease in tants undergo hydrolysis in highly alkaline
the number of EO units of the chain (at a given solutions.
alkyl chain). The viscosity of a nonionic surfac-
tant solution increases gradually with increase
in its concentration, but at a critical concentra- Sorbitan Esters and Their Ethoxylated
tion (which depends on the alkyl and EO chain Derivatives (Spans and Tweens) S
length), the viscosity shows a rapid increase,
and ultimately a gel-like structure appears. The fatty acid esters of sorbitan (generally
This results from the formation of liquid referred to as Spans, an Atlas commercial
crystalline structure of the hexagonal type. trade name) and their ethoxylated derivatives
In many cases, the viscosity reaches a maximum (generally referred to as Tweens) are perhaps
after which it shows a decrease due to the for- one of the most commonly used nonionics.
mation of other structures (e.g., lamellar phases) They were first commercialized by Atlas in
(see below). the USA, which has been purchased by ICI.
The sorbitan esters are produced by reaction
Alkyl Phenol Ethoxylates of sorbitol with a fatty acid at a high tempera-
These are prepared by reaction of ethylene ture (>200 C). The sorbitol dehydrates to
oxide with the appropriate alkyl phenol. 1,4- sorbitan and then esterification takes
S 1250 Surfactants
place. If one mole of fatty acid is reacted with number surfactants). The ethoxylated products
one mole of sorbitol, one obtains a monoester are generally soluble in number, and they
(some diester is also produced as a by-product). have relatively high HLB numbers. One of
Thus, sorbitan monoester has the following the main advantages of the sorbitan esters and
general formula: their ethoxylated derivatives is their approval
CH2
as food additives. They are also widely used in
cosmetics and some pharmaceutical preparations.
H C OH
HO C H O
H C Ethoxylated Fats and Oils

H C OH
A number of natural fats and oils have been
CH2OCOR ethoxylated, for example, lanolin (wool fat) and
castor oil ethoxylates. These products are useful
The free OH groups in the molecule can be for application in pharmaceutical products, for
esterified, producing di- and tri-esters. Several example, as solubilizers.
products are available depending on the nature
of the alkyl group of the acid and whether the
product is a mono-, di-, or tri-ester. Some Amine Ethoxylates
examples are given below:
Sorbitan monolaurate – Span 20 These are prepared by addition of ethylene oxide
Sorbitan monopalmitate – Span 40 to primary or secondary fatty amines. With pri-
Sorbitan monostearate – Span 60 mary amines, both hydrogen atoms on the amine
Sorbitan monooleate – Span 80 group react with ethylene oxide, and therefore the
Sorbitan tristearate – Span 65 resulting surfactant has the structure
Sorbitan trioleate – Span 85
The ethoxylated derivatives of Spans (CH2CH2O)xH
(Tweens) are produced by reaction of ethylene
oxide on any hydroxyl group remaining on the R–N
sorbitan ester group. Alternatively, the sorbitol is
(CH2CH2O)yH
first ethoxylated and then esterified. However,
the final product has different surfactant proper- The above surfactants acquire a cationic char-
ties to the Tweens. Some examples of Tween acter if the EO units are small in number and if
surfactants are given below: the pH is low. However, at high EO levels and
Polyoxyethylene (20) sorbitan monolaurate – neutral pH, they behave very similarly to non-
Tween 20 ionics. At low EO content, the surfactants are not
Polyoxyethylene (20) sorbitan monopalmitate – soluble in water but become soluble in an acid
Tween 40 solution. At high pH, the amine ethoxylates are
Polyoxyethylene (20) sorbitan monostearate – water soluble, provided the alkyl chain length of
Tween 60 the compound is not long (usually a C12 chain is
Polyoxyethylene (20) sorbitan monooleate – adequate for reasonable solubility at sufficient
Tween 80 EO content).
Polyoxyethylene (20) sorbitan tristearate –
Tween 65
Polyoxyethylene (20) sorbitan trioleate – Amine Oxides
Tween 85
The sorbitan esters are insoluble in water These are prepared by oxidizing a tertiary nitro-
but soluble in most organic solvents (low HLB gen group with aqueous hydrogen peroxide at
Surfactants 1251 S
temperatures in the region 60–80 C. (2-hydroxyethyl) amine oxide, and N-alkyl
Several examples can be quoted: N-alkyl dimethyl amine oxide. They have the general
amidopropyl-dimethyl amine oxide, N-alkyl bis formula
CH3
CocoCONHCH2CH2CH2N → O Alkyl amidopropyl-dimethyl amine oxide
CH3
CH2CH2OH
Coco N → O Coco bis (2-hydroxyethyl) amine oxide
CH2CH2OH
CH3
CH12H25N → O Lauryl dimethyl amine oxide
CH3
In acid solutions, the amino group is proton- property. Fluorocarbon surfactants have been
ated and acts as a cationic surfactant. In neutral or prepared with various structures consisting of
alkaline solution, the amine oxides are essentially perfluoroalkyl chains and anionic, cationic,
nonionic in character. Alkyl dimethyl amine amphoteric, and polyethylene oxide polar
oxides are water soluble up to C16 alkyl chain. groups. These surfactants have good thermal
Above pH 9, amine oxides are compatible with and chemical stability, and they are excellent
most anionics. At pH 6.5 and below, some wetting agents for low energy surfaces. Silicone
anionics tend to interact and form precipitates. surfactants, sometimes referred to as
In combination with anionics, amine oxides can organosilicones, are those with polydimethyl-
be foam boosters (e.g., in shampoos). siloxane backbone. The silicone surfactants are
prepared by incorporation of a water-soluble or
Speciality Surfactants hydrophilic group into a siloxane backbone. The S
Fluorocarbon and Silicone Surfactants latter can also be modified by incorporation of
These surfactants can lower the surface tension a paraffinic hydrophobic chain at the end or
of water to values below 20 mNm1. Most sur- along the polysiloxane backbone. The most
factants described above lower the surface ten- common hydrophilic groups are EO/PO, and
sion of water to values above 20 mNm1, the structures produced are rather complex,
typically in the region of 25–27 mNm1. The and most manufacturers of silicone surfactants
fluorocarbon and silicone surfactants are some- do not reveal the exact structure. The mechanism
times referred to as superwetters, as they cause by which these molecules lower the surface ten-
enhanced wetting and spreading of their aqueous sion of water to low values is far from being well
solution. However, they are much more expen- understood. The surfactants are widely applied
sive than conventional surfactants, and they as spreading agents on many hydrophobic sur-
are only applied for specific applications faces. Incorporating organophilic groups into
whereby the low surface tension is a desirable the backbone of the polydimethyl siloxane
S 1252 Surfactants
backbone can give products that exhibit surface- example, esterification of sucrose with fatty
active properties in organic solvents. acids or fatty glycerides to produce sucrose esters
having the following structure:
Gemini Surfactants
CH2COOR
A gemini surfactant is a dimeric molecule
O
consisting of two hydrophobic tails and two H O H CH2OH H
head groups linked together with a short spacer
H
(Zana and Alami 2003). This is illustrated below H OH
OH H
for a molecule containing two cationic head
CH2OH
groups (separated by two methylene groups) OH O
with two alkyl chains: OH OH H
H
The most interesting sugar surfactants are the

Br− H3C-N+-CH2- CH2- H3C-N+- CH3
alkyl polyglucosides (APG) which are synthe-
sized using a two-stage transacetalization process
R R
(von Rybinsky and Hill 2003). In the first stage,
These surfactants show several interesting the carbohydrate reacts with a short-chain alco-
physicochemical properties, such as very high hol, for example, butanol or propylene glycol. In
efficiency in lowering the surface tension the second stage, the short-chain alkyl glucoside
and very low cmc. For example, the cmc of a is transacetalized with a relatively long-chain
conventional cationic dodecyltrimethy- alcohol (C12-14-OH) to form the required alkyl
lammonium bromide is 16 mM whereas that of polyglucoside. This process is applied if oligo-
the corresponding gemini surfactant, having and polyglucoses (e.g., starch, syrups with a low
a two carbon linkage between the head groups, dextrose equivalent, DE) are used. In a simplified
is 0.9 mM. In addition, the surface tension transacetalization process, syrups with high glu-
reached at and above the cmc is lower for gemini cose content (DE > 96 %) or solid glucose types
surfactants when compared to that of can react with short-chain alcohols under normal
the corresponding conventional one. Gemini sur- pressure. The scheme for alkyl polyglucoside
factants are also more effective in lowering the synthesis is shown below. Commercial alkyl
dynamic surface tension (time required polyglucosides (APG) are complex mixtures of
for reaching the equilibrium value is shorter). species varying in the degree of polymerization
These effects are due to the better packing of (DP, usually in the range 1.1–3) and in the length
the gemini surfactant molecules at the air/water of the alkyl chain. When the latter is shorter than
interface. C14, the product is water soluble. The cmc values
of APGs are comparable to nonionic surfactants,
and they decrease with increase in the alkyl chain
Surfactants Derived from Mono- and length.
Polysaccharides APG surfactants have good solubility in
water, and they have high cloud points
Several surfactants were synthesized starting (>100 C). They are stable in neutral and alka-
from mono- or oligosaccharides by reaction line solutions but are unstable in strong acid
with the multifunctional hydroxyl groups: solutions. APG surfactants can tolerate high
alkyl glucosides, alkyl polyglucosides electrolyte concentrations, and they are compat-
(von Rybinsky and Hill 2003), sugar fatty acid ible with most types of surfactants. They are
esters, and sucrose esters (Drummond et al. used in personal care products for cleansing for-
2003), etc. The technical problem is one of join- mulations as well as for skin care and hair prod-
ing a hydrophobic group to the multihydroxyl ucts. They are also used in hard-surface cleaners
structure. Several surfactants were made, for and laundry detergents. Several applications in
Surfactants 1253 S
agrochemical formulations can be mentioned higher cmc than the less hydrophilic dihydroxy
such as wetting agents and penetrating agents derivatives. As expected, the pH of solutions of
for the active ingredient. these carboxylic acid salts has an influence on
micelle formation. At sufficiently low pH, bile
Naturally Occurring Surfactants acids which are sparingly soluble will be precip-
Several naturally occurring amphipathic mole- itated from solution, initially being incorporated
cules (in the body) exist, such as bile salts, phos- or solubilized in the existing micelles. The pH at
pholipids, and cholesterol, which play an which precipitation occurs, on saturation of the
important role in various biological processes. micellar system, is generally about one pH unit
Their interactions with other solutes, such as higher than the pKa of the bile acid. Bile salts play
drug molecules, and with membranes are also important roles in physiological functions and
very important. Bile salts are synthesized in the drug absorption. It is generally agreed that bile
liver, and they consist of alicyclic compounds salts aid fat absorption. Mixed micelles of bile
possessing hydroxyl and carboxyl groups. As an salts, fatty acids, and monoglycerides can act as
illustration, the structure of cholic acid is given vehicles for fat transport. However, the role of
below: bile salts in drug transport is not well understood.
Several suggestions have been made to explain
HO
COOH the role of bile salts in drug transport, such as
facilitation of transport from liver to bile by direct
effect on canicular membranes, stimulation of
micelle formation inside the liver cells, binding
HO OH of drug anions to micelles, etc. The enhanced
H
absorption of medicinals on administration with
It is the positioning of the hydrophilic groups deoxycholic acid may be due to reduction in
in relation to the hydrophobic steroidal nucleus interfacial tension or micelle formation. The
that gives the bile salts their surface activity and administration of quinine and other alkaloids in
determines the ability to aggregate. Figure 3 combination with bile salts has been claimed to
shows the possible orientation of cholic acid at enhance their parasiticidal action. Quinine, taken
the air-water interface, the hydrophilic groups orally, is considered to be absorbed mainly from
being oriented toward the aqueous phase (Barry the intestine, and a considerable amount of bile
and Gray 1975; Oakenfull and Fisher 1977). The salts is required to maintain a colloidal dispersion
steroid portion of the molecule is shaped like of quinine. Bile salts may also influence drug
a “saucer” as the A ring is cis with respect to the absorption either by affecting membrane perme-
B ring. Small (1968) suggested that small or ability or by altering normal gastric emptying
primary aggregates with up to 10 monomers rates. For example, sodium taurocholate S
form above the cmc by hydrophobic interactions increases the absorption of sulfaguanidine from
between the nonpolar sides of the monomers. the stomach, jejunum, and ileum. This is due to
These primary aggregates form larger units by increase in membrane permeability induced by
hydrogen bonding between the primary micelles. calcium depletion and interference with the
This is schematically illustrated in Fig. 4. bonding between phospholipids in the
Oakefull and Fisher (1977, 1978) stressed the membrane.
role of hydrogen bonding rather than hydropho- Another important naturally occurring class of
bic bonding in the association of bile salts. surfactants which are widely found in biological
However, Zana (1978) regarded the association membranes is the lipids, of which are phosphati-
as a continuous process, with hydrophobic inter- dylcholine (lecithin), lysolecithin, phosphatidyl-
action as the main driving force (Figs. 1, 2). ethanolamine, and phosphatidyl inositol. The
The cmc of bile salts is strongly influenced by structure of these lipids is given in Fig. 5. These
its structure; the trihydroxy cholanic acids have lipids are also used as emulsifiers for intravenous
S 1254 Surfactants
Surfactants,
Fig. 1 (a) Structural
formula of cholics acid
showing the cis position of
the A ring; (b) Courtauld
space filling model of
cholic acid; (c) orientation
of cholic acid molecules at
the air-water interface
(hydroxyl groups
represented by filled circles
and carboxylic acid groups
by open circles)
Surfactants,
Fig. 2 Schematic
structure of bile acid salt
micelles
fat emulsions and anesthetic emulsions, as well as phospholipids organized into bilayers which
for production of liposomes or vesicles for drug assume a multilamellar or unilamellar structure.
delivery. The lipids form coarse turbid disper- The multilamellar species are heterogeneous
sions of large aggregates (liposomes) which on aggregates, most commonly prepared by
ultrasonic irradiation form smaller units or vesi- dispersal of a thin film of phospholipid
cles. The liposomes are smectic mesophases of (alone or with cholesterol) into water. Sonication
Surfactants 1255 S
Surfactants, Non-polar
Fig. 3 Structure of lipids CH2O · CO ·R
portion
R1 · COO · CH O CH3
+ Polar
CH2O P O·CH2 ·CH2 · N CH3
portion
O− CH3
Choline
Lecithin
CH2O · CO · R CH2O ·CO·R
HOCH O CH3 R1 · COO· CH O

+
CH2O P O · CH2·CH2· N CH3 CH2O P O ·CH2 ·CH2
O− CH3 O− Ethanolamine
Lysolecithin Phosphatidyl ethanolamine
CH2O·CO·R
OH OH
R1 · CO O ·CH O
CH2O P O
HO
OH
O−
OH
Phosphatidyl inositol
of the multilamellar units can produce the Phospholipids play an important role in lung
unilamellar liposomes, sometimes referred to as functions. The surface-active material to be
vesicles. The net charge of liposomes can be found in the alveolar lining of the lung is
varied by incorporation of a long-chain amine, a mixture of phospholipids, neutral lipids, and
such as stearyl amine (to give a positively proteins. The lowering of surface tension by the
charged vesicle) or dicetyl phosphate (giving lung surfactant system and the surface elasticity S
negatively charged species). Both lipid-soluble of the surface layers assist alveolar expansion and
and water-soluble drugs can be entrapped in lipo- contraction. Deficiency of lung surfactants in
somes. The liposoluble drugs are solubilized in the newborns leads to a respiratory distress
the hydrocarbon interiors of the lipid bilayers, syndrome, and this led to the suggestion that
whereas the water-soluble drugs are intercalated instillation of phospholipid surfactants could
in the aqueous layers. The use of liposomes as cure the problem.
drug carriers has been reviewed by Fendler and
Romero (Lindman 1984), to which the reader Biosurfactants
should refer for details. Liposomes, like micelles, In recent years, there has been a great deal of
may provide a special medium for reactions to concern regarding the use of conventional surfac-
occur between the molecules intercalated in the tants in cosmetic and personal care applications.
lipid bilayers or between the molecules entrapped This is due to the environmental and health con-
in the vesicle and free solute molecules (Fig. 3). cerns when using many of the currently used
S 1256 Surfactants
synthetic surfactants. These concerns can be alle- lipase as a catalyst in a suitable reaction
viated by the use of biosurfactants that are pro- medium. Enzymes can be used to modify
duced from natural raw materials that possess glycerophospholipids by removal of one of the
good biodegradability, low toxicity, and the fatty acids to produce lysophospholipids (a good
desirable functional performance such as good emulsifying agent). This transformation can be
emulsification and dispersion, high physical sta- achieved using phospholipase A2 that removes
bility, and high performance on application. The the fatty acid in the sn-2 position.
biosurfactants are produced using catalysts in the Lipases have been used to synthesize a wide
form of living microorganisms and enzymes range of surfactant esters from fatty acids and
(Desai and Banat 1997; Guatam and Tyagi carbohydrates. The fatty acid can be used either
2006; Straathof and Aldercreutz 2000). Microbial in the free form or as an ester. To make it possi-
biosurfactants are structurally diverse and ble to find a solvent that efficiently dissolves
complex and are produced in a biosynthetic both fatty acid and carbohydrate. The carbohy-
root catalyzed by enzymes. Several classes are drate is made more hydrophobic by converting it
available: glycolipids, for example, to an acetal, leaving one OH group for esterifi-
rhamnolipids, sophorolipids; lipopeptides and cation. The acetal groups are removed by acid
lipoproteins, for example, surfactin, poly- catalysis. This is illustrated in Scheme 1 shown
myxins, gramicidins; phospholipids and fatty below. Alternatively, the carbohydrate is made
acids; and complex combinations of biopoly- more hydrophobic by converting it to a butyl
mers, for example, emulsan, liposan. The enzy- glycoside using glucosidase which is then fur-
matic synthesis of surfactants is essentially ther reacted with fatty acid using lipase as is
chemical reactions in which an enzyme replaces illustrated in Scheme 2 below. Most studies of
a conventional chemical catalyst. The surfac- enzymatic sugar ester synthesis have been
tants produced by a single enzyme are simpler focused on the esterification of monosaccha-
in structure when compared to those produced rides, since the problem with poor substrate mis-
by microorganisms. Enzymes are able to cata- cibility increases significantly with increasing
lyze a wide range of reactions, and they are the size of the carbohydrate. However, by care-
unique in their specificity and selectivity. The ful choice of reaction conditions, it has been
selectivity is recognized at three levels: chemo-, possible to acylate several di- and trisaccharides.
regio-, and enantioselectivity. Thus, reactions Solvent mixtures of 2-methyl-2-butanol and
that cannot be achieved by classical organic dimethylsulfoxide have been used in combina-
synthesis (that may require several steps) can tion with vinyl esters (C8–C18) to make success-
be facilitated by biocatalysis. The high selectiv- ful acylation of maltose, maltotriose, and
ity results in the production of fewer by-prod- leucrose. Ionic liquids are promising
ucts. Development of solvent-free processes is nonaqueous solvents for the dissolution of car-
also possible using enzymes, and hence the bohydrates, and they have been used in several
products are safer and environmentally friendly. studies on enzymatic sugar ester synthesis. Pro-
Biocatalysis has been used for synthesis of sur- tease-catalyzed synthesis of sugar esters has
factants with different hydrophilic groups that been reported. 6-O-butyl-D-glucose was pre-
are attached to the hydrophobic chain via ester, pared using subtilisin (a protease from Bacillus
amide, or glycosidic bonds. Typical examples subtilis) as catalyst and trichloroethyl butyrate
are the mono- and diacyl glycerols. The techni- as acyl donor in anhydrous dimethyl formamide.
cal production uses a natural fat or oil (triglyc- Oligosaccharides as long as maltoheptase were
eride) as the starting material. Enzymatic also acylated under these conditions. Bacillus
methods have the advantage of high selectivity. protease has been used to synthesize sucrose
For example, it is possible to make 2-monoacyl laurate from sucrose and vinyl laurate in
glycerols by selective removal of the fatty acids dimethylformamide. Lipases can accept a wide
in the 1 and 3 positions using a regiospecific variety of nucleophiles for the deacylation of
Surfactants 1257 S
Surfactants, Scheme 1
Fig. 4 Schemes for
synthesis of an alkyl RCOOH Trifluoroacetic acid
glycoside by a two-step Acetal Alkyl Acetal Alkyl glycoside
Lipase
enzymatic root
Scheme 2
BuOH RCOOH
Glucose Butyl Glycoside Alkyl glycoside
Glucosidase Lipase
acyl enzyme synthesis. By proper choice of reac- example, personal care and hard-surface cleaning.
tion conditions, it was possible to acylate a wide They are characterized by their skin tolerance,
good biodegradability, and low toxicity.
range of carbohydrates. The main products are Alkanolamides are produced by condensation of
obtained by esterification of the primary fatty acids or fatty acid esters or triglyceride with
hydroxyl group. In carbohydrates having more alkanolamine, for example, monoethanolamine or
than one primary OH group, the enzyme can diethanolamine, using high temperature or a metal
oxide catalyst. Alkanolamine synthesis can also
selectively acylate one of them. This is the case be achieved using lipase. Alkanolamines are sus-
with maltose. With fructose, mixtures are ceptible to acylation both at the amine and
obtained. Most carbohydrate esters have been hydroxyl groups. The main product using lipase
prepared using monosaccharides as substrates (Novozym 435) is the amide.
(Fig. 4). Amino acid/peptide–lipid conjugates are
Fatty acid derivatives with an amide bond possess interesting class of surfactants with good surface
useful properties for surfactants, for example, activity, excellent emulsifying agents, antimicro-
enhancing foaming properties of cleaning, personal
care products, stabilizing foam, and enhancing
bial activity, low toxicity, and high biodegrad-
detergency. The amide bond increases the hydro- ability. They are attractive for application in
philicity of the fatty acid, and therefore the surfac- personal care products. The large variety of
tant becomes chemically and physically very stable amino acid/peptide structures combined with
under alkaline conditions. Synthesis of amides can
be done by proteases, but these enzymes are very
fatty acids of varying structures and carbon
specific for certain amino acids and are more chain lengths can produce surfactants with wide
sensitive to organic solvents. Lipases have been structural diversity and different physicochemi-
used for synthesis of peptides, fatty amides, and cal and biological properties. Depending on the
N-acylamino acids, and the acylation of
alkanolamines.
free functional groups on the amino acids,
Another important class of biosurfactants is anionic, nonionic, amphoteric, and cationic sur-
the amino-sugar derivatives. Glycamide surfac- factants can be produced. Different forms of S
tants are nonionic, biodegradable in which the amino acid surfactants have been synthesized
hydrophilic moiety (an amino-sugar derivative)
is linked to the fatty acid by an amide bond, for
using enzymes.
example, glucamides and lactobionamides. As mentioned before, carbohydrate ester sur-
A conventional method for preparing sugar fatty factants suffer from the disadvantage of chemical
amide surfactants includes the Schotten– instability at neutral or alkaline pH (due to the
Baumann reaction between an amine and a fatty
acid chloride, where the chloride salt produced has
instability of the ester bond). This instability
to be removed. The regio- and enantioselectivity problem can be overcome by using alkyl glyco-
of enzymes provides a convenient method of acyl- sides which are based on similar building blocks.
ation of sugars and sugar amines. Chemoselective Chemical synthesis is useful for the production of
acylation of a secondary amine, N-methyl
glucamine, with fatty acid is possible using lipases
mixtures of alkyl glycosides as discussed before.
(from Novozym). When pure isomers are needed, enzymatic syn-
Alkanolamides are important fatty acid deri- thesis offers an attractive alternative. Two main
vates for a wide range of applications, for classes of enzymes can be used for coupling the
S 1258 Surfactants
carbohydrate part to a hydrophobic alcohol, technically referred to as poly(vinyl alcohol)

namely, glycosyl transferases and glycosyl (PVA). The polymer is prepared by partial hydro-
hydrolases. An alternative route to alkyl glyco- lysis of PVAc, leaving some residual vinyl ace-
sides is to use transglycosylation reaction with an tate groups. Most commercially available PVA
activated carbohydrate substrate. molecules contain 4–12 % acetate groups. These
acetate groups which are hydrophobic give the
Polymeric Surfactants molecule its amphipathic character. On a hydro-
This subject has been covered in detail in phobic surface such as polystyrene, the polymer
a separate entry, and here a summary is given. adsorbs with preferential attachment of the
The simplest type of a polymeric surfactant is acetate groups on the surface, leaving the more
a homopolymer, that is, formed from the same hydrophilic vinyl alcohol segments dangling
repeating units, such as poly(ethylene oxide) or in the aqueous medium. These partially hydro-
poly(vinyl pyrrolidone). These homopolymers lyzed PVA molecules also exhibit surface
have little surface activity at the o/w interface activity at the o/w interface (Tadros 2003). The
since the homopolymer segments (ethylene most convenient polymeric surfactants are those
oxide or vinylpyrrolidone) are highly water of the block and graft copolymer type.
soluble and have little affinity to the interface. A block copolymer is a linear arrangement of
However, such homopolymers may adsorb sig- blocks of variable monomer composition.
nificantly at the S/L interface. Even if the adsorp- The nomenclature for a diblock is poly-A-
tion energy per monomer segment to the surface block-poly-B and for a triblock is poly-A-block-
is small (fraction of kT, where k is the Boltzmann poly-B-poly-A. One of the most widely used
constant and T is the absolute temperature), the triblock polymeric surfactants are the “Pluronics”
total adsorption energy per molecule may be suf- (BASF, Germany) which consists of two poly-A
ficient to overcome the unfavorable entropy loss blocks of poly(ethylene oxide) (PEO) and one
of the molecule at the S/L interface. Clearly, block of poly(propylene oxide) (PPO). Several
homopolymers are not the most suitable emulsi- chain lengths of PEO and PPO are available.
fiers or dispersants. A small variant is to use Two types may be distinguished: those prepared
polymers that contain specific groups that have by reaction of polyoxypropylene glycol (difunc-
high affinity to the surface. This is exemplified by tional) with EO or mixed EO/PO, giving block
partially hydrolyzed poly(vinyl acetate) (PVAc), copolymers with the structure
HO(CH2CH2O)n-(CH2CHO)m-(CH2CH2)nOH abbreviated (EO)n(PO)m(EO)n
CH3
Various molecules are available, where n and surface, leaving the two PEO chains dangling
m are varied systematically. The second type of in aqueous solution and hence providing steric
EO/PO copolymers is prepared by reaction of repulsion. Trifunctional products are also
polyethylene glycol (difunctional) with PO or available where the starting material is glycerol.
mixed EO/PO. These will have the structure These have the structure
(PO)n(EO)m(PO)n, and they are referred to as
CH2-(PO)m(EO)n
reverse Pluronics. These polymeric triblocks
can be applied as emulsifiers or dispersants, CH -(PO)n(EO)n
whereby the assumption is made that the hydro-
phobic PPO chain resides at the hydrophobic CH2-(PO)m(EO)n
Surfactants 1259 S
Tetrafunctional products are available where H OH
the starting material is ethylene diamine. These H O
have the structures, HO
HO
(EO)n (EO)n H H OH
() N O H O
NCH2CH2N O
r
O OH
(EO)n (EO)n
CH2
(EO)n(PO)m (PO)m(EO)n
OH n
NCH2CH2N O
CH2OH
O
(EO)n(PO)m (PO)m(EO)n
OH
Although these block polymeric surfactants CH2OH

have been widely used in various applications in OH
emulsions and suspensions, some doubt has (GFn)
arisen on how effective these can be. It is gener- 1a
ally accepted that the PPO chain is not suffi- ®
Surfactants, Fig. 5 Structure of INUTEC SP1
ciently hydrophobic to provide a strong
“anchor” to a hydrophobic surface or to an oil
droplet. Indeed, the reason for the surface activity polyethylene oxide graft copolymers. Recently,
of the PEO-PPO-PEO triblock copolymers at the graft copolymers based on polysaccharides
o/w interface may stem from a process of “rejec- (Tadros 2003) have been developed for stabiliza-
tion” anchoring of the PPO chain since it is not tion of disperse systems. One of the most useful
soluble both in oil and in water. Several other di- graft copolymers are those based on inulin that is
and triblock copolymers have been synthesized, obtained from chicory roots. It is a linear
although these are of limited commercial avail- polyfructose chain with a glucose end. When
ability. Typical examples are diblocks of poly- extracted from chicory roots, inulin has a wide
styrene-block-polyvinyl alcohol, triblocks of range of chain lengths ranging from 2 to 65
poly(methyl methacrylate)-block poly(ethylene fructose units. It is fractionated to obtain
oxide)-block poly(methyl methacrylate), a molecule with narrow molecular weight distri-
diblocks of polystyrene block-polyethylene bution with a degree of polymerization >23, and
®
oxide, and triblocks of polyethylene oxide-block this is commercially available as INUTEC N25.
polystyrene-polyethylene oxide (Tadros 2003). The latter molecule is used to prepare a series of
An alternative (and perhaps more efficient) poly- graft copolymers by random grafting of alkyl
S
meric surfactant is the amphipathic graft copoly- chains (using alkyl isocyanate)on the inulin back-
mer consisting of a polymeric backbone bone. The first molecule of this series is
®
B (polystyrene or polymethyl methacrylate) and INUTEC SP1 (Beneo-Remy, Belgium) that is
several A chains (“teeth”) such as polyethylene obtained by random grafting of C12 alkyl chains.
oxide (Tadros 2003) This graft copolymer is It has an average molecular weight of 5,000 Da,
sometimes referred to as a “comb” stabilizer. and its structure is given in Fig. 8. The molecule
This copolymer is usually prepared by grafting is schematically illustrated below which shows
a macromonomer such as methoxy polyethylene the hydrophilic polyfructose chain (backbone)
oxide methacrylate with polymethyl methacry- and the randomly attached alky chains. The
®
late. The “grafting onto” technique has also main advantages of INUTEC SP1 (Fig. 5) as
been used to synthesize polystyrene- stabilizer for disperse systems are (1) strong
S 1260 Surfactants
adsorption to the particle or droplet by multipoint

attachment with several alky chains – this ensures osmotic pressure
lack of desorption and displacement of the mole-
turbidity solubilization
cule from the interface; and (2) strong hydration
of the linear polyfructose chains both in water magnetic resonance
and in the presence of high electrolyte concentra- surface tension

tions and high temperatures – this ensures equivalent
effective steric stabilization. conductivety
self-diffusion
Physical Chemistry of Surfactant CMC Concentration

Solutions
Surfactants, Fig. 6 Variation of solution properties with
The physical properties of surface-active agent surfactant concentration
solutions differ from those of non-amphipathic
molecule solutions (such as sugars) in one major
aspect, namely, the abrupt changes in their prop-
erties above a critical concentration (Lindman
1984). This is illustrated in Fig. 6 which shows
plots of several physical properties (osmotic pres-
sure, surface tension, turbidity, solubilization,
magnetic resonance, equivalent conductivity,
and self-diffusion) as a function of concentration
for an anionic surfactant. At low concentrations,
most properties are similar to those of a simple
electrolyte. One notable exception is the surface
tension, which decreases rapidly with increasing
surfactant concentration. However, all the prop-
erties (interfacial and bulk) show an abrupt
change at a particular concentration, which is
consistent with the fact that at and above this
concentration, surface-active molecules or ions
associate to form larger units. These associated
units are called micelles (self-assembled Surfactants, Fig. 7 Illustration of a spherical micelle for
structures), and the first formed aggregates are dodecyl sulfate (Istraelachvili 1985)
generally approximately spherical in shape.
A schematic representation of a spherical micelle
is given in Fig. 7. measurements, NMR, and fluorescence spectros-
The concentration at which this association copy can be applied. A compilation of cmc values
phenomenon occurs is known as the critical has been given in 1971 by Mukerjee and Mysels
micelle concentration (cmc). Each surfactant (1971), which is clearly not an up-to-date text but
molecules has a characteristic cmc value at is an extremely valuable reference. As an illus-
a given temperature and electrolyte concentra- tration, the cmc values of a number of surface-
tion. The most common technique for measuring active agents are given in Table 1 to show some
the cmc is surface tension, g, which shows break of the general trends (Mukerjee and Mysels
at the cmc, after which g remains virtually con- 1971). Within any class of surface-active agent,
stant with further increase in concentration. How- the cmc decreases with increase in chain length
ever, other techniques such as self-diffusion of the hydrophobic portion (alkyl group).
Surfactants 1261 S
Surfactants, Table 1 CMC of some surfactants
Surface-active agent cmc/mol dm3 10
(A) Anionic
Sodium decyl-l-sulfate 3.32 102 9 C12SO4−Na+
Sodium dodecyl-l-sulfate 8.39 103
CMC (mM)
Sodium tetradecyl-l-sulfate 2.05 103
8
(B) Cationic
1
Octyl trimethyl ammonium 1.30 10
bromide 2
Decetryl trimethyl ammonium 6.46 102
bromide
C10E5
Dodecyl trimethyl ammonium 1.56 102 1
bromide
Hexactecyltrimethyl ammonium 9.20 104 0 10 20 30 40 50 60
bromide Temperature (°C)
(C) Nonionic
Octyl hexaoxyethylene glycol 9.80 103 Surfactants, Fig. 8 Temperature dependence of the cmc
monoether C8E6 of SDS and C10E5 (Elworthy et al. 1968)
Decyl hexaoxyethylene glycol 9.00 104
monoether C10E6
Decyl nonaoxyethylene glycol 1.30 103 range. The shallow minimum around 25 C can
monoether C10E9 be compared with a similar minimum in the sol-
Dodecyl hexaoxyethylene glycol 8.70 105 ubility of hydrocarbon in water (Elworthy et al.
monoether C12E6 1968). However, nonionic surfactants of the
Octylphenyl hexaoxyethylene 2.05 104 ethoxylate type show a monotonic decrease
glycol monoether C8E6
(Elworthy et al. 1968) of cmc with increase of
temperature as is illustrated in Fig. 5 for C10E5.
The effect of addition of cosolutes, for example,
As a general rule, the cmc decreases by a factor of electrolytes and nonelectrolytes, on the cmc, can
2 for ionics (without added salt) and by a factor be very striking. For example, addition of 1:1
of 3 for nonionics on adding one methylene group electrolyte to a solution of anionic surfactant
to the alkyl chain. With nonionic surfactants, gives a dramatic lowering of the cmc, which
increasing the length of the hydrophilic group may amount to an order of magnitude. The effect
(polyethylene oxide) causes an increase in cmc. is moderate for short-chain surfactants but is
In general, nonionic surfactants have lower much larger for long-chain ones. At high electro-
cmc values than their corresponding ionic surfac- lyte concentrations, the reduction in cmc with
tants of the same alkyl chain length. Incorpora- increase in number of carbon atoms in the alkyl S
tion of a phenyl group in the alkyl group increases chain is much stronger than without added elec-
its hydrophobicity to a much smaller extent than trolyte. This rate of decrease at high electrolyte
increasing its chain length with the same number concentrations is comparable to that of nonionics.
of carbon atoms. The valency of the counterion in The effect of added electrolyte also depends on
ionic surfactants has a significant effect on the the valency of the added counterions. In contrast,
cmc. For example, increasing the valency of for nonionics, addition of electrolytes causes only
the counterion from 1 to 2 causes a reduction of small variation in the cmc.
the cmc by roughly a factor of 4. Nonelectrolytes such as alcohols can also
The cmc is, to a first approximation, indepen- cause a decrease in the cmc (Shinoda et al.
dent of temperature. This is illustrated in Fig. 8 1963). The alcohols are less polar than water
which shows the variation of cmc of SDS with and are distributed between the bulk solution
temperature. The cmc varies in a non-monotonic and the micelles. The more preference they have
way by ca 10–20 % over a wide temperature for the micelles, the more they stabilize them.
S 1262 Surfactants
A longer alkyl chain leads to a less favorable hydrocarbon. This is confirmed by the high
location in water and more favorable location in mobility of the alkyl chains and the ability of
the micelles. the micelles to solubilize many water-insoluble
The presence of micelles can account for organic molecules, for example, dyes and agro-
many of the unusual properties of solutions of chemicals. To a first approximation, micelles
surface-active agents. For example, it can can, over a wide concentration range above the
account for the near constant surface tension cmc, be viewed as microscopic liquid hydrocar-
value, above the cmc (see Fig. 6). It also bon droplets covered with polar head groups,
accounts for the reduction in molar conductance which interact strongly with water molecules.
of the surface-active agent solution above the It appears that the radius of the micelle core
cmc, which is consistent with the reduction in constituted of the alkyl chains is close to the
mobility of the micelles as a result of counterion extended length of the alkyl chain, that is, in
association. The presence of micelles also the range 1.5030 nm. As we will see later, the
accounts for the rapid increase in light scattering driving force for micelle formation is the elimi-
or turbidity above the cmc. The presence of nation of the contact between the alkyl chains
micelles was originally suggested by McBain and water. The larger a spherical micelle, then
(1913) who suggested that below the cmc, most the more efficient this is, since the volume-to-
of the surfactant molecules are unassociated, area ratio increases. It should be noted that the
whereas in the isotropic solutions immediately surfactant molecules in the micelles are not all
above the cmc, micelles and surfactant ions extended. Only one molecule needs to be
(molecules) are thought to coexist, the concen- extended to satisfy the criterion that the radius
tration of the latter changing very slightly as of the micelle core is close to the extended length
more surfactant is dissolved. However, the self- of the alkyl chain. The majority of surfactant
association of an amphiphile occurs in molecules are in a disordered state. In other
a stepwise manner with one monomer added to words, the interior of the micelle is close to that
the aggregate at a time. For a long-chain amphi- of the corresponding alkane in a neat liquid oil.
phile, the association is strongly cooperative up This explains the large solubilization capacity of
to a certain micelle size where counteracting the micelle toward a broad range of nonpolar and
factors became increasingly important. Typi- weakly polar substances. At the surface of the
cally, the micelles have a closely spherical micelle, associated counterions (in the region of
shape in a rather wide concentration range 50–80 % of the surfactant ions) are present.
above the cmc. Originally, it was suggested by However, simple inorganic counterions are
Adam (1925) and Hartley (1936) that micelles very loosely associated with the micelle. The
are spherical in shape and have the following counterions are very mobile (see below), and
properties: (1) the association unit is spherical there is no specific complex formed with
with a radius approximately equal to the length a definite counterion-head-group distance. In
of the hydrocarbon chain; (2) the micelle con- other words, the counterions are associated by
tains about 50–100 monomeric units; aggrega- long-range electrostatic interactions.
tion number generally increases with increase in A useful concept for characterizing micelle
alkyl chain length; (3) with ionic surfactants, geometry is the critical packing parameter, CPP.
most counterions are bound to the micelle sur- The aggregation number N is the ratio between
face, thus significantly reducing the mobility the micellar core volume, Vmic, and the volume of
from the value to be expected from a micelle one chain, v:
with non-counterion bonding; (4) micellization
occurs over a narrow concentration range as Vmic ð4=3Þp R3mic
a result of the high association number of sur- N¼ ¼ (1)
v v
factant micelles; (5) the interior of the surfactant
micelle has essentially the properties of a liquid where Rmic is the radius of the micelle.
Surfactants 1263 S
The aggregation number, N, is also equal to Sphere
the ratio of the area of a micelle, Amic, to the

cross-sectional area, a, of one surfactant
Rob-like
molecule:
Amic 4p R2mic
N¼ ¼ (2)
a a
Lamellar micelle
Combining Eqs. 1 and 2,
Surfactants, Fig. 9 Shape of micelles
v 1
¼ (3)
Rmic a 3
Since Rmic cannot exceed the extended length Solubility

of a surfactant alkyl chain, lmax,
lmax ¼ 1:5 þ 1:265 nc (4) Solubility
CMC
This means that for a spherical micelle,
v 1 Kraft temperature
(5)
lmax a 3
Temperature
The ratio v/(lmax a) is denoted as the critical Surfactants, Fig. 10 Variation of solubility and critical
packing parameter (CPP). micelle concentration (CMC) with temperature
Although, the spherical micelle model
accounts for many of the physical properties of
solutions of surfactants, a number of phenomena McBain, Hartley, and Debye is given in Fig. 9.
remain unexplained, without considering other Many ionic surfactants show dramatic tempera-
shapes. For example, McBain (1950) suggested ture-dependent solubility as illustrated in Fig. 10.
the presence of two types of micelles, spherical The solubility first increases gradually with rising
and lamellar, in order to account for the drop in temperature, and then, above a certain tempera-
molar conductance of surfactant solutions. The ture, there is a sudden increase of solubility with
lamellar micelles are neutral, and hence they further increase of temperature. The cmc
account for the reduction in the conductance. increases gradually with increasing temperature. S
Later, Harkins et al. (1946) used McBain’s At a particular temperature, the solubility
model of lamellar micelles to interpret his x-ray becomes equal to the cmc, that is, the solubility
results in soap solutions. Moreover, many mod- curve intersects the cmc, and this temperature is
ern techniques such as light scattering and neu- referred to as the Krafft temperature. At this
tron scattering indicate that in many systems the temperature, an equilibrium exists between slid
micelles are not spherical. For example, Debye hydrated surfactant, micelles, and monomers
and Anacker (1951) proposed a cylindrical (i.e., the Krafft temperature is a “tripple point”).
micelle to explain that light scattering results on Surfactants with ionic head groups and long
hexadecyltrimethyl ammonium bromide in straight alkyl chains have high Krafft tempera-
water. Evidence for disc-shaped micelles has tures. The Krafft temperature increases with
also been obtained under certain conditions. increase of the alkyl chain of the surfactant mol-
A schematic representation of the spherical, ecule. It can be reduced by introducing branching
lamellar, and rod-shaped micelles, suggested by in the alkyl chain. The Krafft temperature is also
S 1264 Surfactants
reduced by using alkyl chains with wide distribu- Thermodynamics of Micellization

tion of the chain length. Addition of electrolytes
causes an increase in the Krafft temperature. The process of micellization is one of the most
With nonionic surfactants of the ethoxylate important characteristics of surfactant solution,
type, increase of temperature for a solution at and hence it is essential to understand its mecha-
a given concentration causes dehydration of the nism (the driving force for micelle formation).
PEO chains, and at a critical temperature, the This requires analysis of the dynamics of
solution becomes cloudy. This is illustrated in the process (i.e., the kinetic aspects) as well as
Fig. 11 which shows the phase diagram of the equilibrium aspects whereby the laws of
C12E6. Below the CP curve, one can identify thermodynamics may be applied to obtain the
different liquid crystalline phases (hexagonal, free energy, enthalpy, and entropy of
cubic, lamellar) which are schematically shown micellization.
in Fig. 12.
Kinetic Aspects
100
Micellization is a dynamic phenomenon in which
80 water + soln n monomeric surfactant molecules associate to
rev
form a micelle Sn, that is,
Temperature (°C)
mic
60
nS , Sn (6)
lam
40
mic Hartley (1936) envisaged a dynamic equilib-
cu
rium whereby surface-active agent molecules are

bb
20
hex
solid constantly leaving the micelles while other mol-
0 ecules from solution enter the micelles. The same
0 20 40 60 80 100
Concentration of C12EO6 (wt%) applies to the counterions with ionic surfactants,
which can exchange between the micelle surface
Surfactants, Fig. 11 Phase diagram of C12E6 and bulk solution. Experimental investigation
water surfactant
Hexagonal
surfactant water
Lamellar
1/2
1/2
1/4 1/4
Surfactants,
Fig. 12 Schematic picture 1/2
1/2 Cubic
of liquid crystalline phases
Surfactants 1265 S
using fast kinetic methods, such as stop flow, Equilibrium Aspects: Thermodynamics
temperature, and pressure jumps, and ultrasonic of Micellization
relaxation measurements have shown that there
are two relaxation processes for micellar equilib- Various approaches have been employed in tack-
rium (Anainsson and Wall 1974; Aniansson et al. ling the problem of micelle formation. The sim-
1976; Rassing et al. 1974; Jaycock and Ottewill plest approach treats micelles as a single phase,
1964; Okub et al. 1979; Phillips 1955; Kahlweit and this is referred to as the phase separation
and Teubner 1980) characterized by relaxation model. In this model, micelle formation is con-
times t1 and t2. The first relaxation time, t1, is sidered as a phase separation phenomenon, and
of the order of 107 s (108 to 103 s) and the cmc is then the saturation concentration of the
represents the lifetime of a surface-active mole- amphiphile in the monomeric state, whereas the
cule in a micelle, that is, it represents the associ- micelles constitute the separated pseudophase.
ation and dissociation rate for a single molecule Above the cmc, a phase equilibrium exists with
entering and leaving the micelle, which may be a constant activity of the surfactant in the micellar
represented by the equation phase. The Krafft point is viewed as the temper-
ature at which solid hydrated surfactant, micelles,
S þ Sn1 , Sn (7) and a solution saturated with undissociated sur-
factant molecules are in equilibrium at a given
where K+ and K represent the association and pressure.
dissociation rates, respectively, for a single mol- Consider an anionic surfactant, in which
ecule entering or leaving the micelle. n surfactant anions, S, and n counterions, M+,
The slower relaxation time t2 corresponds to associate to form a micelle, that is,
a relatively slow process, namely, the
micellization–dissolution process represented by n S þ n Mþ , Sn (8)
Eq. 1. The value of t2 is of the order of millisec-
onds (103–1 s) and hence can be conveniently The micelle is simply a charged aggregate
measured by stopped flow methods. The fast of surfactant ions plus an equivalent number of
relaxation time t1 can be measured using various counterions in the surrounding atmosphere and
techniques depending on its range. For example, is treated as a separate phase. The chemical
t1 values in the range of 108–107 s are acces- potential of the surfactant in the micellar state
sible to ultrasonic absorption methods, whereas is assumed to be constant, at any given tem-
t1 in the range of 105–103 s can be measured perature, and this may be adopted as the stan-
by pressure jump methods. The value of t1 dard chemical potential, mmo, by analogy to
depends on surfactant concentration, chain a pure liquid or a pure solid. Considering the
length, and temperature. t1 increases with equilibrium between micelles and monomer, S
increase of chain length of surfactants, that is, then,
the residence time increases with increase of
chain length. mom ¼ mo1 þ RT ln a (9)
The above discussion emphasizes the dynamic
nature of micelles, and it is important to realize where m1 is the standard chemical potential of the
that these molecules are in continuous motion surfactant monomer and a1 is its activity which is
and that there is a constant interchange between equal to f1 x1, where f1 is the activity coefficient
micelles and solution. The dynamic nature also and x1 the mole fraction. Therefore, the standard
applies to the counterions which exchange rap- free energy of micellization per mol of monomer,
idly with lifetimes in the range 109–108 s. Fur- D Gmo, is given by
thermore, the counterions appear to be laterally
mobile and not to be associated with (single)
specific groups on the micelle surfaces. D Gom ¼ mom mo1 ¼ RT ln a1 RT ln x1 (10)
S 1266 Surfactants
where f1 is taken as unity (a reasonable value in the surfactant in the aqueous phase would be con-
very dilute solution). The cmc may be identified stant above the cmc. If this was the case, the surface
with x1 so that tension of a surfactant solution should remain
constant above the cmc. However, careful mea-
D Gom ¼ RT ln ½c:m:c (11) surements have shown that the surface tension of
a surfactant solution decreases slowly above the
In Eq. 10, the cmc is expressed as a mole frac- cmc, particularly when using purified surfactants.
tion, which is equal to C/(55.5 + C), where C is the A convenient solution for relating DGm to [cmc]
concentration of surfactant in mole dm3, that is, was given by Phillips (Adam 1925) for ionic sur-
factants who arrived at the following expression:
D Gom ¼ RT ln C RT ln ð55:5 þ CÞ (12)
D Gom ¼ f2 ðp=nÞg RT ln ½c:m:c: (15)
o
It should be stated that DG should be calcu-
lated using the cmc expressed as a mole fraction where p is the number of free (unassociated)
as indicated by Eq. 12. However, most cmcs surfactant ions and n is the total number of sur-
quoted in the literature are given in mole dm3, factant molecules in the micelle. For many ionic
and in many cases of DGo, values have been surfactants, the degree of dissociation (p/n) is
quoted when the cmc was simply expressed in 0.2 so that
mol dm3. Strictly speaking, this is incorrect,
since DGo should be based on x1 rather than on D Gom ¼ 1:8 RT ln ½c:m:c (16)
C. The value of DGo, when the cmc is expressed
in mol dm3, is substantially different from Comparison with Eq. 11 clearly shows that for
the DGo value when the cmc is expressed similar DGm, the [cmc] is about two orders of
in mole fraction. For example, dodecyl magnitude higher for ionic surfactants when
hexaoxyethylene glycol, the quoted cmc value compared with nonionic surfactant of the same
is 8.7 105 mol dm3 at 25 C. Therefore, alkyl chain length (see Table 1).
In the presence of excess added electrolyte,
8:7 105
D Go ¼ RT ln with mole fraction x, the free energy of
55:5 þ 8:7 105 (13) micellization is given by the expression
¼ 33:1 KJ mol1
when the mole fraction scale is used. On the other D Gom ¼ RT ln ½c:m:c þ f1 ðp=nÞg ln x
hand, (17)
Equation 17 shows that as x increases, the

D Go ¼ RT ln 8:7 105
(14) [cmc] decreases.
¼ 23:2 KJ mol1 It is clear from Eq. 15 that as p ! 0, that is,
when most charges are associated with
when the molarity scale is used. counterions,
The phase separation model has been
questioned for two main reasons. Firstly, according D Gom ¼ 2 RT ln ½c:m:c (18)
to this model, a clear discontinuity in the physical
property of a surfactant solution, such as surface whereas when p n, that is, the counterions are
tension, turbidity, etc., should be observed at the bound to micelles:
cmc. This is not always found experimentally, and
the cmc is not a sharp break point. Secondly, if two D Gom ¼ RT ln ½c:m:c (19)
phases actually exist at the cmc, then equating the
chemical potential of the surfactant molecule in which is the same equation for nonionic
the two phases would imply that the activity of surfactants.
Surfactants 1267 S
Surfactants, Table 2 Thermodynamic quantities for micellization of octylhexaoxyethylene glycol monoether
DHo/KJ mol1 DHo/KJ mol l1
1

Temp/ C o
DG /KJ mol (from cmc) (from calorimetry) TDSo/KJ mol1
25 21.3 2.1 8.0 4.2 20.1 0.8 41.8 1.0
40 23.4 2.1 14.6 0.8 38.0 1.0
Enthalpy and Entropy of Micellization Surfactants, Table 3 Change of thermodynamic param-

eters of micellization of alkyl sulfoxide with increasing
chain length of the alkyl group
The enthalpy of micellization can be calculated
from the variation of cmc with temperature. This DG/KJ DHo/KJ
Surfactant mol1 mol1 TDSo/KJ mol1
follows from
C6H13S(CH3)O 12.0 10.6 22.6
C7H15S(CH3)O 15.9 9.2 25.1
dln ½c:m:c
D Ho ¼ RT2 (20) C8H17S(CH3)O 18.8 7.8 26.4
dT C9H19S(CH3)O 22.0 7.1 29.1
C10H21S(CH3)O 25.5 5.4 30.9
The entropy of micellization can then be
C11H23S(CH3)O 28.7 3.0 31.7
calculated from the relationship between DGo
and DHo, that is,
D Go ¼ D Ho TDSo (21) negative as the chain length increases. This is to be

expected since the cmc decreases with increase of
Therefore, DHo may be calculated from the the alkyl chain length. However, DHo becomes
surface tension-log C plots at various tempera- less positive, and TDS becomes more positive
tures. Unfortunately, the errors in locating the with increase in chain length of the surfactant.
cmc (which in many cases is not a sharp point) Thus, the large negative free energy of
lead to a large error in the value of DHo. A more micellization is made up of a small positive
accurate and direct method of obtaining DHo is enthalpy (which decreases slightly with increase
microcalorimetry. As an illustration, the thermo- of the chain length of the surfactant) and a large
dynamic parameters, DGo, DHo, and TDSo for positive entropy term TDSo, which becomes more
octylhexaoxyethylene glycol monoether (C8E6), positive as the chain is lengthened. As we will see
are given in Table 2. in the next section, these results can be accounted
It can be seen from Table 2 that DGo is large for in terms of the hydrophobic effect which will
and negative. However, DHo is positive, indicat- be described in some detail. S
ing that the process is endothermic. In addition,
TDSo is large and positive which implies that in
the micellization process, there is a net increase in Driving Force for Micelle Formation
entropy. This positive enthalpy and entropy point
to a different driving force for micellization from Until recently, the formation of micelles was
that encountered in many aggregation processes. regarded primarily as an interfacial energy pro-
The influence of alkyl chain length of the sur- cess, analogous to the process of coalescence of
factant on the free energy, enthalpy, and entropy of oil droplets in an aqueous medium. If this was the
micellization, was demonstrated by Rosen (1978) case, micelle formation would be a highly exo-
who listed these parameters as a function of alkyl thermic process, as the interfacial free energy has
chain length for sulfoxide surfactants. The results a large enthalpy component. As mentioned
are given in Table 3; it can be seen that the standard above, experimental results have clearly shown
free energy of micellization becomes increasingly that micelle formation involves only a small
S 1268 Surfactants
Surfactants, Table 4 Thermodynamic parameters for transfer of hydrocarbons from water to liquid hydrocarbon at
25 C
DGo DHo DSo DCpo Cpo,gas
Hydrocarbon kJmol1 kJmol1 kJmol1 K1 kJmol1 K1 kJmol1 K1
C2H6 16.4 10.5 88.2 – –
C3H8 20.4 7.1 92.4 –
C4H10 24.8 3.4 96.6 273 143
C5H12 28.8 2.1 105.0 403 172
C6H14 32.5 0 109.2 441 197
C6H6 19.3 2.1 58.8 227 134
C6H5CH3 22.7 1.7 71.4 265 155
C6H5C2H5 26.0 2.0 79.8 319 185
C6H5C3H8 29.0 2.3 88.2 395
enthalpy change and is often endothermic. The is also operative on transfer of surfactant mono-
negative free energy of micellization is the result mer to a micelle during the micellization process.
of a large positive entropy. This led to the con- The surfactant monomers will also contain
clusion that micelle formation must be predomi- “structured” water around their hydrocarbon
nantly entropy-driven process. Two main sources chain. On transfer of such monomers to
of entropy may have been suggested. The first is a micelle, these water molecules are released,
related to the so-called hydrophobic effect. The and they have a higher entropy. The second
latter effect was first established from a consid- source of entropy increase on micellization may
eration of the free-energy enthalpy and entropy of arise from the increase in flexibility of the hydro-
transfer of hydrocarbon from water to a liquid carbon chains on their transfer from an aqueous to
hydrocarbon. Some results are listed in Table 4. a hydrocarbon medium (Tanford 1980). The ori-
This table also lists the heat capacity change DCp entations and bendings of an organic chain are
on transfer from water to a hydrocarbon, as well likely to be more restricted in an aqueous phase
as Cpo,gas, that is, the heat capacity in the gas compared to an organic phase. It should be
phase. It can be seen from Table 4 that the prin- mentioned that with ionic and zwitterionic
cipal contribution to the value of DGo is the large surfactants, an additional entropy contribution,
positive value of DSo, which increases with associated with the ionic head groups, must be
increase in the hydrocarbon chain length, considered. Upon partial neutralization of the
whereas DHo is positive, or small, and negative. ionic charge by the counterions when aggregation
To account for this large positive entropy of occurs, water molecules are released. This will be
transfer several authors (Tanford 1980; Stainsby associated with an entropy increase which should
and Alexander 1950; Arnow and Witten 1960) be added to the entropy increase resulting from
suggest that the water molecules around the hydrophobic effect mentioned above. How-
a hydrocarbon chain are ordered, forming “clus- ever, the relative contribution of the two effects is
ters” or “icebergs.” On transfer of an alkane from difficult to assess in a quantitative manner.
water to a liquid hydrocarbon, these clusters are
broken, thus releasing water molecules which
now have a higher entropy. This accounts for Micellization in Surfactant Mixtures
the large entropy of transfer of an alkane from (Mixed Micelles)
water to a hydrocarbon medium. This effect is
also reflected in the much higher heat capacity In most industrial applications, more than one
change on transfer, DCpo, when compared with surfactant molecule is used in the formulation. It
the heat capacity in the gas phase, Cpo. This effect is, therefore, necessary to predict the type of
Surfactants 1269 S
a b
1.2 1
Mole fraction of component 2 in the micelles

CMC2/CMC1=1 0.01
1 0.8
0.1
0.8
CMC/CMC1
0.6
0.6
0.4 CMC2/CMC1=1
0.4
0.1 0.2
0.2
0.01
0 0
0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1
Surfactant composition in solution, k2 Surfactant composition in solution, k2
Surfactants, Fig. 13 Calculated cmc (a) and micellar composition (b) as a function of solution composition for three
ratios of cmcs
possible interactions and whether this leads to The molar composition of the mixed micelle is
some synergistic effects. Two general cases may given by
be considered: surfactant molecules with no net
interaction (with similar head groups) and sys- x1 cmc2
1 ¼
xm (25)
tems with net interaction (Tadros 1984). The first x1 cmc2 þ x2 cmc1
case is that when mixing two surfactants with the
same head group but with different chain lengths. Figure 13 shows the calculated cmc and the
In analogy with the hydrophilic–lipophilic bal- micelle composition as a function of solution
ance (HLB) for surfactant mixtures, one can also composition using Eqs. 24 and 25 for three
assume that the cmc of a surfactant mixture (with cases where cmc2/cmc1 ¼ 1, 0.1, and 0.01. As
no net interaction) to be an average of the two can be seen, the cmc and micellar composition
cmcs of the single components (Tadros 1984): change dramatically with solution composition
when the cmcs of the two surfactants vary con-
cmc ¼ x1 cmc1 þ x2 cmc2 (22) siderably, that is, when the ratio of cmcs is far
from 1. This fact is used when preparing
where x1 and x2 are the mole fractions of the microemulsions where the addition of medium- S
respective surfactants in the system. However, chain alcohol (like pentanol or hexanol) changes
the mole fractions should not be those in the the properties considerably. If component 2
whole system but those inside the micelle. This is much more surface active, that is, cmc2/cmc1
means that Eq. 22 should be modified: 1, and it is present in low concentrations (x2 is
of the order of 0.01), then from Eq. 25 x1m x2m
cmc ¼ xm
1 cmc1 þ x2 cmc2
m
(23) 0.5, that is, at the cmc of the systems, the
micelles are up to 50 % composed of component
The superscript m indicates that the values are 2. This illustrates the role of contaminants in
inside the micelle. If x1 and x2 are the solution surface activity, for example, dodecyl alcohol in
composition, then, sodium dodecyl sulfate (SDS).
Figure 14 shows the cmc as a function of
x1 x2 molar composition of the solution and in the
1cmc ¼ þ (24)
cmc1 cmc2 micelles for a mixture of SDS and nonylphenol
S 1270 Surfactants
Surfactants, Fig. 14 cmc 2.0

as a function of surfactant
composition, x1, or micellar
surfactant composition,
x1m, for the system SDS +
NP-E10 1.5
Using fraction
CMC (mg/ml)
in the micelles
1.0
0.5
Using total
fraction
0.0
0.2 0.4 0.6 0.8 1.0
Surfactant composition in solution, kNP-EO
10
with 10 moles ethylene oxide (NP-E10). If the An expression for the activity coefficients can
molar composition of the micelles is used as the be obtained using the regular solutions theory
x-axis, the cmc is more or less the arithmetic (Tadros 1984):
mean of the cmcs of the two surfactants. If, on
m 2
the other hand, the molar composition in the 1 ¼ ðx2 Þ b
ln f m (27)
solution is used as the x-axis (which at the cmc
m 2
is equal to the total molar concentration), then the 2 ¼ ðx2 Þ b
ln f m (28)
cmc of the mixture shows a dramatic decrease at
low fractions of NP-E10. This decrease is due to where b is an interaction parameter between the
the preferential absorption of NP-E10 in the surfactant molecules in the micelle A positive
micelle. This higher absorption is due to the b value means that there is a net repulsion
higher hydrophobicity of the NP-E10 surfactant between the surfactant molecules in the micelle,
when compared with SDS. whereas a negative b value means a net attraction.
With many industrial formulations, surfac- The cmc of the surfactant mixture and the
tants of different kind are mixed together, for composition x1 are given by the following
example, anionics and nonionics. The nonionic equations:
surfactant molecules shield the repulsion
between the negative head groups in the micelle, 1 x1 x2
¼ m þ m (29)
and hence there will be a net interaction between cmc f 1 cmc1 f 2 cmc2
the two types of molecules. Another example is
the case when anionic and cationic surfactants are x1 f m
2 cmc2
1 ¼
xm (30)
mixed, whereby very strong interaction will take x1 f m
2 cmc2 þ x2 f m
2 cmc1
place between the oppositely charged surfactant
molecules. To account for this interaction, Eq. 25 Figure 15 shows the effect of increasing the b
has to be modified by introducing activity coeffi- parameter on the cmc and micellar composition
cients of the surfactants, f1m and f2m, in the for two surfactants with a cmc ratio of 0.1.
micelle: This figure shows that as b becomes more
negative, the cmc of the mixture decreases.
cmc ¼ xm
1 f 1 cmc1 þ x2 f 2 cmc2
m m m
(26) b values in the region of 2 are typical for
Surfactants 1271 S
a 1 b 1
Mole fraction of component 2 in the micelles

0
0.8 0.8 –2
–10
–20
CMC/CMC1
0.6 0.6
0.4 0.4
0
–2
0.2 0.2
–10
–20
0 0
0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1
Surfactant composition in solution, k2 Surfactant composition in solution, k2
Surfactants, Fig. 15 cmc (a) and micellar composition (b) for various values of b for a system with a cmc ratio cmc2/
cmc1 of 0.1
anionic/nonionic mixtures, whereas values in the a number of reasons. For example, polymers are
region of 10 to 20 are typical of anionic/ added as suspending agents (“thickeners”) to pre-
cationic mixtures. With increasing the negative vent sedimentation or creaming of these systems.
value of b, the mixed micelles tend toward In many other systems, such as in personal care
a mixing ratio of 50:50, which reflects the mutual and cosmetics, water-soluble polymers are added
electrostatic attraction between the surfactant for enhancement of the function of the system, for
molecules. The predicted cmc and micellar com- example, in shampoos, hair sprays, lotions, and
position depends both on the ratio of the cmcs as creams. The interaction between surfactants and
well as the value of b. When the cmcs of the water-soluble polymers results in some synergis-
single surfactants are similar, the predicted tic effects, for example, enhancing the surface
value of the cmc is very sensitive to small varia- activity, stabilizing foams and emulsions, etc. It
tion in b. On the other hand, when the ratio of the is therefore important to study the interaction
cmcs is large, the predicted value of the mixed between surfactants and water-soluble polymers
cmc and the micellar composition are insensitive in a systematic way.
to variation of the b parameter. For mixtures of One of the earliest studies of surfactant/
nonionic and ionic surfactants, the b decreases polymer interaction came from surface tension
with increase in electrolyte concentration. This is measurements. Figure 16 shows some typical S
due to the screening of the electrostatic repulsion results for the effect of addition of polyvinylpyr-
on the addition of electrolyte. With some surfac- rolidone (PVP) on the surface tension g- log
tant mixtures, the b decreases with increase of surfactant concentration C curves of SDS (Robb
temperature, that is, the net attraction decreases 1972). In a system of fixed polymer concentration
with increase of temperature. and varying surfactant concentrations, two criti-
cal concentrations appear, denoted by T1 and T2.
T1 represents the concentration at which interac-
Surfactant–Polymer Interaction tion between the surfactant and polymer first
occurs. This is sometimes termed the critical
Mixtures of surfactants and polymers are very aggregation concentration (CAC), that is, the
common in many industrial formulations. With onset of association of surfactant to the polymer.
many suspension and emulsion systems stabi- Because of this, there is no further increase in
lized with surfactants, polymers are added for surface activity and thus no lowering of surface
S 1272 Surfactants
PVP– concentration
PVP 9.92 – 8.15 × 10–3 mol dm–3

0 g/l
15
Bound surfactant [mol SDS (gm PVP)–1 × 103]

1 g/l
3 g/l
60
10 g/l
γ [dyn/cm]
78.8 – 65.2 × 10–3 mol dm–3

50 10 0.1%+0.1N NaCl
T1
40 T2
5
10–3 10–2 10–1
CT[10–3mol/l]
Region Ι
Surfactants, Fig. 16 g – log C curves for SDS solutions CΙ

0
in the presence of different concentrations of PVP 10–4 10–3 10–2
[SDS]e mol dm–3
tension. T2 represents the concentration at which

Surfactants, Fig. 17 Binding isotherms of a surfactant
the polymer becomes saturated with surfactant. to a water-soluble polymer
Since T1 is generally lower than the cmc of the
surfactant in the absence of polymer, then
“adsorption” or “aggregation” of SDS on or surfactant/polymer solutions: (1) The CAC/cmc
with the polymer is more favorable than normal is only weakly dependent on polymer concentra-
micellization. As the polymer is saturated with tion over wide ranges; (2) CAC/cmc is, to a good
surfactant (i.e., beyond T2), the surfactant mono- approximation, independent of polymer molecu-
mer concentration and the activity start to lar weight down to low values. For very low
increase again, and there is lowering of g until molecular weight, the interaction is weakened;
the monomer concentration reaches the cmc, (3) The plateau binding increases linearly with
after which g remains virtually constant, and nor- polymer concentration; (4) Anionic surfactants
mal surfactant micelles begin to form (Cabane show a marked interaction with most homopoly-
and Duplessix 1982). mers (e.g., PEO and PVP), while cationic surfac-
The above picture is confirmed if the associa- tants show a weaker but still significant
tion of surfactant is directly monitored (e.g., by interaction. Nonionic and zwitterionic surfactants
using surfactant selective electrodes, by equilib- only rarely show a distinct interaction with
rium dialysis, or by some spectroscopic tech- homopolymers.
nique). The binding isotherms are illustrated in A schematic representation of the association
Fig. 17. At low surfactant concentration, there is between surfactants and polymers for a wide
no significant interaction (binding). At the CAC, range of concentration of both components is
a strongly cooperative binding is indicated, and at shown in Fig. 18. It can be seen that at low
higher concentrations, a plateau is reached. Fur- surfactant concentration (region I), there is no
ther increase in surfactant concentration produces significant association at any polymer concentra-
“free” surfactant molecules until the surfactant tion. Above the CAC (region II), association
activity or concentration joins the curve obtained increases up to a surfactant concentration which
in the absence of polymer. The binding isotherms increases linearly with increase in polymer con-
of Fig. 17 show the strong analogy with micelle centration. In region III, association is saturated,
formation and the interpretation of these iso- and the surfactant monomer concentration
therms in terms of a depression of the increases till region IV is reached where there
cmc. Several conclusions could be drawn from is coexistence of surfactant aggregates at the
the experimental binding isotherms of mixed polymer chains and free micelles.
Surfactants 1273 S
Surfactants,
Fig. 18 Association 4×10−3
between surfactant and
homopolymer in different
concentration domains
X2 X′2
PEO concentration y (g/cm3)

3
No association
2
ΙΙ Excess ΙΙΙ Excess

PEO SDS
Ι
0
1 2 3 4 5 6 7 8×10−3
X0
SDS concentration x (g/cm3)
Several factors influence the interaction (6) Polymer molecular weight: a minimum
between surfactant and polymer, and these are molecular weight of 4,000 for PEO and PVP
summarized as follows. (1) Temperature: is required for “complete” interaction.
increase of temperature generally increases the (7) Amount of polymer: the CAC seems to be
CAC, that is, the interaction becomes less favor- insensitive to (or slightly lower) with increase in
able. (2) Addition of electrolyte: this generally polymer concentration. T2 increases linearly with
decreases the CAC, that is, it increases the bind- increase in polymer concentration. (8) Polymer
ing. (3) Surfactant chain length: increase in the structure and hydrophobicity: several uncharged
alkyl chain length decreases the CAC, that is, it polymers such as PEO, PVP, and polyvinyl alco-
increases association. A plot of log CAC versus hol (PVOH) interact with charged surfactants. S
the number of carbon atoms, n, is linear (similar Many other uncharged polymers interact weakly
to the log cmc–n relationship obtained for surfac- with charged surfactants, for example,
tants alone). (4) Surfactant structure: alkyl ben- hydroxyethyl cellulose (HEC), dextran, and poly-
zene sulfonates are similar to SDS, but acrylamide (PAAm). For anionic surfactants, the
introduction of EO groups in the chain weakens following order of increased interaction has been
the interaction. (5) Surfactant classes: weaker listed: PVOH < PEO < MEC (methyl cellulose)
interaction is generally observed with cationics < PVAc (partially hydrolyzed polyvinyl acetate)
when compared to anionics. However, the inter- < PPO PVP. For cationic surfactants, the fol-
action can be promoted by using a strongly lowing order was listed: PVP < PEO < PVOH <
interacting counterion for the cationic (e.g., MEC < PVAc < PPO. The position of PVP can
CNS). Interaction between ethoxylated surfac- be explained by the slight positive charge on the
tants and nonionic polymers is weak. The interchain which causes repulsion with cations and
action is stronger with alkyl phenol ethoxylates. attraction with anionics.
S 1274 Surfactants
NMR data showed that every “bound” surfac- binding sites for surfactant aggregates of average
tant molecule experienced the same environment, size gb, Kb is the intrinsic equilibrium constant
that is, the surfactant molecules might be bound for binding surfactant on the polymer, and Xp is
in micelle-like clusters but with smaller size. the total polymer concentration (mass concentra-
Assuming that each polymer molecules to consist tion is n Xp).
of a number of “effective segments” of mass Ms The polymer–micelle complexation may affect
(minimum molecular weight for interaction to the conformation of the polymer, but this was
occur), then each segment will bind a cluster of assumed not to affect Kb and gb. The relative
n surfactant anions, D, and the binding equilib- magnitudes of Kb, Kf, and gb determine whether
rium may be represented by complexation with the polymer occurs as well as
the critical surfactant concentration exhibited by
P þ n D P Dn
n (31) the system. If kf > kb and gb ¼ gf, then free micelles
occur in preference to complexation. If Kf < Kb
and the equilibrium constant is given by and gb ¼ gf, then micelles bound to polymer occur
first. If Kf < Kb, but gb is much smaller than gf, then
½P Dn
n the free micelles can occur prior to saturation of the
K¼ (32)
½P½D n polymer. A first critical surfactant concentration
(CAC) occurs close to X1 ¼ Kb1. A second crit-
K is obtained from the half saturation condition ical concentration occurs near X1 ¼ Kf1.
Depending on the magnitude of n Xp, one
K ¼ ½D n1=2 (33) may observe only one critical concentration
over a finite range of surfactant concentrations.
By varying n and using the experimental bind- Figure 19 shows the relationship between X1
ing isotherms, one obtains the following values: and Xf for different polymer concentrations
Ms ¼ 1,830 and n ¼ 15. The free energy of (SDS/PEO system) using the following values
binding is given by the expression for Kb, Kf, gf, and gb: Kb ¼ 319, Kf ¼ 120,
gb ¼ 51, and gf ¼ 54.
D Go ¼ RT ln K1=n (34) In the region from 0 to A, the surfactant mol-
ecules remain singly dispersed. In the region from
DGo was found to be 5.07 K cal mol1 which is A to B, polymer-bound micelles occur; X1
close to that for surfactants. increases very little in this region (large size of
Najaragan (1885) introduced a comprehensive polymer-bound micelles). If gb is small (say 10),
thermodynamic treatment of surfactant/polymer then X1 should increase more significantly in this
interaction. The aqueous solution of surfactant region. If n Xp is small, the region AB is confined
and polymer was assumed to contain both free to a narrow surfactant concentration range. IF
micelles and “micelles” bound to the polymer n Xp is very large, the saturation point B may
molecule. The total surfactant concentration, Xt, not be reached. At B, the polymer is saturated
is partitioned into single dispersed surfactant, X1, with surfactant. In the region AC, increase in Xt
surfactant in free micelles, Xf, and surfactant is accompanied by an increase in X1. At C, the
bound as aggregates, Xb: formation of free micelles becomes possible; CD
denotes the surfactant concentration range over
Xt ¼ X1 þ gf ðKf X1 Þgf þ gb n Xp which any further addition of surfactant results in

ðK b X1 Þgb (35) formation of free micelles. The point C depends
on the polymer mass concentration (n Xp).
1 þ ðK b X1 Þgb
The above theoretical predictions were veri-
gf is the average aggregation number of free fied by the results of Guilyani and Wolfram
micelles, Kf is the intrinsic equilibrium constant (1981) using specific ion electrodes. This is illus-
for formation of free micelles, n is the number of trated in Fig. 20.
Surfactants 1275 S
Surfactants, 10
Fig. 19 Variation of X1
D
with Xt for the SDS/PEO
system C
C
8 C
nXp = 10–5 M
nXp = 10–4 M nXp = 10–4 M
6
X1 (mM)
B B B
B
4 A
nXp = 10–3 M
0
4 8 12 16 20 24 28
Xt (mM)
PEO
D 0.25 (g/l)
1.5 0 C
C 0.5
B B B 1.0
4.0
B
1.0 A
X1 (mM)
0.5
S
Surfactants,
Fig. 20 Experimentally
0
measured values of X1 2 4 6 8 10 12
versus Xt for SDS/PEO
system Xt (mM)
The driving force for polymer/surfactant interac- association. For example, aggregation is resisted
tion is the same as that for the process of by the crowding of the ionic head groups at
micellization (see above). As with micelles, the the surface of the micelle. Packing constraints
main driving force is the reduction of hydrocarbon/ also resist association. Molecules that screen the
water contact area of the alkyl chain of the dissolved repulsion between the head groups, for example,
surfactant. A delicate balance between several electrolytes and alcohol, promote association.
forces is responsible for the surfactant/polymer A polymer molecule with hydrophobic and
S 1276 Surfactants
The consequences of the above model are

(1) more favorable free energy of association
(CAC < cmc) and increased ionic dissociation
of the aggregates and (2) an altered environment
of the CH2 groups of the surfactant near the head
group. The micelle sizes are similar with polymer
present and without, and the aggregation num-
bers are typically similar or slightly lower than
those of the micelles forming in the absence of
a polymer. In the presence of a polymer, the
surfactant chemical potential is lowered with
respect to the situation without polymer (Cabane
and Duplessix 1982; Evans and Winnerstrom
1994).
The interaction between surfactants and
hydrophobically modified polymers shows inter-
Surfactants, Fig. 21 Schematic representation of the
esting behavior, resulting from the interaction
topology of surfactant/polymer complexes according to between surfactant micelles and the hydrophobic
Cabane (Istraelachvili 1985) chains of the polymer. Water-soluble polymers
are modified by grafting a low amount of hydro-
phobic groups (of the order of 1 % of the mono-
hydrophilic segments (which is also flexible) can mers reacted in a typical molecule), resulting in
enhance association by ion–dipole association the formation of “associative structures.” These
between the dipole of the hydrophilic groups and molecules are referred to as associative thick-
the ionic head groups of the surfactant. In addition, eners and are used as rheology modifiers in
contact between the hydrophobic segments of the many industrial applications, for example, paints
polymer and the exposed hydrocarbon areas of the and personal care products. An added surfactant
micelles can enhance association. With SDS/PEO will interact strongly with the hydrophobic
and SDS/PVP, the association complexes are groups of the polymer, leading to a strengthened
approximately three monomer units per molecule association between the surfactant molecules and
of aggregated surfactant. the polymer chain. A schematic picture for the
Generally speaking, there are two alternative interaction between SDS and hydrophobically
pictures of mixed surfactant/polymer solutions, modified hydroxyethyl cellulose (HM-HEC) is
one describing the interaction in terms of shown in Fig. 22 which shows the interaction at
a strongly cooperative association or binding of various surfactant concentrations (Goddard).
the surfactant to the polymer chain and one in Initially, the surfactant monomers interact
terms of a micellization of surfactant on or in the with the hydrophobic groups of the HM polymer,
vicinity of the polymer chain. For polymers with and at some surfactant concentration (CAC), the
hydrophobic groups, the binding approach is pre- micelles can cross-link the polymer chains. At
ferred, whereas for hydrophilic homopolymers, higher surfactant concentrations, the micelles
the micelle formation picture is more likely. The which are now abundant will no longer be shared
latter picture has been suggested by Cabane between the polymer chains, that is, the cross-
and Duplessix (1982) who proposed a structure links are broken. These effects are reflected in the
in which the aggregated SDS is surrounded by variation of viscosity with surfactant concentra-
macromolecules in a loopy configuration. tion for HM polymer as is illustrated in Fig. 23.
A schematic picture of this structure that is some- The viscosity of the polymer increases with
times referred to as “pearl–necklace model” is increase of surfactant concentration, reaching
given in Fig. 21. a maximum at an optimum concentration
Surfactants 1277 S
Surfactants,
Fig. 22 Schematic
representation of the A Polymer backbone Hydrophobic group
interaction between
surfactant and HM polymer
Surfactant
Mixed micelle
Surfactant monomer
Surfactant
Free micelles
(maximum cross-links) and then decreases with charge is of special importance in many cosmetic
further increase of surfactant concentration. For formulations, for example, as hair conditioners.
the unmodified polymer, the changes in viscosity As an illustration, the interaction between SDS
are relatively small. and cationically modified cellulosic polymer
The interaction between surfactants and (Polymer JR, Union Carbide) is shown in
polymers with opposite charge (surfactant/ Fig. 24 using surface tension g measurements
polyelectrolyte interaction) involves charge– (Goddard). The g-log C curves for SDS in the
charge interaction. The case of surfactant– presence and absence of the polyelectrolyte are
polymer pairs in which the polymer is a polyion shown in Fig. 24 which also shows the appear-
and the surfactant is also ionic, but of opposite ance of the solutions. At low surfactant
S 1278 Surfactants
in the micellar region. These trends are schemat-

ically illustrated in Fig. 25.
The surfactant/polyelectrolyte interaction has
several consequences on application in hair condi-
tioners. The most important effect is the high
foaming power of the complex. Maximum foaming
occurs in the region of highest precipitation, that is,
h maximum hydrophobization of the polymer. It is
likely that the precipitate stabilize the foam. Direct
determination of the amount of surfactant bound to
the polyelectrolyte chains revealed a number of
interesting features. The binding occurs at very
low surfactant concentration (1/20th of the cmc).
The degree of binding b reached a value of 0.5
(b ¼ 1 corresponds to a bound DS ion for each
Surfactant concentration
ammonium group). b versus SDS concentration
Surfactants, Fig. 23 Viscosity–surfactant concentration curves were identical for polymeric homologues
relationship for HM modified and unmodified polymer with a degree of cationic substitution (CS) >
solutions 0.23. Precipitation occurred when b ¼ 1. The bind-
ing affinity depends on the nature of the polyanion.
Addition of electrolytes increases the steepness of
binding, but the binding occurs at higher surfactant
70
concentration as the electrolyte concentration is
increased. Increasing the alkyl chain length of the
surfactant increases binding, a process that is sim-
60 ilar to micellization. Viscometric measurements
SDS alone
showed a rapid increase in the relative viscosity at
γ (dyne/cm)
a critical surfactant concentration. However, the

50 behavior depends on the type of polyelectrolyte
SDS + 0.1% used. As an illustration, Fig. 26 shows the
polymer
viscosity–SDS concentration curves for two types
40 of cationic polyelectrolyte: JR 400 (cationically
Surfactant/polymer modified cellulosic) and Reten (an acrylamide/
wt.ratio b-methylacryloxytrimethyl ammonium chloride
30 1:8 1:4 1:2 1:1 2:1
Appearance c t p sp c copolymer, ex Hercules).
The difference between the two polyelectro-
–4 –3 –2 –1
LOG surfactant concentration (m/l)
lytes is striking, and it suggests little change in
the conformation of Reten on addition of SDS
Surfactants, Fig. 24 g-log C curves of SDS with and but strong intermolecular association between
without addition of polymer (0.1 % JR 400) – c clear, polymer JR 400 and SDS.
t turbid, p precipitate, sp slight precipitate
concentration, there is a synergistic lowering of Adsorption of Surfactants at the

surface tension, that is, the surfactant/polyelec- Air/Liquid and Liquid/Liquid Interfaces
trolyte complex is more surface active. The low
surface tension is also present in the precipitation Surfactants play a major role in the formulation
zone. At high surfactant concentrations, g of most chemical products. In the first place,
approaches that of the polymer-free surfactant they are used for stabilization of emulsions and
Surfactants 1279 S
Surfactants,
Fig. 25 Schematic
representation of
surfactant/polyelectrolytre
interaction
5000
SDS + 1% polymer JR-400
4000
3000
Relative (to water) viscosity
2000
1000
Precipitation
125 zone
100
75 S
50
C.M.C.
Surfactants,
25 SDS + 1% reten
Fig. 26 Relative viscosity
of 1 % JR and 1 % Reten as
a function of SDS 0.0001 0.001 0.01 0.1 1.0 10.0
concentration SDS concentration %
microemulsions. Secondly, surfactants are added adsorption. The simplest interface is that of the
in emulsifiable concentrates for their spontaneous air/liquid, and in this case, the surfactant will
dispersion on dilution. In all these phenomena, adsorb with the hydrophilic group pointing
the surfactant needs to accumulate at the inter- toward the polar liquid (water), leaving the
face, a process that is generally described as hydrocarbon chain pointing toward the air.
S 1280 Surfactants
This process results in lowering of the surface It is clear from Eq. 36 that surface or interfa-
tension g. Typically, surfactants show a gradual cial tension, that is, the force per unit length
reduction in g till the cmc is reached above which tangentially to the surface measured in units of
the surface tension remains virtually constant. mNm1 is dimensionally equivalent to an
Hydrocarbon surfactants of the ionic, nonionic, energy per unit area measured in mJm2. For
or zwitterionic ionic type lower the surface ten- this reason, it has been stated that the excess
sion to limiting values reaching 30-40 mNm1, surface free energy is identical to the surface
depending on the nature of the surfactant. Lower tension, but this is true only for a single com-
values may be achieved using fluorocarbon sur- ponent system, that is, a pure liquid (where the
factants, typically of the order of 20 mNm1. It is, total adsorption is zero).
therefore, essential to understand the adsorption Gibbs (1928) derived a thermodynamic rela-
and conformation of surfactants at the air/liquid tionship between the surface or interfacial tension
interface. g and the surface excess G (adsorption per unit
With emulsifiable concentrates, emulsions area):
and microemulsion, the surfactant adsorbs at the
oil/water interface, with the hydrophilic head X ns X
dg ¼ i
dmi ¼ Gi dmi (37)
group immersed in the aqueous phase, leaving A
the hydrocarbon chain in the oil phase. Again,
the mechanism of stabilization of emulsions and where Gi ¼ nis/A is the number of moles of
microemulsions depends on the adsorption and component i and adsorbed per unit area that is
orientation of the surfactant molecules at the liq- denoted as the surface excess and mI is the chem-
uid/liquid interface. The surfactant molecule con- ical potential. Equation 37 is the general form for
sists of a small number of units, and it is mostly the Gibbs adsorption isotherm. The simplest case
reversibly adsorbed, allowing one to apply some of this isotherm is a system of two components in
thermodynamic treatments. In this case, it is pos- which the solute denoted by subscript 2 is the
sible to describe the adsorption in terms of vari- surface-active component, that is, it is adsorbed
ous interaction parameters such as chain/surface, at the surface of the solvent denoted by subscript
chain solvent, and surface solvent. Moreover, the 1. For such a case, Eq. 37 may be written as
configuration of the surfactant molecule can be
simply described in terms of these possible dg ¼ Gs1 d m1 þ Gs2 d m2 (38)
interactions.
At constant temperature and composition of If G1 ¼ 0 (no solvent molecules adsorbed at
the interface (i.e., in absence of any adsorption), the interface):
the surface or interfacial tension is given by the
change in the Gibbs free energy Gs with the dg ¼ Gs1;2 d m2 (39)
change in the area of the interface A:
s where G2,1s is the relative adsorption of surfac-
@G tant with respect to solvent. Since
g¼ (36)
@A T;ni
m2 ¼ mo2 þRT ln aL2 (40)
It is obvious from Eq. 36 that for a stable
interface, g should be positive. In other words, or
the free energy should increase if the area of
the interface increases, otherwise the interface d m2 ¼ RTdln aL2 (41)
will become convoluted, increasing the inter-
facial area, until the liquid evaporates (for A/L then,
Case) or the two “immiscible” phases
dissolved in each other (for the L/L case). dg ¼ Gs2;1 RTdln aL2 (42)
Surfactants 1281 S
Surfactants,
Fig. 27 Surface or 72 50
interfacial tension log
C curves
γ/mNm−1
γ/mNm−1
CMC
CMC
log C log C
or From G2, the area per surfactant ion or molecule

can be calculated since

1 dg
Gs2;1 ¼ (43) 1
RT dln aL2 Area=molecule ¼ (44)
G2 Nav
where a2L is the activity of the surfactant in bulk
solution that is equal to C2f2 or x2f2, where C2 is where Nav is the Avogadro’s constant. Determining
the concentration of the surfactant in moles dm3 the area per surfactant molecule is very useful since
and x2 is its mole fraction. it gives information on surfactant orientation at the
Equation 43 allows one to obtain the surface interface. For example, for ionic surfactants such as
excess (abbreviated as G2) from the variation of sodium dodecyl sulfate, the area per surfactant is
surface or interfacial tension with surfactant determined by the area occupied by the alkyl chain
concentration. Note that a2 C2 since in dilute and head group if these molecules lie flat at the
solutions f2 1. This approximation is valid interface, whereas for vertical orientation, the area
since most surfactants have low cmc (usually per surfactant ion is determined by that occupied by
less than 103 mol dm3), but adsorption is com- the charged head group, which at low electrolyte
plete at or just below the cmc. The surface excess concentration will be in the region of 0.40 nm2.
G2 can be calculated from the linear portion of the Such area is larger than the geometrical area
g log C2 curves before the cmc. Such g log occupied by a sulfate group as a result of the lateral
C curves are illustrated in Fig. 27 for the air/water repulsion between the head group. On addition of
and o/w interfaces; [CSAA] denotes the concen- electrolytes, this lateral repulsion is reduced and the S
tration of surface-active agent in bulk solution. It area/surfactant ion for vertical orientation will
can be seen that for the A/W interface, g be lower than 0.4 nm2 (reaching in some case
decreases from the value for water (72 mNm1 0.2 nm2). On the other hand, if the molecules
at 20 C) reaching about 25-30 mNm1 near the lie flat at the interface, the area per surfactant ion
cmc. This is clearly schematic since the will be considerably higher than 0.4 nm2.
actual values depend on the surfactant nature. Another important point can be made from the
For the o/w case, g decreases from a value of g log C curves. At concentration just before the
about 50 mNm1 (for a pure hydrocarbon–water break point, one has the condition of constant
interface) to 1–5 mNm1 near the cmc (again slope, which indicates that saturation adsorption
depending on the nature of the surfactant). has been reached:
As mentioned above, G2 can be calculated
@g
from the slope of the linear position of the curves ¼ constant (45)
@ ln a2 p;T
shown in Fig. 27 just before the cmc is reached.
S 1282 Surfactants
Just above the break point, indifferent electrolyte ions present. For a strong
surfactant electrolyte such as an Na+ R,
@g
¼0 (46)
@ ln a2 1 @g
p;T
G2 ¼ (48)
2RT @ ln a
indicating the constancy of g with log C above the
cmc. Integration of Eq. 46 gives The factor of 2 in Eq. 48 arises because both
surfactant ion and counterion must be adsorbed to
g ¼ constant x ln a2 (47) maintain neutrally, and dg//dlna is twice as large
as for a unionized surfactant. If a non-adsorbed
Since g is constant in this region, then a2 must electrolyte, such as NaCl, is present in large
remain constant. This means that addition of sur- excess, then any increase in concentration of
factant molecules above the cmc must result in Na+R produces a negligible increase in Na+ ion
association to form units (micellar) with low concentration, and therefore dmNa becomes
activity. negligible. Moreover, dmCl is also negligible so
The hydrophilic head group may be unionized, that the Gibbs adsorption equation reduces to
for example, alcohols or poly(ethylene oxide)
alkane or alkyl phenol compounds, weakly ion- 1 @g
G2 ¼ (49)
ized such as carboxylic acids, or strongly ionized RT @ lnCNaR
such as sulfates, sulfonates, and quaternary
ammonium salts. The adsorption of these differ- that is, it becomes identical to that for a nonionic
ent surfactants at the air/water and oil/water inter- surfactant.
faces depends on the nature of the head group.
With nonionic surfactants, repulsion between the
head groups is small, and these surfactants are Adsorption of Surfactants at the
usually strongly adsorbed at the surface of water Solid/Liquid Interface
from very dilute solutions. As mentioned before,
nonionic surfactants have much lower cmc values The use of surfactants (ionic, nonionic, and zwit-
when compared with ionic surfactants with the terionic) to control the stability behavior of sus-
same alkyl chain length. Typically, the cmc is in pensions is of considerable technological
the region of 105 –104 mol dm3. Such non- importance. Surfactants are used in the formula-
ionic surfactants form closely packed adsorbed tion of dyestuffs, paints, paper coatings, agro-
layers at concentrations lower than their cmc chemicals, pharmaceuticals, ceramics, printing
values. The activity coefficient of such surfac- inks, etc. They are particularly robust form of
tants is close to unity and is only slightly affected stabilization which is useful at high disperse vol-
by addition of moderate amounts of electrolytes ume fractions and high electrolyte concentra-
(or change in the pH of the solution). Thus, non- tions, as well as under extreme conditions of
ionic surfactant adsorption is the simplest case high temperature, pressure, and flow. In particu-
since the solutions can be represented by a two lar, surfactants are essential for the stabilization
component system and the adsorption can be of suspensions in nonaqueous media, where elec-
accurately calculated using Eq. 43. With ionic trostatic stabilization is less successful. The key
surfactants, on the other hand, the adsorption to understanding how surfactants function as
process is relatively more complicated since one stabilizers is to know their adsorption and
has to consider the repulsion between the head conformation at the solid/liquid interface. Surfac-
groups and the effect of presence of any indiffer- tant adsorption is relatively simple since the mol-
ent electrolyte. Moreover, the Gibbs adsorption ecules consist of a small number of units and they
equation has to be solved taking into account the mostly are reversibly adsorbed, allowing on to
surfactant ions, the counterion, and any apply thermodynamic treatments. In this case,
Surfactants 1283 S
it is possible to describe the adsorption in terms The rate of desorption is proportional to the
of the various interaction parameters, namely, fraction of surface covered y:
chain–surface, chain–solvent, and surface–
solvent. Moreover, the conformation of the sur- Rate of desorption ¼ kdes y (51)
factant molecules at the interface can be deduced
from these simple interaction parameters. How- At equilibrium, the rate of adsorption is equal
ever, in some cases, these interaction parameters to the rate of desorption, and the ratio of
may involve ill-defined forces, such as hydropho- (kads/kdes) is the equilibrium constant K, that is,
bic bonding, solvation forces, and chemisorption.
In addition, the adsorption of ionic surfactants y C
¼ K (52)
involves electrostatic forces particularly with ð1 yÞ 55:5
polar surfaces containing ionogenic groups. For
that reason, the adsorption of ionic and nonionic The equilibrium constant K is related to the
surfactants will be treated separately. The sur- standard free energy of adsorption by
faces (substrates) can be also hydrophobic or
hydrophilic, and these may be treated separately. D Goads ¼ RT ln K (53)
The adsorption of ionic surfactants on hydro-
phobic surfaces such as carbon black, polymer R is the gas constant, and T is the absolute
surfaces, and ceramics (silicon carbide or silicon temperature. Equation 53 can be written in the
nitride) is governed by hydrophobic interaction form
between the alkyl chain of the surfactant and the or
hydrophobic surface. In this case, electrostatic

interaction will play a relatively smaller role. D Goads
K ¼ exp (54)
However, if the surfactant head group is of the RT
same sign of charge as that on the substrate sur-
face, electrostatic repulsion may oppose adsorp- Combining Eqs. 52 and 54,
tion. In contrast, if the head groups are of opposite
sign to the surface, adsorption may be enhanced.
y C D GO
Since the adsorption depends on the magnitude of ¼ exp ads
(55)
the hydrophobic bonding free energy, the amount 1 y 55:5 RT
of surfactant adsorbed increases directly with
increase in the alkyl chain length in accordance Equation 55 applies only at low surface cov-
with Traube’s rule. The adsorption of ionic sur- erage (y < 0.1) where lateral interaction between
factants on hydrophobic surfaces may be the surfactant ions can be neglected. At high
represented by the Stern–Langmuir isotherm surface coverage (y > 0.1), one should take the S
(Hough and Randall 1983). Consider a substrate lateral interaction between the chains into
containing Ns sites (mol m2) on which G moles account by introducing a constant A, for example,
m2 of surfactant ions are adsorbed. The surface using the Frumkin–Fowler–Guggenheim
coverage y is (G/Ns), and the fraction of uncov- equation:
ered surface is (1 y). The rate of adsorption is
proportional to the surfactant concentration
y C D Goads
expressed in mole fraction, (C/55.5), and the frac- expðAyÞ ¼ exp (56)
ð1 yÞ 55:5 RT
tion of free surface (1 y), that is,

C Various authors (Fuerstenau and Healy
Rate of Adsorption ¼ kads ð1 yÞ (50)
55:5 1972; Somasundaran and Goddard 1979; Healy
1974; Somasundaran and Hannah 1979) have
where kads is the rate constant for adsorption. 7used the Stern–Langmuir equation in a simple
S 1284 Surfactants
form to describe the adsorption of surfactant C1. The suspension is kept stirred for sufficient
ions on mineral surfaces: time to reach equilibrium. The particles are then
removed from the suspension by centrifugation,

D Goads and the equilibrium concentration C2 is deter-
G ¼ 2rCexp (57)
RT mined using a suitable analytical method. The
amount of adsorption G (mole m2) is calculated
Various contributions to the adsorption free as follows:
energy may be envisaged. To a first approxima-
tion, these contributions may be considered to be ðC1 C2 Þ
G¼ (60)
additive. In the first instance, DGads may be taken m As
to consist of two main contributions, that is,
The adsorption isotherm is represented by
D Gads ¼ D Gelec þ D Gspec (58) plotting G versus C2. A range of surfactant con-
centrations should be used to cover the whole
where DGelec accounts for any electrical interac- adsorption process, that is, from the initial values
tions and DGspec is a specific adsorption term low to the plateau values. To obtain accurate
which contains all contributions to the adsorption results, the solid should have a high surface area
free energy that is dependent on the “specific” (usually >1 m2). Several examples may be
(nonelectrical) nature of the system (Fuerstenau quoted from the literature to illustrate the adsorp-
and Healy 1972; Somasundaran and Goddard tion of surfactant ions on solid surfaces. For
1979; Healy 1974; Somasundaran and Hannah a model hydrophobic surface, carbon black has
1979). Several authors subdivided DGspec into been chosen (Greenwood et al. 1968; Day et al.
supposedly separate independent interactions 1967). Figure 28 shows typical results for the
(Fuerstenau and Healy 1972; Somasundaran and adsorption of sodium dodecyl sulfate (SDS) on
Goddard 1979; Healy 1974; Somasundaran and two carbon black surfaces, namely, Spheron 6
Hannah 1979), for example, (untreated) and Graphon (graphitized) which
also describes the effect of surface treatment.
D Gspec ¼ D Gcc þ D Gcs þ D Ghs þ:::::::::: The adsorption of SDS on untreated Spheron 6
(59) tends to show a maximum that is removed on
sive chain–chain interaction between the hydro- at high surfactant concentration. The plateau
phobic moieties of the adsorbed ions, DGcs is the adsorption value is 2 106 mol m2 (2 m
term for chain/substrate interaction, whereas mole m2). This plateau value is reached at 8 m
DGhs is a term for the head group/substrate inter- mole dm3 SDS, that is, close to the cmc of the
action. Several other contributions to DGspec may surfactant in the bulk solution. The area per sur-
be envisaged, for example, ion–dipole, ion– factant ion in this case is 0.7 nm2. Graphitiza-
induced dipole, or dipole–induced dipole tion (Graphon) removes the hydrophilic ionizable
interactions. groups (e.g., –C ¼ O or –COOH), producing
Since there is no rigorous theory that can pre- a surface that is more hydrophobic. The same
dict adsorption isotherms, the most suitable occurs by heating Spheron 6–2,700 C. This
method to investigate adsorption of surfactants leads to a different adsorption isotherm (Fig. 28)
is to determine the adsorption isotherm. Measure- showing a step (inflection point) at a surfactant
ment of surfactant adsorption is fairly straightfor- concentration in the region of 6 m mole dm3.
ward. A known mass m (g) of the particles The first plateau value is 2.3 m mole m2,
(substrate) with known specific surface area As whereas the second plateau value (that occurs at
(m2 g1) is equilibrated at constant temperature the cmc of the surfactant) is 4 m mole m2. It is
with surfactant solution with initial concentration likely in this case that the surfactant ions adopt
Surfactants 1285 S
Surfactants, Graphon in 10−1 mol dm−3 NaCl
Fig. 28 Adsorption 5
isotherms for sodium
dodecyl sulfate on carbon
4
Graphon
Γμmolm−2
3
Spheron 6 after washing

2
1 Spheron 6 before
washing
0
4 8 12 16 20 24
C/mmol dm−3
different orientations at the first and second pla- 10−5 III −40
teaus. In the first plateau region, a more “flat” −30
−20
orientation (alkyl chains adsorbing parallel to the 10−6
Γ −10
Γ mol m−2
surface) is obtained, whereas at the second pla- II 0
ζ/mV
teau, vertical orientation is more favorable, with 10−7 ζ 10
I 20
the polar head groups being directed toward the
10−8 30
solution phase. Addition of electrolyte (101 40
mole dm3 NaCl) enhances the surfactant 50
10−9
adsorption. This increase is due to the reduction 10−5 10−4 10−3 10−2
of lateral repulsion between the sulfate head C/mol dm−3
groups, and this enhances the adsorption.
The adsorption of ionic surfactants on other Surfactants, Fig. 29 Adsorption isotherm for sodium
dodecyl sulfonate on alumina and corresponding zeta (z)
hydrophobic surfaces resembles that for carbon potential
black (Saleeb and Kitchener 1965; Conner and
Ottewill 1971). For example, Saleeb and Kitche-
ner (1965) found similar limiting area for techniques such as microcalorimetry and various
cetyltrimethyl ammonium bromide on Graphon spectroscopic methods to obtain a full picture on S
and polystyrene (0.4 nm2). As with carbon surfactant adsorption.
black, the area per molecule depends on the The adsorption of ionic surfactants on polar
nature and amount of added electrolyte. surfaces that contain ionizable groups may show
This can be accounted for in terms of reduction characteristic features due to additional interac-
of head group repulsion and/or counterion bind- tion between the head group and substrate and/or
ing. Surfactant adsorption close to the cmc may possible chain–chain interaction. This is best
appear Langmuirian, although this does not auto- illustrated by the results of adsorption of sodium
matically imply a simple orientation. For exam- dodecyl sulfonate (SDSe) on alumina at pH ¼ 7.2
ple, rearrangement from horizontal to vertical obtained by Fuersetenau (1971) and shown in
orientation or electrostatic interaction and coun- Fig. 29. At the pH value, the alumina is positively
terion binding may be masked by simple adsorp- charged (the isoelectric point of alumina is at pH
tion isotherms. It is essential, therefore, to 9), and the counterions are Cl from the added
combine the adsorption isotherms with other supporting electrolyte. In Fig. 29, the saturation
S 1286 Surfactants
adsorption G1 is plotted versus equilibrium sur- type is that based on a poly(oxyethylene) glycol
factant concentration C1 in logarithmic scales. group, that is, (CH2CH20)nOH (where n can vary
The figure also shows the results of zeta potential from as little as 2 units to as high as 100 or more
(z) measurements (which are a measure of the units) linked either to an alkyl (CxH2x+1) or alkyl
magnitude sign of charge on the surface). Both phenyl (CxH2x+1–C6H4–) group. These surfac-
adsorption and zeta potential results show three tants may be abbreviated as CxEn or CxjEn
distinct regions. The first region which shows (where C refers to the number of C atoms in the
a gradual increase of adsorption with increase in alkyl chain, j denotes C6H4, and E denotes eth-
concentration, with virtually no change in the ylene oxide). These ethoxylated surfactants are
value of the zeta potential, corresponds to an characterized by a relatively large head group
ion-exchange process (Schick 1966). In other compared to the alkyl chain (when n > 4). How-
words, the surfactant ions simply exchange with ever, there are nonionic surfactants with small
the counterions (Cl) of the supporting electro- head group such as amine oxides (–N ! 0) head
lyte in the electrical double layer. At a critical group, phosphate oxide (–P ! 0), or sulfinyl–
surfactant concentration, the desorption increases alkanol (–SO–(CH2)n–OH) (Schonfeldt 1970).
dramatically with further increase in surfactant Most adsorption isotherms in the literature are
concentration (region II). In this region, the pos- based on the ethoxylated type surfactants.
itive zeta potential gradually decrease, reaching The adsorption isotherm of nonionic surfactants
a zero value (charge neutralization)) after which is in many cases Langmuirian, like those of most
a negative value is obtained which increases rap- other highly surface-active solutes adsorbing
idly with increase in surfactant concentration. from dilute solutions, and adsorption is generally
The rapid increase in region II was explained in reversible. However, several other adsorption
terms of “hemi-micelle formation” that was orig- types are produced (Clunie and Ingram 1983),
inally postulated by Gaudin and Fuerestenau and those are illustrated in Fig. 30. The steps in
(Wakamatsu and Fuerstenau 1968; Gaudin and the isotherm may be explained in terms of
Fuertsenau 1955). In other words, at a critical the various adsorbate–adsorbate, adsorbate–
surfactant concentration (to be denoted the cmc adsorbent, and adsorbate–solvent interactions.
of “hemi-micelle formation” or better the critical These orientations are schematically illustrated
aggregation concentration CAC), the hydropho- in Fig. 31.
bic moieties of the adsorbed surfactant chains are In the first stage of adsorption (denoted by I in
“squeezed out” from the aqueous solution by Figs. 30 and 31), surfactant–surfactant interac-
forming two-dimensional aggregates on the tion is negligible (low coverage), and adsorption
adsorbent surface. This is analogous to the pro- occurs mainly by van der Waals interaction. On
cess of micellization in bulk solution. However, a hydrophobic surface, the interaction is domi-
the CAC is lower than the cmc, indicating that the nated by the hydrophobic portion of the surfac-
substrate promotes surfactant aggregation. At tant molecule. This is mostly the case with
a certain surfactant concentration in the hemi- agrochemicals which have hydrophobic surfaces.
micellization process, the isoelectric point is However, if the chemical is hydrophilic in nature,
exceeded, and, thereafter, the adsorption is the interaction will be dominated by the EO
hindered by the electrostatic repulsion between chain. The approach to monolayer saturation
the hemi-micelles, and hence the slope of the with the molecules lying flat is accompanied by
adsorption isotherm is reduced (region III). a gradual decrease in the slope of the adsorption
The adsorption of nonionic surfactants on isotherm (region II in Fig. 30). Increase in the size
hydrophobic and hydrophilic surfaces depends of the surfactant molecule, for example, increas-
on the surfactant structure and its interaction ing the length of the alkyl or EO chain, will
with the surface. Several types of nonionic sur- decrease adsorption (when expressed in moles
factants exist, depending on the nature of the per unit area). On the other hand, increasing tem-
polar (hydrophilic) group. The most common perature will increase adsorption as a result of
Surfactants 1287 S
Surfactants,
Fig. 30 Adsorption
isotherms corresponding to
the three adsorption V
sequences shown in Fig. 31
Γ
IV
IV
III III
I II I II III
I II
cmc cmc cmc
X2
I extended conformation. The larger the surfactant

alkyl chain, the greater will be the cohesive
II forces, and hence the smaller the cross-sectional
area. This may explain why saturation adsorption
A B C increases with increasing alkyl chain length. The
III interaction occurring in the adsorption layer dur-
ing the fourth and subsequent stages of adsorp-
IV tion is similar to those that occur in bulk solution.
In this case aggregate units, as shown in Fig. 31
Surfactants, Fig. 31 Model for adsorption of nonionic V (hemi-micelles or micelles) may be formed.
surfactants This picture was supported by Kleminko et al.
(1974) who found close agreement between sat-
uration adsorption and adsorption calculated
desolvation of the EO chains, thus reducing their based on the assumption that the surface is cov-
size. Moreover, increasing temperature reduces ered with close-packed hemi-micelles. Kleminco
the solubility of the nonionic surfactant, and this (1978) developed a theoretical model for the
enhances adsorption. The subsequent stages of three stages of adsorption of nonionic surfactants.
adsorption (region III and IV) are determined by In the first stage (flat orientation), a modified
surfactant–surfactant interaction, although sur- Langmuir adsorption equation was used. In the
factant–surface interaction initially determines second stage of horizontal orientation, the surface
adsorption beyond stage II. This interaction concentration increases by an amount that is
depends on the nature of the surface and the determined by the displacement of the ethoxy S
hydrophilic–lipophilic balance of the surfactant chain by the alkyl group. Finally, in the region
molecule (HLB). For a hydrophobic surface, of the hemi-micelle formation, the adsorption can
adsorption occurs via the alkyl group of the sur- be described by a simple Langmuir equation of
factant. For a given EO chain, the adsorption will the form
increase with increase in the alkyl chain length.
On the other hand, for a given alkyl chain length, G2
C2 K
a ¼ infinity
(61)
adsorption increases with decrease of the PEO ðG2 G2 Þ
chain length. As the surfactant concentration
approaches the cmc, there is a tendency for aggre- where G21 is the maximum surface excess, that
gation of the alkyl groups. This will cause vertical is, the surface excess when the surface is covered
orientation of the surfactant molecules (stage IV). with close-packed hemi-micelles, Ka* is
This will compress the head group, and for an EO a constant, that is, inversely proportional to the
chain, this will result in a less coiled more cmc, and C2 is the equilibrium concentration.
S 1288 Surfactants
Toxicological and Environmental expressed as LC50 (for fish) or EC50 (for daphnia
Aspects of Surfactants and algea), where LC and EC stand for lethal and
effective concentration, respectively. Values
A large fraction of dermatological problems in below 1 mg l1 after 96 h testing on fish and
normal working life can be related to exposure of algae and 48 h on daphnia are considered
unprotected skin to surfactant solutions. Several toxic. Environmentally benign surfactants should
formulations contain significant amount of sur- preferable be above 10 mg l1.
factants, for example, cutting fluids, rolling oil Biodegradation is a process carried out by
emulsions, some household cleaning formula- bacteria in nature. By enzymatic reactions,
tions, and some personal care products. Skin irri- a surfactant molecule is ultimately converted to
tation of various degrees of seriousness is carbon dioxide, water, and oxides of the other
common, and in some cases, allergic reactions elements. If the surfactant does not undergo nat-
may also appear. The physiological aspects of ural biodegradation, then it is stable and persists
surfactants on the skin are investigated by various in the environment. For surfactants, the rate of
dermatological laboratories, starting with the sur- biodegradation varies from 1 to 2 h for fatty acids,
face of the skin and progressing via the horny 1–2 days for linear alkyl benzene sulfonates, and
layer and its barrier function to the deeper layer for several months for branched alkyl benzene
of the basal cells. Surfactant classes that are gen- sulfonate. The rate of biodegradation depends
erally known to be mild to the skin include polyol on the surfactant concentration, pH, and temper-
surfactants (alkyl polyglucosides), zwitterionic ature. The temperature effect is particularly
surfactants (betaines, amidobetaines, and important, since the rate can vary by as much
isethionates), and many polymeric surfactants. a factor of five between summer and winter in
Alcohol ethoxylates are relatively mild but not northern Europe. Two criteria are important
as mild as the polyol-based nonionics (the alkyl when testing for biodegradation: (1) primary
polyglucosides). In addition, alcohol ethoxylates degradation that results in loss of surface activity
may undergo oxidation giving by-products and (2) ultimate biodegradation, that is, conver-
(hyperoxides and aldehydes) which are skin sion to carbon dioxide which can be measured
irritants. These classes are commonly used in using closed-bottle tests. The rate of biodegrada-
personal care and cosmetic formulations. For tion depends on the surfactant structure. For
a homologous series of surfactants, there is usu- example, the surfactant must be water soluble.
ally a maximum in skin irritation at a specific Lipophilic amphiphiles such as fluorocarbon sur-
alkyl chain length; maximum irritation usually factants may accumulate in lipid compartments
occurs at C12 chain length. This reflects the max- of the organism and break down very slowly. The
imum in surface activity at this chain length and initial degradation may also lead to intermediates
the reduction in the cmc. Anionic surfactants are with much lower water solubility, and these
generally more skin irritants than nonionics. For degrade very slowly. An example of this is the
example, sodium dodecyl sulfate which is com- alkyl phenol ethoxylates which degrade by oxi-
monly used in toothpaste has a relatively high dative cleavage from the hydroxyl end of the
skin toxicity. In contrast, the ether sulfates polyoxyethylene chain. This leads to a compound
are milder, and they are recommended for use with much smaller EO groups that is very lipo-
in hand dishwashing formulations. Sometimes philic and degrades at a very slow rate. A third
addition of a mild surfactant (such as alkyl important factor in biodegradation is the presence
polyglucoside) can cause large improvement in of cleavable bonds in the alkyl chain which
the dermatological properties. Some amphoteric depends on branching. Extensive branching of
surfactants such as betaines can also reduce the the alkyl chain tends to reduce the rate of biodeg-
skin irritation of anionic surfactants. radation. This is probably due to steric hindrance
Aquatic toxicity is usually measured on fish, preventing close approach of the surfactant mole-
daphnia, and algae. The toxicity index is cule into the active site of the enzyme.
Surfactants 1289 S
Cross-References Healy TW (1974) J Macromol Sci Chem 118:603
Holmberg K, Jonsson B, Kronberg B, Lindman B (2003a)
Surfactants and polymers in solution, 2nd edn. Wiley,
▶ Surfactant Molecule Chichester
Holmberg K, Jonsson B, Kronberg B, Lindman B (2003b)
Surfactants and polymers in aqueous solution,
References 2nd edn. Wiley, Chichester
Hough BH, Randall HM (1983) In: Parfitt GD, Rochester
Adam NK (1925) J Phys Chem 29:87 CH (eds) Adsorption from solution at the solid/liquid
Anainsson EAG, Wall SN (1974) J Phys Chem 78:1024; interface. Academic, London, p 247
79, 857 (1975) Istraelachvili J (1985) Intermolecular and surface forces,
Aniansson EAG, Wall SN, Almagren M, Hoffmann H, with special applications to colloidal and biological
Ulbricht W, Zana R, Lang J, Tondre C (1976) J Phys systems. Academic, London, p 251
Chem 80:905 Jaycock MJ, Ottewill RH (1964) Fourth Int Congr surf Act
Arnow RH, Witten L (1960) J Phys Chem 64:1643 2:545
Barry BW, Gray GMT (1975) J Colloid Interface Sci Jungermana E (1970) Cationic surfactants. Marcel
52(314) Dekker, New York
Buestein BR, Hiliton CL (1982) Amphoteric surfactants. Kahlweit M, Teubner M (1980) Adv Colloid Interface Sci
Marcel Dekker, New York 13:1
Cabane B (1977) J Phys Chem 81:1639 Kleminko NA (1978) Kolloid Zh 40:1105; 1979, 41: 78
Cabane B, Duplessix R (1982) J Phys (paris) 43:1529 Klimenko NA, Tryasorukova AA, Permilouskayan AA
Clunie JS, Ingram BT (1983) In: Parfitt GD, Rochester CH (1974) Kolloid Zh 36:678
(eds) Adsorption from solution at the solid/liquid Lindman B (1984) In: Tadros TF (ed) Surfactants.
interface. Academic, London, p 105 Academic, London
Conner P, Ottewill RH (1971) J Colloid Interface Sci Linfield WM (ed) (1967) Anionic surfactants. Marcel
37:642 Dekker, New York
Day RE, Greenwood FG, Parfitt GD 4th (1967) Int Congr Lucasssen-Reynders EH (1981) Anionic surfactants –
Surf Act Subst 18:1005 physical chemistry of surfactant action. Marcel Dek-
Debye P, Anaker EW (1951) J Phys Colloid Chem 55:644 ker, New York
Desai JD, Banat IM (1997) Microbial production of McBain JW (1913) Trans Faraday Soc 9:99
surfactants and their commercial potential. Microbiol McBain JW (1950) Colloid science. Heath, Boston
Mol Biol Rev 61(47–64) McCutcheon Detergents and emulsifiers. Allied
Drummond CJ, Fong C, Krodkiewska I, Boyd BJ, Publishing, New Jersey (published annually)
Baker IJA (2003) In: Holberg K (ed) Novel surfac- Mukerjee P, Mysels KJ (1971) Critical micelle concentra-
tants. Marcel Dekker, New York, Chapter 3 tions of aqueous surfactant systems. National Bureau
Elworthy PH, Florence AT, Macfarlane CB (1968) of Standards Publication, Washington, DC
Solubilization by surface active agents. Cahpman and Nagaarajan R (1885) Colloids surf 13:1
Hall, London Oakenfull DG, Fisher LR (1977) J Phys Chem 81:1838
Evans DF, Winnerstrom H (1994) The colloidal domain. Oakenfull DG, Fisher LR (1978) J Phys Chem
Where physics, chemistry, biology and technology 82:2443
meet. Wiley INC. VCH, New York, p 312 Okub T, Kitano H, Ishiwatari T, Isem N (1979) Proc R Soc
Fuerestenau DW (1971) In: Hair ML (ed) The chemistry A36:81
of biosurfaces. Marcel Dekker, New York, p 91 Phillips JN (1955) Trans Faraday Soc 51:561
Fuerstenau DW, Healy TW (1972) In: Lemlich R (ed) Porter MR (1994) Handbook of surfactants. Blackie/ S
Adsorptive bubble separation techniques. Academic, Chapman and Hall, New York
London, p 91 Rassing J, Sams PJ, Wyn-Jones E (1974) J Chem Soc,
Gaudin AM, Fuertsenau DW (1955) Trans AIME 202:958 Faraday II 70:1247
Gibbs JW (1928) Collected works, vol 1. Longman, Robb ID (1972) Chem Ind 13:530–535
Harlow Rosen ML (1978) Surfactants and interfacial phenomena.
Gilyani T, Wolfram E (1981) Colloids Surf 3:181 Wiley-Interscience, New York
Goddard ED Colloids Surf 19:301 Rubingh N, Holland PM (eds) (1991) Cationic srfactants –
Greenwood FG, Parfitt GD, Picton NH, Wharton DG physical chemistry. Marcel Dekker, New York
(1968) Adv Chem Ser 79:135 Saleeb FZ, Kitchener JA (1965) J. Chem Soc 911
Guatam KK, Tyagi VK (2006) Microbial surfactants; Schick MJ (ed) (1966) Nonionic surfactants. Marcel
a review. J Oleo Sci 55(155–166) Dekker, New York
Harkins WD, Mattoon WD, Corrin ML (1946) J Am Chem Schick MJ (ed) (1987) Nonionic surfactants: physical
Soc 68:220; J Colloid Sci, 1, 105 (1946) chemistry. Marcel Dekker, New York
Hartley GS (1936) Aqueous solutions of paraffin chain Schonfeldt N (1970) Surface active ethylene oxide
salts. Hermann and Cie, Paris adducts. Pergamon Press, Oxford
S 1290 Surfactants, Phase Behavior
Shinoda K, Nagakawa T, Tamamushi BI, phase behavior; phase diagram; phase equilibria;
Isemura T (1963) Colloidal surfactants, some phase rule; phase transition; polarized light
physicochemical properties. Academic, London
Small DM (1968) Adv Chem Ser 84:31 microscopy; polymer; segregation; self-assem-
Somasundaran P, Goddard ED (1979) Mod Asp bly; small-angle x-ray scattering; surfactant;
Electrochem 13:207 ternary system; vesicle
Somasundaran P, Hannah HS (1979) In: Shah DO,
Schechter RS (eds) Improved oil recovery by surfactant
and polymer flooding. Academic, London, p 205
Stainsby G, Alexander AE (1950) Trans Faraday Soc Definition
46:587
Straathof AJJ, Aldercreutz P (2000) Applied biocatalysis. Phase is any part of a system that is uniform in
Harwood Academic, Amsterdam
Tadros TF (ed) (1984) Surfactants. Academic, London physical state, chemical composition, and inter-
Tadros T (2003) In: Holberg K (ed) Novel surfactants. nal structure. The study of surfactant phase
Marcel Dekker, New York, Chapter 16 behavior consists in the determination of the
Tanford C (1980) The hydrophobic effect, 2nd edn. New number, composition, and structure of phases
York, Wiley
van Os NM, Haak JR, Rupert LAM (1993) Physico- formed by surfactant systems at a given set of
chemical properties of selected anionic, cationic and conditions (pressure, temperature, and system
nonionic surfactants. Elsevier Publishing, Amsterdam composition), in observance of Gibbs phase
von Rybinsky W, Hill K (2003) In: Holberg K (ed) Novel rule. Phase behavior studies rely on the investi-
surfactants. Marcel Dekker, New York, Chapter 2
Wakamatsu T, Fuerstenau DW (1968) Adv Chem Ser gation and analysis of phase diagrams. For binary
71:161 surfactant-water systems, the phase diagram is
Zana R (1978) J Phys Chem 82:2440 a temperature versus composition graph, at con-
Zana RR, Alami E (2003) Gemini surfactants. In: stant pressure. Ternary systems can be of several
Holberg K (ed) Novel surfactants. Marcel Dekker,
New York, Chapter 12 kinds, typically comprising surfactant-water-oil,
surfactant 1-surfactant 2-water, and surfactant-
polymer-water systems. For these systems, the
phase diagram is usually an equilateral triangle
Surfactants, Phase Behavior where the concentrations of the three components
are indicated, at constant temperature and pres-
Eduardo F. Marques1 and Bruno F. B. Silva2 sure. Surfactant phases are essentially of one of
1
Department of Chemistry and Biochemistry, the following types: solution phases (e.g., micel-
Faculty of Science, University of Porto, lar solutions and microemulsions), liquid-
Porto, Portugal crystalline phases (e.g., cubic, hexagonal, and
2
Physical Chemistry, Centre for Chemistry and lamellar phases), and crystalline phases. In
Chemical Engineering, Lund University, binary- and ternary-phase diagrams, single-
Lund, Sweden phase regions are separated by appropriate
multiphase regions, where two or three phases
coexist in equilibrium.
Synonyms
Bulk behavior of surfactants; Phase behavior of Overview

surfactants; Phase science of surfactants
Alongside fundamental and theoretical develop-
ments, colloid and interface science has always
Keywords been deeply involved with solving practical prob-
lems. From oil recovery to household-cleaning
Association; bilayer; binary system; deuterium products and from foodstuffs to information stor-
splitting NMR; differential scanning calorimetry; age, colloidal and interfacial systems have had
liquid crystal; micelle; microemulsion; phase; a continuous impact in everyday life and
Surfactants, Phase Behavior 1291 S
technology. Within this field, surfactants have we will then cover polymer-surfactant systems in
played a pivotal role owing to their unique molec- the more general aspects of their phase behavior.
ular structure (amphiphilicity), which allows Finally, a section will be devoted to the most
them the capacity to disperse, solubilize, mix, relevant experimental methods and techniques
compartmentalize, encapsulate, and thicken, employed in the investigation of surfactant
among others. phase diagrams.
Therefore, the physicochemical description We note that a detailed review of the structure
and understanding of how surfactants behave in and properties of the various surfactant phases is
solution, both macroscopically and at nano- and given in the entry “▶ Surfactant Self-assembly,”
microscale lengths, is important not only for the to which the reader is referred, as complement to
general advance of scientific knowledge but also the contents of this entry.
as a necessary step for the development of tech-
nical applications. Moreover, because polar
lipids and other relevant biomolecules are also Surfactant Self-organization
amphiphilic and take part in many structures and
processes of living cells, surfactant phase science Driving Forces
has direct relevance to biological, biomedical, The amphiphilic, or dual solubility/polarity,
and pharmaceutical fields. character of surfactant molecules implies that
Phase diagrams are the most common and surfactants can adsorb at any polar-apolar inter-
useful way to condense and represent the phase face and hence reduce interfacial tension, that is,
behavior of surfactants, for binary surfactant- the amount of reversible work required to
solvent systems and for more complex increase the area of the interface, at constant
multicomponent mixtures (e.g., ternary surfac- temperature and pressure. Furthermore, the
tant-water-oil systems). Hence, these diagrams amphiphilic structure means that they can also
are an important end result in the physicochemi- self-organize in bulk media into aggregates
cal characterization of surfactants. Conversely, containing many molecules, of discrete or virtu-
they help us to gain insight into colloidal interac- ally unlimited size, in a process designated by
tions and structure-function relationships, surfactant self-assembly.
providing as well important information and The thermodynamic driving force for both
clues for practical applications. adsorption and self-assembly in water is the
In this entry, we will start by describing some same and is the so-called hydrophobic effect,
basic solution properties of surfactants, followed the large Gibbs energy decrease associated with
by a description of the general principles behind the sequestering of surfactant chains from the
phase equilibria and phase diagrams. Because in contact with water (Tanford 1980). The hydro-
most cases the solvent or dispersing medium is phobic effect is characterized by a subtle inter- S
water or water-based, emphasis will be put on play between a large positive entropic term and
aqueous systems, but one should also bear in an enthalpic term whose sign is temperature-
mind that surfactant molecules self-associate dependent (Evans and Wennerström 1999;
in other polar solvents. The typical sequences of Shinoda 1974). However, the hydrophobic effect
phases formed by amphiphiles in aqueous per se does not explain the fact that surfactants
systems, as composition or temperature is varied, associate into particular types of aggregates, as
will be discussed. We will then proceed to the one must also take into account headgroup-
main phase behavior features of representative headgroup and chain-chain interactions in the
binary surfactant-water systems (belonging to aggregates.
the most common classes of surfactants). Ternary The particular shape and size of
and other more complex multicomponent sys- the self-assembled aggregate formed are the
tems involving surfactants will be dealt with in result of a Gibbs energy minimization
the following section. Owing to their importance, (Israelachvili 1985). This Gibbs energy involved
with the self-association from unimers to a given concentration and liquid-crystalline phases at
equilibrium aggregate can be split into different higher concentration. Moreover, at the
contributions: nanoscopic level, the different structures can be
• A transfer term, related to the removal of the generally divided into normal ones or reverse
chains from water into the liquid-like apolar ones, depending, respectively, on whether the
core of the aggregate. surfactant film adopts a positive curvature (curv-
• Headgroup repulsion terms, associated with ing away from water) or a negative curvature
electrostatic and/or steric repulsions between (curving toward water).
the hydrated headgroups. Solution phases encompass micellar solutions,
• A packing term, associated to the packing vesicle solutions, and microemulsions.
constraints of the chains in the hydrophobic In micellar solutions, the aggregates may be nor-
core. mal micelles or reverse micelles, and the same
• An interfacial term, related to the minimiza- applies to discrete-type microemulsions, which
tion of oil-water interfacial contact in the contain oil-in-water nanodroplets or water-in-oil
aggregate surface. nanodroplets. In bicontinuous microemulsions,
• An entropy of mixing term. water and oil are dispersed continuously in one
It is the contact of the headgroups with water another, without a defined shape, with the amphi-
and the repulsions between them that ensures the phile lying at the interface with zero mean curva-
formation of aggregates and not a macroscopic ture. Vesicles are self-folded bilayer aggregates
phase separation. In the vast majority of cases, that enclose an aqueous interior. They can be
the solvent is water or water-based, but obtained in the form of solutions or dispersions
self-assembly can also take place in other bulk resulting from the breakup of a lamellar phase in
polar solvents, such as formamide and ethylene excess solvent.
glycol (Jönsson et al. 1998). Liquid-crystalline phases possess some char-
acteristic long-range periodicity, being interme-
Self-assembled Structures diate in molecular ordering between liquids and
The structure and dynamics of surfactant aggre- crystals. They can be structurally anisotropic,
gates vary considerably, as well as the phase such as the hexagonal phase, the lamellar phase,
behavior from system to system. One can never- and the intermediate phases, or isotropic, such as
theless extract some general principles when it cubic phases (Tiddy 1980). Reverse phases
comes to describing surfactant aggregation appear most commonly in ternary mixtures
(Jönsson et al. 1998; Laughlin 1994): containing water, surfactant, and a third compo-
• Monomolecular films and bilayers constitute nent, usually oil or a cosurfactant (short-chain
the two basic types of aggregation state. alcohol). They are characterized by having the
• Either discrete aggregates or unlimited assem- hydrophobic chains in contact with oil (the con-
blies form the building blocks of surfactant tinuous phase) and the polar headgroups
phases. surrounding discrete aqueous domains. Thus,
• Infinite assemblies can have a different type of reverse micelles, reverse hexagonal, and reverse
continuity: one-dimensional, such as the cubic phases can be formed. Reverse vesicles can
hexagonal liquid-crystalline phase; two- also be formed in certain surfactant and lipid
dimensional, such as the lamellar liquid- systems, having inner and outer oleic domains
crystalline phase; or three-dimensional, such separated by a reverse surfactant bilayer, which
as bicontinuous cubic phases, sponge phases, may or may not contain water.
and bicontinuous microemulsions. Less common phases can also be found in
For most surfactant-water systems, at surfactant binary and ternary systems, among
a temperature above the Krafft boundary in the which are the sponge phase and intermediate
phase diagram, the hydrated crystallites dis- phases (Seddon and Templer 1995; Tiddy
solve to yield solution phases at low 1980). The sponge phase is an isotropic liquid
Surfactants, Phase Behavior, Table 1 Different sur- Phase Science of Surfactants
factant homogeneous and heterogeneous systems
Homogeneous (one-phase) systems Phase science deals at first with the influence that
Liquid-crystalline Solid temperature, pressure, and composition exert on
Solution phases phases phases the number of phases that exist for a given system
Micelles (normal or Hexagonal (normal or Crystalline of choice. It then proceeds to the compositions
reverse) reverse)
and internal structure of those phases. The phase
Microemulsions Lamellar Gel
science of surfactants deals exactly with the same
Vesicles Cubic (micellar or
bicontinuous) issues but now applied to systems of varying
Heterogeneous (multiphase) systems complexity containing amphiphilic molecules
Emulsions (Evans and Wennerström 1999; Laughlin 1994;
Suspensions gels Tadros 2005).
Vesicles
Foams Phase Diagrams
Gels Phase diagrams of surfactant systems are invalu-
able tools in colloid science and technology.
They are used for a number of important reasons:
to gather and summarize information regarding
phase consisting of a surfactant bilayer forming the phase behavior of surfactants, to infer struc-
a connected 3-D network in a solvent; both water ture-property-function relationships, to study
and hydrophobic regions are connected over colloidal interactions, to predict new forms of
macroscopic distances, thus forming behavior, and, in industry, to conceive new
a bicontinuous structure. Intermediate phases technical formulations of interest.
occur between the hexagonal and the lamellar According to Gibbs’s definition, phase is any
phases in some amphiphilic systems and fall portion of a system that is physically uniform
into three broad classes: ribbon, mesh, or (i.e., all constituent parts are in the same physi-
bicontinuous phases. cal state) and whose chemical composition and
The surfactant self-assembled structures so far structure are homogeneous. In the previous sec-
described all fall in the category of single (homo- tion, we have seen the different types of phases
geneous) phases (Table 1). In addition, surfactant that surfactants can form, namely, solutions, liq-
molecules can self-assemble in heterogeneous or uid crystals, or solid phases. Phase diagrams
multiphasic systems, which are particularly rele- consist of graphics depicting the regions of ther-
vant for practical applications – examples of modynamic stability of these different phases in
these are dispersions, suspensions, emulsions, terms of the intensive variables pressure, tem-
and foams. perature, and composition. A phase diagram S
Mainly because of historical reasons, there is should then:
not a standard notation for surfactant phases in • Show the existence range of single-phase
the colloid and interface literature. The systems regions and give information on their structure.
of phase nomenclature most commonly adopted • Indicate multiphase regions (typically
are listed in Table 2, taking the names of two-and three-phase equilibria regions), in
their original proponents, namely, V. Luzzati, observance to Gibbs phase rule.
P. A. Winsor, and P. Ekwall (Laughlin 1994). • Be able to inform on how the variation of
Often, they are used in intermixed way. Through- a parameter, such as temperature or composi-
out this entry, in different figures containing tion, affects the number and structure of the
phase diagrams, different types of notations will phases at equilibrium.
be apparent (reflecting the authors’ original Phase diagrams for one-component systems,
choice), so it is important to keep track of their or unary-phase diagrams, are temperature-
meaning through Table 2. pressure diagrams, since composition is constant.
Surfactants, Phase Behavior, Table 2 Different systems of notation for surfactant phases
Phase Luzzati’s Winsor’s Ekwall’s Others
Normal micellar solution – S1 L1 M, mic
Reverse micellar solution – S2 L2 Rev mic
Normal hexagonal liquid crystal H1 M1 E Hex
Reverse hexagonal liquid crystal H2 M2 F Rev hex
Lamellar liquid crystal La G D Lam
Cubic (normal micellar) liquid crystal Q V1 I10 I, cub (for all)
Cubic (reverse micellar) liquid crystal Q V2 I20
Bicontinuous cubic (normal) liquid crystal F1b, S1c I100
Bicontinuous cubic (reverse) liquid crystal – I200
Vesicular – – L4, V, ves
Sponge – – L3
Gel Lb, Lb0 , Pb0 – – Gel
A A
0.5
Surfactants, Phase 0.5
Behavior,
Fig. 1 Triangular-phase T T1
diagram used for
three-component systems
(left), at constant pressure
and temperature, and
B 0.5 C B C
a phase prism (right) used
for temperature variation Ternary phase diagram Prismatic phase diagram
Phase diagrams for two-component (binary) without altering the number and composition of
systems, such as surfactant-water systems, typi- phases at equilibrium), C is the number of com-
cally consist of a temperature versus composition ponents of the system, and P is the number of
diagram, because pressure is kept constant (as it phases at equilibrium.
has little effect on condensed phases), often at Phase boundaries are lines that delimit the
p 1 atm. For three-component (ternary) sys- existence range of one-phase regions. Tie lines
tems, such as surfactant-oil-water mixtures, are also important, as they belong to two-phase
phase diagrams consist of equilateral triangles, regions and connect two points, each in
where the composition of the system is varied at a one-phase region, with a given composition.
constant temperature and pressure (Fig. 1). In all For instance, in a binary surfactant-water system
these phase diagrams, for thermodynamic (C ¼ 2), at constant pressure (one degree of
reasons, the phase rule has to be followed. freedom used, F0 ¼ F1), the number of degrees
The phase rule is given by the expression of freedom available for a two-phase region is 1.
That means that in the phase diagram – usually
F¼CPþ2 (1) temperature versus surfactant concentration in
weight percentage – the tie lines are horizontal,
where F is the number of degrees of freedom since F0 ¼ 1, and thus, temperature cannot be
(i.e., variables that can be changed independently varied; only composition (surfactant weight
percentage) can. In tie lines, the proportions somewhere in the range 0.5–1 but closer to unity.
of the two coexisting phases are given by the If the surfactant has the shape of a truncated
lever rule. inverted cone, then it tends to form reverse struc-
In triangular phase diagrams, the composi- tures for which PS > 1.
tions of the three components are commonly Another model, with more physical insight,
expressed in weight or molar percentage. In which accounts for and predicts surfactant
these diagrams, three-phase regions can be self-assembling patterns is the flexible surface
found but they are necessarily triangles. If tem- model (Helfrich 1973; Safran 1999). It considers
perature is varied, a 3-D-phase diagram is the surfactant film to be a surface with an intrinsic
required, and this can be done with a triangular interfacial tension and elastic properties, that is,
prism. Such diagrams are often found, for exam- capable of bending and being laterally com-
ple, in the study of microemulsion systems. pressed or expanded. The bending properties are
characterized by the curvature bending constants.
Sequence of Surfactant Phases A mean curvature H at a point on the surface
A convenient way to rationalize the sequence of covered with surfactant molecules is defined by
phases formed by surfactants as a function of
a given variable, and consequently the main H ¼ 1=2ðH1 þ H2 Þ ¼ 1=2ð1=R1 þ 1=R2 Þ (3)
features of phase behavior, is through the
concepts of surfactant packing parameter and where R1 and R2 are the principle radii of curva-
spontaneous film curvature (within the flexible ture, that is, the radii of curvature in two perpen-
surface model). dicular directions, and H1 ¼ 1/R1 and H2 ¼ 1/R2
The surfactant packing parameter PS relates are the principle curvatures. By convention, if the
the geometric shape of a surfactant molecule with film bends away from water, it has a positive
the preferred type of aggregate formed curvature, while if it bends toward water, it has
(Israelachvili et al. 1976). The PS is a negative curvature. A spontaneous mean curva-
a dimensionless quantity defined by the ture, Ho, can be assigned with the surfactant film,
expression defined by the mean curvature that the film
adopts without any mechanical constraints, that
Ps ¼ Vhc = ahg lhc (2) is, the preferred curvature. If the surface bends
away from Ho, there is a Gibbs energy cost – the
where Vhc is the volume of the hydrocarbon curvature energy. The central quantity in this
chain, ahg is the optimal area occupied by the model is thus the curvature Gibbs energy per
headgroup in the polar/apolar interface, and lhc unit area, defined up to second order in H, by
is the length of the fully extended chain. Alterna- the expression
tively, PS can be seen as the ratio between ahc, the S
cross-sectional area of the hydrocarbon part, and c K
gc ¼ 2kc ðH Ho Þ2 þ k (4)
that of the headgroup. PS does not contain any
actual molecular sizes but rather average quanti- where kc is the mean curvature (or normal
ties that relate to the presence of the surfactant in bending) modulus, k c is the Gaussian curvature
a particular aggregate. If the amphiphile has the (or saddle splay) modulus, and K is the Gaussian
shape of a cone, then it tends to form spherical curvature (K ¼ H1H2).
micelles, for which the value of PS should be A connection between the two models
around 1/3. If the shape is more similar to described above can be established. The PS
a truncated cone, it forms cylindrical micelles defines a preferred aggregate structure, and this
since for these PS 1/3–1/2. For a cylinder- one is the aggregate for which the spontaneous
shaped amphiphile, the most favorable aggregate curvature Ho is adopted by the film. In Fig. 2
is the bilayer, where PS 1. In disk-shaped a schematic view of the relation between Ho and
micelles and vesicles, the values of PS should be PS for different surfactant geometries is shown.
Surfactants, Phase R1= R2 = R R1= R2 = −R

Behavior, Fig. 2 Relation R2 = ∞ R1=R2 = ∞
between spontaneous R
curvature, Ho, and R R
surfactant packing
parameter, Ps
Ho =1/R Ho =1/(2R) Ho = 0 Ho = −1/R

Ps =1/3 1/3<Ps <1/2 Ps =1 Ps >1
Surfactants, Phase Behavior, Fig. 3 The idealized phase sequence. The gray bars represent regions where different
cubic phases may occur
In a binary surfactant-water system, as the In this sequence, cubic phases may occur at
amphiphile concentration increases in the sys- any point, after the micellar phase. Intermediate
tem, the value of PS increases, and thus the pre- phases are much less common, forming between
ferred aggregate shape changes. Alternatively, it the normal hexagonal and lamellar phases.
can be stated that the spontaneous curvature of It is important to recognize that this phase
the surfactant films decreases with concentration, sequence is only ideal and a given system does
from positive (normal phases) via zero (planar not usually show the complete sequence of
bilayers) to negative values (reverse structures). structures. Single-chained micelle-forming
In the case of ionic surfactants, for instance, the surfactants present richer phase behavior than
increase in PS is caused by counterion condensa- double-chained bilayer-forming surfactants, as
tion on the aggregate surface that reduces the the latter have a higher starting value of PS. The
headgroup area. Similar trends are observed for importance of Fontell scheme lies in that it pro-
some nonionic surfactants with oxyethylene vides a way of predicting and rationalizing the
groups in water when temperature is increased, phase behavior when a given parameter changes,
as a result of headgroup dehydration. and this can be done by tracking the effect caused
Interaggregate repulsions and available volume on PS (or Ho) by a change in that parameter.
for aggregate packing also become critical, and
hence, phase transitions ultimately take place. An Main Types of Surfactant-Phase Behavior
ideal (or idealized) sequence of preferred phases, From a phase-behavior point of view, surfactants
pictorially depicted in Fig. 3, can thus be envis- can be broadly and fundamentally divided as
aged, known as Fontell scheme (Fontell 1990). It follows:
basically consists of the following sequence: • Soluble or micelle-forming surfactants, those
• Micellar solution ! hexagonal phase ! having PS around 1/3–1/2.
lamellar phase ! reverse phases (hexagonal • Insoluble (swelling) or bilayer-forming
and micellar). surfactants, those having PS around 1.
For soluble surfactants, two cases can happen
as follows:
• As the concentration increases, there is
a smooth variation of the physicochemical
properties of the micellar solution until satu-
ration, indicating that the micelles do not grow
much and remain essentially small and
spheroidal.
• As the concentration increases, there are dra-
matic changes in properties of the micellar
solution, indicating strong micellar growth.
For the case of insoluble surfactants, the satu-
ration concentration is reached at very low con-
centration, and often this solution is basically
a unimer solution, without formation of any
type of micelles. Surfactants, Phase Behavior, Fig. 4 Phase diagram of
a surfactant-water system near the cmc
In all cases, when saturation is reached, phase
separation takes place. The new phase formed for
insoluble or soluble surfactants can be
(1) a liquid-crystalline phase (the most common As the surfactant concentration changes
case), (2) a hydrated crystalline phase of the within the micellar phase limits, changes can
surfactant, and (3) a more concentrated surfactant happen in micellar size (Jönsson et al. 1998).
solution. Short-chain surfactants (up to 8–10 carbon
chains) do not show significant micellar growth
Micelle Formation and Growth up to phase separation (occurring somewhere
Micelles are the simplest aggregates that surfac- between 10 and 40 surfactant wt%), meaning
tants can form. A micellar phase is, from that micelles remain small and spheroidal. Inter-
a thermodynamic point of view, a solution mediate chains only show micelle deformation at
phase, even though its properties are rather dif- relatively high volume fraction, of the order of
ferent from a simple molecular solution; in other 0.3. Longer chain surfactants, typically C14 and
words, it is a colloidal solution (Lindman and higher, show considerable deformation and
Wennerström 1980). Figure 4 depicts the phase growth already at low or moderate concentration,
diagram of a surfactant-water system at high first to short prolates or cylinders and then to long
dilution (Shinoda 1974) in the vicinity of the cylinders or threadlike (also named wormlike)
critical micelle concentration (cmc). micelles. These micelles have a circular cross S
The solubility-phase boundary separating the section equal to roughly twice the length of the
unimer/micellar solution from the two-phase fully stretched alkyl-chain length of the surfac-
region crystals + solution is designated as the tant, but the long length can vary from a few tens
Krafft boundary. The Krafft point in turn is of nanometers to several hundreds. In less
defined as the temperature at which the surfactant encountered cases, the micelles can grow from
concentration reaches the cmc. Above the Krafft small to long threadlike micelles already at very
point, the line separating the unimer from the low surfactant concentration.
micellar solution is the cmc curve. The cmc of The most important factors affecting micellar
a surfactant has a weak dependence on tempera- growth are: (1) salt concentration, for the case of
ture. For nonionic surfactants, it increases slightly ionic surfactants, with observation of an increase
with temperature, while for ionic surfactants, it in micellar size with increasing ionic strength;
often shows a nonmonotonic parabolic depen- (2) temperature, with micelles growing upon
dence with a minimum near room temperature. heating for nonionic surfactants and upon
temperature decrease for ionic surfactants; model, as described above and in light of Fontell
(3) addition of nonpolar solubilisates, which can scheme. The PS model is more insightful for
have opposing effects depending on the polarity dilute systems where aggregates do not interact.
of the molecule; and (4) counterion type, with, for The addition of salt to a micellar phase, for exam-
instance, organic counterions promoting micellar ple, results in a drastic increase in PS and may
growth in long-chain ionic surfactants, even at consequently induce a transition to a lamellar
low surfactant concentration (e.g., salicylate in phase. The spontaneous curvature Ho of the sur-
cetyltrimethylammonium bromide solutions). factant film is changed from a high positive value
to a value in the vicinity of zero due to the salt
Bilayer Formation effect.
For bilayer-forming surfactants, the most com- The main factors that can induce phase
mon case is that once saturation is reached, changes in surfactant systems are as follows:
a two-phase region is entered, where the solution • Total surfactant concentration.
phase is in equilibrium with a lamellar liquid- • Salt, for ionic surfactants.
crystalline phase. Because of the presence of • Temperature, for nonionic surfactants.
such an excess of water, the lamellar phase is • Addition of organic cosolutes, such as oil,
often dispersed in the form of discrete bilayer other nonpolar solubilisates, or a cosurfactant
aggregates, that is, vesicles (Lasic 1993). (e.g., a short chain alcohol).
Depending on surfactant structure, on surfactant • Mixing molar ratio between surfactants in
concentration, and on the history of the sample surfactant mixtures.
(namely, the dispersion procedure), these vesi- The type of change induced largely depends
cles can be more or less polydisperse in size on the surfactant structure or on the composition
(ranging from tens of nm to mm) and unilamellar, of the system.
oligolamellar, or multilamellar. The single- In the following sections, we will illustrate and
lamellar phase is built up of infinite stacks of discuss the main features of phase behavior for
parallel bilayers, but at lower concentrations different classes of surfactants, on the basis of the
and/or depending on mechanical shear applied, PS and spontaneous curvature concepts, and the
it may consist of giant lamellar “droplets” involv- effect of interactions at high concentrations.
ing many concentric bilayers (often called
onionsomes). Similar to the Krafft phenomenon
for micelles, bilayers present a characteristic tem- Binary Ionic Surfactant-Water Systems
perature, termed chain-melting temperature (Tm)
or gel-to-liquid crystal transition temperature Ionic surfactants represent the largest and most
(Evans and Wennerström 1999). At this temper- important class of amphiphiles, and their phase
ature, they undergo a first-order phase transition behavior has been studied over a longer period
from a solid-like state, with restricted rotational than that of nonionic surfactants. For these sur-
and translational molecules motions and chain factants, long-range electrostatic interactions
conformational order, to a liquid-crystalline play a vital role in determining the lyotropic
state, with molecular in-plane translational behavior. Figure 5 shows examples of ionic sur-
mobility and chain conformational disorder factants, whose phase behavior will be discussed
(Nagle and Tristram-Nagle 2000). in the next sections.
Modification of Phase Behavior Single-Chained Surfactants

The effect of surfactant concentration and other The cationic surfactants dodecyltrimethy-
changeable parameters (addition of salt for ionic lammonium chloride (DTAC) and cetyltrimethy-
surfactants and temperature for nonionic surfac- lammonium bromide (CTAB), and the anionic
tants) on phase behavior can be interpreted both surfactant sodium dodecylsulfate (SDS), will be
within the PS model and the flexible surface used to illustrate phase-behavior features that are
Cl−
N+
DTAC
Br− o
N+ o
o
CTAB o s oo
o Na+
+ −o AOT
o s o− Na
o
o SDS HO
N −
SO3 Na
+
−
H
+
N Br
DDAB HO
STDC
Surfactants, Phase Behavior, Fig. 5 Examples of sodium dodecylsulfate, SDS; didodecyldimethy-

ionic surfactants: dodecyltrimethylammonium chloride, lammonium bromide, DDAB; sodium bis(2-ethylhexyl)
DTAC; hexadecyltrimethylammonium bromide, CTAB; sulfosuccinate, AOT; sodium taurodeoxycholate, STDC
a b 100
200
80
150 Lα
V1
60 L1 H1
T/°C
T/°C
Lα
100 V1 + X
40 H1 + X
50 L1 l1 V1 20 l1 + X
H1
(mic)
0 0
0 20 40 60 80 100 0 20 40 60 80 100
DTAC wt% CTAB wt%
Surfactants, Phase Behavior, Fig. 6 Phase diagrams for (a) DTAC-water system (Adapted from Balmbra et al.
1969); (b) CTAB-water system (Adapted from Laughlin 1994)
S
common for many other ionic single-chained H1, forms. Subsequently, a second-cubic phase
surfactants (see Fig. 5). forms, now of the bicontinuous type, V1, preced-
DTAC is a surfactant with a very low Krafft ing a lamellar phase, La. Further increase in con-
point. At room temperature, a solution L1, centration leads to the solid hydrated crystals.
consisting of spherical micelles, exists until very At higher temperatures the cubic phase I1 is
high concentrations (ca 40 wt%), as can be seen the first to vanish (ca. 85 C), while the remaining
in Fig. 6a. Near this point, micelles deform phases are more stable with temperature. The
slightly into prolates and organize into hexagonal vanishes at ca. 185 C and the lamellar
a discrete cubic phase, I1. Between both phases at ca. 205 C. When all the liquid-crystalline
there is a small two-phase region. When the con- phases vanish, an isotropic region is obtained,
centration increases, the packing of discrete but major changes are expected to occur as com-
micelles is progressively more difficult and position changes. The phase boundaries between
beyond ca 53 wt%, a hexagonal phase, the different liquid-crystalline phases are nearly
Surfactants, Phase
Behavior, Fig. 7 Phase
diagram for the SDS-water
system (Adapted from
Kekicheff et al. 1989)
independent of temperature. This is hexagonal and lamellar phases, a number of less

a consequence of the fact that these phases are common intermediate liquid-crystalline phases
stabilized by electrostatic effects. are formed.
If one increases the surfactant chain by four
methylenes and replaces the counterion with bro- Double-Chained Surfactants
mide, in CTAB, some changes are brought about Because of their hydrophobicity and PS value,
in the phase behavior (Fig. 6b). One of the main double-chained surfactants are poorly soluble in
differences for CTAB is the higher Krafft point water, tending to swell in and give rise to bilayer
(about 27–28 C), which turns the surfactant structures, namely, lamellar phases. This will be
insoluble at room temperature. At higher temper- illustrated by sodium bis(2-ethylhexyl)
atures, isotropic and liquid-crystalline phases sulfosuccinate, also known as aerosol OT or
form. Unlike DTAC, the L1 phase shrinks in AOT, and didodecyldimethylammonium
stability range, extending only until ca. 25 wt%. bromide, DDAB (cf. Fig. 5, for their molecular
The discrete cubic phase I1 is now absent, and the structures).
hexagonal phase forms immediately after L1. The AOT is an anionic double-chained surfactant
L1 phase also differs structurally from that of of common usage, with a relatively short-
DTAC. Globular micelles exist only until ca. branched apolar region. At room temperature
14 wt%, and at this concentration the micelles and very high dilution, it forms an isotropic solu-
grow into elongated micelles, forming a strongly tion that readily enters a two-phase region with
viscoelastic solution. After the hexagonal phase, a lamellar phase coexisting with the isotropic
another bicontinuous cubic phase (V1) forms, just solution, L1 + La region (Fig. 8a). The composi-
prior to a lamellar phase (La). tion of this lamellar phase is ca. 12 wt% surfac-
The phase behavior of the anionic surfactant tant, and beyond this concentration the La phase
SDS follows similar trends with the ones above, enters a single-phase region displaying linear
but some noticeable differences emerge as well swelling until ca. 80 wt%. At this point a narrow
(Fig. 7). SDS has a Krafft point of about 17 C, bicontinuous cubic phase (V2) region is entered,
and in comparison with DTAC, hydrated solid which is replaced by a reverse hexagonal phase
phases appear closer to room temperature and at H2 at ca. 83 wt%.
increasingly higher temperatures, as concentra- With DTAB as a starting point, the replace-
tion increases. Moreover, between the common ment of one of the methyl groups by another
Surfactants, Phase
diagram for:
(a) AOT-water system
(Adapted from Rogers and
Winsor 1969); (b) DDAB-
water system (Adapted
from Zemb et al. 1993)
a b
N+ N+
o o
o s o− s o−
o o
Surfactants, Phase Behavior, Fig. 9 Catanionic surfactants: (a) dodecyltrimethylammonium dodecylsulfate (sym-
metric chain length); (b) hexadecyltrimethylammonium octylsulfonate (asymmetric chain length)
12-carbon alkyl chain gives rise to DDAB and, as between the attractive and repulsive forces con-
expected, to significant changes in phase behav- trolling the swelling of the bilayers.
ior (Fig. 8b). In a demonstration that the PS value
has changed from ca. 1/3 to ca. 1, bilayer struc- Pseudodouble-Chained Surfactants:
tures are now the single morphology. Small Salt-Free Catanionic Surfactants
bilayer aggregates form at the critical aggrega- When ionic surfactants of opposite charge are
tion concentration (cac), which is of the order of mixed in equimolar amounts and the inorganic
0.05 mM at room temperature. Vesicular struc- salt is removed, a novel surfactant, usually des-
tures (of different sizes and morphology regimes) ignated in the literature by catanionic surfactant,
are found slightly above this point up to about is produced. This compound is thus made up by
3 wt%, and they constitute a fine dispersion in two amphiphilic ions in a 1:1 molar ratio (Khan
excess solvent of a swollen-lamellar phase that and Marques 1998). Generically, we shall refer to
has its maximum swelling at ca. 3 wt%. Strik- this class of complex organic salts as C+C, and S
ingly, however, instead of showing linear swell- two examples are given in Fig. 9.
ing over the whole-composition range (as AOT), Catanionic surfactants have been found to
a miscibility gap in the lamellar phase appears. At possess a lyotropic behavior that is strongly
25 C, this phase separation occurs between ca. influenced by the degree of difference in length
25 and 80 wt% (Fig. 8b). Two lamellar phases, between the two chains or, in other words, by
naturally with different periodic spacings, are their differential solubility. When a catanionic
found to coexist: the previously mentioned swol- surfactant is mixed with water, a simple binary
len La0 and a collapsed La00 phase. Beyond 80 wt surfactant-water system is obtained. The com-
%, only the La00 exists again. As can be seen in the pounds with equal chain lengths or with similar
phase diagram, a critical point exists where ion solubilities (symmetric catanionics) typically
the gap vanishes. The demixing of the lamellar form lamellar phases with limited swelling
phase is a rare phenomenon in binary (Jokela et al. 1987). They often have gel-to-liquid
systems, and it is linked with a complex balance crystal-transition temperatures above room
electrostatic repulsive force and the spontaneous

curvature (Silva et al. 2011). Therefore,
a relatively complex interplay between aggregate
shape (as dictated by Ho) and aggregate-
aggregate interactions is expected when changing
the mismatch.
The phase behavior of alkyltrimethy-
lammonium alkylsulfonates, denoted by TAmSon,
where m and n are the chain lengths of the cationic
and anionic parts, respectively, serves to illustrate
these concepts (Fig. 11). The relative solubility of
the two ions of salt-free catanionics has profound
effects on the aqueous phase behavior (Silva et al.
2011). Depending on the asymmetry between
both ions, three regimes can be identified:
(1) weakly asymmetric, (2) approximately sym-
metric, and (3) highly asymmetric.
For the case of TA16Son compounds with
n ¼ 8 and 9 (weakly asymmetric), the surfactant
forms a lamellar phase in water but, with
a striking miscibility gap (coexisting La0 and
La00 phases), reminiscent of the DDAB-water
system described above. This miscibility gap is
ultimately a consequence of the fact that the
Surfactants, Phase Behavior, Fig. 10 Phase diagrams bilayer charge density is concentration depen-
for catanionic surfactant-water systems: (a)
dodecylammonium dodecanoate, AD; dent. For n ¼ 8, also a temperature-driven transi-
(b) dodecylammonium dodecylsulfate, AS; (c) dodecyltri- tion between vesicles and elongated micelles is
methylammonium dodecylsulfate, TAS; and observed at low surfactant concentration. When
(d) dodecylethyldimethylammonium dodecylsulfate, the mismatch is low, for n ¼ 10 (approximately
EDAS, in water (Adapted from Jokela et al. 1987)
symmetric regime), only one lamellar phase is
formed and the surfactant behaves like a normal
temperature (Fig. 10) unless both chains are rel- double-chained surfactant. For high mismatch,
atively small (<10 carbons). These surfactants n ¼ 6 and 7 (highly asymmetric regime), an
behave almost identically to zwitterionic ones, extensive micellar phase is obtained, with
because the catanionic interface should have TA16So6 behaving like a normal soluble surfac-
zero or very low surface charge density. tant. Conversely, for the TAmSo8 compounds,
However, if the two individual ions have very where m ¼ 12 and 14, the unusual lamellar mis-
different solubilities in water (e.g., if their chain cibility gap (and vesicle-micelle transition at low
lengths are very different), then the catanionic concentration) is again present.
film is left with an excess surface charge and the Full temperature-composition phase diagram
catanionic compound behaves effectively as an is also available for a series of N-alkypyridinium
ionic surfactant in solution, with the peculiarity alkylsulfonates (Fig. 12), where the number of
that the counterion is an organic amphiphilic ion. alkyl carbons m of the cationic part is varied from
The aggregation properties of catanionics with 0 to 8. All these catanionics behave as soluble
very asymmetric chain lengths, in particular, surfactants showing an extended L1 micellar
can be very interesting. Higher solubility region and, at higher surfactant volume fractions,
mismatch increases the surfactant film charge a sequence of liquid-crystalline phases, which
density, which in turn will strengthen the depends on the value of m.
a
L1 H1 Hydr. TA So
16 6
Cr.
Lα′D
Hydr. TA So
L1 R Lα″D Lα 16 7
Cr.
Lα′
Hydr. TA So
+ Lα′ Lα′ + Lα″ Lα ′ 16 8
Cr.
L1
L1 + Lα′
Hydr. TA So
Lα′ + Lα″ Lα″ 16 9
Cr.
Lα′
L1 + Lα Hydr. TA So
Lα Cr. 16 10
0 20 40 60 80 100 wt%
b L′
α
Hydr. TA So
+ Lα′ Lα′ + Lα″ Lα″ 16 8
Cr.
L1
L 1 + Lα ′
Hydr. TA So
Lα′ Lα′ + Lα″ Lα″ 14 8
Cr.
L1 + Lα′
Hydr. TA So
Lα′ Lα′ + Lα″ Lα″ 12 8
Cr.
0 20 40 60 80 100 wt%
Surfactants, Phase Behavior, Fig. 11 Phase diagrams compounds; (b) TAmSo8 compounds (TA16So10 at 40 C)
for alkyltrimethylammonium alkylsulfonate with varying (From Silva et al. 2011)
chain length mismatch in water, at 25 C: (a) TA16Son
a b
100 100 S
80 80
60 Lα 60 L1 Lα
T/°C
T/°C
40 L1 40
Lα + S
H1 Lα + S
20 20
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Wt% Wt%
Surfactants, Phase Behavior, Fig. 12 Phase diagrams for (a) heptylpyridinium octylsulfonate, HepPOS, and
(b) octylpyridinium octylsulfonate, OPOS, in water (Adapted from Persson et al. 2004)
a c 40
air
30
Temperature /°C
water
solution liquid
20 crystal
b solution
hydrated crystals
+
liquid
crystal
10
0
0 20 40 60 80
STDC wt%
Surfactants, Phase Behavior, Fig. 13 Bile salts in (From Madenci and Egelhaaf 2010); (c) phase diagram
water: (a) position at the air-water interface (b) proposed for the STDC-water system, showing the region of liquid
models for bile salt micellar aggregates, including disklike crystal formation (From Marques et al. 2000)
micelles (second to right) and helical structures (right)
Bile Salts conventional (Fig. 13b). Some features are simi-

The vast majority of surfactants are composed by lar to the behavior of classical amphiphiles, while
flexible alkyl chains (hydrogenated or fluori- others are very different and reminiscent of other
nated) that, in micelles and liquid crystals, are types of molecules, for example, low-molecular-
in liquid-like conformational disorder. Moreover, weight gelators or chromonic materials. Their
in those surfactants there is a clear-cut separation association is weakly cooperative, as shown by
between the polar part (the headgroup) and the the relatively poorly defined cmcs. At this critical
nonpolar moiety. These features are absent in bile aggregation region, bile salts are thought to pack
acid salts, naturally occurring amphiphiles that through “back-to-back” hydrophobic interactions
have a vital role in the human body and into very small micellar aggregates, with aggre-
self-aggregate in a peculiar manner (Kratohvil gation number of 4–10 (primary micelles). As the
1986; Small 1971). These molecules are pivotal amphiphile concentration rises, these micelle-
in the process of fat emulsification in the diges- like aggregates associate into larger and more
tive tract and also aid in the excretion of lecithin complex structures where intermolecular hydro-
and cholesterol from the body. gen bonding (secondary micelles), in addition to
The bile salts are derived from cholic acid and hydrophobic interactions, takes place. In more
so their hydrophobic part consists of a rigid, concentrated solutions, reversible formation of
nonplanar, steroid ring, as can be seen for sodium fibrils has been observed, and through the aggre-
taurodeoxycholate (STDC) in Fig. 5. The ring gation of these fibrils, tubules are formed, yield-
usually contains two or three OH groups imbed- ing viscous or gel-like solutions.
ded in the hydrophobic moiety and lying in the In mixed systems, bile salts can be incorpo-
same side, so the molecule has a concave polar rated into a variety of liquid-crystalline phases, as
side and a convex nonpolar one. The molecules is the case with lecithin (Almgren 2000), but it
lie preferentially in a flat position at the air-water was assumed for a long time that they alone in
interface (i.e., parallel to the interface), with the water did not self-assemble into mesophases.
convex apolar side exposed to the air. However, further investigations showed that for
The rigidity and the dual character of the ring temperatures ranging between 5 C and 40 C,
imply that the packing of these amphiphiles into several types of bile salt molecules form more
micelles or other supramolecular structures is not or less extensive regions of liquid crystal,
O OH
HO O
O
HO O O HO
O O OH
C12E5 8-β-Glu
O O
O OH O O
OH + O P O O
N −
O
O
MO DPPC
Surfactants, Phase Behavior, Fig. 14 Neutral surfac- octyl-b-D-glucoside (8-b- glu), and the lipid monoolein
tants including: nonionic surfactants pentaethylene glycol (MO); zwitterionic lipid, dipalmitoylphosphatidylcholine
monododecyl ether, C12E5, sugar-based surfactant n- (DPPC)
as exemplified in Fig. 13c for the STDC-water they can be subject to electrostatic effects
system (Marques et al. 2000). The mesophases (though, naturally, much more weakly
are thought to be formed by reverse micelle-like than ionics).
aggregates arranged with hexagonal symmetry,
bearing some relation to the helical structures Nonionic Surfactants
present in the solution phase. Bile salts serve to Nonionic surfactants involve a variety of
illustrate a class of important amphiphiles, for headgroups and structures, but the most important
which the concept of amphiphilicity becomes group is fatty alcohol ethoxylates, for which the
somewhat more fluid (owing to molecular struc- polar group is a polyether consisting of
ture) and which therefore behave in solution in oxyethylene units, formed by the condensation
a different (and intriguing) way with respect to of ethylene oxide. These compounds are usually
classical ionic surfactants. referred to by the designation CmEn, where m is
the alkyl-chain length and n is the number of
oxyethylene groups (typically n ¼ 3–10). One of
Binary Neutral Surfactant-Water the main characteristics of CmEn nonionics is the
Systems progressive dehydration of the headgroups when
temperature is raised, which leads to a continuous
Neutral surfactants include nonionic and zwitter- increase of the packing parameter and decrease of
ionic molecules, and some examples are given in Ho (Jönsson et al. 1998). This is general for the S
Fig. 14. The phase behavior of nonionic surfac- whole class excluding very long headgroups.
tants should be easier to rationalize than that of Therefore, and in contrast with ionics, one finds
ionic surfactants because there are no long-range that temperature drives the formation of phases
electrostatic forces. The intra- and intermicellar roughly in accordance with Fontell scheme.
headgroup interactions occur over a much shorter These features are illustrated in Fig. 15a,
range, and so the headgroups can pack more where the phase diagram for C12E5-water system
tightly. The effective headgroup volume is is shown (Strey et al. 1990). The L1 phase
given by its actual size, together with about one displayed in the phase diagram is micellar. At
hydration layer (Tiddy 1980). A wide range of low temperatures and concentrations, these
mesophases can occur. Zwitterionic surfactants micelles are spherical or approximately spheri-
broadly resemble nonionic surfactants, but cal. With increasing temperature, the change in
because two bulky charged groups are usually spontaneous curvature Ho first leads to growth
involved, the headgroups can be large; moreover, into flexible and long wormlike micelles.
Surfactants, Phase
diagram for nonionic
surfactants: (a) C12E5-
water system (Adapted
from Strey et al. 1990);
(b) b-octylglucoside-water
system (Adapted from
Nilsson et al. 1996); here,
extremely narrow two-
phase regions exist between
the single phases
(not apparent in the figure)
Further decrease of Ho at higher temperatures rev mic + water

100
leads to the appearance of branch points in the
rev hex + water rev hex
micelles and formation of a network. The local Temperature / °C
80
curvature is smaller in the branch points than in rev mic
the rest of the cylindrical micelle, and thus, the
system adjusts the overall aggregate curvature to 60
Ho by adjusting the density of branches. The den- cub´ + water cub´´
sity of branches is, however, coupled with surfac- 40
tant volume fraction, and as a consequence phase
lam
separation into a micelle-rich and a micelle-poor cub´ cr
region occurs. The system has therefore a lower 0 60 70 80 90 100

consolute temperature (critical point), commonly monolein wt %
called cloud point, due to the fact that samples
become turbid. For the C12E5-water system, this Surfactants, Phase Behavior, Fig. 16 Phase diagram
for the monoolein-water system (Adapted from Qiu and
critical temperature Tc is 31.9 C. With continuous Caffrey 2000)
temperature increase, Ho reaches a value of ca.
zero, and depending on concentration, a phase
transition to a L3 sponge or La phase is observed. nonionic surfactants, because they display some
Continuous T raising leads to reverse structures. properties that are different from alkyl
At low temperatures, the increase in concen- ethoxylates and, in addition, are made from
tration first leads to a hexagonal mesophase renewable raw materials and are easily biode-
(H1) at ca. 35 wt% and to a narrow cubic phase gradable. One of the distinguishing features of
(V1) around ca. 60 wt%, just before the La phase the phase behavior of sugar surfactants is that
is entered. At higher temperatures, the high the mesophases are very stable with temperature,
branch point density does not allow packing of as illustrated by n-octyl-b-D-glucoside in
micelles into a hexagonal arrangement, and this Fig. 15b, which results in simpler phase diagrams
phase vanishes at around 20 C. Hence, only the than those of ethylene oxide surfactants.
La phase forms at ca. 55 wt%. A distinction A single-chained, nonionic surfactant of bio-
should be made from the La phase formed at logical importance, and one of the mostly widely
much higher dilutions (and higher temperatures). studied lipids, is 1-monoolein, a monoglyceride
The latter is favored by a change in Ho, while the that has the particularity of having a double
former is favored by concentration effects. C–C bond in the n-9 position of the alkyl chain.
Alkyl glucosides and other sugar-based sur- The phase behavior of this lipid (Fig. 16) is rather
factants have also become an important group of unusual and illustrates the complexity that the
a b
100
80 90
80
Temperature / °C
multi-phase
Temperature / °C
V1
lam+rev mic
70
60
cub 60 L2
Lα
+ L2 H1
L1 hex
H1 Lα 50 I1
hex+lam
40 40 L1
mic+liq
V1 30
20 20
0 20 40 60 80 100 0 20 40 60 80 100
(EO)37(PO)58(EO)37 wt% (EO)37(PO)58(EO)37 wt%
Surfactants, Phase Behavior, Fig. 17 Binary phase diagram for a triblock copolymer in (a) water (Adapted from
Alexandridis et al. 1996); (b) formamide (Adapted from Alexandridis 1998)
presence of an unsaturated chain may induce in these macromolecular surfactants are parti-
surfactant-phase behavior (Qiu and Caffrey cularly sensitive to temperature.
2000). Monoolein forms two reverse The starting materials for the polymerization
bicontinuous cubic phases, cub0 (Pn3m type) are ethylene oxide (EO), propylene oxide (PO),
and cub00 (Ia3d type) in Fig. 11, and, at higher and butylene oxide (BO). There are many varia-
concentration, a lamellar phase. The cubic phases tions of EO-PO type copolymers, but by far the
have a big extension in the phase diagram both as most studied and used ones are the so-called
a function of temperature and concentration; Pluronics, of general formula (EO)n(PO)m(EO)n,
moreover, cub0 exists also in a wide two-phase and inverse Pluronics, (PO)n(EO)m(PO)n.
region with excess solvent, leading at high water Figure 17 shows the phase diagrams of one such
content to the formation of dispersed particulates triblock copolymer, (EO)37(PO)58(EO)37, in
termed cubosomes. The phase behavior is also water and in formamide. As can be seen, a rich
strongly affected by temperature. sequence of phases occurs in both solvents, and
is in striking resemblance to the sequence of
Macromolecular Surfactants with Fontell scheme.
Oxyethylene Groups A critical micellar concentration can be
Block copolymers containing oxyethylene defined that shows a dramatic decrease as the
groups in their backbone also behave as surfac- temperature is increased. The cloud point S
tants capable of self-assembling in water and of increases with the fraction of EO units, and it
yielding a phase behavior that is very rich and decreases with increasing molecular weight.
fundamentally similar to that of common A relevant feature is that a large number of
low-molecular-weight surfactants, in particular liquid-crystalline phases is usually formed, span-
to CmEn nonionics (Jönsson et al. 1998). For ning a wide range of PS values or spontaneous
these poly(alkylene glycol) copolymers, the curvatures. Moreover, a higher fraction of EO
hydrophilic block is composed by poly(ethylene groups promotes high positive curvature phases,
glycol) units, –OCH2CH2–, whereas the whereas a higher fraction of PO groups promotes
hydrophobic block is composed by poly(propyl- lower and negative curvature phases.
ene glycol), –OCH2CH(CH3)–, or, less com-
monly, by poly(butylene glycol), –OCH2CH Zwitterionic Surfactants
(CH2CH3)–. Because of the large number Zwitterionic surfactants are amphiphiles that
of oxyethylene groups that they contain, under typical conditions contain simultaneously
Surfactants, Phase
Behavior, Fig. 18 Phase 70
diagram for the
DPPC-water system
60 Lα
(Adapted from
Grabiellemadelmont and
Lα + H2O Pβ ’
Perron 1983) 50
+
Temperature / °C
Lβ ’
40
Pβ ’
Pβ ’ + H2O
Pβ ’ + Lβ ’
30
20 Lβ ’ + H2O Lβ ’
10
10 20 30 40 50 60 70 80 90
DPPC wt%
a positive and a negative charge in their polar Ternary Surfactant-Water-Oil Systems,

groups. Single-chained zwitterionics show simi- Mixed Surfactant Systems, and Other
lar phase-behavior features to nonionic surfac- Multicomponent Systems
tants. Double-chained zwitterionic lipids, like
DPPC (Fig. 6), also behave similarly to other The study of the phase behavior of
double-chained surfactants, in that they form multicomponent systems of the type surfactant-
a lamellar phase La (Chernik 1995). However, water-oil, surfactant-surfactant-water, or surfac-
as exemplified by DPPC and in contrast to ionics, tant-polymer-water has been very important for
the lamellar swelling is very limited (Fig. 18), technical applications, like those related with
because there is no long-range electrostatic force enhanced interfacial performance, oil recovery,
to counteract the van der Waals attractive force. and control of rheological properties of materials,
A wide two-phase region exists between to cite but a few examples. From a fundamental
the lamellar phase and the solution phase, where point of view, these systems are a natural exten-
the lipid solubility is extremely small (so that the sion of the binary surfactant-water where the
liquid is virtually neat water). It is in this added cosolute can be an oil, a weak amphiphile,
two-phase region, at high water content, that an apolar molecular, another surfactant, or
liposomes of varying size, polydispersity, and a macromolecule. The obtained mixture can be
kinetic stability can be prepared by using strong a three-, four-, and even five-component system,
mechanical methods, such as sonication and depending on the number and interactions
extrusion (Lasic 1993). This phase diagram also (physical or chemical) of the constituents.
highlights the occurrence of the so-called gel Ternary systems include surfactant-oil-
phases, in this case Lb0 and Pa0 , which are water, and surfactant-surfactant-water systems
hydrated bilayer-based phases where alkyl chains of the type nonionic-nonionic, ionic-nonionic,
are solid-like and molecular motions highly or ionic-ionic of similarly charged headgroups.
restricted. Lb0 corresponds to a tilted phase, Examples of commonly found complex sys-
while Pa0 is a rippled phase (Larsson 1994; tems with more than three components include
Nagle and Tristram-Nagle 2000). cationic surfactant-anionic surfactant-water,
Surfactants, Phase decanol (oil)
diagram for a ternary C10OH
surfactant-oil-water microemulsion
system. As an example, (water-in-oil)
rev mic
20
a 3-phase triangle 80
(coexisting micellar, rev hex
lamellar, and normal 40
hexagonal phases) and its 60
two neighboring two-phase
regions are indicated mic+ 60
(Adapted from Laughlin lam 40
rev mic+lam
1994)
80
20
hex
mic
H2O 20 40 60 80 C7CO2Na
sodium octanoate
(surfactant)
microemulsion
(oil-in-water)
surfactant-cosurfactant-water-oil, and surfac- goals involved, for dilute and concentrated

tant-cosurfactant-brine-oil. systems can be rather different. We will
The same models for rationalizing and under- analyze below the case of surfactant-water-oil
standing self-assembly and phase behavior as and surfactant-surfactant-water systems, leaving
described so far can be applied here. However, polymer-surfactant-water systems for the next
due to the number of components, complications section.
may arise concerning the role of thermodynamic
and kinetic factors on the attainment of equilib- Surfactant-Water-Oil Systems
rium, the determination of the number and nature Figures 19 and 20 show two examples of three-
of equilibrium phases, and so forth. component systems of the surfactant-water-oil
A broad and somewhat arbitrary division can type. In Fig. 19 one can see the sodium
be made when it comes to these multicomponent octanoate-decanol-water phase diagram, in
mixtures: which different phases occur. At 20 wt% surfac-
• Dilute systems – typically systems below tant, for instance, progressive addition of oil S
1–5 wt% in total amphiphile content, where induces the following phase sequence: normal
mixed micelle solutions, vesicle solutions, micelles ! lamellar phase ! reverse hexagonal
microemulsions, liquid crystal dispersions phase ! reverse micelles. On the other hand, the
(such as lamellar and cubic dispersions), crys- binary surfactant-water systems show that only
talline precipitates, and coacervates frequently a micellar and a hexagonal phase form. Different
occur. two- and three-phase coexistence regions can be
• Concentrated systems – above 5 wt% in sur- seen in the triangle, in keeping with the phase
factant concentration where liquid-crystalline rule.
phases (normal or reversed), microemulsions, The phase diagram of the triblock copolymer-
and concentrated reverse micellar solutions xylene-water system in Fig. 20 shows an even
can be found. wider diversity of self-assembled phases. In the
The experimental methods and techniques binary surfactant-water axis, the sequence of
used for determining phase diagrams, and the phases follows Fontell scheme as the surfactant
cubb cubb
lam
triblock copolymer
(EO)37(PO)58(EO)37
hex rev hex
cubb
20
80
cubb cubb
40
rev h
lam 60
ex
cubb cub m
x
he
60
40
cubm rev
mic mic mic rev mic
20 40
water oil
(p-xylene)
Surfactants, Phase Behavior, Fig. 20 Phase diagram for a ternary block copolymer-oil-water system (From
Alexandridis et al. 1996)
concentration increases. Upon addition of oil one Nonionic Surfactants

finds, for instance at 80 wt% surfactant, Nonionic-surfactant (CmEn)-water-oil systems
a sequence of lamellar ! cubic ! reverse are one of the most widely studied ternary sys-
hexagonal ! cubic ! reverse micellar phases. tems that form microemulsions, due to the ease
The two cubic phases and the reverse hexagonal with which Ho can be adjusted by changing the
only form upon addition of oil. temperature T. Increasing temperature causes the
headgroups to become less and less hydrophilic
Microemulsion Systems and therefore originates a monotonic decrease of
Microemulsions are homogeneous and thermo- Ho. Because both temperature and composition
dynamically stable solutions containing oil and are important variables, four axes are needed for
water and where an amphiphile acts as the emul- a full thermodynamic representation of the sys-
sifier. They can be either discrete, containing tem. This can be obtained with a prism, in which
monodisperse oil-in-water (o/w) or water-ion-oil the two independent compositional degrees of
(w/o) nanodroplets, or have a bicontinuous struc- freedom are represented in the triangle base, and
ture (Evans and Wennerström 1999). They can be temperature is the height of the prism (Fig. 21).
formed by using nonionic or ionic surfactants, These three degrees of freedom make determina-
and their formation and structure illustrate well tion of the complete phase diagram a difficult
how phase behavior can be suitably manipulated task. As a simplification, one of the parameters
by changing temperature, composition, salt, and can be fixed, while the remaining two are system-
other parameters. atically changed and investigated.
Fs
T
Tu
w/o + water
PIT
2
3 1
To
L
T
bic + water + oil
2
o/w + oil
Tl
Fs,o
Fs
Fw Fw=Fo Fo
Surfactants, Phase Behavior, Fig. 21 Schematic rep- bicontinuous (granite texture) phases. PIT is the phase
resentation of a microemulsion phase diagram in the form inversion point, Tl is the temperature of the lower critical
of prism and the two-dimensional fish-cut (shaded area) endpoint, To is the temperature at which Ho is zero
represented in detail in the right. The test tubes represent (the microemulsion is balanced), and Tu is the temperature
the relative amounts of aqueous (white), oily (gray), o/w of the upper critical endpoint (Adapted from Stubenrauch
(white with droplets), w/o (gray with droplets), and 2009)
A common simplification that enables The same behavior is to be expected at high

a good understanding of the whole system is the temperatures where Ho is negative. Here the
“fish-cut” diagram shown in Fig. 21. Here, oil and microemulsion phase is of the w/o type, and the
water have the same volume fraction, and only two-phase region of w/o plus excess water is
surfactant volume fraction, Fs, and temperature denoted as the 2 region.
are changed. A nonionic surfactant with If Fs is now fixed and T is gradually increased
a positive spontaneous curvature, Ho > 0, will in such a way that Ho is gradually changed from
form o/w droplets in the water-rich phase slightly positive to negative, the microemulsion formed S
above Fs,o (the surfactant volume fraction at To, will shift from o/w droplets to a bicontinuous
the temperature where Ho ¼ 0), plus oil in excess phase (when Ho 0) and then to w/o droplets.
at low temperatures. This two-phase region is If Fs is not high enough to emulsify the whole
denoted as 2. Increasing the surfactant volume water/oil volume, the system will be in a three-
fraction, at constant temperature T, will result in phase region (region 3), consisting of excess
an increase in the fraction of the o/w phase rela- water and oil, and the bicontinuous phase.
tive to excess oil, because there is more surfactant Increasing Fs will bring the system into
available to emulsify oil. Continuous increase of a one-phase region when Fs is high enough to
Fs will eventually lead to full emulsification emulsify both water and oil. This happens at the
of the oil, and a single microemulsion phase is phase inversion temperature (PIT) point, which is
reached (region 1 in Fig. 21). Further increase the point of the one-phase region (region 1) in the
of Fs will bring the system to a lamellar phase diagram of lowest Fs. The PIT point occurs
phase, La. at To (the temperature at which Ho = 0), because it
Surfactants, Phase
Behavior,
Fig. 22 Fish-cut phase
diagrams for the SDS-
hexanol-dodecane-brine
system. Here,
W1 ¼ Fcos/(Fs + Fcos) and
X ¼ (Fs + Fcos)/(Fs + Fcos
+ Fw + Fo). Fw ¼ Fo, and
salt concentration is 3.0
(straight line), 5.0 (dotted
line), and 7.5 wt% (dashed
line) (Adapted from Li et al.
1999)
is when the surfactant film is balanced that it monotonically. Figure 22 shows a schematic
is able to emulsify the highest amount of water fish-cut diagram for a typical micellizing surfac-
and oil. tant, SDS, by addition of hexanol and salt. The
addition of hexanol and salt is able to bring the
Ionic Surfactants system from a o/w to a w/o microemulsion, pass-
In single-chain ionic surfactants, the curvature of ing through a region where a bicontinuous phase
the film is usually very high to incorporate any oil is formed, in a very similar way to nonionic
in it, and one generally has to decrease the spon- surfactants, with the difference that the fish cut
taneous curvature to form microemulsions. This is skewed. This occurs due to the fact that
can be done by the addition of a cosurfactant, a significant fraction of hexanol is solubilized in
such as a long-chain alcohol. The alkyl chain the oil phase instead of residing only at the oil-
will increase the hydrophobic portion of the water interface.
film, while the hydroxyl headgroup will diminish Double-chained ionic surfactants such as AOT
the film surface area, by weakening the electro- and DDAB, due to their larger hydrophobic por-
static repulsions between the charged surfactant tion, have a tendency to form reverse w/o
headgroups. This will raise the surfactant param- microemulsions. In the AOT case, without
eter PS and lower Ho, enabling the film to form added salt, the microemulsion region has
microemulsions. The addition of salt also screens a small stability range due to the fact that
the electrostatic interactions of the surfactant a lamellar phase is the preferred structure
headgroups, decreasing even further the sponta- because of the electrostatic repulsive forces
neous curvature. from the charged AOT headgroups (Arleth and
In this case, a fish-cut phase diagram can Pedersen 2001). The addition of salt leads to
also be constructed. The temperature axis can screening of this charge and a weakening of the
be replaced by a cosurfactant concentration axis repulsive force, favoring high entropy w/o
(W1 ¼ Fcos/(Fs + Fcos), where Fcos is the microemulsion droplets over the lamellar
cosurfactant volume fraction), because through phase. Increase of T leads to an increase in Ho,
an increase in W1, Ho will also decrease causing the system to undergo phase inversion at
a 100 b DDAB
XNacl=1.0%
o/w
80
XNacl=0.6%
60
XNacl=0.4%
T/ °C
40 A
w/o
20
w/o B
0
0 10 20 30 H2O Dodecane
AOT wt%
Surfactants, Phase Behavior, Fig. 23 Microemulsion water-dodecane system, from Warr et al. (1988); along
systems with double-chained surfactants: (a) fish-cut the A-B line, the system is composed of bicontinuous
phase diagrams for the AOT-brine-decane system. branched micelles, while addition of water moves the
Fw ¼ Fo, XNaCl ¼ mNaCl/(mNaCl + mW). (b) DDAB- system progressively to discrete w/o droplets
higher temperatures, in an opposite way to surfactant molecules, but chain-chain and

nonionic CmEn systems (Fig. 23). headgroup-headgroup interactions also play an
DDAB does not need addition of salt to form important role in some systems. An example of
w/o microemulsions (Arleth et al. 2001). Inter- the contribution of chain-chain interactions in
estingly, along the A-B line in Fig. 23, the aggregation is provided by mixtures of surfac-
microemulsion phase is a bicontinuous network tants with hydrocarbon and fluorocarbon alkyl
of branched micelles. The micellar core is com- chains. Here, in a micellar solution phase,
posed of water. Further addition of water leads to microsegregation can occur into fluorocarbon-
a transition to discrete droplets, as expected when rich and hydrocarbon-rich micelles, and two
approaching the emulsification failure boundary. cmcs can be detected (Jönsson et al. 1998). With
respect to headgroup-headgroup interactions,
Surfactant-Surfactant-Water Systems: they can be discussed in terms of packing effects
General Trends and electrostatic interactions. The former effect is S
Aqueous mixtures of two surfactants apparent when headgroups of very different sizes
almost invariably show associative phase behav- are mixed, with the addition of small headgroups
ior. Phases containing mixed aggregates are into a film with large headgroups being energet-
formed, while separation into macroscopic ically favorable.
phases enriched in each surfactant is not known Electrostatic interactions are clearly important
to occur (Piculell and Lindman 1992). However, for nonionic-ionic surfactant mixtures and, in
it is not uncommon that within one macroscopic particular, for cationic-anionic surfactant
phase, namely, solution phase, there is coexis- mixtures (analyzed in more detail in section
tence of small aggregates that can be enriched “Cationic-Anionic Surfactant Mixtures”). For
in one of the surfactants; in other words, segre- these two types of mixtures, several interfacial
gation at the microscopic level can occur. properties show a large negative deviation
The driving force for surfactant miscibility from ideal mixing (synergism) due to a net favor-
in solution is the entropy of mixing of able interaction between the two molecules
Surfactants, Phase DMPC

Behavior,
Fig. 24 Ternary phase 10
diagram for 90
DMPC-CTAB-water 20
system at 30 C (Adapted 80
from Rydhag et al. 1982) 30
70
40
60
Weight % H2O 50
50 Weight % DMPC
60
40
70 Lα
L1+Lα
80
L1+Lα
90
L1
H1
H2O 10 20 30 40 50 CTAB
Weight % CTAB
(Holland and Rubingh 1992; Ogino and 20 wt% CTAB, due to the curvature effects
Masahiko 1993). This is seen for instance by imposed by this single-chained surfactant.
a strong decrease in the cmc of the mixtures and
by enhanced surface activity. Cationic-Anionic Surfactant Mixtures
For surfactant-concentrated mixtures, mixed Cationic-anionic surfactant mixtures in water,
aggregates will also form, in particular liquid- commonly designated by catanionic mixtures,
crystalline phases. Often the phases present in have been subject to extensive investigations
each binary surfactant-water system will accom- not only because of their synergistic interfacial
modate the second surfactant more or less exten- properties (Holland and Rubingh 1992; Ogino
sively, but new phases can also be induced, and Masahiko 1993) – a fact long recognized
depending on the mixing ratio between the two and explored – but, in more recent time, espe-
amphiphiles. Again, a qualitative understanding cially due to the rich features of phase behavior
of the phase behavior can be done on the basis of they present (Kaler et al. 1989; Khan and
geometric packing effects and, when appropriate, Marques 1998; Tondre and Caillet 2001). The
electrostatic effects. Figure 24 illustrates one most important factors affecting phase behavior
such example, a mixture composed by the zwit- in these mixtures are the packing parameter of
terionic lipid dimyristoylphosphatidylcholine, initial surfactants, the mixing molar ratio, total
DMPC, and CTAB (Rydhag et al. 1982). The surfactant concentration, and salt. By varying the
lipid forms by itself a lamellar phase that can ratio between ionic amphiphiles, for instance, one
take up a maximum of 40 wt% water. When is able to tune the charge of the aggregates and the
small amounts of ionic surfactant CTAB are effective packing parameter of the mixture, and
added, the lamellar phase swells extensively to this in turn can originate a diversity of aggregates
the water apex (Fig. 24), as now repulsive double- and phases. Additionally, it is possible to vary the
layer electrostatic interactions, together with molecular structure of both surfactants from
thermal undulation forces, stabilize the lamellar mixture to mixture, allowing further manipula-
phase at large periodic spacings. This lamellar tion of the packing parameter Ps and spontaneous
phase is only stable up to a maximum of ca. curvature Ho of the resulting mixture.
Surfactants, Phase H2O
Behavior,
Fig. 25 Pyramidal phase
diagram for the full phase C+C− C−X+
behavior representation of H2O
mixtures of cationic and
anionic surfactants
C−X+
C+C−
C+X− X−X+
X−X+
C+X−
A fundamental distinction can be made regions on the plane, but two such cuts, for exam-
between dilute and concentrated systems. In ple, are not distinct homogeneous phases;
dilute mixtures, micellar growth and spontaneous (3) phase-separating samples may have constitu-
formation of stable vesicle aggregates can occur, ent phases absent from the triangular plane; and
as well as phase separation phenomena. At high (4) the shape of three-phase regions is not limited
concentrations, a wide variety of mixed liquid- to triangles. However, experimental investiga-
crystalline phases form, some of which entirely tions of phase behavior of catanionic mixtures
distinct from those of the binary surfactant-water over the full-concentration range, and above the
system. Krafft temperature of catanionic solid com-
It should be recognized that when a system is plexes, can be satisfactorily represented in trian-
composed of a cationic amphiphile C+X, an gles, especially in what concerns the depiction of
anionic amphiphile CX+, and water, with no single-phase regions.
common ion, it has in principle four components.
This is due to the possibility of having the salts Concentrated Systems
C+C (catanionic amphiphile, or complex We will illustrate concentrated catanionic sys-
salt; cf. section “Pseudodouble-Chained Surfac- tems by considering the mixing of different
tants: Salt-Free Catanionic Surfactants”) and anionic surfactants with DDAB, a cationic
X+X (simple electrolyte), in addition to the double-chained surfactant whose phase behavior
cationic and anionic surfactant salts, and taking in water has already been discussed (cf. section
also in consideration the condition of “Double-Chained Surfactants”). Significant
electroneutrality. In principle, all four of these changes in phase behavior are brought about S
salts can form, and a triangular phase diagram is when an anionic surfactant is added to the
not sufficient to represent the full phase behavior DDAB-water system (Fig. 26). Solutions and liq-
of the system. This can be achieved by using uid crystals that are not present in the individual
a pyramidal representation (Marques et al. systems usually appear as a result of the interplay
1993), and thus, the conventional mixing plane between packing and electrostatic interactions
depicted (ternary diagram) constitutes within the aggregates and electrostatic interac-
a particular 2-D slice through the 3-D diagram tions between different aggregates (Marques
(Fig. 25). et al. 2003). The cases of adding a single-chained
When inspecting a triangular phase diagram surfactant (SDS), a double-chained surfactant
for such four-component systems, one should (AOT), and a bile salt (STDC) with a rigid ring
then bear in mind that (1) one-phase regions in system as backbone structure are considered
principle lie in cuts on 3-D one-phase bodies; below. Anionic surfactants associate strongly
(2) thus, the same phase can be cut at distinct with DDAB, as expected, because of the
electrostatic interactions of the headgroups and Studies of this type show that there are
the large entropic gain related with counterion possibilities to tailor new aggregate structures
liberation from the catanionic interface. and phase equilibria by a suitable blending of
The particular nature of the phases formed in surfactants (Fig. 26).
the mixed system results from the geometric
packing effects induced by the shape of the Dilute Systems
anionic surfactant. The single-chained SDS and One of the most appealing features of dilute
the branched double-chained AOT are expected catanionic mixtures is the appearance of regions
to induce different phase behaviors, as indeed in the phase diagrams where unilamellar vesicles
observed. The same applies to the bile salt spontaneously form (i.e., form in the absence of
STDC, a special case, since its structure differs strong shear forces), often showing very long-
from a conventional amphiphile: the hydrophobic term stability that is thought to be true thermody-
part is a rigid steroidal ring system in which namic stability (Kaler et al. 1989; Khan and
hydrophilic OH groups are imbedded. Marques 1998; Tondre and Caillet 2001). The
In both DDAB-AOT and DDAB-SDS mix- description of the general phase behavior of
tures, the phases formed by the anionic surfactant these mixtures can be illustrated with a few
alone have a limited capability of solubilizing the examples from the literature, shown in Fig. 27
cationic surfactant. Moreover, a new lamellar (phase diagrams in weight percent of constitu-
phase, DIII, and a new sponge phase, L3, appear ents, with indication of straight lines of surfactant
for the SDS and AOT system, respectively, still equimolarity, n(C+)/n(C) ¼ 1).
with very limited stability in both cases. The One can start with the DDAB-SDS-water
situation is somewhat different for the bile salt phase diagram (Herrington et al. 1993), presented
system, where an extensive L1 region is formed, in Fig. 27a. When the two surfactants are mixed
reflecting the versatility of SDTC to accommo- together, the association is so strong that a solid
date the double-chained amphiphile. phase dominates most of the phase diagram. This
On the DDAB-rich side, several interesting is mainly due to the fact that the symmetry
things happen as well. The water-swollen lamel- between the chain lengths allows a very effective
lar phase DI is destabilized by very small frac- and stable packing of the chains in a crystalline
tions of any of the anionic surfactants, lattice. Even so, far from the region
a consequence of double-layer electrostatic of equimolarity, in the SDS-rich side, elongated
effects. The collapsed lamellar phase DII, how- micelles and vesicles can be found. At low
ever, incorporates a large fraction of SDS, a small cationic-to-anionic ratios, the micelles are small
fraction of STDC, but negligible fraction of AOT. but grow considerably to very large threadlike
For the latter system, this is an indication of the micelles as the ratio increases. At
packing constraints imposed by the branched and C+/C ¼ 35/65 w/w, a multiphase region is
bulky hydrophobic region of AOT in the mixed entered, and at C+/C ¼ 40/60 w/w, a two-phase
film. The lamellar phase is easily destabilized to region is entered consisting of vesicles and crys-
yield isotropic phases of negative mean curva- talline precipitate. Clearly, the continuous asso-
ture, the cubic (I0 ), and the solution phase L2. ciation of cationic surfactant in the ionic
At equimolar and near-equimolar mixing threadlike micelles increases Ps enough as to
ratios, and high concentration, the strong asso- induce the formation of bilayers. However, bila-
ciation leads to a novel phase: a bicontinuous yers have to compete with the solid crystal that is
cubic phase, I, for DDAB-SDS and a reverse also favored at these mixing ratios. Increase of
hexagonal phase, F20 , for DDAB-AOT. C+/C weight ratio will favor more solid precip-
A unique observation in the DDAB-AOT system itate, but if the ratio C+/C ¼ 40/60 w/w is
is the presence of two different reverse hexago- maintained and the whole mixture is diluted,
nal phases F2 and F20 , found to coexist in a single-phase vesicle region is reached. These
multiphase regions. vesicles are thought to be in thermodynamic
DDAB DDAB
DII
G1
+ +
L2
DII
I′
equimolar line
I 50 50
DIV –
–
F2′
DI DI
DIII L3 I F2
GII DIII
L1 E
2H 50
2H O
SDS 2O AOT
2
DDAB
DII
+
DI
–
LI
DI
lc hc
2H O
2 STDC
Surfactants, Phase Behavior, Fig. 26 Phase diagrams AOT-water system, (c) DDAB-STDC (bile salt)-water
for three catanionic mixtures over the full composition system (Adapted, respectively, from Marques et al.
range at 25 C: (a) DDAB-SDS-water system, (b) DDAB- 1993; Caria and Khan 1996; Sjobom et al. 2005)
equilibrium. The transition from micelles to ves- negative charge, and the second rich in CTAB,
icles is abrupt and involves a small two-phase with excess positive charge. The bilayer proper- S
region where elongated micelles and ties of the vesicles depend on the ratio of CTAB
multilamellar vesicles are found to coexist. to SOS, with CTAB-rich bilayers being stiffer
Turning our attention to the cetyltrimethy- than SOS-rich ones, probably due to a greater
lammonium bromide (CTAB) – sodium octyl thickness of the bilayer. In the CTAB-rich side
sulfate (SOS) – water system (Yatcilla et al. as SOS is added to CTAB-rich micelles, there is
1996), displayed in Fig. 27b, the same general strong rodlike micellar growth. SOS-rich
features can be deduced with the important micelles on the other hand are spherical at low
difference that now the chain length mismatch amounts of added CTAB, growing only to rodlike
(m 6¼ n) effectively destabilizes the catanionic micelles at higher ratios of CTAB to SOS and
solid, now found only at equimolarity. Away higher concentrations. At moderate concentra-
from this value, liquid-crystalline and solution tions, the transition between rodlike micelles
phases are formed. Now, two vesicle regions are and vesicles is first order, and so there is macro-
also found, the first rich in SOS, thus with excess scopic phase separation for both CTAB-rich and
a water b water
1% 99% 1% 99%
Ves+
Ves−
2% 98% 2% 98%
Ves−
L1(el.mic.)
+Ves−
multiphase
3% 3% multiphase
97% 97%
4% 4%
96% 96%
Ves−
L1 L1+X + L1 L1 L1+Lα Ves−+Lα L1 L1
X (el.mic.) (el.mic.) (sp.mic.)
DTAB 1% 2% 3% 4% SDS CTAB 1% 2% 3% 4% SOS

equimolar equimolar
c water d water
blue
turb sol
L αl
L1
sol+Lα+X sol+Lα+X
1% 99%
H1
sol
Lαlll
2% 98%
sol+X
V
sol+Lα sol+Lα
Lαll
DDAB LαlV SDS DDAB 1% SDS

2%
equimolar equimolar
Surfactants, Phase Behavior, Fig. 27 Ternary-phase SDS-water (40 C) systems, and (d) dilute corner of the
diagrams for the (a) DTAB-SDS-water at 25 C (Adapted DDAB-SDS-water system in detail at 25 C (Adapted
from Herrington et al. 1993), (b) CTAB-SOS-water at from Khan and Marques 1998)
25 C (Adapted from Yatcilla et al. 1996), (c) DDAB-
SOS-rich solutions. At low concentrations, mix- apart from the fact that the alkyl chains have the
tures rich in SOS exhibit no phase separation. same length, the extension of the solid precipitate
Instead, small micelles abruptly transform into is reduced, and two vesicle lobes are observed
vesicles over a narrow range of surfactant con- (Marques et al. 1998). However, in this case,
centration. Increasing the surfactant concentra- owing to the much higher PS of DDAB, just
tion at C+/C w/w ratios that favor vesicles, a slight addition of cationic surfactant to the
a swollen lamellar phase is found, either with SDS micelles is enough to form bilayer phases,
excess negative or positive charge, depending and hence, rodlike micelles are not observed.
on which side of the equimolar line it is situated. Interestingly, in the SDS-rich vesicle lobe, vesi-
If, instead of changing the asymmetry in the cles coexist with globular micelles at C+/C w/w
length of the hydrophobic tails, a single-chain between 10/90 and 24/76, whereas between 24/76
surfactant is mixed with a double-chain one, and 53/47, vesicles coexist with disk-shaped
novel features appear. Considering the DDAB- micelles. In the range C+/C w/w 53/47–80/20,
SDS-water system (already discussed above), only vesicles are found in the vesicle lobe.
This range corresponds to a cationic-to-anionic trimethylammonium. If one adds a cosurfactant
molar ratio of 1:3 to 1:2, indicating that vesicles to the catanionic system to reduce Ho, the neces-
form only for a relatively small excess of anionic sary amount is smaller than what would be
molecules in the mixed film and for a narrow needed in an equivalent ionic system (since part
composition range. Furthermore, in this system, of the charge is already neutralized).
the transition between micelles and vesicles is If the catanionic headgroups are smaller, as in
continuous, and both aggregates coexist within the case of dodecylammonium dodecanoate (AD)
the same thermodynamic phase. If the concentra- and octylammonium octanoate (AO), even
tion of the micellar phase (C+/C w/w < 10/90) is though the crystals are insoluble in water, they
increased, a hexagonal H1 phase is found. form reverse w/o microemulsion droplets when
The appearance of disk-micelles is interesting, oil is added (Jönsson et al. 1991). Furthermore, it
and they are suggested to be enriched by DDAB has been observed that by adding a net charge to
in the flat region, with the rims stabilized by SDS. the surfactant film, the curvature increases, and the
On the DDAB-rich side, no micelles are droplets incorporate more water (Fig. 28). This is
observed, because DDAB already forms bilayers the result of an increase in curvature toward water.
in the neat system. Addition of SDS however Since the catanionic is now slightly charged, the
induces important changes in the bilayers, and headgroup area should increase as a result of ionic
a stable phase of vesicles with excess positive repulsions, increasing the curvature.
charge is observed. Interestingly, within both In catanionic systems, thus, the ratio between
vesicle lobes, moving the C+/C molar ratio both ions, the headgroup volume, and the inclu-
toward equimolarity, vesicles are seen to grow sion of a cosurfactant, constitute additional
in size. This may be related to a decrease in parameters in the adjustment of the surfactant
charge that reduces the spontaneous curvature film spontaneous curvature at room temperature.
Ho. If concentration is increased at ratios favoring Hence, not only the nature of the microemulsion
vesicle formation, swollen-lamellar phases can phase (o/w, w/o, or bicontinuous) can be manip-
appear, as seen before. ulated, but also the radius of the discrete droplets.
Microemulsions in Catanionic Mixtures

As seen previously, ionic micellizing surfactants Surfactant-Polymer Phase Behavior
have too high spontaneous curvature to allow the
formation of microemulsion droplets. If, instead Surfactant-polymer mixtures display a wide
of adding a cosurfactant, an oppositely charged range of interesting colloidal phenomena
surfactant is added, one should also be able to (Goddard and Anathapadmanabhan 1993;
reduce the spontaneous curvature, since some of Jönsson et al. 1998; Piculell and Lindman 1992)
the charge is neutralized and the hydrophobic and find a wide variety of technical and industrial S
portion of the film is increased. In the cases applications. They can be used for colloidal sta-
where these mixtures have a net charge, o/w bility, emulsification, structuring and suspending
microemulsion droplets have been observed properties, and control of rheological properties
(Li and Kunieda 2003). of diverse materials. The physicochemical prin-
However, adding more oppositely charged ciples underlying the phase behavior and proper-
surfactant to decrease Ho further (Ho will ties of surfactant-polymer systems are thus of
decrease as the surfactant mixture approaches enormous interest for various fields of science
equimolarity/neutrality) will lead to precipitation and technology.
if the surfactant headgroups are bulky, thus
preventing formation of bicontinuous and Critical Aggregation Concentration and
reversed microemulsions. This problem can be Mixed Micelle Formation
solved by adding a cosurfactant or by using If there is a net attractive interaction between
smaller headgroups, like ammonium instead of a water-soluble polymer and a surfactant, the
Surfactants, Phase a octane b octane

Behavior,
Fig. 28 Ternary phase
diagram for the
octylammonium octanoate-
octane-water system
(a) and pseudo ternary- L2 L2
phase diagrams for the
octylammonium octanoate-
sodium octanoate-octane-
water systems, (b)-(d). The
weight ratio of sodium water Octylammonium water Octylammonium octanoate
octanoate +sodium octanoate
octanoate to
octylammonium octanoate
is 1/9 in (b), 2/8 in (c), and c octane d octane
3/7 in (d) (Adapted from

Jönsson et al. 1991)
L2
L2
water Octylammonium octanoate water Octylammonium octanoate

+sodium octanoate +sodium octanoate
Surfactants, Phase
Behavior,
Fig. 29 Polymer effects
on surfactant self-
association: (a) cac for the
surfactant in the mixed
solution, (b) binding
isotherm for the surfactant
in presence (left) or absence
(right) of polymer
surfactant will display a critical aggregation micelle formation picture is more relevant in
concentration, cac, in the presence of polymer these systems. Electrostatic interactions between
that is lower than the cmc for the surfactant polymer and surfactant, in oppositely charged
alone in water (Fig. 29). The polymer-surfactant systems, will greatly enhance the association.
interactions can be pictured in two alternative Nonionic surfactants only rarely have distinct
ways: (1) as a strongly cooperative binding of interactions with water-soluble homopolymers,
the surfactant to the polymer chain and (2) as but they may interact with slightly hydrophobic
a micellization process of the surfactant in the homopolymers or if the polymer has grafted
vicinity of the polymer chains (pearl-necklace hydrophobic parts.
model). Both models are equally meaningful. If a water-soluble polymer bears
Ionic surfactants show a marked interaction hydrophobic chains grafted onto its
with most water-soluble homopolymers, in par- backbone – hydrophobically modified polymer
ticular anionic surfactants, while cationics have (HMP) – there will be a strong association
a weaker but still significant interaction. The between any type of surfactant, ionic or nonionic,
between surfactants and polymers. The knowl-
edge of the phase behavior of mixed polymer-
polymer systems can be qualitatively conveyed to
surfactant-polymer systems, as essentially the
same driving forces are present in both types of
systems (Piculell and Lindman 1992). Indeed,
a micelle can be regarded to some extent as
a macromolecule, except that it has variable
“degree of polymerization” since the micelle
aggregation number may change with different
factors such as temperature and ionic strength.
The main observations on phase behavior will
Surfactants, Phase Behavior, Fig. 30 Binding be briefly described.
isotherm for an ionic surfactant and a hydrophobically Just as for two different polymers in a solvent,
modified polymer the entropic driving force for mixing a polymer
and a surfactant is weak. Whether miscibility or
and polymer. The polymer alone may have phase separation takes place depends on the net
a tendency to self-associate (weakly) and interaction between the polymer and surfactant:
increase the viscosity of the solution. Upon addi- • No attractive interaction leads to separation
tion of surfactant, formation of mixed micelles into a phase enriched in polymer and another
between polymer and surfactant occurs. A cac is phase enriched in surfactant – segregative
also obtained for the surfactant-polymer mixture, phase separation.
and the binding model is rather useful for pictur- • A moderate attractive interaction leads to
ing the interaction. At low surfactant concentra- a complete miscibility with the formation of
tion, the mixed micelles will reinforce the cross- a single phase.
links between the polymers, and there is • A strong attractive interaction leads to the
a pronounced increase of viscosity. At suffi- formation of one phase concentrated in both
ciently high surfactant concentration, there is polymer and surfactant and another phase
about one polymer hydrophobe per mixed diluted in both components – associative
micelle, and thus, no cross-links exist between phase separation.
the chains; the viscosity is reduced. The tie lines of the two-phase region in ternary
For the case of the binding of an ionic surfac- phase diagrams will be different in the case of
tant to a water-soluble HMP, three distinct segregation or association, as can be seen in Fig. 31.
regions can be seen in a binding isotherm: The general trends for polymer-surfactant
a noncooperative, an anticooperative, and interactions depending on charge are summarized S
a cooperative region, as shown in Fig. 30. For in Table 3.
the corresponding homopolymer, where no If the polymers contain hydrophobic groups,
hydrophobic modification exists, there is only some special features appear. Thus, nonionic
cooperative binding, as mentioned above. If the HMP will even be more segregative with non-
HMP is a polyelectrolyte, the association with an ionic surfactants. Also, an HMP that contains
oppositely charged surfactant will have stronger charged groups will associate with a similarly
viscosity effects as compared to nonionic or charged surfactant. The case of HMPs and oppo-
similarly charged surfactants. sitely charged surfactants is discussed below.
Association Vs. Segregation Mixtures of Polyelectrolytes and Oppositely

Comparative phase-behavior studies of different Charged Surfactants
types of surfactant-polymer systems have greatly A polyelectrolyte dissolves in water into
improved our understanding of the interactions a polyion and n counterions. The dissociation of
Surfactants, Phase
Behavior, Fig. 31 The
two types of phase
separation in polymer-
surfactant-water systems:
segregation and
association. The tie lines in
the two-phase region are
indicated
the counterions gives a large contribution to the Surfactants, Phase Behavior, Table 3 Phase behavior
entropy of mixing, since n + 1 particles exist in trends in different types of surfactant-polymer mixtures
solution as compared with 1 for the dissolution of Type of
a nonionic polymer. Thus, polyelectrolytes surfactant/
polymer
derive their higher solubility in water, as com- mixture Phase behavior Driving forces
pared with uncharged polymers, from the entro- Nonionic/ Segregationa Low entropy of
pically favorable release of counterions (Piculell nonionic mixing, no net
and Lindman 1992). The confinement of the large attractions
number of counterions in one phase would Nonionic/ Total miscibility Entropic penalty for
involve a large entropic penalty. Hence, poly- ionic without salt counterion
confinement,
electrolytes tend to mix with both nonionic sur- hydrophobic
factants and nonionic polymers, in the absence of interactions
salt. This interaction is termed as the polyelectro- Ionic/ Association or Variable
lyte effect. Addition of high amounts of salt elim- nonionic segregation may
occur with salt
inates the gain in counterion entropy associated
Ionic/ionic Segregation Low entropy of
with miscibility, and therefore, the polyelectro- of similar mixing, no net
lyte usually segregates from the nonionic surfac- charge attraction, no
tant/polymer. Two phases can be formed with counterion entropy
similar concentrations of small ions. gain
Mixtures of polyelectrolytes and oppositely Ionic/ionic Association without Large counterion
of opposite salt entropy
charged surfactants show similar effects. charge Association- Salt eliminates
A strong associative phase separation occurs, miscibility- contribution of
that is, a concentrated phase forms which is in segregation with counterion entropy
equilibrium with a very dilute phase in both com- increasing salt
a
ponents. The driving force for this strong poly- Association may take place with slightly hydrophobic
mer-surfactant association is again the release of water-soluble polymers
the counterions, which involves a large entropy
gain. The charged surfactant micelles and the
polyelectrolytes are enriched with counterions at The polymer-surfactant complex that sepa-
the surface (counterion condensation). With rates in the concentrated phase can be a precipi-
association, the counterions of both solutes are tate (crystalline or amorphous), a coacervate (a
transferred to the bulk with a consequent large concentrated liquid phase), or a gel, depending on
increase in entropy. the specific composition of the polymer. In recent
The cac is much lower than the cmc for the years, the phase behavior of solid surfactant-
surfactant alone, by orders of magnitude in polyelectrolyte complexes (free of inorganic
the case of long-chain surfactants (more than counterions), called complex salts, has been stud-
12 carbon atoms). ied in both binary and ternary systems, because of
the importance of such complexes in a number of scanning calorimetry. A brief description of these
applications, such as surface coating and nano- techniques will be given below.
structured materials (Piculell et al. 2009). The These techniques can be further
complex is seen as an ionic surfactant for which complemented by other ones typically more
the counterions are polymeric (polyions). Studies directed toward structural characterization of
of alkyltrimethylammonium polyacrylates have aggregates. The latter comprise spectroscopic
shown that these amphiphiles also form micelles techniques (self-diffusion and relaxation NMR,
and mesophases in water, namely, cubic and hex- fluorescence, Raman, and infrared spectroscopy),
agonal phases. imaging methods (electron microscopy), scatter-
In conventional surfactant-polyion mixtures ing methods (light and neutron scattering),
of opposite charge, addition of salt has large and rheology.
consequences for phase behavior. The following
effects are typically observed: Visual Inspection and Birefringence
• Reduction of the electrostatic interactions Visual inspection usually gives a preliminary
between surfactant and polyelectrolyte, with indication of the number and type of phases pre-
an increase in the cac of the system. This sent in samples, as well as helping to ascertain
effect is opposite to that observed for surfac- phase equilibration. Further inspection of the
tant micellar solutions, where addition of salt sample between crossed polarizers can be easily
lowers the CMC and induces micellar growth. done in order to check for the occurrence of
• Stabilization of the surfactant micelles. At birefringence.
moderate addition of soluble salt, this may Birefringence is the phenomenon by which
render soluble the polymer-surfactant com- a ray of light decomposes into two rays
plex. Thus, at enough salt addition, there can (the ordinary ray and the extraordinary ray)
be formation of a one-phase region over the when it passes through optically anisotropic
entire phase diagram (total miscibility). materials. Additionally, the vibration plane of
• Further addition of salt to high concentrations light is altered. If the material has a single axis
totally eliminates the counterion entropy of anisotropy (it is uniaxial), the birefringence
effects, and thus, the system may evolve to phenomenon can be quantified according to
segregation if the polymer and surfactant are
intrinsically segregating. Both phases have Dn ¼ no ne (5)
large concentrations of small counterions,
and thus, no entropic penalty arises. in which no and ne are the refractive indices of the
polarization planes perpendicular (no) and paral-
lel (ne) to the major axis of anisotropy. Birefrin-
Investigation Methods for Phase gence is easily seen when an optically anisotropic S
Diagrams sample is placed between crossed polarizers. The
polarized light coming from the first polarizer
The investigation of the phase behavior of will not pass through the second one unless the
surfactant-containing systems, in particular the vibration plane is changed by the sample. This
detailed mapping of phase diagrams, requires happens if the sample is birefringent, and under
the use of several complementary methods, these conditions it will appear bright against
since no single method is usually conclusive for a dark background.
phase assignment or the determination of the The intensity of birefringence may be
phases at equilibrium. A set of experimental tech- observed when the sample is at rest (static bire-
niques can be considered more or less routine for fringence) or when it is subject to mechanical
these types of studies, namely polarized light shear (flow birefringence). During flow, the
microscopy, small-angle X-ray scattering, deute- structure of the fluid is momentarily altered
rium quadrupolar splitting NMR, and differential from that which exists at rest, and these structural
changes may enhance the optical anisotropy of to the propagation path of the radiation, as
the phase. Some liquid-crystalline phases have a consequence of their structural anisotropy.
different refractive indices along different space The refractive index is given by
directions and hence are also optically aniso-
tropic and birefringent. Placing a sample between c
n¼ (6)
crossed polarizers provides thus c0
a straightforward way of checking for optically
anisotropic liquid-crystalline phases. where c is the light speed propagation in vacuum
Phases that appear dark between crossed and c0 is the light speed propagation in a given
polarizers, such as micellar solutions and cubic medium. c0 is related to the substance by the
phases, are optically isotropic or nonbirefringent. following equation:
Phases that appear bright are optically anisotropic
c
or birefringent, and this is the case for hexagonal, c0 ¼ pffiffiffiffiffi (7)
lamellar, and nematic liquid-crystalline phases. em
Fluid or viscous isotropic samples that display
transient birefringence under crossed polarizers where e is the dielectric constant (or electric
when sheared are often composed of elongated permittivity) of the medium and m is the magnetic
micelles, which undergo partial alignment caused permeability of the sample. Since all the studied
by the shear. systems are nonmagnetic materials, then m 1
and so Eq. 6 can be simplified to
Polarized Light Microscopy pffiffi
For common observations, normal unpolarized n¼ e (8)
light is used. Nevertheless, when the birefrin-
gence of materials is to be studied under the If a given sample shows three-dimensional
microscope, polarized light is required. Polariza- structural homogeneity in term of its molecular
tion of light is achieved by placing a polarizing composition or molecular distribution, a single
filter, usually of crystalline nature with a highly dielectric constant will be observed for the
aligned structure that interferes with the electric whole sample. On the other hand, if a sample
component of the electromagnetic radiation. shows structural anisotropy, different regions
Only those waves that vibrate in a parallel direc- will have different refraction index values as
tion to the transmission axis (crystal axis) of the a consequence of the varying dielectric constants
polarizer are able to pass. observed in different directions.
Common polarizers allow the transmission of Liquid-crystalline phases exhibit characteris-
only ca. 25 % of the incident radiation but tic optical textures when observed in a polarized
achieve a polarization amount of almost 99 % of light microscope, as has been long recognized
the transmitted light. When polarized light prop- (Rosevear 1954, 1968). These textures result
agates through an anisotropic material, the vibra- mainly from the birefringence derived from
tion plane is altered, and the waves no longer structural defects in the confined liquid-
vibrate parallel to each other. If a second crystalline phase, and because these defects
polarizer is placed between the sample and the depend on the nature of the phase, the overall
eyepiece of the microscope in a position such that texture is characteristic of that phase and helps
the transmission axis is in a 90 position regard- in its assignment (Demus et al. 1998). Factors
ing the first polarizer, the only radiation that will such as the film thickness, alignment of the sam-
be able to cross it will be that that has undergone ple relative to the slide and coverslip, and
polarization by the sample. heating-cooling cycling also affect the observed
The light vibration plane shift by anisotropic textures.
substances is due to the existence of different Hexagonal liquid-crystalline phases display
refractive indices in the same material according easily recognizable fanlike and angular textures
Surfactants, Phase
Behavior,
Fig. 32 Striated
nongeometric texture
(a) and oily streaks texture
(b), characteristic of
a hexagonal and lamellar
liquid-crystalline phase,
respectively
(magnification: 100)
a b
cr lam cub′′ cub′ + w cr Lα cub′ cub′ + w

cub′ cub′′
Surfactants, Phase Behavior, Fig. 33 Phase scanning experiment for monoolein in water at 25 C: (a) image under
polarized light and (b) image under normal light. The arrow indicates the downward gradient of surfactant
in some cases or nongeometric, striated textures into the solid film. The sequence of different
in others (Fig. 32a). Lamellar liquid-crystalline phases is formed, identified through characteris-
phases usually show mosaic textures at high sur- tic birefringent bands (for anisotropic phases) or
factant concentration; at lower concentration, dark bands (for isotropic ones). In Fig. 33 one
they often appear with characteristic oily streaks such experiment is presented for monoolein in
in a planar (homeotropic) matrix (Fig. 32b). water at 25 C, showing a sequence of cubic and
Lamellar dispersions containing large lamellar phases (cf. phase diagram in Fig. 10).
multilamellar vesicles (onions) show characteris- Here a combination of observations under polar-
tic focal conic domains (Maltese crosses). Isotro- ized and normal light is essential for the detection S
pic phases appear as dark under polarized light, of the optically isotropic cubic phases and their
with cubic liquid crystals being distinguished by phase boundaries.
their high viscosity from solution phases. Light
microscopy can also be very effective in the Differential Interference Contrast with
detection of crystals alone or in coexistence Video-Enhancement
with other phases. Many surfactant aggregates like vesicles and
The phase scanning experiment, in connection microtubules are inherently of low contrast with
with polarized light microscopy, is a very useful the dispersing medium (water), and so they can-
and simple method, based on the formation of not be easily distinguished from the background
a surfactant concentration gradient upon con- if normal light is used in a light microscope.
trolled hydration of a powder sample (Laughlin Besides, the use of visible light puts a lower
1994). A drop of water is placed at one end of limit to resolution. If a microscope is equipped
a coverslip and, by capillary action, seeps slowly with polarizers, differential interference contrast
Surfactants, Phase Behavior, Fig. 34 Light micro- (b) under crossed polarizers, and (c) under differential
graphs for the same surfactant-water sample showing interference constrast. Scale bar: 50 mm
multilamellar vesicle structures: (a) under bright field,
lenses, and a video camera connected to the into a single one. At this point, the two compo-
microscope, both contrast and resolution can be nents propagating perpendicularly will interfere
improved in what is termed differential interfer- due to the phase shifts experienced when cross-
ence contrast (DIC) video-enhanced light micros- ing the sample through different regions. This
copy (Kachar et al. 1984). The DIC lenses per se gives the appearance of a three-dimensional
enhance contrast between object and background physical relief corresponding to the variation of
and allow a resolution down to about 0.5 mm. optical density of the sample, where lines and
Further contrast enhancement comes from the edges are emphasized, resulting in enhanced
video camera, which responds to small differ- contrast.
ences in light intensity no matter what the back- A comparison between the different imaging
ground light level is, allowing the detection of techniques is provided in Fig. 34 where the
aggregates not visible to the naked eye (resolu- contrast enhancement provided by the DIC sys-
tion limit of 0.3 mm). It should be noted that tem is clearly visible when comparing to bright
mathematical processing of the digital image field imaging.
can extend this technique to its absolute limits,
giving a resolution limit of 0.1 mm. Small-Angle X-Ray Scattering
The contrast system works by separating the In lyotropic liquid-crystalline phases, typical
incident polarized light into two beams, by use of structural length scales are in the order of
a Nomarski prism (situated between the polarizer 1–50 nm, and thus, small-angle X-ray scattering
and the condenser). Since these two beams (SAXS) can often be used to determine the type
vibrate at an angle of 90 relative to each other, of phase present in a sample (Lindner and
their waves do not experience interference. Addi- Zemb 2002). Different liquid-crystalline phases
tionally, they also propagate with a minor path (lamellar, hexagonal, cubic, and others),
difference relative to one another (always smaller possessing some type of long-range structural
than the lenses resolving power). The perpendic- order, show well-defined Bragg peaks
ular components of the split wave cross the sam- (reflections), resulting from the constructive
ple in different regions due to the path difference; interference of reflected X-rays. Given the typical
hence, they will experience different propagation length scales, the Bragg peaks appear at low
conditions (thickness of the sample, refractive scattering angles, and in practice only a limited
index) that will influence their transmission. number of peaks are observed in the
Immediately after the objective lenses, above diffractogram. SAXS also allows the determina-
the revolving nosepiece, a second Nomarski tion of length scales associated to the lattice
prism is placed that recombines the split waves periodicity in the liquid crystals.
Surfactants, Phase
Behavior,
Fig. 35 Schematic
representation of Bragg’s
law of diffraction for
parallel planes. l is the
wavelength of the
radiation; d is the distance
between planes; y is the
scattering angle; and A, B,
C, and D are points used for
the calculation of the path
length
Bragg’s Law
Figure 35 shows a graphical representation of the
Bragg’s law of diffraction and how constructive
interference allows the extraction of structural
data from the diffraction patterns. For simplicity,
we will consider the case of parallel plane waves.
The extra path length traveled by the radiation of
l wavelength and scattering angle y is given by

y
AB þ BC ¼ 2d sin (9)
2
The scattered waves from the different planes

are in phase if the path length AB + BC is Surfactants, Phase Behavior, Fig. 36 Schematic rep-
resentation of a hexagonal phase with the characteristic
a multiple of l (AB + BC ¼ n l). Hence, Eq. 9 lattice spacing d, nearest neighbor distance d*, and rod
takes the form known as Bragg’s law: radius R

y
nl ¼ 2d sin (10)
2 rods d* is d* ¼ 2d/√3. A more complete list of n
periodicities for different liquid-crystalline S
Combining it with the modulus of the scatter- phases is given in Table 4.
ing vector q yields the Bragg’s law in terms of q: Thus, in a diffraction measurement, the differ-
ent planes give rise to specific Bragg peaks. Since
q ¼ n2p=d (11) different liquid-crystalline phases give rise to
different diffraction patterns (due to the different
where d is the repeat distance between the planes reflection sequences for n) from the observation
of a lattice and n is the order of the scattering of this pattern, the nature of the phase can be
peak. For the case of parallel planes, such as identified. It should just be added that as
a lamellar phase, n is a series of integers q increases, the higher order reflections are
(1:2:3:4:. . .), whereas for a hexagonal phase more difficult to observe. Through Eq. 11, the
n ¼ 1:√2:2:√7:3 (Figs. 36 and 37). In a hexagonal lattice spacing can also be determined. If
phase, d is also the distance between adjacent a dilution series is studied, additional structural
planes of rods, whereas the distance between the parameters, such as the radius R of the rods in the
Surfactants, Phase
Behavior,
Fig. 37 Schematic
representation of SAS
curves with reflections
corresponding to lamellar
(straight), hexagonal
(dashed), and simple cubic
(dotted line) lattices
Surfactants, Phase Simple Cubic Cubic Cubic Cubic

Behavior, Table 4 Ratios Lamellar Hexagonal cubic Pn3m Fd3m Im3m Ia3d
between q values and liquid
crystal identity √1 √1 √1 √2 √3 √2 √6
√4 √3 √2 √3 √8 √4 √8
√9 √4 √3 √4 √11 √6 √14
√16 √7 √4 √6 √12 √8 √16
√25 √9 √5 √8 √16 √10 √20
√36 √12 √6 √9 √19 √12 √22
√49 √13 √8 √10 √24 √14 √24
√64 √16 √9 √11 √27 √16 √26
hexagonal phase, or the lamellar thickness d of Ka lines of a copper anode are employed, as these
a lamellar phase can be determined: give a reasonable transmission for organic mate-
rial and solvents for samples with 1-mm thick-
sffiffiffiffiffiffiffiffiffiffiffiffiffi ness. X-rays can also be produced with very high
2f
R¼ pffiffiffi d (12) intensities at synchrotron sources, where the
p 3 beam intensity (and thus the scattered intensity)
is much higher than in laboratory apparatuses,
d ¼ fd (13) together with a better q resolution.
Commercial SAXS instruments can be used in Deuterium Quadrupolar-Splitting NMR

the laboratory. Conventional laboratory sources The structure, dynamics, and phase behavior of
are sealed tubes and rotating anodes. Usually the surfactant systems can be elucidated to a great
Surfactants, Phase Behavior, Fig. 38 Deuterium (c) coexistence of one (or two) isotropic phases and
NMR spectra for different compositions in a surfactant- a hexagonal phase, (d) splittings of two anisotropic liquid
oil-water system, the calcium octyl sulfate-decanol-water crystals (hexagonal and lamellar phases), (e) three-phase
system (Khan et al. 1982): (a) singlet of an isotropic phase sample with one isotropic phase (solution) and two aniso-
(micellar solution), (b) splitting of a hexagonal phase, tropic liquid crystals (hexagonal and lamellar)
extent by use of NMR techniques (Söderman and the samples, but perdeuterated surfactants or spe-
Stilbs 1994). The NMR parameters obtained are cifically deuterated amphiphiles can also be used.
both static properties – such as chemical shifts The physical basis of this method lies in
and quadrupolar splittings – and dynamic quanti- the fact that the 2H nucleus has a spin quantum
ties, such as relaxation rates and self-diffusion number I ¼ 1, and therefore, it possesses an
coefficients. However, the information to electric quadrupole moment. In anisotropic
be extracted almost always regards static liquid-crystalline phases, the interaction of the
properties – examples are micellar size and quadrupole moment with nearby electric field
shape, degree of micellization and counterion gradients gives rise to a splitting of the 2H NMR
binding, and microstructure of microemulsions. signal into 2I peaks. In isotropic phases, the inter-
Typical magnetic nuclei investigated in surfac- action is averaged to zero by the rapid isotropic
tant systems are 1H, 2H, and 13C, present in the molecular motions, and a single signal results. In
alkyl chains; 14N and 31P in the headgroups; multiphase samples, due to the slow deuteron
and many counterions such as 23Na, 35,37Cl, exchange (on the NMR time scale) between the
and 79,81Br. phases, the resultant 2H NMR spectrum is S
In what concerns phase behavior more specif- a superposition of the spectra of the independent
ically, the 2H quadrupolar-splitting method for phases. The magnitude of the 2H quadrupolar
phase-diagram studies is the most relevant and splittings (Dn), measured as peak-to-peak separa-
informative method. It is useful for the assign- tion in Hz, depends on the degree of anisotropy of
ment of phases, in combination with polarizing the phase. For symmetry reasons, it is consider-
microscopy and SAXS, and for delineating ably higher in lamellar phases than in hexagonal
multiphase regions involving two or more phases, and under identical local molecular con-
liquid-crystalline phases, which are often very ditions and composition, it should be for
difficult to separate macroscopically. It is also a lamellar phase twice as large as for
effective in the detection of minute amounts of a hexagonal one (Fig. 38).
isotropic phases present in heterogeneous sam- In surfactant-water systems, water associates
ples. In general, deuterated water (D2O) is used in to the headgroups of the amphiphilic molecules,
which are organized into different aggregate keep the sample and a reference at the same
structures. A two-site model can thus be temperature when heated at a constant rate.
applied with free and bound deuterated water The thermograms obtained in a DSC measure-
molecules. Assuming a fast exchange between ment are expressed in terms of heat
the sites and Dn ¼ 0 for free water, the observed flow as a function of temperature, and it is there-
Dn for a given liquid-crystalline phase is related fore appropriate to convert heat flow,
to the molar ratio between surfactant and water F ¼ (∂qdiff/∂t)p, in differential heat capacity at
through constant pressure, Cp,diff ¼ (∂qdiff/∂T)p. The term
differential relates to the fact that all the measure-
XS ments are done relative to a reference. To obtain
Dv ¼ vQ;D Sn (14)
Xw Cp,diff, it is only necessary to divide the heat flow
by the heating rate, b ¼ (dT/dt)p.
where nQ,D is the deuteron quadrupolar coupling Most of the phase transitions are first order, so
constant; n is the average hydration number of the in pure substances an infinite value of Cp should
surfactant; XS and Xw are, respectively, the molar exist at the transition. In practice, there is a peak
fractions of surfactant and water XS/Xw; and S is of finite dimensions, which is in principle
a local order parameter. As can be seen from narrower for lower heating rates. In other words,
Eq. 14, this technique can be further used for low heating rate measurements yield higher res-
the characterization of a given medium in terms olutions. In contrast, higher heating rates might
of its structural anisotropy. For instance, the enhance the sensitivity because F is proportional
swelling behavior of a lamellar liquid crystal to b. Care should be taken to certify that the
can be studied from the XS/Xw relation, and mod- transition temperature and enthalpy are indepen-
ifications in the local order of the aggregates can dent of the heating rate. Otherwise, the transition
be inferred through the S dependence. is controlled kinetically, and kinetic models
have to be used in the data analysis. Sometimes
Differential Scanning Calorimetry second-order transitions also occur in soft
Differential scanning calorimetry or DSC is materials. These are more difficult to identify,
a technique widely used in the study of phase and it is necessary to use more precise techniques
transitions, since it enables the determination of (such as adiabatic scanning calorimetry) for
the number of phase transitions and their thermo- this purpose.
dynamic parameters, namely, temperature and The enthalpy change associated with
enthalpy and entropy change (if the transforma- a temperature change is the integral of the heat
tion is reversible). Therefore, this technique capacity between the start and end temperatures:
can also be applied to the study of phase behavior,
Tð2
for instance, in the determination of the
thermodynamic parameters associated with the DH ¼ Cp;diff dT (15)
gel-to-liquid crystal-phase transition in bilayers T1
(Blandamer et al. 1995), and the delineation of
phase boundaries (Chernik 1995). Conversely, in a DSC measurement, the
In a typical DSC measurement, the sample and enthalpy change (DH) is the integral of the dif-
a reference are heated at a constant heating rate, ference between the heat capacity associated with
in such a way that the temperatures of both sam- the peak and the baseline, as shown in Fig. 39.
ple and reference are practically the same. When The corresponding temperature associated
the sample undergoes physical or chemical with the transition is usually defined as the onset
changes, it releases or absorbs heat, which will temperature of the peak. If the process is revers-
manifest in an increase or decrease of the sample ible, the change in entropy, DtrsS, associated with
temperature, respectively. This technique there- the phase transition is easily calculated through
fore measures the differential heat flow needed to DtrsH/Ttrs.
a b
Tpeak
tilted rippled
Cp,diff /J·K−1
gel phase (Lb¢ ) gel phase (Pb¢ )
0.1 J·K−1
liquid-crystalline
Cp,diff
phase (La )
ΔtransH
Ti Tf
25 30 35 40 45 50 55 60
Tonset T/K q/°C
Surfactants, Phase Behavior, Fig. 39 DSC studies in liposomes, with peaks associated with the different phase
surfactant systems: (a) endothermic peak associated with transitions in the bilayer (cf also Fig. 18)
a phase transition; (b) example of a thermogram for DPPC
Conclusion systems. Through the manipulation of simple

experimental parameters (chemical structure,
Surfactants in aqueous and nonaqueous solvents composition, and temperature), one can then
form a wealth of colloidal structures, from orga- change self-assembly and macroscopic behavior
nized solutions to various types of liquid- in a controlled way. This fact has large repercus-
crystalline phases and crystal hydrates. When sions for practical uses of surfactant-based
other components are added to a surfactant- colloidal systems, lying on the basis of many
water system – for instance, a second amphiphile, products and processes currently employed in
an oil, or a polymer – more or less complex phase technology and everyday life. In a reciprocal
behavior emerges, depending on the number and way, observations on phase behavior can give
nature of components present. Phase equilibria important clues on how molecular aggregates
can be experimentally determined by comple- interact in solution.
mentary experimental techniques and depicted
through phase diagrams, in conformity to basic S
thermodynamics. Cross-References
The shape and size of self-assembled aggre-
gates can be rationalized to a great extent through ▶ Dispersions
simple concepts like the surfactant packing ▶ Electrical Double Layers
parameter and spontaneous curvature, while ▶ Emulsion Formation
phase transitions and phase equilibria require in ▶ Emulsion Stabilization
nontrivial manner the consideration of both intra- ▶ Food Colloids
aggregate interactions and long-range ▶ Microemulsions
interaggregate interactions. By studying phase ▶ Polymeric Surfactant
behavior and phase microstructure at ▶ Surface Forces
a fundamental level, one can understand, or at ▶ Surfactant Self-assembly
least keep track of, the macroscopic features and ▶ Surfactants
properties displayed by many surfactant-based ▶ Suspensions
References and sodium dodecyl-sulfate (SDS). J Phys Chem

B 97:13792–13802
Alexandridis P (1998) Structural polymorphism of poly Holland PM, Rubingh DN (eds) (1992) Mixed surfactant
(ethylene oxide)-poly(propylene oxide) block copoly- systems. ACS, Washington, DC
mers in nonaqueous polar solvents. Macromolecules Israelachvili J (1985) Intermolecular and surface forces
31:6935–6942 with application to colloidal and biological systems.
Alexandridis P, Zhou DL, Khan A (1996) Lyotropic Academic, London
liquid crystallinity in amphiphilic block copolymers: Israelachvili JN, Mitchell DJ, Ninham BW (1976) Theory
temperature effects on phase behavior and structure of self-assembly of hydrocarbon amphiphiles into
for poly(ethylene oxide)-b-poly(propylene oxide)-b- micelles and bilayers. J Chem Soc – Faraday Trans II
poly(ethylene oxide) copolymers of different compo- 72:1525–1568
sition. Langmuir 12:2690–2700 Jokela P, Jönsson B, Khan A (1987) Phase equilibria of
Almgren M (2000) Mixed micelles and other structures in catanionic surfactant-water systems. J Phys Chem
the solubilization of bilayer lipid membranes by sur- 91:3291–3298
factants. Biochim Biophys Acta-Biomembr Jönsson B, Jokela P, Khan A, Lindman B,
1508:146–163 Sadaghiani A (1991) Catanionic surfactants:
Arleth L, Pedersen JS (2001) Droplet polydispersity and phase behavior and microemulsions. Langmuir 7:
shape fluctuations in AOT bis(2-ethylhexyl) 889–895
sulfosuccinate sodium salt microemulsions studied Jönsson B, Lindman B, Holmberg K, Kronberg B (1998)
by contrast variation small-angle neutron scattering. Surfactants and polymers in aqueous solution, 1st edn.
Phys Rev E 63:1–18 Wiley, Chichester
Arleth L, Marčelja S, Zemb T (2001) Gaussian random Kachar B, Evans DF, Ninham BW (1984) Video enhanced
fields with two level-cuts – model for asymmetric differential interference contrast microscopy:
microemulsions with nonzero spontaneous curvature? a new tool for the study of association colloids
J Chem Phys 115:3923–3936 and prebiotic assemblies. J Colloid Interface Sci
Balmbra RR, Clunie JS, Goodman JF (1969) Cubic meso- 100:287
morphic phases. Nature 222:1159–1160 Kaler EW, Murthy AK, Rodriguez BE, Zasadzinski JAN
Blandamer MJ, Briggs B, Cullis PM, Engberts J (1995) (1989) Spontaneous vesicle formation in aqueous mix-
Gel to liquid-crystal transitions in synthetic amphi- tures of single-tailed surfactants. Science
phile vesicles. Chem Soc Rev 24:251–257 245:1371–1374
Caria A, Khan A (1996) Phase behavior of catanionic Kekicheff P, Grabiellemadelmont C, Ollivon M (1989)
surfactant mixtures: sodium bis(2-ethylhexyl) Phase-diagram of sodium dodecyl-sulfate water-sys-
sulfosuccinate-didoceyldimethylammonium bromide- tem. 1. A calorimetric study. J Colloid Interface Sci
water system. Langmuir 12:6282–6290 131:112–132
Chernik GG (1995) Phase-equilibria in phos- Khan A, Marques E (1998) Catanionic surfactants. In:
pholipid water-systems. Adv Colloid Interface Sci Robb ID (ed) Specialist surfactants. Blackie Academic
61:65–129 & Professional, London, pp 37–76
Demus D, Goodby J, Gray GW, Spiess H-W, Vill V (1998) Khan A, Fontell K, Lindblom G, Lindman B (1982)
Handbook of liquid crystals. Wiley-VCH Verlag Liquid crystallinity in a calcium surfactant
GmbH, Weinheim system – phase-equilibria and phase structures in the
Evans DF, Wennerström H (1999) The colloidal domain: system calcium octyl sulfate decan-1-ol water. J Phys
where physics, chemistry, biology and technology Chem 86:4266–4271
meet, 2nd edn. Wiley, New York Kratohvil J (1986) Size of bile salt micelles: techniques,
Fontell K (1990) Cubic phases in surfactant and surfac- problems and results. Adv Colloid Interface Sci
tant-like lipid systems. Colloid Polym Sci 26:131–154
268:264–285 Larsson K (1994) Lipids – molecular organization, phys-
Goddard ED, Anathapadmanabhan KP (1993) Interactions ical functions and technical applications. The Oily
of surfactants with polymer and proteins. CRC Press, Press, Dundee
Boca Raton Lasic DD (1993) Liposomes: from physics to applications.
Grabiellemadelmont C, Perron R (1983) Calorimetric Elsevier, Amsterdam
studies on phospholipid-water systems. 1. DL- Laughlin RG (1994) The aqueous phase behavior of
dipalmithoylphosphatidylcholine (DPPC)-water sys- surfactants. Academic, London
tem. J Colloid Interface Sci 95:471–482 Li XF, Kunieda H (2003) Catanionic surfactants:
Helfrich W (1973) Elastic properties of lipid bilayers: microemulsion formation and solubilization. Curr
theory and possible experiments. Z Naturforsch Opin Colloid Interface Sci 8:327–336
28:693–703 Li XF, Ueda K, Kunieda H (1999) Solubilization and
Herrington KL, Kaler EW, Miller DD, Zasadzinski JA, phase behavior of microemulsions with mixed
Chiruvolu S (1993) Phase-behavior of aqueous mix- anionic-cationic surfactants and hexanol. Langmuir
tures of dodecyltrimethylammonium bromide (DTAB) 15:7973–7979
Lindman B, Wennerström H (1980) Micelles: topics in Rydhag L, Stenius P, Odberg L (1982) Phase-
current chemistry. Springer, Berlin equilibria and vesicle stability in dimyristoyl phospha-
Lindner P, Zemb T (2002) Neutrons, X-rays and light: tidylcholine-cetyltrimethylammonium bromide sys-
scattering methods applied to soft condensed matter, tems. J Colloid Interface Sci 86:274–276
1st edn. Elsevier Science B.V, Amsterdam Safran SA (1999) Curvature elasticity of thin films. Adv
Madenci D, Egelhaaf SU (2010) Self-assembly in aqueous Phys 48:395–448
bile salt solutions. Curr Opin Colloid Interface Sci Seddon JM, Templer RH (1995) Polymorphism of lipid-
15:109–115 water systems. In: Lipowsky R, Sackmann E (eds)
Marques E, Khan A, Miguel MG, Lindman B (1993) Self Handbook of biological physics. Elsevier Science
assembly in mixtures of a cationic and an anionic B.V, Amsterdam
surfactant: the sodium dodecyl sulfate- Shinoda K (1974) Principles of solution and solubility.
didodecyldimethylammonium bromide-water system. Marcel Dekker, New York
J Phys Chem 97:4729–4736 Silva BFB, Marques EF, Olsson U (2011) Aqueous
Marques EF, Regev O, Khan A, Miguel MG, Lindman B phase behavior of salt-free catanionic surfactants:
(1998) Vesicle formation and general phase the influence of solubility mismatch on spontaneous
behavior in the catanionic mixture SDS-DDAB- curvature and balance of forces. Soft Matter
water. The anionic-rich side. J Phys Chem B 7:225–236
102:6746–6758 Sjobom MB, Marques EF, Edlund H, Khan A (2005)
Marques EF, Edlund H, La Mesa C, Khan A (2000) Liquid Phase equilibria of the mixed didodecyl-
crystals and phase equilibria binary bile salt-water dimethylammonium bromide-sodium taurodeox-
systems. Langmuir 16:5178–5186 ycholate-water system with a large solution
Marques EF, Regev O, Khan A, Lindman B (2003) Self- region. Colloids Surf-Physicochem Eng Asp 269:
organization of double chained and pseudodouble 87–95
chained surfactants: conterion and geometry effects. Small D (1971) The physical chemistry of cholanic acids.
Adv Colloid Interface Sci 100–102:83–104 In: Nair P, Kritchevsky D (eds) The bile acids: chem-
Nagle JF, Tristram-Nagle S (2000) Structure of lipid istry, physiology and metabolism. Plenum Press,
bilayers. Biochim Biophys Acta-Rev Biomembr New York
1469:159–195 Söderman O, Stilbs P (1994) NMR studies of complex
Nilsson F, Soderman O, Johansson I (1996) Physical- surfactant systems. Prog Nucl Magn Reson Spectrosc
chemical properties of the n-octyl beta-D-gluco- 26:445–482
side/water system. A phase diagram, self-diffusion Strey R, Schomacker R, Roux D, Nallet F, Olsson U
NMR, and SAXS study. Langmuir 12:902–908 (1990) Dilute lamellar and L3 phases in the binary
Ogino K, Masahiko A (eds) (1993) Mixed surfactant sys- water-C12E5 system. J Chem Soc – Faraday Trans
tems. Marcell Dekker, New York 86:2253–2261
Persson G, Edlund H, Hedenstrom E, Lindblom G (2004) Stubenrauch S (2009) Microemulsions: background, new
Phase behavior of 1-alkylpyridinium octane-1-sulfo- concepts, applications, perspectives. Wiley, Chichester
nates. Effect of the 1-alkylpyridinium counterion size. Tadros TF (2005) Applied surfactants. Wiley-VCH
Langmuir 20:1168–1179 Verlag GmbH, Weinheim
Piculell L, Lindman B (1992) Association and segregation Tanford C (1980) The hydrophobic effect: formation
in aqueous polymer/polymer, polymer/surfactant, of micelles and biological membranes. Wiley,
surfactant/surfactant mixtures – similarities and New York
differences Adv. Colloid Interface Sci 41:149–178 Tiddy GJT (1980) Surfactant-water liquid-crystal phases.
Piculell L, Norrman J, Svensson AV, Lynch I, Phys Rep-Rev Sect Phys Lett 57:1–46
Bernardes JS, Loh W (2009) Ionic surfactants with Tondre C, Caillet C (2001) Properties of the amphiphilic S
polymeric counterions. Adv Colloid Interface Sci films in mixed cationic/anionic vesicles: a comprehen-
147–148:228–236 sive view from a literature analysis. Adv Colloid
Qiu H, Caffrey M (2000) The phase diagram of the Interface Sci 93:115–134
monoolein/water system: metastability and equilib- Warr GG, Sen R, Evans DF, Trend JE (1988)
rium aspects. Biomaterials 21:223–234 Microemulsion formation and phase-behavior of
Rogers J, Winsor PA (1969) Change in optic sign of dialkyldimethylammonium bromide surfactants.
lamellar phase (G) in aerosol OT/wtare system with J Phys Chem 92:774–783
composition or temperature. J Colloid Interface Sci Yatcilla MT, Herrington KL, Brasher LL, Kaler EW,
30:247 Chiruvolu S, Zasadzinski JA (1996) Phase behavior
Rosevear FB (1954) The microscopy of the liquid crystal- of aqueous mixtures of cetyltrimethylammonium
line neat and middle phases of soaps and synthetic bromide (CTAB) and sodium octyl sulfate (SOS).
detergents. J Am Oil Chem Soc 31:628–639 J Phys Chem 100:5874–5879
Rosevear FB (1968) Liquid crystals – mesomorphic Zemb T, Gazeau D, Dubois M, Gulikkrzywicki T (1993)
phases of surfactants compositions. J Soc Cosmet Critical behavior of lyotropic liquid-crystals.
Chem 19:581–594 Europhys Lett 21:759–766
S 1334 Suspension Concentrates
(2) a good dispersing agent to break such aggre-

Suspension Concentrates gates or agglomerates into smaller units and subse-
quently help in the milling process (one usually
Tharwat Tadros aims at a dispersion with a volume mean diameter
Wokingham, Berkshire, UK of 1–2 mm), and (3) it should provide good stability
in the colloid sense (this is essential for maintaining
the particles as individual units once formed). Pow-
Synonyms erful dispersing agents are particularly important
for the preparation of highly concentrated suspen-
Solid/liquid dispersions sions (sometimes required for seed dressing). Any
flocculation will cause a rapid increase in the vis-
Definition cosity of the suspension, and this makes the wet
milling of the agrochemical a difficult process. The
Suspension concentrates (SCs) are dispersions of resulting dispersion, referred to as mill base, must
agrochemical solids in aqueous solution. Several be stabilized against flocculation, Ostwald ripening
advantages can be quoted for the advantages of (crystal growth), and sedimentation. Flocculation
formulating a solid agrochemical as a SC. Firstly, can be prevented by the use of dispersants that
one may control the particle size by controlling provide an effective repulsive barrier. Ostwald rip-
the milling conditions and proper choice of the ening is reduced by the addition of materials that
dispersing agent. Secondly, it is possible to incor- strongly adsorb on the surface of the particles, thus
porate high concentrations of surfactants in the blocking the “active sites” and hence preventing
formulation which is sometimes essential for transfer of the molecules from the small to the large
enhancing wetting, spreading, and penetration. particles. Sedimentation is prevented by addition of
Stickers may also be added to enhance adhesion “thickeners” (such as xanthan gum or finely
and in some cases to provide slow release. In divided inert materials such as clays and oxides)
recent years, there has been considerable research which produce a three-dimensional “gel network”
into the factors that govern the stability of sus- in the continuous phase with a very high residual
pension concentrates. The theories of colloid sta- (zero-shear) viscosity.
bility could be applied to predict the physical
states of these systems on storage. In addition,
analysis of the problem of sedimentation of SCs Cross-References
at a fundamental level has been undertaken. Since
the density of the particles is usually larger than ▶ Agrochemical Formulations
that of the medium (water), SCs tend to separate ▶ Colloid stability
as a result of sedimentation. The sedimented par- ▶ Dispersant
ticles tend to form a compact layer at the bottom ▶ Surfactants
of the container (sometimes referred to as clay or ▶ Wetting
cake), which is very difficult to redisperse. It is,
therefore, essential to reduce sedimentation and
formation of clays by incorporation of an Suspensions
antisettling agent. Suspension concentrates are
usually formulated using a wet milling process Tharwat Tadros
which requires the addition of a surfactant/ Wokingham, Berkshire, UK
dispersing agent. The latter should satisfy the
following criteria: (1) A good wetting agent for
the agrochemical powder (both external and Synonyms
internal surfaces of the powder aggregates or
agglomerates must be spontaneously wetted), Suspensions (solid/liquid dispersions)
Suspensions 1335 S
Keywords agrochemicals, and ceramics, or hydrophilic, for
example, silica, titania, and clays. The liquid can
Assessment; bulk properties; charge separation; be aqueous or nonaqueous. It is essential to
dispersants; dispersion; double layers; electro- understand the process of dispersion of powders
static stabilization; nucleation and growth; prep- in liquids (to prepare suspensions) at
aration of suspensions; rheological properties; a fundamental level: “Dispersion is a process
sedimentation; solid/liquid dispersions; steric whereby aggregates and agglomerates of pow-
stabilization; surfactants ders are dispersed into ‘individual’ units, usu-
ally followed by a wet milling process (to
subdivide the particles into smaller units) and
Definition stabilization of the resulting dispersion against
aggregation and sedimentation” (Parfitt 1977;
Suspensions are dispersions of solid particles Tadros 1987).
(the disperse phase) in aqueous or nonaqueous In this entry, I will describe the main pro-
medium (the continuous phase). They are pre- cedures of the preparation of solid/liquid disper-
pared by condensation methods (nucleation and sions (suspensions). This is followed by sections
growth) or dispersion of the bulk solid followed on the origin of charge in suspension particles,
by wet milling. The control of their properties is the electrical double layer, and the concept of
very important for the long-term physical stabil- zeta potential. The adsorption of surfactants and
ity as well as on application. Two stabilizing polymers at the solid/liquid interface is
mechanisms can be applied, namely, electro- described. The stabilization of suspensions by
static repulsion (resulting from the interaction surfactants both electrostatically and sterically
of double layers) or steric repulsion (due to the will be described at a fundamental level. This is
presence of adsorbed surfactants or polymers). followed by a section on the different states of
Several states are reached depending on the suspensions on standing and how the three-
particle size and shape distribution, the magni- dimensional structures formed can be accounted
tude of interaction forces, and the presence of for in terms of the various interaction forces that
rheology modifiers (thickeners). The properties occur between the particles compact sediments
of the suspension can be assessed by measuring (clays or cakes). A summary of the various pro-
the rate of sedimentation, the rate of floccula- cedures that may be used for characterization of
tion, and the rate of crystal growth. The rheo- suspensions will be given. Particular attention
logical properties of the concentrated will be given to the application of rheological
suspension give information on the long-term techniques for the assessment and prediction of
physical stability of the suspension. Various the long-term stability of suspensions. This is
methods can be applied to prevent sedimenta- followed by a section on settling of suspensions S
tion of suspensions and formation of “hard” and prevention of formation of compact sedi-
(dilatant) sediments. ments (clays or cakes). The methods that can
be applied for prevention of settling are briefly
described. As far as possible, the fundamental
Overview principles involved in each of the above pro-
cesses will be given.
Suspensions are solid/liquid dispersions that find
application in almost every industrial prepara-
tion, for example, paints, dyestuffs, paper Preparation of Suspensions
coatings, printing inks, agrochemicals, pharma-
ceuticals, cosmetics, food products, detergents, There are two main processes for the preparation
and ceramics. The powder particles can be hydro- of solid/liquid dispersions. The first depends on
phobic, for example, organic pigments, the “buildup” of particles from molecular units,
S 1336 Suspensions
that is, the so-called condensation method, which S=0

involves two main processes, namely, nucleation
Increasing S
and growth. In this case, it is necessary first to ΔG
prepare a molecular (ionic, atomic, or molecular)
distribution of the insoluble substances; then by ΔG*
changing the conditions, precipitation is caused
leading to the formation of nuclei that grow to the
particles in question. In the second procedure, r* r
usually referred to as a dispersion process, larger Irreversible
“lumps” of the insoluble substances are growth
subdivided by mechanical or other means into
smaller units. The role of surfactants in the prep-
aration of suspensions by these two methods will Suspensions, Fig. 1 Schematic representation of the
effect of supersaturation on particle growth
be described separately.

4p r3 r S
DG ¼ 4p r2 s RT ln (2)
Condensation Methods: Nucleation 3M So
and Growth
r is the density, R is the gas constant, and T is the
Nucleation is the spontaneous process of the absolute temperature.
appearance of a new phase from a metastable In the initial stages of nucleation, DGs
(supersaturated) solution of the material in ques- increases faster with increase in r when compared
tion (Parfitt 1973). The initial stages of nucleation to DGv and DG remains positive, reaching
result in the formation of small nuclei where the a maximum at a critical radius r*, after which it
surface-to-volume ratio is very large, and hence, decreases and eventually becomes negative.
the role of specific surface energy is very impor- This occurs since the second term in Eq. 2
tant. With the progressive increase of the size of rises faster with increase in r than the first term
the nuclei, the ratio becomes smaller, and even- (r3 vs. r2). When DG becomes negative, growth
tually large crystals appear, with a corresponding becomes spontaneous, and the cluster grows very
reduction in the role played by the specific sur- fast. This is illustrated in Fig. 1 which shows
face energy. Addition of surfactants can be used the effect of increasing the supersaturation. This
to control the process of nucleation and the size of figure shows the critical size of the nucleus r*
the resulting nucleus. above which growth becomes spontaneous. The
According to Gibbs (1906) and Volmer free energy maximum DG* at the critical radius
(1939), the free energy of formation of represents the barrier that has to be overcome
a spherical nucleus, DG, is given by the sum of before growth becomes spontaneous. Both r*
two contributions: A positive surface energy term and DG* can be obtained by differentiating
DGs which increases with increase in the radius Eq. 2 with respect to r and equating the result to
of the nucleus r and a negative contribution DGv zero. This gives the following expressions:
due to the appearance of a new phase, which also
increases with increase in r: 2sM
r
¼ (3)
rRT lnðS= So Þ
DG ¼ D Gs þD Gv (1)
16 p s3 M2
D G
¼ (4)
DGs is given by the product of area of the nucleus 3 ðrRT Þ2 ½lnðS= So Þ 2
and the specific surface energy, solid/liquid inter-
facial tension s; DGv is related to the relative It is clear from Eqs. 1–4 that the free energy of
supersaturation (S/So): formation of a nucleus and the critical radius r*
Suspensions 1337 S
above which the cluster formation grows sponta- step coagulative nucleation model has been pro-
neously depends on two main parameters, s and posed by Napper and coworkers (Litchi et al.
(S/So), both of which are influenced by the pres- 1983; Feeney et al. 1984). In this process, the
ence of surfactants. s is influenced in a direct way oligomers grow by propagation, and this is
by adsorption of surfactant on the surface of the followed by a termination process in the contin-
nucleus; this adsorption lowers g, and this uous phase. A random coil is produced which is
reduces r* and DG*. In other words, spontaneous insoluble in the medium, and this produces
formation of clusters occurs at smaller critical a precursor oligomer at the y-point. The precursor
radius. In addition, surfactant adsorption stabi- particles subsequently grow primarily by coagu-
lizes the nuclei against any flocculation. The lation to form true latex particles. Some growth
presence of micelles in solution also affects the may also occur by further polymerization. The
process of nucleation and growth directly and colloidal instability of the precursor particles
indirectly. The micelles can act as “nuclei” on may arise from their small size, and the slow
which growth may occur. In addition, the rate of polymerization can be due to reduced
micelles may solubilize the molecules of the swelling of the particles by the hydrophilic mono-
material, thus affecting the relative supersatura- mer (Litchi et al. 1983; Feeney et al. 1984). The
tion, and this can have an effect on nucleation and role of surfactants in these processes is crucial
growth. since they determine the stabilizing efficiency,
and the effectiveness of the surface active agent
ultimately determines the number of particles
Emulsion Polymerization formed. This was confirmed by using surface
active agents of different nature. The effective-
In emulsion polymerization, the monomer, for ness of any surface active agent in stabilizing the
example, styrene or methyl methacrylate that is particles was the dominant factor, and the number
insoluble in the continuous phase, is emulsified of micelles formed was relatively unimportant.
using a surfactant that adsorbs at the monomer/ According to the theory of Smith and Ewart
water interface (Blakely 1975). The surfactant (1948) of the kinetics of emulsion polymeriza-
micelles in bulk solution solubilize some of the tion, the rate of propagation Rp is related to the
monomer. A water-soluble initiator such as number of particles N formed in a reaction by the
potassium persulfate K2S2O8 is added, and this equation
decomposes in the aqueous phase forming free
radicals that interact with the monomers forming d½M
¼ Rp kp N nav ½M (5)
oligomeric chains. It has long been assumed that dt
nucleation occurs in the “monomer-swollen
micelles.” The reasoning behind this mechanism where [M] is the monomer concentration in S
was the sharp increase in the rate of reaction the particles, kp is the propagation rate constant,
above the critical micelle concentration, and and nav is the average number of radicals
that the number of particles formed and their per particle.
size depend to a large extent on the nature of the According to Eq. 5, the rate of polymerization
surfactant and its concentration (which deter- and the number of particles are directly related to
mines the number of micelles formed). However, each other, that is, an increase in the number of
later, this mechanism was disputed, and it was particles will increase the rate. This has been
suggested that the presence of micelles means found for many polymerizations, although there
that excess surfactant is available and molecules are some exceptions. The number of particles is
will readily diffuse to any interface. The most related to the surfactant concentration [S] by the
accepted theory of emulsion polymerization is Eq. 8:
referred to as the coagulative nucleation theory
(Litchi et al. 1983; Feeney et al. 1984). A two- N ½S 3=5 (6)
S 1338 Suspensions
Using the coagulative nucleation model, Dispersion Polymerization

Napper et al. (Litchi et al. 1983; Feeney et al.
1984) found that the final particle number In this case, the reaction mixture consisting of
increases with increase in surfactant concentra- monomer, initiator, and solvent (aqueous or
tion with a monotonically diminishing expo- nonaqueous) for both is usually homogeneous,
nent. The slope of d(log Nc)/d(log t) varies but as polymerization proceeds, polymer sepa-
from 0.4 to 1.2. At high surfactant concentra- rates out, and the reaction continues in
tion, the nucleation time will be long in dura- a homogeneous manner (Barrett 1975).
tion since the new precursor particles will be A dispersant, sometimes referred to as “protec-
readily stabilized. As a result, more latex parti- tive agent,” is added to stabilize the particles once
cles are formed and eventually will outnumber formed. This mechanism for the preparation of
the very small precursor particles at long times. polymer particles is usually applied for prepara-
The precursor/particle collisions will become tion of nonaqueous dispersions (latex particles
more frequent, and fewer latex particles are dispersed in a nonaqueous medium). As men-
produced. The dNc/dt will approach zero, and tioned above, the two main criteria for this type
at long times, the number of latex particles of polymerization are the insolubility of the
remains constant. This shows the inadequacy formed polymer in the continuous phase and
of the Smith-Ewart theory which predicts the solubility of the monomer and initiator in
a constant exponent (3/5) at all surfactant con- the dispersion medium. Initially, the polymeriza-
centrations. For this reason, the coagulative tion starts as a homogeneous system, but after
nucleation mechanism has now been accepted polymerization proceeds to some extent, the
as the most probable theory for emulsion poly- insolubility of the formed polymer chains causes
merization. In all cases, the nature and concen- their precipitation. The process can be visualized
tration of surfactant used are very crucial, and as starting with the formation of polymer chains
this is very important in the industrial prepara- by free radical initiation, followed by formation
tion of latex systems. of nuclei which then grow into polymer particles.
Most reports on emulsion polymerization In the early production of nonaqueous latex dis-
have been limited to commercially available persions, the continuous medium was chosen to
surfactants, which in many cases are relatively be a hydrocarbon solvent. However, later mixed
simple molecules such as sodium dodecyl sul- solvents with polar components were used.
fate and simple nonionic surfactants. However, Indeed, the process of dispersion polymerization
studies on the effect of surfactant structure on has been applied in many cases using completely
latex formation have revealed the importance of polar solvents such as alcohol, water, or alcohol-
the structure of the molecule. Block and graft water mixtures (Barrett 1975). The mechanism of
copolymers (polymeric surfactants) are dispersion polymerization has been discussed in
expected to be better stabilizers when compared detail in the book edited by Barrett (Barrett
to simple surfactants. Studies on styrene poly- 1975). A distinct difference between emulsion
merization using an A-B block of polystyrene and dispersion polymerization may be considered
with polyethylene oxide (PS-PEO) with various in terms of the rate of reaction. As mentioned
ratios of the molecular weight of the two blocks above, with emulsion polymerization, the rate of
showed that an optimum composition is required reaction depends on the number of particles
(Piirma 1992). For efficient anchoring to the formed. However, with dispersion polymeriza-
latex particles, the block length need not be tion, the rate is independent on the number of
more than 10 units, and the PEO block with particles formed. This is to be expected, since in
a molecular weight of 3,000 was sufficient to the latter case, polymerization initially occurs in
stabilize the particles. The results also showed the continuous phase whereby both monomer and
that using a higher molecular weight stabilizer initiator are soluble, and the continuation of poly-
could be counterproductive. merization after precipitation is questionable.
Suspensions 1339 S
Although in emulsion polymerization the initial formation of one colloidally stable particle. This
monomer initiation reaction also occurs in the step depends on the nature of the stabilizer and
continuous medium, the particles formed become how many particles have to heterocoagulate to
swollen with the monomer, and polymerization decrease the total surface area to a size that the
may continue in these particles. A comparison of stabilizer in the system is capable of stabilizing.
the rate of reaction for dispersion and solution
polymerization showed a much faster rate for the
former process (Barrett 1975). Dispersion Methods
To prevent aggregation of the formed polymer
particles, one needs a dispersant (polymer surfac- Dispersion methods are used for the preparation
tant) which much satisfy a number of criteria. The of suspension of preformed particles. The term
most effective dispersants are those of the block dispersion is used to refer to the complete process
(A-B or A-B-A) or graft (BAn) type. The B chain of incorporating the solid into a liquid such that
is chosen to be insoluble in the medium and has the final product consists of fine particles distrib-
high affinity to the surface of the polymer parti- uted throughout the dispersion medium. The role
cles (or become incorporated within its matrix). of surfactants (or polymers) in the dispersion can
This is usually referred to as the “anchor” chain. be seen from consideration of the stages involved
The A chain(s) is chosen to be highly soluble in (Parfitt 1977). Three stages have been considered
the medium and strongly solvated with its mole- (Parfitt 1973): Wetting of the powder by the liq-
cules. It should give a Flory-Huggins interaction uid, breaking of the aggregates and agglomerates,
parameter w < 0.5 in order to ensure effective and comminuting (milling) of the resulting parti-
steric stabilization. The nature and concentration cles into smaller units. Wetting is a fundamental
of the stabilizer determine the number of particles process in which one fluid phase is displaced
formed in dispersion polymerization. In general, completely or partially by another fluid phase
increasing dispersant concentration increases the from the surface of a solid. A useful parameter
number of particles formed (at any given mono- to describe wetting is the contact angle y of
mer content), that is, smaller latex particles are a liquid drop on a solid substrate (Blake 1984).
produced. This is not surprising since smaller If the liquid makes no contact with the solid, that
particles have larger surface area, and this is, y ¼ 180o, the solid is referred to as non-
requires a higher dispersant concentration. The wettable by the liquid in question. This may be
particles in dispersion polymerization were con- the case for a perfectly hydrophobic surface
sidered to be formed by two main steps (Piirma with a polar liquid such as water. However,
1992): (i) Initiation of monomer in the continuous when 180o > y > 90o, one may refer to a case
phase and subsequent growth of the oligomeric of poor wetting. When 0o < y <90o, partial
chains until insolubility occurs and (ii) growing (incomplete) wetting is the case, whereas when S
oligomeric chains associate with each other to y ¼ 0o, complete wetting occurs, and the liquid
form aggregates, which below a certain critical spreads on the solid substrate forming a uniform
size are unstable but gain stability through dis- liquid film.
persant adsorption. However, several other pro- The utility of contact angle measurements
cesses may take place, for example, depends on equilibrium thermodynamic argu-
homocoagulation (collision with other precursor ments (static measurements) using the well-
particles), growth by propagation, adsorption of known Young’s Eq. 12. The value depends on
stabilizer, and swelling by monomer. It should be the following: (i) The history of the system and
pointed out, however, that the number of particles (ii) whether the liquid is tending to advance across
in the final latex cannot be dependent on particle or recede from the solid surface (advancing
nucleation only, since there is another step angle yA, receding angle yR; usually, yA > yR).
involved which determines how many of the pre- Under equilibrium, the liquid drop takes the shape
cursor particles created are involved in the that minimizes the free energy of the system.
S 1340 Suspensions
γLVcos θ If there is strong interaction between solid and

liquid (maximum wetting), the latter spreads until
Young’s equation is satisfied:
γSL θ γSV gLV ¼ gSV gSL (11)
Suspensions, Fig. 2 Schematic representation of the that is, cos y ¼ 1 or y ¼ 0o; the liquid is
contact angle described to spread spontaneously on the solid
surface.
Three interfacial tensions can be identified: gSV, When the surface of the solid is in equilibrium
solid/vapor area ASV; gSL, solid/liquid area ASL; with the liquid vapor, we can consider the spread-
and gLV, liquid/vapor area ALV. A schematic ing pressure p. The solid surface tension is
representation of the balance of tensions at the lowered as a result of adsorption of vapor
solid/liquid/vapor interface is shown in Fig. 2. molecules:
The contact angle is that formed between the
planes tangent to the surfaces of the solid and p ¼ gs gSV (12)
liquid at the wetting perimeter. Here, solid and
liquid are simultaneously in contact with each Young’s equation can be written as
other and the surrounding phase (air or vapor of
the liquid). The wetting perimeter is referred to as gLV cos y ¼ gS gSL p (13)
the three-phase line or wetting line. In this region,
there is an equilibrium between vapor, liquid, and There is no direct way by which gSV or gSL can
solid (Blake 1984). be measured. The difference between gSV and gSL
gSV ASV + gSL ASL + gLV ALV should be can be obtained from contact angle measure-
a minimum at equilibrium, and this leads to the ments (¼ gLV cos y). This difference is referred
well-known Young’s equation: to as wetting tension or adhesion tension:
gSV ¼ gSL þ gLV cos y (7) Adhesion tension ¼ gSV gSL

¼ gLV cos y (14)
g gSL
cos y ¼ SV (8)
gLV Gibbs defined the adhesion tension t as the
difference between the surface pressure of the
The contact angle y depends on the balance solid/liquid and that between the solid/vapor
between the solid/vapor (gSV) and solid/liquid interface (Blake 1984):
(gSL) interfacial tensions. The angle which
a drop assumes on a solid surface is the result of t ¼ pSL pSV (15)
the balance between the adhesion force between
solid and liquid and the cohesive force in the pSV ¼ gs gSV (16)
liquid:
pSL ¼ gs gSL (17)
gLV cos y ¼ gSV gSL (9)
t ¼ gSV gSL ¼ gLV cos y (18)
If there is no interaction between solid and
liquid, The work of adhesion is a direct measure of
the free energy of interaction between solid and
gSL ¼ gSV þ gLV (10) liquid:
that is, cos y ¼ 1 or y ¼ 180o. Wa ¼ ðgLV þ gSV Þ gSL (19)

Suspensions 1341 S
Using Young’s equation, yA is measured by advancing the periphery of
a drop over a surface (e.g., by adding more liquid
Wa ¼ gLV þ gSV gLV cos y ¼ gLV ðcos y þ 1Þ to the drop). yR is measured by pulling the liquid
(20) back. (yA–yR) is referred to as contact angle hys-
teresis. Two main effects can cause contact angle
The work of adhesion depends on the follow- hysteresis: (1) Penetration of wetting liquid into
ing: gLV, the liquid/vapor surface tension and y, pores during advancing contact angle measure-
the contact angle between liquid and solid. The ments and (2) surface roughness: The first and
work of cohesion Wc is the work of adhesion rear edges both meet the liquid with some intrin-
when the two surfaces are the same: sic angle yo (microscopic contact angle). The
macroscopic angles yA and yR vary significantly.
Wc ¼ 2 gLV (21) yo values are determined by contact of liquid with
the “rough” valleys (microscopic contact angle).
For adhesion of a liquid on a solid, Wa Wc yA and yR are determined by contact of liquid
or y ¼ 0o (cos y ¼ 1). with arbitrary parts on the surface (peak or val-
Harkins (Blake 1984) defined the spreading ley). Surface roughness can be accounted for by
coefficient as the work required to destroy unit comparing the “real” area of the surface A with
area of SL and LV and leaves unit area of bare that of the projected
solid SV.
Spreading coefficient S ¼ Surface energy of A
r¼ (25)
final state – Surface energy of initial state: A0
S ¼ gSV ðgSL þ gLV Þ (22) A ¼ area of surface taking into account all
peaks and valleys. A0 ¼ Apparent area (same
Using Young’s equation, macroscopic dimension); r > 1.
gSV ¼ gSL þ gLV cos y (23) cos y ¼ r cos yo (26)

S ¼ gLV ðcos y 1Þ (24)
y is the macroscopic contact angle, whereas yo
If S is zero (or positive), that is, y ¼ 0, the liquid is the microscopic contact angle.
will spread until it completely wets the solid. If S is
negative, that is, y > 0, only partial wetting occurs. ðg gSL Þ
cos y ¼ r SV (27)
Alternatively, one can use the equilibrium (final) gLV
spreading coefficient. For dispersion of powders
into liquids, one usually requires complete spread- If cos y is negative on a smooth surface S
ing, that is, y should be zero. (y > 90o), it becomes more negative on a rough
surface (y is larger), and surface roughness
reduces wetting. If cos y is positive on a smooth
Contact Angle Hysteresis surface (y < 90o), it becomes more positive on
a rough surface (y is smaller), and roughness
For a liquid spreading on a uniform, non- enhances wetting (Blake 1984).
deformable solid (idealized case), there is only Another factor that causes hysteresis is surface
one contact angle-equilibrium value. With real heterogeneity (Blake 1984). Most practical sur-
systems (practical solids), a number of stable faces are heterogeneous consisting of “islands” or
contact angles can be measured. Two relatively “patches” with different surface energies. As the
reproducible angles can be measured: Largest – drop advances on such a surface, the edge of the
advancing angle yA and smallest – receding drop tends to stop at the boundary of the “island.”
angle yR. The advancing angle will be associated with the
S 1342 Suspensions
γc γLV (30 mNm1). Very low energy solids such as

Teflon (polytetrafluoroethylene, PTFE) give
1.0
lower gc (<25 mNm1). This concept explains
0.8 the degree of wettability of the above surfaces
0.6 with water, glass being the easiest to wet and
cos ?
Teflon being the most difficult to wet.

0.4
Surfactants lower the surface tension of water,
0.2 g, and they adsorb at the solid/liquid interface.
0.0 A plot of gLV versus log C (where C is the sur-
10 20 30 40 50 60 70
factant concentration) results in a gradual reduc-
tion in gLV followed by a linear decrease of gLV
Suspensions, Fig. 3 Variation of cos y with g LV with log C (just below the critical micelle
concentration, cmc), and when the cmc is
reached, gLV remains virtually constant. From
intrinsic angle of the high contact angle region. the slope of the linear portion of the g – log
The receding angle will be associated with the C curve (just below the cmc), one can obtain the
low contact angle region. If the heterogeneities surface excess (number of moles of surfactant per
are very small compared with the dimensions of unit area at the L/A interface). Using Gibbs
the liquid drop, one can define a composite con- adsorption isotherm,
tact angle using Cassie’s equation:
dg
¼ 2:303 RTG (30)
cos y ¼ Q1 cos y1 þ Q2 cos y2 (28) dlogC
Q1 ¼ Fraction of surface having contact angle G ¼ surface excess (moles m2); R ¼ gas
y1; Q2 ¼ Fraction of surface having contact constant; and T ¼ absolute temperature.
angle y2. y1 and y2 are the maximum and mini- From G, one can obtain the area per molecule:
mum possible angles.
A systematic way of characterizing “wettabil- 1 1018
Area per molecule ¼ ðm2 Þ ¼ ðnm2 Þ
ity” of a surface was introduced by Fox and G Nav G Nav
Zisman (Zisman 1964). For a given substrate (31)
and for a series of related liquids (e.g., n-alkanes,
siloxanes, and dialkyl ethers), a plot of cos y Most surfactants produce a vertically ori-
versus gLV gives a straight line. This is schemat- ented monolayer just below the cmc. The area/
ically shown in Fig. 3. molecule is usually determined by the cross-
Extrapolation of the straight line to cos y ¼ 1 sectional area of the head group. For ionic sur-
(y ¼ 0) gives the critical surface tension of wet- factants containing say -OSO3 or -SO3 head
ting gc. Any liquid with gLV < gc will give y ¼ 0, group, the area per molecule is in the region of
that is, it wets the surface completely; gc is the 0.4 nm2. For nonionic surfactants containing
surface tension of a liquid that just spreads on the several moles of ethylene oxide (8–10), the
substrate to give complete wetting. This linear area per molecule can be much larger
relationship can be represented by the following (1–2 nm2). Surfactants will also adsorb at the
empirical equation: solid/liquid interface. For hydrophobic surfaces,
the main driving force for adsorption is by
cos y ¼ 1 þ bðgLV gc Þ (29) hydrophobic bonding. This results in lowering
of the contact angle of water on the solid surface.
High energy solids, for example, glass, give For hydrophilic surfaces, adsorption occurs via
high gc (>40 mNm1). Low energy solids, for the hydrophilic group, for example, cationic sur-
example, hydrophobic surfaces, give lower gc factants on silica. Initially, the surface becomes
Suspensions 1343 S
more hydrophobic, and the contact angle y channels between the particles and inside any
increases with increase in surfactant concentra- “capillaries.” For wetting of hydrophobic pow-
tion. However, at higher cationic surfactant con- ders into water, anionic surfactants, for example,
centration, a bilayer is formed by hydrophobic alkyl sulfates or sulfonates or nonionic surfac-
interaction between the alkyl groups, and the tants of the alcohol or alkyl phenol ethoxylates,
surface becomes more and more hydrophilic, are usually used.
and eventually, the contact angle reaches zero A useful concept for choosing wetting agents
at high surfactant concentrations. of the ethoxylated surfactants is the hydrophilic-
Smolders (1961) suggested the following rela- lipophilic balance (HLB) concept:
tionship for change of y with C:
% of hydrophilic groups
HLB ¼ (34)
d gLV cos y d gSV d gSL 5
¼ (32)
dlnC dlnC dlnC
Most wetting agents of this class have an HLB
Using the Gibbs equation, number in the range 7–9.
The process of wetting of a solid by a liquid
dg involves three types of wetting: adhesion wetting,
sin y ¼ RTðGSV GSL gLV cos yÞ
dlnC Wa; immersion wetting, Wi; and spreading wet-
(33) ting, Ws. In every step, one can apply Young’s
equation:
Since gLV sin y is always positive, (dy/dln C)
will always have the same sign as the RHS of gSV ¼ gSL þ gLV cos y (35)
Eq. 33. Three cases may be distinguished:
(dy/dln C) < 0; GSV < GSL + GLV cos y; addition Wa ¼ gSL ðgSV þ gLV Þ ¼ gLV ðcos y þ 1Þ
of surfactant improves wetting. (dy/dln C) ¼ 0; (36)
GSV ¼ GSL + GLV cos y; surfactant has no effect
on wetting. (dy/dln C) > 0; GSV > GSL + GLV Wi ¼ 4 gSL 4 gSV ¼ 4 gLV cos y (37)
cos y; surfactant causes dewetting.
Ws ¼ ðgSL þ gLV Þ gSV ¼ gLV ðcos y 1Þ
(38)
Wetting of Powders by Liquids
The work of dispersion Wd is the sum of Wa,
Wetting of powders by liquids is very important Wi, and Ws:
in their dispersion, for example, in the prepara-
tion of concentrated suspensions. The particles in Wd ¼ Wa þ Wi þ Ws ¼ 6 gSV gSL S
a dry powder form either aggregates (where the ¼ 6 gLV cos y (39)
particles are joined by their crystal faces forming
compact structures) or agglomerates (where the Wetting and dispersion depend on the follow-
particles are joined by their edges or corners ing: gLV, liquid surface tension; y, contact angle
forming loose structures). It is essential in the between liquid and solid. Wa, Wi, and Ws are
dispersion process to wet both external and inter- spontaneous when y < 90o. Wd is spontaneous
nal surfaces and displace the air entrapped when y ¼ 0. Since surfactants are added in suffi-
between the particles. Wetting is achieved by cient amounts (gdynamic is lowered sufficiently),
the use of surface active agents (wetting agents) spontaneous dispersion is the rule rather than the
of the ionic or nonionic type which are capable of exception.
diffusing quickly (i.e., lower the dynamic surface Wetting of the internal surface requires pene-
tension) to the solid/liquid interface and displace tration of the liquid into channels between and
the air entrapped by rapid penetration through the inside the agglomerates. The process is similar to
S 1344 Suspensions
forcing a liquid through fine capillaries. To force complex path formed by the channels between
a liquid through a capillary with radius r, the particles, that is,
a pressure p is required that is given by
kt gLV cos y
l2 ¼ (42)
2 gLV cos y 2ðgSV gSL Þ 2Z
p¼ ¼ (40)
r r gLV
Thus, a plot of l2 versus t gives a straight line,
gSL has to be made as small as possible; rapid and from the slope of the line, one can obtain y.
surfactant adsorption to the solid surface, low y. The Rideal-Washburn (Rideal 1922; Washburn
When y ¼ 0, p 1 gLV. Thus, for penetration into 1921) equation can be applied to obtain the con-
pores, one requires a high gLV. Thus, wetting of tact angle of liquids (and surfactant solutions) in
the external surface requires low contact angle y powder beds. K should first be obtained using
and low surface tension gLV . Wetting of the a liquid that produces zero contact angle.
internal surface (i.e., penetration through A packed bed of powder is prepared say in
pores) requires low y but high gLV. These two a tube fitted with a sintered glass at the end
conditions are incompatible, and a compromise (to retain the powder particles). It is essential to
has to be made: gSV – gSL must be kept at pack the powder uniformly in the tube (a plunger
a maximum. gLV should be kept as low as possi- may be used in this case). The tube containing the
ble but not too low. These conclusions illustrate bed is immersed in a liquid that gives spontane-
the problem of choosing the best dispersing ous wetting (e.g., a lower alkane), that is, the
agent for a particular powder. This requires liquid gives a zero contact angle and cos y ¼ 1.
measurement of the above parameters as well By measuring the rate of penetration of the liquid
as testing the efficiency of the dispersion (this can be carried out gravimetrically using,
process. e.g., a microbalance or a Kruss instrument), one
The rate of liquid penetration has been inves- can obtain K. The tube is then removed from the
tigated by Rideal and Washburn (Rideal 1922; lower alkane liquid and left to stand for evapora-
Washburn 1921). For horizontal capillaries tion of the liquid. It is then immersed in the liquid
(gravity neglected), the depth of penetration l in in question, and the rate of penetration is mea-
time t is given by the Rideal-Washburn sured again as a function of time. Using Eq. 27,
Equation (Rideal 1922; Washburn 1921): one can calculate cos y and hence y.

rt gLV cos y 1=2
l¼ (41) Structure of the Solid/Liquid Interface
2Z
To enhance the rate of penetration, gLV has to A great variety of processes occur to produce
be made as high as possible, y as low as possible, a surface charge (Lyklema 1987):
and Z as low as possible. For dispersion of pow- (1) Surface ions: These are ions that have such
ders into liquids, one should use surfactants that a high affinity to the surface of the particles that
lower y while not reducing gLV too much. The they may be taken as part of the surface, for
viscosity of the liquid should also be kept at example, Ag+ and I for AgI. For AgI in
a minimum. Thickening agents (such as poly- a solution of KNO3, the surface charge so is
mers) should not be added during the dispersion given by the following expression:
process. It is also necessary to avoid foam forma-
tion during the dispersion process. For a packed sp ¼ FðGAgþ GI Þ ¼ F GAgNO3 GKI (43)
bed of particles, r may be replaced by K, which
contains the effective radius of the bed and where F is the Faraday constant (96,500 C mol1)
a tortuosity factor, which takes into account the and G is the surface excess of ions (mol m2).
Suspensions 1345 S
Similarly for an oxide such as silica or alumina s0 sd
in KNO3, H+ and OH may be taken as part of the Ψ0
surface: s0 = sd
sd = diffuse double Surface
+
so ¼ FðGHþ GOH Þ ¼ FðGHCl GKOH Þ (44) layer charge (negatively
+ –
+
charged) +
– +
–
+ +
The ions which determine the charge on the +
– –
surface are termed potential determining ions. +
–
+ +
The charge depends on the pH of the solution: + –
Ψx
Below a certain pH, the surface is positive, x
and above a certain pH, surface is negative.
At a specific pH (GH ¼ GOH), surface is Suspensions, Fig. 4 Schematic representation of
the diffuse double layer according to Gouy and Chapman
uncharged; this is referred to as the point of zero
charge (pzc).
The pzc depends on the type of the oxide: For Note that when x ¼ 1/k, cx ¼ co/e. 1/k is
an acidic oxide such as silica, the pzc is pH referred to as the “thickness of the double
2–3. For a basic oxide such as alumina, pzc is layer.” The double layer extension depends on
pH 9. For an amphoteric oxide such as titania, electrolyte concentration and valency of the
the pzc is pH 6. In some cases, specifically counterions:
adsorbed ions (that have non-electrostatic

affinity to the surface) “enrich” the surface but 1 er eo kT 1=2
¼ (46)
may not be considered as part of the surface, for k 2 no Z2i e2
example, bivalent cations on oxides, cationic, and
anionic surfactants on most surfaces (Lyklema er is the permittivity (dielectric constant), 78.6 for
1987). water at 25 C. eo is the permittivity of free space.
(2) Isomorphic substitution, for example, with k is the Boltzmann constant, and T is the absolute
sodium montmorillonite, that is, replacement of temperature. no is the number of ions per unit
cations inside the crystal structure by cations of volume of each type present in bulk solution,
lower valency, for example, Si4+ replaced by Al3+. and Zi is the valency of the ions. e is the electronic
The deficit of one positive charge gives one charge.
negative charge. The surface of Na montmoril- For 1:1 electrolyte (e.g., KCl)
lonite is negatively charged with Na+ as coun-
terions. The surface charge + counterions form C/mol dm3 105 104 103 102 101
(1/k)/nm 100 33 10 3.3 1
the electrical double layer. Several pictures
were proposed for the structure of the double The double layer extension increases with S
layer: decrease in electrolyte concentration.
(a) Diffuse double layer (Gouy and Chapman): (b) Stern-Grahame model of the double layer:
The surface charge so is compensated by unequal Stern (Lyklema 1987) introduced the concept of
distribution of counterions (opposite in charge the non-diffuse part of the double layer for spe-
to the surface) and co-ions (same sign as the cifically adsorbed ions, the rest being diffuse in
surface) which extend to some distance from the nature. This is schematically illustrated in Fig. 5.
surface (Lyklema 1987). This is schematically The potential drops linearly in the Stern region
represented in Fig. 4. The potential decays expo- and then exponentially. Grahame distinguished
nentially with distance x. At low potentials, two types of ions in the Stern plane, physically
adsorbed counterions (outer Helmholtz plane)
and chemically adsorbed ions (that loose part of
c ¼ co exp ðkxÞ (45)
their hydration shell) (inner Helmholtz plane).
S 1346 Suspensions
s0 ss
ψd ψd
sd
Ψ0 + Ψd
s0 = ss + sd +
+ +
+
+
– + ψH/2
ss = Charge due to specifically +
+ – x=o x = H/2 x=H
adbsorbed counter ions –
+
+ – + Suspensions, Fig. 6 Schematic representation of double
–
layer interaction for two flat plates
x
Stem plane
most important since it occurs for polar and non-
Suspensions, Fig. 5 Schematic representation of the polar molecules. It arises from fluctuations in the
double layer according to Stern and Grahame electron density distribution. At small distances
of separation r in vacuum, the attractive energy
(c) Electrical double layer repulsion: When between two atoms or molecules is given by
charged colloidal particles in a dispersion
approach each other such that the double layers b11
Gaa ¼ (48)
begin to overlap (particle separation becomes less r6
than twice the double layer extension), repulsion
occurs. The individual double layers can no lon- b11 is the London dispersion constant.
ger develop unrestrictedly since the limited space For colloidal particles which are made of atom
does not allow complete potential decay or molecular assemblies, the attractive energies
(Bijesterbosch 1987). This is illustrated in Fig. 6 may be added, and this results in the following
for two flat plates. expression for two spheres (at small h):
The potential cH/2 half way between the plates
is no longer zero (as would be the case for iso- AR
GA ¼ (49)
lated particles at x ! 1). For two spherical 12h
particles of radius R and surface potential co
and condition kR < 3, the expression for the where A is the effective Hamaker constant:
electrical double layer repulsive interaction is
1=2 1=2
given by A ¼ ðA11 A22 Þ2 (50)
4p er eo R2 c2o exp ðkhÞ A11 is the Hamaker constant between particles

Gel ¼ (47)
2R þ h in vacuum and A22 Hamaker constant for equiv-
alent volumes of the medium
where h is the closest distance of separation
between the surfaces. This expression shows the A ¼ p2 q2 bii (51)
exponential decay of Gel with h. The higher the
value of k (i.e., the higher the electrolyte concen- q is number of atoms or molecules per unit
tration), the steeper the decay. volume.
(d) Van der Waals attraction: As is well GA decreases with increase of h and becomes
known, atoms or molecules always attract each very large at short distances of separation. At
other at short distances of separation. The attrac- very short distances, the Born repulsion appears.
tive forces are of three different types: Dipole- (e) Total energy of interaction: Derjaguin-
dipole interaction (Keesom), dipole-induced Landau-Verwey-Overbeek (DLVO) theory
dipole interaction (Debye), and London disper- (Deryaguin and Landau 1941; Verwey and
sion force. The London dispersion force is the Overbeek 1948): Combination of Gel and GA
Suspensions 1347 S
Suspensions,
Fig. 7 Total energy–
distance curve according to
the DLVO theory
G GT Ge
Gmax
h
GA Gsec
Gprimary
results in the well-known theory of stability of (1/κ) = 1000 nm

colloids (DLVO Theory) (Deryaguin and Landau 10−7 mol dm−3
1941; Verwey and Overbeek 1948):
(1/κ) = 100 nm
10−5 mol dm−3
GT ¼ Gel þ GA (52) G (1/κ) = 10 nm
10−3 mol dm−3
A plot of GT versus h is shown in Fig. 7, which
represents the case at low electrolyte concentra-
tions, that is, strong electrostatic repulsion h
between the particles. Gel decays exponentially
with h, that is, Gel – > 0 as h becomes large. GA (1/κ) = 1 nm
is 1 1/h, that is, GA does not decay to 0 at large h. 10−1 mol dm−3
At long distances of separation, GA > Gel,
resulting in a shallow minimum (secondary min- Suspensions, Fig. 8 Energy–distance curves at various
imum). At very short distances, GA Gel, 1:1 electrolyte concentrations
S
resulting in a deep primary minimum. At inter-
mediate distances, Gel > GA, resulting in energy may occur. This flocculation is weak and revers-
maximum, Gmax, whose height depends on co (or ible. By increasing the electrolyte concentration,
cd) and the electrolyte concentration and Gmax decreases till at a given concentration it
valency. At low electrolyte concentrations vanishes and particle coagulation occurs. This is
(< 102 mol dm3 for a 1:1 electrolyte), Gmax illustrated in Fig. 8 which shows the variation of
is high (> 25 kT), and this prevents particle GT with h at various electrolyte concentrations.
aggregation into the primary minimum. The Coagulation occurs at a critical electrolyte
higher the electrolyte concentration (and concentration, the critical coagulation concentra-
the higher the valency of the ions), the lower tion (ccc) which depends on the electrolyte
the energy maximum. Under some conditions valency. At low surface potentials, ccc 1 1/Z2.
(depending on electrolyte concentration and parti- This is referred to as the Schulze-Hardy rule.
cle size), flocculation into the secondary minimum One can define a rate constant for flocculation:
S 1348 Suspensions
2:2 Electrolyte 1:1 Electrolyte experimental techniques for measuring the zeta
potential, it is essential to consider the electroki-
log W
netic effects in some detail, describing the theo-
ries that can be used to calculate the zeta potential
from the particle electrophoretic mobility
(Hunter 1981). Electrokinetic effects are the
W=1 0
direct result of charge separation at the interface
between two phases. Consider a negatively
charged surface; positive ions (counterions) are
10−3 10−2 10−1 attracted to the surface, whereas negative ions
log C
(co-ions) are repelled. This was schematically
Suspensions, Fig. 9 Log W–log C curves shown in Fig. 4. The accumulation of excess
positive ions causes a gradual reduction in the
potential from its value co at the surface to 0 in
ko ¼ rapid rate of flocculation (in the absence of bulk solution. At a point p from the surface, one
an energy barrier) and k ¼ slow rate of floccula- can define a potential cx. Electrokinetic effects
tion (in the presence of an energy barrier): arise when one of the two phases is caused to
move tangentially past the second phase. Tangen-
ko tial motion can be caused by electric field, forcing
¼ WðThe stability ratioÞ (53)
k a liquid in a capillary or by gravitational field on
the particles. The most commonly used proce-
Note that W increases as Gmax increases. The dure for suspensions is electrophoresis where
stability of colloidal dispersions can be quantita- movement of one phase is induced by application
tively assessed from plots of log W versus log C, of an external electric field. One measures the
as illustrated in Fig. 9. particle velocity v from which the electrophoretic
(f) Criteria for stabilization of suspensions mobility u can be calculated:
with double layer interaction: The two main
v
criteria for stabilization are the following: (i) u¼ m2 V1 s1 (54)
High surface or Stern potential (zeta potential), ðE=lÞ
high surface charge and (ii) low electrolyte con-
centration and low valency of counter and co- where E is the applied potential, l is the distance
ions. One should ensure that an energy maximum between the two electrodes, and E/l is the field
in excess of 25 kT should exist in the energy– strength. In all electrokinetic phenomena (Hunter
distance curve. When Gmax > > kT, the particles 1981), a fluid moves with respect to a solid sur-
in the dispersion cannot overcome the energy face. One needs to derive a relationship between
barrier, thus preventing coagulation. In some fluid velocity (which varies with distance from
cases, particularly with large and asymmetric the solid) and the electric field in the interfacial
particles, flocculation into the secondary mini- region. The most important concept is the surface
mum may occur. This flocculation is usually of shear, an imaginary surface close to the surface
weak and reversible and may be advantageous within which the fluid is stationary. This is illus-
for preventing the formation of hard sediments. trated in Fig. 10 which shows the position of the
surface potential co, the shear plane, and zeta
Electrokinetic Phenomena and the Zeta potential (that is close to the Stern potential cd).
Potential Although measurement of particle mobility is
As mentioned above, one of the main criteria for fairly simple (particularly with the development
electrostatic stability is the high surface or zeta of automated instruments), the interpretation of
potential which can be experimentally measured the results is not simple. The calculation of zeta
as will be discussed below. Before describing the potential from particle mobility is not
Suspensions 1349 S
Permittivity ε⬘ < εb (iii) Henry’s treatment (Henry 1948): Henry
accounted for the discrepancy between
η⬘ >> ηb Liquid Smoluchowski and Huckel’s treatment. Huckel
Bulk Immobile disregarded the deformation of the electric field
Viscosity
Particle Viscosity η⬘
by the particle, whereas Smoluchowski assumed
ηb
the field to be uniform and everywhere parallel
to the particle surface. These two assumptions
Surface of
are justified in the extreme cases of kR 1 and
Particle
Shear kR 1, respectively. For the intermediate
surface
Electrokinetic cases where kR is not too small or too large,
Potential ψo
Potential Henry derived the following expression (which
ζ potential can be applied at all kR values):
Suspensions, Fig. 10 Surface (plane) of shear
2 er eo z
u¼ fðkRÞ (58)
straightforward since this depends on the particle 3 Z
size and shape as well as the electrolyte concen-
tration. Several theories are available for calcu-
The function f(kR) depends also on the parti-
lation of zeta potential from the particle mobility:
cle shape. Values of f(kR) at various values of kR
(i) Von Smoluchowski (Classical) Treatment
are given below.
(von Smoluchowski 191): This applies to the case
where the particle radius R is much larger than
the double layer thickness (1/k), that is, kR 1.
Henry’s Correction Factor f(kR)
This generally applies to particles that are greater
than 0.5 mm (when the 1:1 electrolyte concentra- kR f(kR)
tion is lower than 103 mol dm3, i.e., kR > 10): 0 1.0
1 1.027
er eo z
u¼ (55) 2 1.066
Z 3 1.101
4 1.133
where er is the relative permittivity of the 5 1.160
medium, 78.6 for water at 25o C; eo is the permit- 10 1.239
tivity of free space, 8.85 1012 Fm1; and Z is 25 1.370
the viscosity of the medium, 8.9 104 Pas for 100 1.460
water at 25o C. z is the zeta potential in volts. 1 1.500
For water at 25o C, S
Henry’s calculations are based on the assump-
z ¼ 1:282 106 u (56) tion that the external field can be superimposed on
the field due to the particle, and hence, it can only
u is expressed in m2 V1 s1. be applied for low potentials (z < 25 mV). It also
(ii) The Huckel equation (Huckel 1924): This does not take into account the distortion of the field
applies for the case kR < 1: induced by the movement of the particle (relaxa-
tion effect). Wiersema, Loeb, and Overbeek
2 er eo z (Wiersema et al. 1967) introduced two corrections
u¼ (57)
3 Z for Henry’s treatment, namely, the relaxation and
retardation (movement of the liquid with the double
The above equation applies for small particles layer ions) effects, and a numerical tabulation of
(< 100 nm) and thick double layers (low electro- the relation between mobility and zeta potential
lyte concentration). has been given by Ottewill and Shaw (1972).
S 1350 Suspensions
Such tables are useful for conversion of u to z at all particles will show intensity fluctuation as a result
practical values of kR. of the Brownian diffusion (Doppler shift). By
Two main procedures can be applied for mea- application of an electric field as the particles
suring the zeta potential (Hunter 1981): (i) Ultra- undergo Brownian motion and measuring the
microscopic technique (microelectrophoresis): fluctuation in intensity of the scattered light
This is the most commonly used method since it (using a correlator), one can measure the particle
allows direct observation of the particles using an mobility. Two laser beams of equal intensity are
ultramicroscope (suitable for particles that are allowed to cross at a particular point within the
larger than 100 nm). Basically, a dilute suspen- cell containing the suspension of particles. At
sion is placed in a cell consisting of a thin-walled the intersection of the beam, which is focused at
(100 mm) glass tube that is attached to two the stationary level, interferences of known spac-
larger bore tubes with sockets for placing the ing are formed. The particles moving through the
electrodes. The cell is immersed in a thermostat fringes under the influence of the electric field
bath (accurate to + 0.1o C) that contains attach- scatter light whose intensity fluctuates with
ment for illumination and a microscope objective a frequency that is related to the mobility of the
for observing the particles. It is also possible to particles. The photons are detected by
use a video camera for directly observing the a photomultiplier, and the signal is fed to the
particles. Since the glass walls are charged (usu- correlator. The resulting correlation function is
ally negative at practical pH measurements), the analyzed to determine the frequency (Doppler)
solution in the cell will in general experience spectrum, and this is converted to the particle
electroosmotic flow. Only where the electroos- velocity v:
motic flow is zero, that is, at the stationary level,
can the electrophoretic mobility of the particles v ¼ Dns (59)
be measured. The stationary level is located at
a distance of 0.707 of the radius from the center where Dv is the Doppler shift frequency and s is
of the tube or 0.146 of the internal diameter from the spacing between the interference fringes in
the wall. By focusing the microscope objective at the region where the beams cross. s is given by
the top and bottom of the walls of the tube, one the relationship
can easily locate the position of the stationary
levels. The average particle velocity is measured l
s¼ (60)
at the top and bottom stationary levels by aver- 2 sinða=2Þ
aging at least 20 measurements in each direction
(the eye piece of the microscope is fitted with where l is the laser wavelength and a is the angle
a graticule). Several commercial instruments are between the crossing laser beams.
available (e.g., Rank Brothers, Bottisham, Cam- The velocity spectrum is then converted to
bridge, England, and Pen Kem in USA). For large a mobility spectrum (allowing one to obtain the
particles (> 1 mm and high density), sedimenta- mobility distribution), and the mobility is
tion may occur during the measurement. In this converted to zeta potential using Huckel’s equa-
case, one can use a rectangular cell and observe tion. Several commercial instruments are avail-
the particles horizontally from the side of the able: Malvern Zeta Sizer and Coulter Delsa Sizer.
glass cell. Microelectrophoresis has many advan-
tages since the particles can be measured in their
normal environment. It is preferable to dilute the General Classification of Dispersing
suspension with the supernatant liquid which can Agents
be produced by centrifugation.
(ii) Laser velocimetry technique: This method As mentioned above, for dispersing powders into
is more suitable for small particles that undergo liquids, one usually requires the addition of
Brownian motion. The scattered light by small a dispersing agent which satisfies the following
Suspensions 1351 S
requirements: (i) Lowers the surface tension of naphthalene formaldehyde sulfonated conden-
the liquid to aid wetting of the powder, (ii) sates, and lignosulphonates.
adsorbs at the solid/liquid interface to lower the
solid/liquid interfacial tension, (iii) lowers the
contact angle of the liquid on the solid surface Adsorption of Surfactants at the
(zero contact angle is very common), (iv) helps to Solid/Liquid Interface
break up the aggregates and agglomerates as well
as in subdivision of the particles into smaller Surfactant adsorption may be described in terms
units, and (v) stabilizes the particles formed of simple interaction parameters. However, in
against any aggregation (or rejoining). All dis- some cases, these interaction parameters may
persing agents are surface active, and they can be involve ill-defined forces, such as hydrophobic
simple surfactants (anionic, cationic, zwitter- bonding, solvation forces, and chemisorption. In
ionic, or nonionic), polymers, or polyelectrolytes. addition, the adsorption of ionic surfactants
The dispersing agent should be soluble (or at least involves electrostatic forces particularly with
dispersible) in the liquid medium, and it should polar surfaces containing ionogenic groups. For
adsorb at the solid/liquid interface. that reason, the adsorption of ionic and nonionic
In this section, I will describe the general surfactants will be treated separately. The sur-
classification of dispersing agents. This is faces (substrates) can be also hydrophobic or
followed by a section on the adsorption of sur- hydrophilic, and these may be treated separately.
factants and polymers at the solid/liquid (i) Adsorption of ionic surfactants on hydropho-
interface. bic surfaces: The adsorption of ionic surfactants
Surfactants: Ionic and anionic, for example, on hydrophobic surfaces such as carbon black,
sodium dodecyl sulfate C12H25OSO3-Na+; polymer surfaces, and ceramics (silicon carbide
cationic, for example cetyltrimethylammonium or silicon nitride) is governed by hydrophobic
chloride C16H33-N+ (CH3)3 Cl-; zwitterionic, interaction between the alkyl chain of the surfac-
for example, 3-dimethyldodecylamine propane tant and the hydrophobic surface. In this case,
sulfonate (betaine C12H25-N+(CH3)2-CH2-CH2- electrostatic interaction will play a relatively
CH2-SO3); nonionic, alcohol ethoxylates CnH2n+1- smaller role. However, if the surfactant head
O-(CH2-CH2-O)n-H, alkyl phenol ethoxylates group is of the same sign of charge as that on
CnH2n+1-C6H4-O-(CH2-CH2-O)n-H; amine oxides, the substrate surface, electrostatic repulsion may
for example, decyl dimethyl amine oxide C10H21-N oppose adsorption. In contrast, if the head groups
(CH3)2 ! O, amine ethoxylates. are of opposite sign to the surface, adsorption
Nonionic polymers: Polyvinyl alcohol (with may be enhanced. Since the adsorption depends
polyvinyl acetate blocks, usually 4–12 %). on the magnitude of the hydrophobic bonding
-(CH2-CH-OH)x-(CH2-CH(OCOCH3)-OH)y-(CH2- free energy, the amount of surfactant adsorbed S
CH-OH)x-. Block copolymers of ethylene oxide- increases directly with increase in the alkyl chain
propylene oxide (ABA block of PEO-PPO-PEO), length in accordance with Traube’s rule. The
Pluronics (BASF), Synperonic PE (ICI), adsorption of ionic surfactants on hydrophobic
H-(O-CH2-CH2)n-(CH2-CH (CH3)-O)m-(CH2- surfaces may be represented by the Stern-
CH2-O)n-H. Graft copolymers, for example, Langmuir isotherm (Hough and Randall 1983).
polymethylmethacrylate (PMMA) backbone Consider a substrate containing Ns sites (mol m2)
(with some polymethacrylic acid) with grafted on which G moles m2 of surfactant ions are
PEO chains, and Atlox 4913 Hypermer CG6 (ICI.) adsorbed. The surface coverage y is (G/Ns), and
Polyelectrolytes: Polyacrylic acid, -(CH2-CH- the fraction of uncovered surface is (1 – y). The
COOH)n-. At pH > 5, ionization of the carbox- Stern-Langmuir equation is given by Eq. 61:
ylic acid groups occurs forming an anionic
polyelectrolyte, polyacrylic acid/polymethacrylic y C D GO
¼ exp ads
(61)
acid (CH2-CH-COO-)n-(CH2-C (CH3)-COO-)m, 1 y 55:5 RT
S 1352 Suspensions
Equation 61 applies only at low surface cov- Since there is no rigorous theory that can pre-
erage (y < 0.1) where lateral interaction between dict adsorption isotherms, the most suitable
the surfactant ions can be neglected. At high method to investigate adsorption of surfactants
surface coverage (y > 0.1), one should take the is to determine the adsorption isotherm. Measure-
lateral interaction between the chains into ment of surfactant adsorption is fairly straightfor-
account, by introducing a constant A, for exam- ward. A known mass m (g) of the particles
ple, using the Frumkin–Fowler–Guggenheim (substrate) with known specific surface area
Equation (Fuerstenau and Healy 1972): As (m2 g1) is equilibrated at constant tempera-
ture with surfactant solution with initial concen-
y C D Goads tration C1. The suspension is kept stirred for
expðAyÞ ¼ exp (62)
ð1 yÞ 55:5 RT sufficient time to reach equilibrium. The particles
are then removed from the suspension by centri-
Various authors (Somasundaran and Goddard fugation, and the equilibrium concentration C2 is
1979; Healy 1974) have used the Stern–Lang- determined using a suitable analytical method.
muir equation in a simple form to describe the The amount of adsorption G (mol m2) is calcu-
adsorption of surfactant ions on mineral surfaces: lated as follows:

D Goads ðC1 C2 Þ
G ¼ 2rC exp (63) G¼ (66)
RT m As
Various contributions to the adsorption free The adsorption isotherm is represented by

energy may be envisaged. To a first approxima- plotting G versus C2. A range of surfactant con-
tion, these contributions may be considered to be centrations should be used to cover the whole
additive. In the first instance, DGads may be taken adsorption process, that is, from the initial values
to consist of two main contributions, that is, low to the plateau values. To obtain accurate
results, the solid should have a high surface area
D Gads ¼ D Gelec þD Gspec (64) (usually > 1 m2). Several examples may be
quoted from the literature to illustrate the adsorp-
where DGelec accounts for any electrical interaction of surfactant ions on solid surfaces. For
tions and DGspec is a specific adsorption term a model hydrophobic surface, carbon black has
which contains all contributions to the adsorption been chosen (Healy 1974; Somasundaran and
free energy that are dependent on the “specific” Hannah 1979). Figure 11 shows typical results
(nonelectrical) nature of the system for the adsorption of sodium dodecyl sulfate
(Somasundaran and Goddard 1979; Healy 1974). (SDS) on two carbon black surfaces, namely,
Several authors subdivided DGspec into supposedly Spheron 6 (untreated) and Graphon (graphitized)
separate independent interactions (Healy 1974; which also describes the effect of surface treat-
Somasundaran and Hannah 1979), for example, ment (Greenwood et al. 1968; Day et al. 1967).
The adsorption of SDS on untreated Spheron 6
D Gspec ¼ D Gcc þD Gcs þD Ghs þ:::::::::: (65) tends to show a maximum that is removed on
sive chain–chain interaction between the hydro- at high surfactant concentration. The plateau
phobic moieties of the adsorbed ions, DGcs is the adsorption value is 2 106 mol m2 (2 m
term for chain/substrate interaction, and DGhs is mol m2). This plateau value is reached at 8 m
the term for the head group/substrate interaction. mol dm3 SDS, that is, close to the cmc of the
Several other contributions to DGspec may be surfactant in the bulk solution. The area per sur-
envisaged, for example, ion–dipole, ion–induced factant ion in this case is 0.7 nm2. Graphitiza-
dipole, or dipole–induced dipole interactions. tion (Graphon) removes the hydrophilic ionizable
Suspensions 1353 S
Graphon in 10–1 mol dm–3 NaCI 10−5 III −40
5 −30
−20
10−6
4 Γ −10
Γ Mol m=2
Graphon
II 0
ζ/mV
m molm–2
3 10−7 ζ 10
Spheron 6 after washing
G
I 20
2
10−8 30
Spheron 6 before 40
1
washing 50
0 10−9
4 8 12 16 20 24 10−5 10−4 10−3 10−2
–3 −3
C/mmol dm C/mol dm
Suspensions, Fig. 11 Adsorption isotherms of SDS on Suspensions, Fig. 12 Adsorption isotherm for sodium
carbon black dodecyl sulfonate on alumina and corresponding zeta (z)
potential
groups (e.g.,-C═O or -COOH), producing

a surface that is more hydrophobic. The same this does not automatically imply a simple orien-
occurs by heating Spheron 6–2,700 C. This tation. For example, rearrangement from hori-
leads to a different adsorption isotherm (Fig. 1) zontal to vertical orientation or electrostatic
showing a step (inflection point) at a surfactant interaction and counterion binding may be
concentration in the region of 6 m mol dm3. masked by simple adsorption isotherms. It is
The first plateau value is 2.3 m mol m2, essential, therefore, to combine the adsorption
whereas the second plateau value (that occurs at isotherms with other techniques such as micro-
the cmc of the surfactant) is 4 m mol m2. It is calorimetry and various spectroscopic methods to
likely in this case that the surfactant ions adopt obtain a full picture on surfactant adsorption.
different orientations at the first and second pla- (ii) Adsorption of ionic surfactants on polar
teaus. In the first plateau region, a more “flat” surfaces: The adsorption of ionic surfactants on
orientation (alkyl chains adsorbing parallel to the polar surfaces that contain ionizable groups may
surface) is obtained, whereas at the second show characteristic features due to additional
plateau, vertical orientation is more favorable, interaction between the head group and substrate
with the polar head groups being directed and/or possible chain–chain interaction. This is
toward the solution phase. Addition of electrolyte best illustrated by the results of adsorption of
(101 mol dm3 NaCl) enhances the surfactant sodium dodecyl sulfonate (SDSe) on alumina at
adsorption. This increase is due to the reduction pH ¼ 7.2 obtained by Fuerstenau (Fuerestenau
of lateral repulsion between the sulfate head 1971) and shown in Fig. 12. At the pH value,
groups, and this enhances the adsorption. the alumina is positively charged (the isoelectric S
The adsorption of ionic surfactants on other point of alumina is at pH 9), and the counter-
hydrophobic surfaces resembles that for carbon ions are Clfrom the added supporting electro-
black (Greenwood et al. 1968; Day et al. 1967). lyte. In Fig. 12, the saturation adsorption G1 is
For example, Saleeb and Kitchener (Saleeb and plotted versus equilibrium surfactant concentra-
Kitchener 1965; Conner and Ottewill 1971) tion C1 in logarithmic scales. The figure also
found similar limiting area for cetyltrimethy- shows the results of zeta potential (z) measure-
lammonium bromide on Graphon and polysty- ments (which are a measure of the magnitude
rene (0.4 nm2). As with carbon black, the area sign of charge on the surface). Both adsorption
per molecule depends on the nature and amount and zeta potential results show three distinct
of added electrolyte. This can be accounted for in regions. The first region which shows a gradual
terms of reduction of head group repulsion and/or increase of adsorption with increase in concen-
counterion binding. Surfactant adsorption close tration, with virtually no change in the value of
to the cmc may appear Langmuirian, although the zeta potential corresponds to an ion-exchange
S 1354 Suspensions
Suspensions,
Fig. 13 Adsorption
isotherms corresponding to
the three adsorption V
sequences shown in Fig. 4
Γ
IV
IV
III III
I II I II III
I II
cmc cmc cmc
X2
process (Fuerestenau 1971). In other words, the group. The most common type is that based on
surfactant ions simply exchange with the coun- a poly(oxyethylene) glycol group, that is,
terions (Cl) of the supporting electrolyte in the (CH2CH20)nOH (where n can vary from as little
electrical double layer. At a critical surfactant as 2 units to as high as 100 or more units) linked
concentration, the desorption increases dramati- either to an alkyl (CxH2x+1) or alkyl phenyl
cally with further increase in surfactant concen- (CxH2x+1-C6H4-) group. These surfactants may
tration (region II). In this region, the positive zeta be abbreviated as CxEn or CxjEn (where
potential gradually decreases, reaching a zero C refers to the number of C atoms in the alkyl
value (charge neutralization), after which chain, j denotes C6H4, and E denotes ethylene
a negative value is obtained which increases rap- oxide). These ethoxylated surfactants are charac-
idly with increase in surfactant concentration. terized by a relatively large head group compared
The rapid increase in region II was explained in to the alkyl chain (when n > 4). However, there
terms of “hemi-micelle formation” that was orig- are nonionic surfactants with small head group
inally postulated by Gaudin and Fuerestenau such as amine oxides (-N ! 0) head group,
(Wakamatsu and Fuerstenau 1968; Gaudin and phosphate oxide (-P ! 0), or sulphinyl-alkanol
Fuertsenau 1955). In other words, at a critical (-SO-(CH2)n-OH). Most adsorption isotherms in
surfactant concentration (to be denoted the cmc the literature are based on the ethoxylated-type
of “hemi-micelle formation” or better the critical surfactants. The adsorption isotherm of nonionic
aggregation concentration CAC), the hydropho- surfactants is in many cases Langmuirian, like
bic moieties of the adsorbed surfactant chains are those of most other highly surface active solutes
“squeezed out” from the aqueous solution by adsorbing from dilute solutions, and adsorption is
forming two-dimensional aggregates on the generally reversible. However, several other
adsorbent surface. This is analogous to the pro- adsorption types are produced (Clunie and
cess of micellization in bulk solution. However, Ingram 1983), and these are illustrated in Fig. 13.
the CAC is lower than the cmc, indicating that the The steps in the isotherm may be explained
substrate promotes surfactant aggregation. At in terms of the various adsorbate–adsorbate,
a certain surfactant concentration in the hemi- adsorbate–adsorbent, and adsorbate–solvent
micellization process, the isoelectric point is interactions. These orientations are schematically
exceeded, and, thereafter, the adsorption is hin- illustrated in Fig. 14. In the first stage of
dered by the electrostatic repulsion between the adsorption (denoted by I in Figs. 13 and 14),
hem-micelles, and hence, the slope of the adsorp- surfactant–surfactant interaction is negligible
tion isotherm is reduced (region III). (low coverage), and adsorption occurs mainly
(iii) Adsorption of nonionic surfactants: Sev- by van der Waals interaction. On a hydrophobic
eral types of nonionic surfactants exist, surface, the interaction is dominated by the
depending on the nature of the polar (hydrophilic) hydrophobic portion of the surfactant molecule.
Suspensions 1355 S
I extended conformation. The larger the surfactant
alkyl chain, the greater will be the cohesive forces
II and hence the smaller the cross-sectional area.
This may explain why saturation adsorption
A B C increases with increasing alkyl chain length.
III The interaction occurring in the adsorption
layer during the fourth and subsequent stages of
IV adsorption is similar to those that occur in bulk
solution. In this case aggregate units, as shown in
Suspensions, Fig. 14 Model for adsorption of nonionic Fig. 14, V (hemi-micelles or micelles) may be
surfactants formed. This picture was supported by Kleminko
et al. (Klimenko 1974) who found close agree-
ment between saturation adsorption and adsorp-
This is mostly the case with hydrophobic sur- tion calculated based on the assumption that the
faces. However, if the chemical is hydrophilic surface is covered with close-packed hemi-
in nature, the interaction will be dominated by micelles. Kleminko (1978, 1979) developed
the EO chain. The approach to monolayer satura- a theoretical model for the three stages of adsorp-
tion with the molecules lying flat is accompanied tion of nonionic surfactants. In the first stage (flat
by a gradual decrease in the slope of the adsorp- orientation), a modified Langmuir adsorption
tion isotherm (region II in Fig. 13). Increase in the equation was used. In the second stage of hori-
size of the surfactant molecule, for example, zontal orientation, the surface concentration
increasing the length of the alkyl or EO chain, increases by an amount that is determined by
will decrease adsorption (when expressed in the displacement of the ethoxy chain by the
moles per unit area). On the other hand, increas- alkyl group. Finally, in the region of the hemi-
ing temperature will increase adsorption as micelle formation, the adsorption can be
a result of desolvation of the EO chains, thus described by a simple Langmuir equation of the
reducing their size. Moreover, increasing temper- form
ature reduces the solubility of the nonionic
surfactant, and this enhances adsorption. G2
C2 K
a ¼ infinity
(67)
The subsequent stages of adsorption (region ðG2 G2 Þ
III and IV) are determined by surfactant–
surfactant interaction, although surfactant– where G21 is the maximum surface excess, that
surface interaction initially determines adsorption is, the surface excess when the surface is covered
beyond stage II. This interaction depends on with close-packed hemi-micelles, Ka* is
the nature of the surface and the hydrophilic– a constant that is inversely proportional to the S
lipophilic balance of the surfactant molecule cmc, and C2 is the equilibrium concentration.
(HLB). For a hydrophobic surface, adsorption
occurs via the alkyl group of the surfactant. For
a given EO chain, the adsorption will increase with Polymer (Polymeric Surfactant)
increase in the alkyl chain length. On the other Adsorption
hand, for a given alkyl chain length, adsorption
increases with decrease of the PEO chain length. The simplest type of a polymeric surfactant is
As the surfactant concentration approaches a homopolymer that is formed from the same
the cmc, there is a tendency for aggregation of repeating units: Poly(ethylene oxide) (PEO) and
the alkyl groups. This will cause vertical orienta- poly(vinylpyrrolidone) (PVP). Homopolymers
tion of the surfactant molecules (stage IV). This have little surface activity at the oil/water
will compress the head group, and for an EO (O/W) interface. However, homopolymers may
chain, this will result in a less coiled more adsorb significantly at the solid/liquid (S/L)
S 1356 Suspensions
interface even if the adsorption energy per mono- information on their conformation; properties in
mer segment is small (fraction of kT, where k is a solvent which is a non-solvent for one of the
the Boltzmann constant and T is the absolute blocks but a good solvent for the other block.
temperature); the total adsorption energy per Graft copolymers also aggregate in solution to
molecule may be sufficient (several segments form micelles, with small aggregation numbers.
are adsorbed at the surface) to overcome the A dimer may be the form of aggregation. Most
unfavorable entropy loss of the molecule at the block and graft copolymers have low critical
S/L interface. Homopolymers may also adsorb at micelle concentrations (cmc), and in many
the solid surface by some specific interaction, for cases, it is not easy to measure the cmc for these
example, hydrogen bonding (e.g., adsorption of block and graft copolymers. The aggregation pro-
PEO or PVP on silica). In general, homopolymers cess is also affected by temperature and solvency
are not the most suitable emulsifiers or disper- of the medium for the A chains. One of the most
sants. A small variant is to use polymers that useful methods to follow the aggregation of block
contain specific groups that have high affinity to and graft copolymers is to use time-average light
the surface, for example, partially hydrolyzed scattering. By measuring the intensity as
poly(vinyl acetate) (PVAc) – technically referred a function of concentration, one can extrapolate
to as poly(vinyl alcohol) (PVA) – commercially the results to zero concentration and obtain the
available PVA molecules contain 4–12 % acetate molecular weight of the micelle – this allows one
groups. The acetate groups give the molecule its to obtain the aggregation number from
amphipathic character; on a hydrophobic surface a knowledge of the molecular weight of the
(such as polystyrene), the polymer adsorbs with monomer.
preferential attachment of the acetate groups on Several examples of block and graft copoly-
the surface, leaving the more hydrophilic vinyl mers may be quoted: triblock polymeric surfac-
alcohol segments dangling in the aqueous tants “Pluronics” (BASF), two poly-A blocks of
medium. Partially hydrolyzed PVA molecules PEO, and one poly-B block of polypropylene
exhibit surface activity at the O/W interface. oxide (PPO). Several chain lengths of PEO and
The most convenient polymeric surfactants are PPO are available. Triblocks of PPO-PEO-PEO
those of the block and graft copolymer type. (inverse “Pluronics”) are also available. Poly-
A block copolymer is a linear arrangement of meric triblock surfactants can be applied as emul-
blocks of varying composition (Piirma 1992; sifiers and dispersants. The hydrophobic PPO
Strobel 2006): chain resides at the hydrophobic surface, leaving
the two PEO chains dangling in aqueous solution
Diblock - Poly A - Block Poly B A (providing steric stabilization). These triblocks
B
are not the most efficient emulsifiers or disper-
Triblock - Poly A - Block Poly B - Poly A
sants; the PPO chain is not sufficiently hydropho-
AB
A bic to provide a strong “anchor” to a hydrophobic
surface. Several other di- and triblock copoly-
A graft copolymer is a nonlinear array of one mers have been synthesized: Diblocks of
B block on which several A polymers are grafted polystyrene block-polyvinyl alcohol; triblocks
of poly(methyl methacrylate)-block polyethylene
B oxide-poly(methyl methacrylate); diblocks of
| | | | | polystyrene-polyethylene oxide; triblocks of
A A A A A polyethylene oxide-polystyrene-polyethylene
oxide. An alternative (and perhaps more effi-
Two types of investigations are essential to cient) polymeric surfactant is the amphipathic
unravel the behavior of block and graft copoly- graft copolymer consisting of a polymeric back-
mers: Properties in a solvent in which both the bone B (polystyrene or polymethylmethacrylate)
A and B blocks are soluble – this gives and several A chains (“teeth”) such as
Suspensions 1357 S
polyethylene oxide. The graft copolymer is stabilizers. Most basic ideas on adsorption and
referred to as a “comb” stabilizer – the polymer conformation of polymers have been developed
forms a “brush” at the solid/liquid interface. The for the solid/liquid interface (Tadros 1985). The
copolymer is usually prepared by grafting a process of polymer adsorption involves a number
macromonomer such as methoxy polyethylene of various interactions that must be separately
oxide methacrylate with polymethylmethacrylate. considered. Three main interactions must be
In most cases, some polymethacrylic acid is taken into account, namely, the interaction of
incorporated with the polymethylmethacrylate the solvent molecules with the surface (or oil in
backbone; this leads to reduction of the glass tran- the case of o/w emulsions) which need to be
sition of the backbone which makes the chain displaced for the polymer segments to adsorb,
more flexible for adsorption at the solid/liquid the interaction between the chains and the sol-
interface. The “grafting into” technique has also vent, and the interaction between the polymer and
been used to synthesize polystyrene-polyethylene the surface. Apart from knowing these interac-
oxide graft copolymers. These molecules are not tions, one of the most fundamental considerations
commercially available. is the conformation of the polymer molecule at
Recently, a novel graft copolymer based on the interface. These molecules adopt various con-
a naturally occurring polysaccharide, namely, formations, depending on their structure. The
Inulin (polyfructose), has been synthesized simplest case to consider is that of
(Stevens et al. 2001). Inulin is a polydisperse a homopolymer that consists of identical seg-
polysaccharide consisting mainly, if not exclu- ments (e.g., poly(ethylene oxide)), which shows
sively, of b(2 ! 1) fructosyl fructose unite (Fm) a sequence of loops, trains, and tails, as is illus-
with normally, but not necessarily, one trated in Fig. 15a. It should be mentioned at this
glucopyranose unit at the reducing end (GFn) stage that for such a polymer to adsorb, the reduc-
(Hirst et al. 1950; Suzuki 1993). To produce the tion in entropy of the chain as it approaches the
amphipathic graft copolymer, the chains were interface must be compensated by an energy of
modified by introduction of alkyl groups adsorption between the segments and the surface.
(C4—C18) on the polyfructose backbone through In other words, the chain segments must have
isocyanates. The alkyl groups represent the a minimum adsorption energy, ws; otherwise, no
B chains (that are randomly distributed on the adsorption occurs. With polymers that are highly
sugar backbone on primary hydroxyl functions water soluble, such as poly(ethylene oxide)
as well as on the secondary ones) which become (PEO), the interaction energy with the surface
strongly adsorbed on a hydrophobic solid such as may be too small for adsorption to occur, and if
carbon black, polystyrene, or an oil droplet. The this takes place, the whole molecule may not be
sugar chain forms the stabilizing chain as this is strongly adsorbed to the surface. For this reason,
highly water soluble. These graft copolymers are many commercially available polymers that are S
surface active, and they lower the surface tension described as hompolymers, such as poly(vinyl
of water and the interfacial tension at the alcohol) (PVA), contain some hydrophobic
oil/water interface. They will also adsorb on groups or short blocks (vinyl acetate in the case
hydrophobic surfaces with the alkyl groups of PVA) that ensured their adsorption to hydro-
strongly attached (multipoint anchoring) leaving phobic surfaces. This is illustrated in Fig. 15b.
the polyfructose chains dangling in solution and Clearly, if all the segments have a high affinity to
probably forming large loops. These graft copol- the surface, the whole molecule may lie flat on
ymers can produce highly stable suspensions and the surface, as illustrated in Fig. 5c. This situation
emulsions, in particular at high electrolyte con- is rarely the case since the molecule will have
centrations (Tadros et al. 2004). very low solubility in the continuous medium.
Understanding the adsorption and conforma- The most favorable structures for polymeric sur-
tion of polymeric surfactants at interfaces is key factants are those represented in Fig. 15d, e, and f,
to knowing how these molecules act as referred to as block and graft copolymers.
S 1358 Suspensions
Suspensions, a Loops b c
Fig. 15 Various
conformations of
polymeric surfactants Tail
adsorbed on a plane
surface: (a) random
conformation of loops-
trains-tails (hompolymer); Trains
(b) preferential adsorption
Homopolymer Chains with “blocks” that Chain lying
of “short blocks”; (c) chain
sequence of loops- have higher affinity to flat on the
lying flat on the surface;
tails and trains the surface surface
(d) AB block copolymer
with loop-train
configuration of B and long d e f
tail of A; (e) ABA block as
in d; (f) BAn graft with
backbone B forming small
loops and several tails of
A (“teeth”)
A-B block A-B-A BAn graft one

B forms small block B chain (small
loops and A loops) and
are tails several A chains
The molecule shown in Fig. 15d is an A-B block, solution, that is, the adsorbed layer thickness d.
consisting of a B chain that has a high affinity to It is important to know how these parameters
the surface (or soluble in the oil phase), referred change with polymer coverage (concentration),
to as the “anchoring” chain and an A chain that the structure of the polymer, and its molecular
has very low affinity to the surface and strongly weight. It is also essential to know how these
solvated by the medium. As will be discussed in parameters change with the environment such as
the section on stabilization, this is the most con- solvency of the medium fort the chains and
venient structure since the forces that ensure temperature.
strong adsorption are opposite to those that Several theories exist that describe the process
ensure stability. A variance on the structure of polymer adsorption, which have been devel-
shown in Fig. 15d is the A-B-A block copolymer oped either using statistical mechanical approach
shown in Fig. 15e. In this case, the anchor chain or quasi-lattice models. In the statistical mechan-
B contains two stabilizing chains (tails). Another ical approach, the polymer is considered to con-
variance is that shown in Fig. 5f, which is sist of three types of structures with different
described as a graft copolymer (“comb”-type energy states, trains, loops, and tails (Silberberg
structure) with one B chain and several A chains 1968; Hoeve 1970, 1971). The structures close to
(tails or “teeth”). the surface (trains) are adsorbed with an internal
It is clear from the above description of poly- partition function determined by short-range
mer configurations that for full characterization forces between the segment and surface (assigned
of the process of adsorption, it is necessary to an adsorption energy per segment ws). The seg-
know the following parameters, namely, the ments in loops and tails are considered to have an
amount of polymer adsorbed per unit area of the internal partition function equivalent to that of
surface, G (mol m2 or mg m2); the fraction of segments in bulk solution, and these are assigned
segments in close contact with the surface, p; and a segment-solvent interaction parameter w
the distribution of polymer segments, r(z), from (Flory–Huggins interaction parameter). By
the surface toward the bulk solution. It is essential equating the chemical potential of the macromol-
to know how far the segments extend into ecule in the adsorbed state and in bulk solution,
Suspensions 1359 S
χs = 1 kT r = 5000
Surface 3
Tail r = 1000
Monolayers
Equivalent
Train
2
χ = 0.5
θ
r = 100
1 r = 1000 χ=0
Loop r = 100 χ=0 r = 20
r = 10
r=1
10–4 5 x 10–4 10–3
ϕ*
Suspensions, Fig. 17 Adsorption isotherms for oligo-

Suspensions, Fig. 16 Schematic representation of
mers and polymers in the dilute region based on the SF
a polymer molecule adsorbing on a flat surface quasicrys-
theory
talline lattice with segments filling layers that are parallel
to the surface (random mixing of segment and solvent
molecules in each layer is assumed)
monolayers) versus polymer volume fraction
j∗ in bulk solution (j∗ was taken to vary
the adsorption isotherm can be determined. In the between 0 and 103 which is the normal experi-
earlier theories, the case of an isolated chain on mental range). The results in Fig. 17 show the
the surface (low coverage) was considered, but effect of increasing the chain length r and effect
later, the theories were modified to take into of solvency (using two values for the Flory–
account the lateral interaction between the Huggins interaction parameter, i.e., w ¼ 0
chains, that is, at high coverage. (athermal solvent) and w ¼ 0.5 (theta solvent)).
The quasi-lattice model was developed by Roe As the number of segments in the chain increases
(1974) and by Scheutjens and Fleer (Scheutjens from a low (with few segments) to high
and Fleer 1979, 1980, 1982; Fleer et al. 1993). (large number of segments) values, the adsorp-
The basis procedure was to describe all chain tion isotherm changes from a Langmuirian type
conformations as step-weighted random walks (characteristic for surfactant adsorption) to high-
on a quasicrystalline lattice which extends in affinity type. In the latter case, the first addition of
parallel layers away from the surface. This is polymer chains to the solution results in their
illustrated in Fig. 16 which shows a possible con- virtual complete adsorption. The adsorption
formation of a polymer molecule at a surface. isotherms for chains with r ¼ 100 and above
The partition function was written in terms of are typical of those obtained experimentally for
number of chain configurations which were most polymers that are not too polydisperse, that S
treated as connected sequences of segments. In is, showing a steep rise followed by a nearly
each layer, random mixing (Bragg–William or horizontal pseudoplateau (which only increases
mean field approximation) between segments few percent per decade of j∗). Adsorption in
and solvent molecules was assumed. Each step this case is described as being “irreversible,”
in the random walk was assigned a weighting that is, the equilibrium between adsorbed and
factor pi that was considered to consist of three free polymer is shifted toward the surface. This
contributions, namely, the adsorption energy ws, explains the strong anchoring of the polymer
the configurational entropy of mixing, and the chains to the surface. As the solvency of the
segment–solvent interaction parameter w. medium for the chains decreases, the amount of
The above theories gave a number of predic- polymer adsorbed increases. This is clearly illus-
tions for polymeric surfactant adsorption. trated in Fig. 17 when comparing the results
Figure 17 shows typical adsorption isotherms obtained when w ¼ 0 (very good solvent) with
plotted as surface coverage y (in equivalent those obtained using a poor solvent with w ¼ 0.5.
S 1360 Suspensions
In good solvents (dashed lines in Fig. 17), y is region, the conformational entropy term domi-
much smaller and levels off for long chains to nates the free energy, disfavoring the adsorption
attain an adsorption plateau which is essentially of long chains.
independent of molecular weight. This explains The bound fraction, p, is high at low polymer
the relatively “weaker” adsorption of homopoly- concentrations (j∗ < j∗c) approaching unity,
mers that are highly solvated by the medium. It is and it is relatively independent of molecular
now clear from these theories why block and graft weight when r > 20. p also increases with
copolymers are preferred for stabilization of dis- increase of the adsorption energy, ws, but it
persions. The poor solubility of the anchor chain decreases with increase in surface coverage and
B in the medium and its strong affinity to the increase of the molecular weight of the polymer.
surface ensure the strong adsorption of the mole- The structure of the adsorbed layer is described in
cule. In contrast, the high solubility of the stabi- terms of the segment density distribution, r(z).
lizing chain A ensures effective steric Determination of the segment density distribu-
stabilization. Another prediction from the theo- tion is not easy, and one usually assigns a value
ries is that the higher the molecular weight of the for the adsorbed layer thickness d. This increases
polymer, the higher the amount of adsorption with increase of the molecular weight of the
when the latter is expressed in mg m2. polymer and increase of solvency of the medium
Another point that emerges from Scheutjens for the chains.
and Fleer’s theory (Scheutjens and Fleer 1979,
1980, 1982; Fleer et al. 1993) is the difference in
shape between experimental and theoretical Experimental Methods for
adsorption isotherms in the low concentration Measurement of Adsorption
region. The experimental isotherms are usually Parameters of Polymeric Surfactants
rounded, whereas those predicted from theory are
flat. This is accounted for in terms of the molec- Amount of Polymer Adsorbed, G: The
ular weight distribution (polydispersity) which is Adsorption Isotherms
encountered in many practical systems. This The amount of polymer adsorbed, G, can be
effect has been explained by Cohen-Stuart et al. directly determined in a similar way as described
(Fleer et al. 1993). With polydisperse polymer for surfactants, except in this case one has to
fractions, the larger molecules adsorb preferen- consider the relatively slow adsorption process
tially over the smaller ones. At low polymer which may take several hours or even days to
concentrations, nearly all molecular weights are reach equilibrium. In addition, one needs very
adsorbed, leaving only a small fraction of poly- sensitive analytical methods for determination
mer with the lowest molecular weight in solution. of polymer concentration in the early stages of
As the polymer concentration is increased, the adsorption (which can be in the ppm range). As
higher molecular weight fractions displace the mentioned before, the amount of adsorption G
lower ones on the surface, which are now can be calculated from a knowledge of the initial
released in solution, thus shifting the molecular polymer concentration C1, and that after reaching
weight distribution of the polymer in bulk solu- equilibrium C2, the mass of the solid m and the
tion to lower values. This process continues with specific surface area As are given by Eq. 11.
further increase in polymer concentration leading As an illustration, Fig. 18 shows the adsorp-
to a fractionation process whereby the higher tion isotherms at 25o C for poly(vinyl alcohol)
molecular weight fractions are adsorbed at the (PVA) (containing 12 % acetate). The polymer
expense of the lower fractions which are released was fractionated using preparative gel perme-
in the bulk. However, in very concentrated solu- ation chromatography (Garvey et al. 1974) or by
tions, monomers adsorb preferentially with a sequential precipitation technique using ace-
respect to polymers and short chains with respect tone (van den Boomgaard et al. 1978). The frac-
to larger ones. This is due to the fact that in this tions were characterized for their molecular
Suspensions 1361 S
3 bonds between the vinyl alcohol units and water).
It shows systematic increase of adsorption with
M = 67000
M = 45000 increase of temperature, that is, with reduction of
2 M = 43000 solvency (increase in the value of the w parame-
Γ/mgm2
Unfractionated
M = 28000 ter), as expected from theory. Addition of elec-
M = 17000
trolyte increases adsorption of PVA, again as
1
a result of the reduction of solvency of the
M = 8000
medium for the chains.
The above polymer (PVA) is a “blocky”
100 200 300 400 500 600 copolymer (containing short vinyl acetate
C2 blocks), and hence, it does not represent the
case for adsorption of homopolymers. The latter
Suspensions, Fig. 18 Adsorption isotherms of PVA
with different molecular weights on polystyrene latex at case is exemplified by poly(ethylene oxide)
250C (PEO) (Obey and Griffiths 1999) for adsorption
on polystyrene latex using three different molec-
ular weight PEOs. As with PVA, the isotherms
weight using ultracentrifugation and later by are of the high-affinity type, and the adsorbed
intrinsic viscosity measurements. amount increases with increase of the molecular
The intrinsic viscosity [Z] could be related to weight of the polymer. However, the amount of
the weight average molecular weight of the poly- adsorption is much lower than that obtained using
mer (determined using ultracentrifugation) using PVA. This reflects the difference between the two
the Mark–Houwink relationship: polymers.
½Z ¼ K Ma (68) Polymer-Bound Fraction, p

The bound fraction p represents the ratio of the
The constants K and a were established from number of segments in close contact with the
knowledge of [Z] and M. The latter values could surface (i.e., in trains) to the total number of
also be used to calculate the molecular segments in the polymer chain. The polymer-
dimensions (radius of gyration), and the poly- bound fraction, p, can be directly determined
mer-solvent interaction parameter w was also using spectroscopic methods such as infrared
determined. The polystyrene latex used for the (IR), electron spin resonance (ESR), and nuclear
adsorption measurements was a model system magnetic resonance (NMR). The IR method
prepared using surfactant-free polymerization, depends on measuring the shift in some absorp-
and the particles were fairly monodisperse. tion peak for a polymer and/or surface group
Hence, the specific surface area of the particles (Killmann et al. 1977; Fontana and Thomas S
could be estimated from simple geometry using 1961). The ESR and NMR methods depend on
electron microscopy. the reduction in the mobility of the segments that
Figure 18 shows the high-affinity isotherms are in close contact with the surface (larger rota-
for the polymers and the increase in adsorption tional correlation time for trains when compared
of the polymer with increase of the molecular to loops). By using a pulsed NMR technique, one
weight. To study the effect of solvency on can estimate p (Robb and Smith 1974; Barnett
adsorption, measurements were carried out as et al. 1981). An indirect method for estimation of
a function of temperature (van den Boomgaard p is to use microcalorimetry. Basically, one com-
et al. 1978) and addition of electrolyte (KCl or pares the enthalpy of adsorption per molecule
Na2SO4) (Tadros and Vincent 1978). Increasing with that per segment (Cohen-Stuart et al.
temperature and/or addition of electrolyte 1982). The latter may be obtained by using
reduces the solvency of the medium for the small molecules of similar structure to
PVA chains (due to breakdown of the hydrogen a polymer segment.
S 1362 Suspensions
Adsorbed Layer Thickness, d, and Segment r(z) with z at long distances is in qualitative
Density Distribution, r(z) agreement with Scheutjens and Fleers’s theory
Three direct methods can be applied for determi- (Scheutjens and Fleers 1980, 1982) which pre-
nation of adsorbed layer thickness: ellipsometry, dicts the presence of long tails. The shallow max-
attenuated total reflection (ATR), and neutron imum at intermediate distances suggests that the
scattering. Both ellipsometry and ATR (Abeles observed segment density distribution is
1964) depend on the difference between refrac- a summation of a fast monotonic decay due to
tive indices, between the substrate, the adsorbed loops and trains together with the segment den-
layer, and bulk solution, and they require flat sity for tails which is a maximum density away
reflecting surface. Ellipsometry (Abeles 1964) is from the surface. The latter maximum was clearly
based on the principle that light undergoes observed for a sample which had PEO grafted to
a change in polarizability when it is reflected at a deuterated polystyrene latex (where the config-
a flat surface (whether covered or uncovered with uration is represented by tails only).
a polymer layer).
The above limitations when using Hydrodynamic Thickness Determination
ellipsometry or ATR are overcome by the appli- Several methods may be applied to determine the
cation technique of neutron scattering, which can hydrodynamic thickness of adsorbed polymer
be applied to both flat surfaces as well as partic- layers of which viscosity, sedimentation coeffi-
ulate dispersions. The basic principle of neutron cient (using an ultracentrifuge), and dynamic
scattering is to measure the scattering due to the light scattering measurements are the most con-
adsorbed layer when the scattering length density venient. A less accurate method is from zeta
of the particle is matched to that of the medium potential measurements.
(the so-called contrast-matching method). The viscosity method (Einstein 1906) depends
Contrast matching of particles and medium can on measuring the increase in the volume fraction
be achieved by changing the isotopic composi- of the particles as a result of the presence of an
tion of the system (using deuterated particles and adsorbed layer of thickness dh. The volume frac-
mixture of D2O and H2O). It was also used tion of the particles j plus the contribution of
for measurement of the adsorbed layer thickness the adsorbed layers is usually referred to as the
of polymers, for example, PVA or poly(ethylene effective volume fraction jeff. Assuming
oxide) (PEO) on polystyrene latex (Cosgrove the particles behave as hard spheres, then the
et al. 1987). Apart from obtaining d, one measured relative viscosity Zr is related to
can also determine the segment density the effective volume fraction by the Einstein’s
distribution r(z). equation, that is,
The results show a monotonic decay of r(z)
with distance z from the surface, and several Zr ¼ 1 þ 2:5 feff (69)
regions may be distinguished. Close to the sur-
face (0 < z < 3 nm), the decay in r(z) is rapid, jeff and j are related from simple geometry by
and assuming a thickness of 1.3 nm for the bound 3
layer, p was calculated to be 0.1, which is in close dh
feff ¼f 1þ (70)
agreement with the results obtained using NMR R
measurements. In the middle region, r(z) shows
a shallow maximum followed by a slow decay where R is the particle radius. Thus, from
which extends to 18 nm, that is, close to the a knowledge of Zr and j, one can obtain dh
hydrodynamic layer thickness dh of the polymer using the above equations.
chain (see below). dh is determined by the longest The sedimentation method depends on
tails and is about 2.5 times the radius of gyration measuring the sedimentation coefficient (using
in bulk solution (7.2 nm). This slow decay of an ultracentrifuge) of the particles So0
Suspensions 1363 S
(extrapolated to zero concentration) in the pres- Assuming the particles obey Stokes’ law, So0 is
ence of the polymer layer (Garvey et al. 1976). given by the expression
ð4=3Þp R3 ðr rs Þ þ ð4=3Þp½ðR þ dh Þ3 R3 ðrads

s rs Þ
So0 ¼ ð71Þ
6pZðR þ dh Þ
where r and rs are the mass density of the solid comparable to the wavelength of the light
and solution phase, respectively, and rads is the beam, i.e., the light is coherent with the
average mass density of the adsorbed layer which interparticle distance), the intensity of the
may be obtained from the average mass concen- scattered light will, at any instant, appear as
tration of the polymer in the adsorbed layer. In random diffraction or “speckle” pattern. As the
order to apply the above methods, one should use particles undergo Brownian motion, the random
a dispersion with monodisperse particles with configuration of the speckle pattern changes.
a radius that is not much larger than dh. Small The intensity at any one point in the pattern
model particles of polystyrene may be used. will, therefore, fluctuate such that the time
A relatively simple sedimentation method for taken for an intensity maximum to become
determination of dh is the slow speed centrifuga- a minimum (i.e., the coherence time) corre-
tion applied by Garvey et al. (1976). Basically, sponds approximately to the time required for
a stable monodisperse dispersion is slowly a particle to move one wavelength. Using
centrifuged at low g values (< 50 g) to form a photomultiplier of active area about the size
a close-packed (hexagonal or cubic) lattice in of a diffraction maximum, that is, approximately
the sediment. From a knowledge of j and the one coherence area, this intensity fluctuation can
packing fraction (0.74 for hexagonal packing), be measured. A digital correlator is used to mea-
the distance of separation between the center of sure the photocount or intensity correlation
two particles Rd may be obtained, that is, function of the scattered light. The photocount
correlation function can be used to obtain the
o:74V r1 R3 diffusion coefficient D of the particles. For
Rd ¼ R þ dh ¼ (72)
W monodisperse noninteracting particles (i.e., at
sufficient dilution), the normalized correlation
where V is the sediment volume, r1 is the density function [g(1)(t)] of the scattered electric field
of the particles, and W is their weight. is given by the equation
The most rapid technique for measuring dh is
photon correlation spectroscopy (PCS) (some- ½gð1Þ ðtÞ ¼ exp ðGtÞ (73) S
times referred to as quasi-elastic light scattering)
which allows one to obtain the diffusion coeffi- where t is the correlation delay time and G is the
cient of the particles with and without the decay rate or inverse coherence time. G is related
adsorbed layer (Dd and D, respectively). This is to D by the equation
obtained from measurement of the intensity fluc-
tuation of scattered light as the particles undergo G ¼ D K2 (74)
Brownian diffusion (Pusey 1973). When a light
beam (e.g., monochromatic laser beam) passes where K is the magnitude of the scattering vector
through a dispersion, an oscillating dipole is that is given by
induced in the particles, thus reradiating the
light. Due to the random arrangement of the 4n y
K¼ sin (75)
particles (which are separated by a distance lo 2
S 1364 Suspensions
where n is the refractive index of the solution, l is obtain dh from electrophoretic mobility measure-
the wavelength of light in vacuum, and y is the ments, results should be obtained at various elec-
scattering angle. trolyte concentrations and de should be plotted
From D, the particle radius R is calculated versus the Debye length (1/k) to obtain the limit-
using the Stokes–Einstein equation: ing value at high (1/k) (i.e., low electrolyte con-
centration) which now corresponds to dh.
kT
D¼ (76)
6pZR
Steric Stabilization of Suspensions
where k is the Boltzmann constant and T is the
absolute temperature. For a polymer-coated par- The use of natural and synthetic polymers
ticle, R is denoted Rd which is equal to R + dh. (referred to as polymeric surfactants) for stabili-
Thus, by measuring Dd and D, one can obtain dh. zation of solid/liquid dispersions plays an impor-
It should be mentioned that the accuracy of the tant role in industrial application (Tadros 1981)
PCS method depends on the ratio of dd/R since dh such as in paints, cosmetics, agrochemicals, and
is determined by difference. Since the accuracy ceramics. Polymers are particularly important for
of the measurement is 1 %, dh should be at least preparation of concentrated dispersions, that is, at
10 % of the particle radius. This method can only high volume fraction j of the disperse phase:
be used with small particles and reasonably thick j ¼ (volume of all particles)/(total volume of
adsorbed layers. Electrophoretic mobility, u, dispersion)
measurements can also be applied to measure dh Polymers are also essential for stabilization of
(Cohen-Stuart and Mulder 1985). From u, the nonaqueous dispersions since in this case, elec-
zeta potential z, that is, the potential at the slip- trostatic stabilization is not possible (due to the
ping (shear) plane of the particles, can be calcu- low dielectric constant of the medium).
lated. Adsorption of a polymer causes a shift in To understand the role of polymers in disper-
the shear plane from its value in the absence of sion stability, it is essential to consider the
a polymer layer (which is close to the Stern plane) adsorption and conformation of the macromole-
to a value that depends on the thickness of the cule at the solid/liquid interface (Tadros 1981).
adsorbed layer. Thus, by measuring z in the pres- When two particles, each with a radius R and
ence (zd) and absence (z) of a polymer layer, one containing an adsorbed polymer layer with
can estimate dh. Assuming that the thickness of a hydrodynamic thickness dh, approach each
the Stern plane is D, then zd may be related to the other to a surface-surface separation distance
z (which may be assumed to be equal to the Stern h that is smaller than 2 dh, the polymer layers
potential cd) by the equation interact with each other resulting in two main
situations (Tadros 1981; Napper 1981): The poly-
e cd ez mer chains may overlap with each other. The
tanh ¼ tanh exp½kðdh DÞ
4kT 4kT polymer layer may undergo some compression.
(77) In both cases, there will be an increase in the local
segment density of the polymer chains in the
where k is the Debye parameter that is related to interaction region. This is schematically illus-
electrolyte concentration and valency. It should trated in Fig. 19.
be mentioned that the value of dh calculated using The real situation is perhaps in between the
the above simple equation shows a dependence above two cases, that is, the polymer chains may
on electrolyte concentration, and hence, the undergo some interpenetration and some com-
method cannot be used in a straightforward pression. Provided the dangling chains (the
manner. Cohen-Stuart et al. showed that the mea- A chains in A-B, A-B-A block, or BAn graft
sured electrophoretic thickness de approaches dh copolymers) are in a good solvent, this local
only at low electrolyte concentrations. Thus, to increase in segment density in the interaction
Suspensions 1365 S
Suspensions,
Fig. 19 Schematic
interaction between
particles containing
adsorbed polymer layers
Interpenetration without Compression without

compression interpenetration
zone will result in strong repulsion as a result of

two main effects: (i) increase in the osmotic pres-
R
sure in the overlap region as a result of the unfa- b
vorable mixing of the polymer chains, when these
are in good solvent conditions (Tadros 1981;
ma
Napper 1981) – this is referred to as osmotic i h b
dV( m i )
repulsion or mixing interaction, and it is Chemical
described by a free energy of interaction Gmix; potential
of solvent
(ii) reduction of the configurational entropy of the
chains in the interaction zone – this entropy
reduction results from the decrease in the volume
available for the chains when these are either
overlapped or compressed. This is referred to as h
volume restriction interaction, entropic, or elastic
Suspensions, Fig. 20 Schematic representation of the
interaction, and it is described by a free energy of mixing interaction
interaction Gel. Combination of Gmix and Gel is
usually referred to as the steric interaction free
energy, Gs, that is, each containing an adsorbed polymer layer with
thickness d. Before overlap, one can define in
Gs ¼ Gmix þ Gel (78) each polymer layer a chemical potential for the
solvent mia and a volume fraction for the polymer
The sign of Gmix depends on the solvency of in the layer j2. In the overlap region (volume
the medium for the chains. If in a good solvent, element dV), the chemical potential of the solvent
that is, the Flory–Huggins interaction parameter is reduced to mib; this results from the increase in
w is less than 0.5, then Gmix is positive, and the polymer segment concentration in this overlap S
mixing interaction leads to repulsion (see below). region.
In contrast, if w > 0.5 (i.e., the chains are in In the overlap region, the chemical potential of
a poor solvent condition), Gmix is negative, and the polymer chains is now higher than in the rest
the mixing interaction becomes attractive. Gel is of the layer (with no overlap). This amounts to an
always positive, and hence, in some cases, one increase in the osmotic pressure in the overlap
can produce stable dispersions in a relatively poor region As a result, solvent will diffuse from the
solvent (enhanced steric stabilization) (Napper bulk to the overlap region, thus separating the
1981). particles, and hence, a strong repulsive energy
The mixing interaction, Gmix, results from the arises from this effect.
unfavorable mixing of the polymer chains when The above repulsive energy can be calculated
these are in a good solvent condition. This is by considering the free energy of mixing of two
schematically shown in Fig. 20. Consider two polymer solutions, as, for example, treated by
spherical particles with the same radius and Flory and Krigbaum (1950). The free energy of
S 1366 Suspensions
Suspensions,
Fig. 21 Schematic Volume lost
representation of
configurational entropy
loss on approach of h∞
a second particle
h
No. of configurations No. of configurations

Ω∞ Ω(h)
mixing is given by two terms: an entropy term The sign of Gmix depends on the value of
that depends on the volume fraction of polymer the Flory–Huggins interaction parameter w:
and solvent and an energy term that is determined If w < 0.5, Gmix is positive, and the interaction
by the Flory–Huggins interaction parameter w: is repulsive; if w > 0.5, Gmix is negative, and the
interaction is attractive. The condition w ¼ 0.5
dðGmix Þ ¼ kTðn1 ln f1 þ n2 ln f2 þw n1 f2 Þ and Gmix ¼ 0 is termed the y condition. The latter
(79) corresponds to the case where the polymer
mixing behaves as ideal, that is, mixing of the
where n1 and n2 are the number of moles of chains does not lead to increase or decrease of the
solvent and polymer with volume fractions j1 free energy of the system.
and j2, k is the Boltzmann constant, and T is The elastic interaction, Gel, arises from the loss
the absolute temperature. in configurational entropy of the chains on the
The total change in free energy of mixing for approach of a second particle. As a result of this
the whole interaction zone, V, is obtained by approach, the volume available for the chains
summing over all the elements in V: becomes restricted, resulting in loss of the number
of configurations. This can be illustrated by con-

2kT V22 1 sidering a simple molecule, represented by a rod
Gmix ¼ n2 w Rmix ðhÞ (80)
V1 2 that rotates freely in a hemisphere across a surface
(Mackor and van der Waals 1951) (Fig. 21).
where V1 and V2 are the molar volumes of sol- When the two surfaces are separated by an
vent and polymer, respectively, v2 is the number infinite distance 1, the number of configurations
of chains per unit area, and Rmix(h) is geometric of the rod is O(1) which is proportional to the
function which depends on the form of the seg- volume of the hemisphere. When a second parti-
ment density distribution of the chain normal to cle approaches to a distance h such that it cuts the
the surface, r(z). k is the Boltzmann constant, and hemisphere (losing some volume), the volume
T is the absolute temperature. available to the chains is reduced and the number
Using the above theory, one can derive an of configurations becomes O (h) which is less
expression for the free energy of mixing of two than O(1) For two flat plates, Gel is given by
polymer layers (assuming a uniform segment the following expression:
density distribution in each layer) surrounding
two spherical particles as a function of the sepa- Gel OðhÞ
¼ 2 n2 ln ¼ 2 n2 Rel ðhÞ (82)
ration distance h between the particles (Fischer kT Oð1Þ
1958). The expression for Gmix is
2 where Rel(h) is a geometric function that depends
2 V2 1 h
Gmix ¼ n2 w 3R þ 2d þ on the segment density distribution.
V1 2 2 It should be stressed that Gel is always positive
(81) and could play a major role in steric stabilization.
Suspensions 1367 S
GT
Increasing δ/R
Gel GT
Gmix
G
h
h 2δ h
Gmin
Suspensions, Fig. 23 Variation of GT with h with d/R
increases with increase of A and R. At a given

A and R, Gmin increases with decrease in d (i.e.,
Suspensions, Fig. 22 Energy–distance curves for steri-
with decrease of the molecular weight, Mw, of the
cally stabilized systems
stabilizer). This is illustrated in Fig. 23 which
shows the energy–distance curves at increasing
It becomes very strong when the separation dis- values of d/R.
tance between the particles becomes comparable Several criteria can be established for effec-
to the adsorbed layer thickness d. tive steric stabilization of suspensions: (i) The
Combination of Gmix and Gel with GA gives particles should be completely covered by the
the total energy of interaction GT (assuming there polymer (the amount of polymer should corre-
is no contribution from any residual electrostatic spond to the plateau value). Any bare patches
interaction) (6), that is, may cause flocculation either by van der Waals
attraction (between the bare patches) or by bridg-
GT ¼ Gmix þ Gel þ GA (83) ing flocculation (whereby a polymer molecule
will become simultaneously adsorbed on two or
A schematic representation of the variation of more particles). (ii) The polymer should be
Gmix, Gel, GA, and GT with surface-surface sepa- strongly “anchored” to the particle surfaces, to
ration distance h is shown in Fig. 22. Gmix prevent any displacement during particle
increases very sharply with decrease of h when approach. This is particularly important for con- S
h < 2d. Gel increases very sharply with decrease centrated suspensions. For this purpose A-B,
of h when h < d. GT versus h shows a minimum, A-B-A block, and BAn graft copolymers are the
Gmin, at separation distances comparable to 2d; most suitable where the chain B is chosen to be
when h < 2d, GT shows a rapid increase with highly insoluble in the medium and has a strong
further decrease in h (Hesselink et al. 1971). affinity to the surface. Examples of B groups for
Unlike the GT-h curve predicted by the DLVO hydrophobic particles in aqueous media are poly-
theory (which shows two minima and one energy styrene and polymethylmethacrylate. (iii) The
maximum), the GT-h for systems that are steri- stabilizing chain A should be highly soluble in
cally stabilized shows only one minimum, Gmin, the medium and strongly solvated by its mole-
followed by sharp increase in GT with decrease of cules. Examples of A chains in aqueous media are
h (when h, 2d). The depth of the minimum poly(ethylene oxide) and poly(vinyl alcohol).
depends on the Hamaker constant A, the particle (iv) d should be sufficiently large (> 5 nm) to
radius R, and adsorbed layer thickness d – Gmin prevent weak flocculation.
S 1368 Suspensions
Suspensions, GT
Fig. 24 Influence of
reduction in solvency on Gmix Gel
Gel GT
the energy–distance curve
G
Reduce
Solvency
δ h
δ h 2δ 2δ
Gmix
GA
Gmin χ > 0.5
χ < 0.5
Flocculation of Sterically Stabilized are adsorbed by multipoint attachment. This sit-

Suspensions uation has been described by Napper (Napper
1981) who referred to it as “enhanced” steric
Two main types of flocculation may be distin- stabilization. Thus, by measuring the y-point
guished: (i) Weak flocculation: This occurs when (CFT or CFV) for the polymer chains (A) in the
the thickness of the adsorbed layer is small (usu- medium under investigation (which could be
ally < 5 nm), particularly when the particle obtained from viscosity measurements), one can
radius and Hamaker constant are large. (ii) Incip- establish the stability conditions for a dispersion
ient flocculation: This occurs when the solvency before its preparation. This procedure helps also
of the medium is reduced to become worse than in designing effective steric stabilizers such as
y-solvent (i.e., w > 0.5). This is illustrated in block and graft copolymers.
Fig. 24 where w was increased from < 0.5
(good solvent) to > 0.5 (poor solvent).
When w > 0.5, Gmix becomes negative (attrac- Properties of Concentrated Suspensions
tive) which when combined with the van der
Waals attraction at this separation distance gives One of the main features of concentrated suspen-
a deep minimum causing flocculation. In most sions is the formation of three-dimensional struc-
cases, there is a correlation between the critical ture units, which determine their properties and in
flocculation point and the y condition of the particular their rheology. The formation of these
medium. Good correlation is found in many units is determined by the interparticle interac-
cases between the critical flocculation tempera- tions, which need to be clearly defined and quan-
ture (CFT) and y-temperature of the polymer in tified. It is useful to define the concentration
solution (with block and graft copolymers, one range above which a suspension may be consid-
should consider the y-temperature of the stabiliz- ered concentrated. The particle number concen-
ing chains A) (Napper 1981). Good correlation is tration and volume fraction, j, above which
also found between the critical volume fraction a suspension may be considered concentrated
(CFV) of a non-solvent for the polymer chains are best defined in terms of the balance between
and their y-point under these conditions. How- the particle translational motion and interparticle
ever, in some cases, such correlation may break interaction. At one extreme, a suspension may be
down, particularly the case for polymers which considered dilute if the thermal motion
Suspensions 1369 S
(Brownian diffusion) of the particles predomi- a b c
nates over the imposed interparticle interaction
(Ottewill 1982, 1983, 1987). In this case, the
particle translational motion is large, and only
occasional contacts occur between the particles,
that is, the particles do not “see” each other until
Stable colloidal Stable coarse Stable coarse
collision occurs, giving a random arrangement of suspension suspension suspension
particles. In this case, the particle interactions can (uniform size) (size distribution)
be represented by two body collisions. In such d e f

“dilute” systems, gravity effects may be
neglected, and if the particle size range is within
the colloid range (1 nm–1 mm), no settling occurs.
The properties of the suspension are time inde-
pendent, and, therefore, any time-average quan- Coagulated Coagulated Coagulated
tity such as viscosity or scattering may be suspension suspension suspension
(chain aggregates) (compact clusters) (open structure)
extrapolated to infinite dilution. As the particle
number concentration is increased in g h i
a suspension, the volume of space occupied by
the particles increases relative to the total vol-
ume. Thus, a proportion of the space is excluded
in terms of its occupancy by a single particle.
Weakly Bridging Depletion
Moreover, the particle–particle interaction flocculated flocculation flocculation
increases, and the forces of interaction between structure
the particles play a dominant role in determining
Suspensions, Fig. 25 States of the suspension
the properties of the system. With further
increase in particle number concentration, the
interactive contact between the particles
increases until a situation is reached where the that is, the vibrational motion of the particles is
interaction produces a specific order between the large compared with particle size. A time-
particles, and a highly developed structure is dependent system arises in which there will be
reached. With solid in liquid dispersions, such spatial and temporal correlation.
a highly ordered structure, which is close to the On standing, concentrated suspensions reach
maximum packing fraction (j ¼ 0.74 for hexag- various states (structures) that are determined
onally closed packed array of monodisperse par- by the following: (i) magnitude and balance of
ticles), is referred to as “solid” suspension. In the various interaction forces, electrostatic repul- S
such a system, any particle in the system interacts sion, steric repulsion, and van der Waals attrac-
with many neighbors, and the vibrational ampli- tion; (ii) particle size and shape distribution; (iii)
tude is small relative to particle size; the density difference between disperse phase and
properties of the system are essentially time inde- medium which determines the sedimentation
pendent (Ottewill 1982, 1983, 1987). In between characteristics; (iv) conditions and prehistory of
the random arrangement of particles in “dilute” the suspension, for example, agitation which
suspensions and the highly ordered structure of determines the structure of the flocs formed
“solid” suspensions, one may easily define “con- (chain aggregates, compact clusters, etc.); and
centrated” suspensions. In this case, the particle (v) presence of additives, for example, high
interactions occur by many body collisions, and molecular weight polymers that may cause bridg-
the translational motion of the particles is ing or depletion flocculation. An illustration of
restricted. However, this reduced translational some of the various states that may be produced is
motion is not as great as with “solid” suspensions, given in Fig. 25. These states may be described in
S 1370 Suspensions
terms of three different energy–distance curves: barrier). Due to the small distances between the
(a) electrostatic, produced, for example, by the particles in the sediment, it is very difficult to
presence of ionogenic groups on the surface of redisperse the suspension by simple shaking.
the particles or adsorption of ionic surfactants; With case (c) consisting of a wide distribution
(b) steric, produced, for example, by adsorption of particle sizes, the sediment may contain larger
of nonionic surfactants or polymers; and (c) elec- proportions of the larger size particles, but still
trostatic + steric (electrosteric), produced, for a hard “clay” is produced. These “clays” are
example, by polyelectrolytes. dilatant (i.e., shear thickening), and they can be
A brief description of the various states shown easily detected by inserting a glass rod in the
in Fig. 25 is given as follows (Tadros 1980, suspension. Penetration of the glass rod into
1983): States (a)–(c) correspond to a suspension these hard sediments is very difficult. States
that is stable in the colloid sense. The stability is (d)–(f) represent the case for coagulated suspen-
obtained as a result of net repulsion due to the sions which either have a small repulsive energy
presence of extended double layers (i.e., at low barrier or its complete absence. State (d) repre-
electrolyte concentration), the result of steric sents the case of coagulation under no stirring
repulsion produced adsorption of nonionic sur- conditions in which case chain aggregates are
factants or polymers, or the result of combination produced that will settle under gravity forming
of double layer and steric repulsion a relatively open structure. State (e) represents
(electrosteric). State (a) represents the case of the case of coagulation under stirring conditions
a suspension with small particle size (submicron) whereby compact aggregates are produced that
whereby the Brownian diffusion overcomes the will settle faster than the chain aggregates and the
gravity force producing uniform distribution of sediment produced is more compact. State (f)
the particles in the suspension, that is, represents the case of coagulation at high volume
fraction of the particles, j. In this case, the whole
kT ð4=3Þp R3 Drgh (84) particles will form a “one-floc” structure that is
formed from chains and cross chains that extend
where k is the Boltzmann constant, T is the abso- from one wall to the other in the container. Such
lute temperature, R is the particle radius, Dr is the coagulated structure may undergo some compres-
buoyancy (difference in density between the par- sion (consolidation) under gravity leaving a clear
ticles and the medium), g is the acceleration due supernatant liquid layer at the top of the container.
to gravity, and h is the height of the container. This phenomenon is referred to as syneresis.
A good example of this case is a latex suspension State (g) represents the case of weak and
with particle size well below 1 mm that is stabi- reversible flocculation. This occurs when the sec-
lized by ionogenic groups, by an ionic surfactant ondary minimum in the energy–distance curve is
or nonionic surfactant, or by a polymer. This deep enough to cause flocculation. This can occur
suspension will show no separation on storage at moderate electrolyte concentrations, in partic-
for long periods of time. ular with larger particles. The same occurs with
States (b) and (c) represent the case of suspen- sterically and electrosterically stabilized suspen-
sions whereby the particle size range is outside sions. This occurs when the adsorbed layer thick-
the colloid range (> 1 mm). In this case, the ness is not very large, particularly with large
gravity force exceeds the Brownian diffusion. particles. The minimum depth required for caus-
With state (b), the particles are uniform, and ing weak flocculation depends on the volume
they will settle under gravity forming a hard sed- fraction of the suspension. The higher the volume
iment (technically referred to as “clay” or fraction, the lower the minimum depth required
“cake”). The repulsive forces between the parti- for weak flocculation. This can be understood if
cles allow them to move past each other till they one considers the free energy of flocculation that
reach small distances of separation (that are consists of two terms, an energy term determined
determined by the location of the repulsive by the depth of the minimum (Gmin) and an
Suspensions 1371 S
Suspensions, µ µ
Fig. 26 Schematic
representation of depletion µ µ µ µ µ µ µ µ
flocculation Polymer coil
µ µ µ µ
µ µ µ µ µ
µ µ µ µ µ
Depletion zone
fp+ µ µ µ µ
µ µ µ
µ µ µ µ µ µ µ µ
µ µ µ Osmotic Pressure
µ µ µ µ
entropy term that is determined by reduction in occurs, leading to bridging flocculation. If the
configurational entropy on aggregation of polymer adsorption is weak (low adsorption
particles: energy per polymer segment), the flocculation
could be weak and reversible. In contrast, if the
D Gflocc ¼ D Hflocc TD Sflocc (85) adsorption of the polymer is strong, tough flocs
are produced, and the flocculation is irreversible.
With dilute suspension, the entropy loss on The last phenomenon is used for solid/liquid sep-
flocculation is larger than with concentrated sus- aration, for example, in water and effluent treat-
pensions. Hence, for flocculation of a dilute sus- ment. Case (i) represents a phenomenon, referred
pension, a higher energy minimum is required to as depletion flocculation, produced by addition
when compared with the case with concentrated of “free” non-adsorbing polymer (38). In this
suspensions. This flocculation is weak and revers- case, the polymer coils cannot approach the parti-
ible, that is, on shaking the container, redispersion cles to a distance D (that is determined by the
of the suspension occurs. On standing, the dis- radius of gyration of free polymer RG) since
persed particles aggregate to form a weak “gel.” the reduction of entropy on close approach of the
This process (referred to as sol ↔ gel transforma- polymer coils is not compensated by an adsorption S
tion) leads to reversible time dependence of vis- energy. The suspension particles will be
cosity (thixotropy). On shearing the suspension, surrounded by a depletion zone with thickness D.
the viscosity decreases, and when the shear is Above a critical volume fraction of the free poly-
removed, the viscosity is recovered. This phenom- mer, jp+, the polymer coils are “squeezed out”
enon is applied in paints. On application of the from between the particles, and the depletion
paint (by a brush or roller), the gel is fluidized, zones begin to interact. The interstices between
allowing uniform coating of the paint. When the particles are now free from polymer coils,
shearing is stopped, the paint film recovers its and hence, an osmotic pressure is exerted outside
viscosity, and this avoids any dripping. the particle surface (the osmotic pressure outside is
State (h) represents the case whereby the par- higher than in between the particles) resulting in
ticles are not completely covered by the polymer weak flocculation (Asakura and Oosawa 1954,
chains. In this case, simultaneous adsorption of 1958). A schematic representation of depletion
one polymer chain on more than one particle flocculation is shown in Fig. 26.
S 1372 Suspensions
The magnitude of the depletion attraction free these methods, care should be taken in sampling
energy, Gdep, is proportional to the osmotic pres- the suspension, which should cause as little dis-
sure of the polymer solution, which in turn is turbance as possible for the “structure” to be
determined by jp and molecular weight M. The investigated. For example, when one investigates
range of depletion attraction is proportional to the the flocculation of a concentrated suspension,
thickness of the depletion zone, D, which is dilution of the system for microscopic investiga-
roughly equal to the radius of gyration, RG, of tion may lead to breakdown of the flocs, and
the free polymer. A simple expression for Gdep is a false assessment is obtained. The same applies
(Asakura and Oosawa 1954, 1958) when one investigates the rheology of
a concentrated suspension since transfer of the

2pR D2 2D system from its container to the rheometer may
Gdep ¼ ðm1 m1 Þ 1 þ
o
(86)
V1 R lead to breakdown of the structure. For these
reasons, one must establish well-defined proce-
where V1 is the molar volume of the solvent, m1 is dures for every technique, and this requires
the chemical potential of the solvent in the pres- a great deal of skill and experience. It is advisable
ence of free polymer with volume fraction jp, in all cases to develop standard operation pro-
and m1o is the chemical potential of the solvent in cedures for these investigations.
the absence of free polymer. (m1 – m1o) is propor-
tional to the osmotic pressure of the polymer
solution. Assessment of the Structure
of the Solid/Liquid Interface
Characterization of Suspensions Double layer investigations: Two procedures

and Assessment of Their Stability may be applied to investigate the charge and
potential distribution at the solid/liquid interface,
For the full assessment of the properties of sus- which is important in the assessment of electro-
pensions, three main types of investigations are static stabilization:
needed: (i) fundamental investigation of the sys- (i) Electrokinetic studies: This allows one to
tem at a molecular level, (ii) investigations into obtain the particle mobility as function of the
the state of the suspension on standing, and (iii) system variables such as pH and electrolyte con-
bulk properties of the suspension. All these inves- centration. From the electrophoretic mobility,
tigations require a number of sophisticated tech- one can calculate the zeta potential, provided
niques such as zeta potential measurements, information is available on particle size and elec-
surfactant and polymer adsorption and their con- trolyte concentration. In case when the above
formation at the solid/liquid interface, measure- information is not available (as is the case with
ment of the rate of flocculation and crystal many practical systems), one should use the elec-
growth, and several rheological measurements. trophoretic mobility for relative comparison
Apart from these practical methods which are between various systems; the assumption can be
present in most industrial laboratories, more fun- made that the higher the mobility, the higher the
damental information can be obtained using mod- surface charge and the more likely the system is
ern sophisticated techniques such as small-angle stable against flocculation if the charge is the
x-ray, neutron scattering measurements, and main stabilizing factor. Clearly for systems sta-
ultrasonic absorption techniques. Several other bilized by nonionic surfactants and polymers,
modern techniques are also now available for electrophoretic mobility measurements are less
investigation of the state of the suspension: informative. However, zeta potential measure-
Freeze fracture and electron microscopy, atomic ments can be qualitatively used to obtain infor-
force microscopy, scanning tunneling micros- mation on the adsorbed layer thickness for
copy, and confocal laser microscopy. In all nonionic surfactants and polymers, as discussed
Suspensions 1373 S
before. When a nonionic surfactant or polymer In most cases (particularly with surfactants),
adsorbs at the solid/liquid interface, a shift in the a plot of G versus C2 gives a Langmuir-type
shear plane occurs, and this results in reduction in isotherm, as described before. The data can be
the zeta potential. If the zeta potential of the fitted using the Langmuir equation:
particles is measured in the presence and absence
of nonionic surfactant or polymer, then the G1 b C2
G¼ (88)
adsorbed layer thickness can be roughly esti- ð1 þ b C2 Þ
mated from the reduction in zeta potential. This
procedure requires measurement at various elec- where b is a constant that is related to the free
trolyte concentrations and extrapolation of the energy of adsorption:
results to infinitely dilute electrolyte concentra-
tion. In addition, electrophoretic mobility mea- D Gads
b1 (89)
surements can be used to investigate specifically RT
adsorbed ions, which lead to significant change in
zeta potential, and in some cases with chemisorbed Most polymers (particularly those with high
counterions, charge reversal may occur. molecular weight) give a high-affinity isotherm,
(ii) Analytical determination of surface as described before.
charge: This can be applied for particles that (iv) Assessment of the state of the dispersion:
contain ionogenic groups such as oxides. The Two general techniques may be applied for mea-
oxide (such as silica or alumina) is covered with suring the rate of flocculation of suspensions,
-OH groups which can be titrated with acid or both of which can only be applied for dilute
alkali using a cell of the type: systems. The first method is based on measuring
E1 | Oxide suspension in Indifferent Electro- the scattering of light by the particles. For mono-
lyte at a given Concentration | E2 disperse particles with a radius that is less than
where E1 is an electrode reversible to H+ and l/20 (where l is the wavelength of light), one can
OH ion, such as a glass electrode, and E2 is apply the Rayleigh equation, whereby the turbid-
a reference electrode. From a knowledge of the ity to is given by
amount of H+ and OHadded and the amount
remaining in solution, GH and GOH (the number to ¼ A0 no V21 (90)
of moles of H+ and OHadsorbed) can be calcu-
lated, and hence, one can establish surface where A0 is an optical constant (which is related
charge-pH isotherms. Measurements are usually to the refractive index of the particle and medium
made at various electrolyte concentrations. In and the wavelength of light) and no is the number
addition, the nature of the electrolyte is changed of particles, each with a volume V1. By combin-
to investigate any possible specific adsorption. ing the Rayleigh theory with the Smoluchowski– S
(iii) Surfactant and polymer adsorption: Fuchs theory of flocculation kinetics (18), one
A representative sample of the solid with known can obtain the following expression for the vari-
mass m and surface area A per gram is equili- ation of turbidity with time:
brated with a surfactant or polymer concentration
with concentration C1. After equilibrium is t ¼ A0 no V21 ð1 þ 2 no ktÞ (91)
reached (at a given constant temperature), the
solid is removed by centrifugation and the equi- where k is the rate constant of flocculation.
librium concentration C2 is determined analyti- The second method for obtaining the rate con-
cally. The amount of adsorption G in moles m2 stant of flocculation is by direct particle counting
is given by as a function of time. For this purpose, optical
microscopy or image analysis may be used, pro-
ðC1 C2 Þ DC vided the particle size is within the resolution
G¼ ¼ (87)
mA mA limit of the microscope. Alternatively, the
S 1374 Suspensions
particle number may be determined using elec- (v) Bulk properties of suspensions:
tronic devices such as the Coulter counter or the A qualitative method for assessment of the bulk
flow ultramicroscope. The rate constant of floc- properties of suspensions is to measure the equi-
culation is determined by plotting 1/n versus t, librium sediment volume (or height) and
where n is the number of particles after time t, redispersion. For a “structured” suspension,
that is, obtained by “controlled” flocculation or addition
of “thickeners” (such as polysaccharides, clays,
1 1 or oxides), the “flocs” sediment at a rate
¼ þ kt (92)
n no depending on their size and porosity of the aggre-
gated mass. After this initial sedimentation, com-
The rate constant k of slow flocculation is paction and rearrangement of the floc structure
usually related to the rapid rate constant ko (the occur, a phenomenon referred to as consolida-
Smoluchowski rate) by the stability ratio W: tion. Normally in sediment volume measure-
ments, one compares the initial volume Vo (or
k height Ho) with the ultimately reached value
W¼ (93)
ko V (or H). A colloidally stable suspension gives
a “close-packed” structure with relatively small
One usually plots log W versus log C (where sediment volume (dilatant sediment referred to as
C is the electrolyte concentration) to obtain the clay). A weakly “flocculated” or “structured”
critical coagulation concentration (ccc), which is suspension gives a more open sediment and
the point at which log W ¼ 0. hence a higher sediment volume. Thus, by com-
The incipient flocculation for sterically stabi- paring the relative sediment volume V/Vo or
lized suspensions, when the medium for the height H/Ho, one can distinguish between
chains becomes a y-solvent, can be measured by a clayed and flocculated suspension.
following the change of suspension turbidity with (vi) Rheological measurements: Three differ-
temperature. For example, on heating an aqueous ent rheological measurements may be applied
suspension stabilized with poly(ethylene oxide) (Ferry 1980; Whorlow 1980; Goodwin and
(PEO) or pol(vinyl alcohol) chains, then above Hughes 2000; Tadros 2010): (i) steady state
a certain temperature (the y-temperature) that shear stress–shear rate measurements (using
depends on electrolyte concentration, floccula- a controlled shear rate instrument), (ii) constant
tion of the suspension occurs. The temperature stress (creep) measurements (carried out using
at which this occurs is defined as the critical a constant stress instrument), and (iii) dynamic
flocculation temperature (CFT). Above the CFT, (oscillatory) measurements (preferably carried
the turbidity of the suspension rises very sharply. out using a constant strain instrument). These
For this purpose, the cell in the spectrophotome- rheological techniques can be used to assess sed-
ter that is used to measure the turbidity is placed imentation and flocculation of suspensions.
in a metal block that is connected to a temperature As will be shown in the next section, the rate
programming unit (which allows one to increase of sedimentation decreases with increase of the
the temperature rise at a controlled rate). volume fraction of the disperse phase, j, and
To obtain a measure of the rate of crystal ultimately, it approaches zero at a critical volume
growth, the particle size distribution of the sus- fraction jp (the maximum packing fraction).
pension is followed as a function of time, using However, at j jp, the viscosity of the system
either a Coulter counter, a Master sizer, or an approaches 1. Thus, for most practical emul-
optical disc centrifuge. One usually plots the sions, the system is prepared at j values below
cube of the average radius versus time which jp and then add “thickeners” to reduce sedimen-
gives a straight line from which the rate of crystal tation. These “thickeners” are usually high
growth can be determined (the slope of the linear molecular weight polymers (such as Xanthan
curve). gum, hydroxyethyl cellulose, or associative
Suspensions 1375 S
thickeners), finely divided inert solids (such as system), the compliance shows a linear increase
silica or swelling clays), or a combination of the with time (i.e., the system reaches a steady state
two. In all cases, a “gel” network is produced in with constant shear rate). If after the steady state
the continuous phase which is shear thinning (i.e., is reached, the stress is removed, elastic recovery
its viscosity decreases with increase of shear rate) occurs, and the strain changes sign. This behavior
and viscoelastic (i.e., it has a viscous and elastic (usually referred to as “creep”) is schematically
components of the modulus). If the viscosity of represented in Fig. 27. The slope of the linear part
the elastic network, at shear stresses (or shear of the creep curve gives the value of the viscosity
rates) comparable to those exerted by the parti- at the applied stress, Zs,
cles, exceeds a certain value, then sedimentation
is completely eliminated. J Pa1 1
The shear stress, sp, exerted by a particle ¼ ¼ ¼ Zs (96)
t s Pas
(force/area) can be simply calculated:
The recovery curve will only give the elastic
ð4=3Þp R Drg DrRg
3 component, which if superimposed on the
sp ¼ ¼ (94) ascending part of the curve will give the viscous
4p R2 3
component. Thus, one measures creep curves as
For a 10-mm radius particle with density dif- a function of the applied stress (starting from
ference Dr ¼ 0.2, sp is equal to a very small stress of the order of 0.01 Pa). This
is illustrated in Fig. 28. The viscosity Zs (which
0:2 103 10 106 9:8 is equal to the reciprocal of the slope of the
sp ¼ 6 103 straight portion of the creep curve) is plotted as
3
(95) a function of the applied stress. This is schemat-
ically shown in Fig. 29.
For smaller droplets, smaller stresses are Below a critical stress, scr, the viscosity
exerted. Thus, to predict sedimentation, one has reaches a limiting value, Z(o), namely, the residual
to measure the viscosity at very low stresses (or zero shear) viscosity. scr may be denoted as the
(or shear rates). These measurements can be carried “true yield stress” of the emulsion, that is, the
out using a constant stress rheometer (Carrimed, stress above which the “structure” of the system
Bohlin, Rheometrics, or Physica). A constant is broken down. Above scr,Zs decreases rapidly
stress s (using, e.g., a drag cup motor that can with further increase of the shear stress (the shear
apply very small torques, and using an air bearing thinning regime). It reaches another Newtonian
system to reduce the frictional torque) is applied value Z1, which is the high shear limiting viscos-
on the system (which may be placed in the gap ity. Z(o) could be several orders of magnitudes
between two concentric cylinders or a cone-plate (104–108) higher than Z1. Usually, one obtains S
geometry), and the deformation (strain g or com- good correlation between the rate of sedimentation
pliance J ¼ g/s ¼ Pa1) is followed as a function v and the residual viscosity Z(o) (Buscall et al.
of time (Ferry 1980; Whorlow 1980; Goodwin 1982). This is illustrated in Fig. 30.
and Hughes 2000; Tadros 2010). For Above a certain value of Z(o), v becomes equal
a viscoelastic system, the compliance shows to 0. Clearly, to minimize sedimentation, one has
a rapid elastic response Jo at t ! 0 (instantaneous to increase Z(o); an acceptable level for the high
compliance Jo ¼ 1/Go, where Go is the instanta- shear viscosity Z1 must be achieved, depending
neous modulus that is a measure of the elastic on the application. In some cases, a high Z(o) may
(i.e., “solid-like”) component). At t > 0, be accompanied by a high Z1 (which may not be
J increases slowly with time, and this corresponds acceptable for application, e.g., if spontaneous
to the retarded response (“bonds” are broken and dispersion on dilution is required). If this is the
reformed but not at the same rate). Above case, the formulation chemist should look for an
a certain time period (that depends on the alternative thickener.
S 1376 Suspensions
Suspensions, J = Pa–1 = 1 = h
Fig. 27 Typical creep s
t s Pas
curve for a viscoelastic
system
t=0 t=t
J
Strain Recovery
Compliance
Strain
m2 N–1
Stress Slow elastic deformation
Rapid elastic deformation
t=0 t
σc
increasing η(0)
s
Residual or
J Zero shear
ησ
Viscosity
slope ∝ 1
ησ
Suspensions, Fig. 28 Creep curves as a function of σ

applied stress Critical stress is a useful parameter (related
to yield stress) as denotes the stress at which
structure “breaks down”.
Another problem encountered with many sus- Suspensions, Fig. 29 Variation of viscosity with
pensions is that of “syneresis,” that is, the appear- applied stress
ance of a clear liquid film at the top of the
suspension. “Syneresis” occurs with most “floc- above techniques will be discussed in detail
culated” and/or “structured” (i.e., those below. It is sufficient to state in this section that
containing a thickener in the continuous phase) when a network of the suspension particles
suspensions. “Syneresis” may be predicted from (either alone or combined with the thickener) is
measurement of the yield value (using steady produced, the gravity force will cause some con-
state measurements of shear stress as a function traction of the network (which behaves as
of shear rate) as a function of time or using a porous plug) thus causing some separation of
oscillatory techniques (whereby the storage and the continuous phase which is entrapped between
loss modulus are measured as a function of strain the droplets in the network. As mentioned before,
amplitude and frequency of oscillation). The flocculation of suspensions is the result of the
Suspensions 1377 S
where Z is the Newtonian viscosity (that is inde-
pendent of the applied shear rate). For most prac-
tical suspensions (with j > 0.1 and containing
thickeners to reduce sedimentation), a plot of s
versus g is not linear (i.e., the viscosity depends
v/cm s–1
on the applied shear rate). The most common

flow curve is shown in Fig. 31 (usually described
as a pseudoplastic or shear thinning system). In
this case, the viscosity decreases with increase in
shear rate, reaching a Newtonian value above
a critical shear rate.
0 Several models may be applied to analyze the
h(0)
results of Fig. 31:
Suspensions, Fig. 30 Variation of sedimentation rate (a) Power law model (Tadros 2010):
with the residual viscosity
s ¼ k gn (98)
long-range van der Waals attraction. Flocculation
can be weak (and reversible) or strong, depending where k is the consistency index of the emulsion
on the magnitude of the net attractive forces. and n is the power (shear thinning) index (n < 1);
Weak flocculation may result in reversible time the lower the value of n, the more shear thinning
dependence of the viscosity, that is, on shearing the suspension is. This is usually the case with
the emulsion at a given shear rate, the viscosity weakly flocculated suspensions or those to which
decreases, and on standing, the viscosity recovers a “thickener” is added. By fitting the results of
to its original value. This phenomenon is referred Fig. 31 to Eq. 98 (this is usually in the soft ware of
to as thixotropy (solgel transformation). the computer connected to the rheometer), one
Rheological techniques are most convenient can obtain the viscosity of the emulsion at a given
to assess suspension flocculation without the shear rate:
need of any dilution (which in most cases result
in breakdown of the floc structure). In steady state s
Zðat a given shear rateÞ ¼ ¼ k gn1 (99)
measurements, the suspension is carefully placed g
in the gap between concentric cylinder and cone-
and-plate platens. For the concentric cylinder (b) Bingham model (Tadros 2010):
geometry, the gap width should be at least 10X
larger than the largest droplet size (a gap width s ¼ sb þ Zpl g (100)
that is greater than 1 mm is usually used). For the S
cone-and-plate geometry, a cone angle of 4o or where sb is the extrapolated yield value
smaller is usually used. A controlled rate instru- (obtained by extrapolation of the shear stress–
ment is usually used for the above measurements; shear rate curve to g ¼ 0). Again, this is provided
the inner (or outer) cylinder, the cone (or the in the soft ware of the rheometer. Zpl is the slope
plate), is rotated at various angular velocities of the linear portion of the s-g curve (usually
(which allows one to obtain the shear rate g), referred to as the plastic viscosity). Both sb and
and the torque is measured on the other element Zpl may be related to the flocculation of the sus-
(this allows one to obtain the stress s). pension. At any given volume fraction of the emul-
For a Newtonian system (such as the case of sion and at a given particle size distribution, the
a dilute suspension, with a volume fraction j less higher the value of sb and Zpl, the more the floc-
than 0.1), s is related to g by the equation culated the suspension is. Thus, if one stores
a suspension at any given temperature and makes
s ¼ Zg (97) sure that the particle size distribution remains
S 1378 Suspensions
Suspensions,
Fig. 31 Shear stress and
viscosity versus shear rate
for a pseudoplastic system
h / Pa s
s / Pa
. .
g /s−1 g /s−1
constant (i.e., no Ostwald ripening occurred), an Note that sb is not equal to sC. Equation 102
increase in the above parameters indicates floccu- shows that a plot of s1/2 versus g1/2 gives a
lation of the suspension on storage. Clearly, if straight line from which sC and ZC can be
Ostwald ripening occurs simultaneously, sb and evaluated.
Zpl may change in a complex manner with storage In all the above analyses, the assumption was
time. Ostwald ripening results in a shift of the made that a steady state was reached. In other
particle size distribution to higher diameters; this words, no time effects occurred during the dura-
has the effect of reducing sb and Zpl. If flocculation of the flow experiment. Many suspensions
tion occurs simultaneously (having the effect of (particularly those that are weakly flocculated or
increasing these rheological parameters), the net “structured” to reduce sedimentation) show time
effect may be an increase or decrease of the rheo- effects during flow. At any given shear rate, the
logical parameters. This trend depends on the viscosity of the suspension continues to decrease
extent of flocculation relative to Ostwald ripening. with increasing the time of shear; on stopping the
Therefore, following sb and Zpl with storage time shear, the viscosity recovers to its initial value.
requires knowledge of Ostwald ripening and/or This reversible decrease of viscosity is referred to
coalescence. Only in the absence of this latter as thixotropy. The most common procedure of
breakdown process, one can use rheological mea- studying thixotropy is to apply a sequence of
surements as a guide of assessment of flocculation. shear stress–shear rate regimes within controlled
(c) Herschel–Bulkley model (Herschel and periods. If the flow curve is carried out within
Buckley 1926a, b) a very short time (say increasing the rate from 0 to
In many cases, the shear stress–shear rate say 500 s1 in 30 s and then reducing it again
curve may not show a linear portion at high from 500 to 0 s1 within the same period), one
shear rates. In this case, the data may be fitted finds that the descending curve is below the
with a Hershel–Bulkley model: ascending one. This behavior can be explained
from consideration of the structure of the system.
s ¼ sb þk gn (101) If, for example, the suspension is weakly floccu-
lated, then on applying a shear force on the sys-
(d) Casson’s model (Casson 1959) tem, this flocculated structure is broken down
This is another semiempirical model that may (and this is the cause of the shear thinning behav-
be used to fit the data of Fig. 42: ior). On reducing the shear rate back to zero, the
1=2 1=2
structure builds up only in part within the dura-
s1=2 ¼ sC þ ZC g1=2 (102) tion of the experiment (30 s). The ascending and
Suspensions 1379 S
Cycle 1 Constant stress (Creep) measurements can
Cycle 2 also be used to study flocculation. This method
has been described in detail in the section on
Stress
s sedimentation. Basically, a constant stress s is
applied on the system, and the compliance
J (Pa1) is plotted as a function of time as was
Cycle 3 illustrated in Fig. 27 (creep curve). This experi-
ment is repeated several times increasing the
.
stress from the smallest possible value (that can
Shear rate g be applied by the instrument), increasing the
Suspensions, Fig. 32 Flow curves for a thixotropic
stress in small increments. A set of creep curves
system are produced at various applied stresses as was
illustrated in Fig. 28. From the slope of the linear
portion of the creep curve (after the system
descending flow curves show hysteresis that is reaches a steady state), the viscosity at each
usually referred to as “thixotropic loop.” If now, applied stress, Zs, is calculated. A plot of Zs
the same experiment is repeated within a longer versus s (as illustrated in Fig. 29) allows one to
time experiment (say 120 s for the ascending and obtain the limiting (or zero shear) viscosity Z(o)
120 s for the descending curves), the hysteresis and the critical stress scr (which may be identi-
decreases, that is, the “thixotropic loop” becomes fied with the “true” yield stress of the system).
smaller. By repeating these experiments within The values of Z(o) and scr may be used to assess
various time periods, one obtains a series of the flocculation of the suspension on storage. If
thixotropic loops as illustrated in Fig. 32. This flocculation occurs on storage (without any Ost-
study may be used to investigate the state of wald ripening or coalescence), the values of Z(o)
flocculation of a suspension. Weakly flocculated and scr may show a gradual increase with
suspensions usually show thixotropy, and the increase of storage time. As discussed in the
change of thixotropy with applied time may be previous section (on steady state measurements),
used as an indication of the strength of this weak the trend becomes complicated if Ostwald ripen-
flocculation. This analysis is only qualitative, and ing occurs simultaneously (both have the effect of
one cannot use the results in a quantitative man- reducing Z(o) and scr). These measurements
ner. This is due to the possible breakdown of the should be supplemented by particle size distribu-
structure on transferring the suspension to the tion measurements of the diluted suspension
rheometer and also during the uncontrolled (making sure that no flocs are present after dilu-
shear experiment. A very important point that tion) to assess the extent of Ostwald ripening.
must be considered during rheological measure- Another complication may arise from the nature S
ments is the possibility of “slip” during the mea- of the flocculation. If the latter occurs in an irreg-
surements. This is particularly the case with ular way (producing strong and tight flocs), Z(o)
highly concentrated suspensions, whereby the may increase, while scr may show some
flocculated system may form a “plug” in the gap decrease, and this complicates the analysis of
of the platens leaving a thin liquid film at the the results. Inspite of these complications, con-
walls of the concentric cylinder or cone-and- stant stress measurements may provide valuable
plate geometry. To reduce “slip,” one should information on the state of the suspension on
use roughened walls for the platens. Strongly storage. Carrying out creep experiments and
flocculated suspensions usually show much less ensuring that a steady state is reached can be
thixotropy than weakly flocculated systems. time consuming. One usually carries out a stress
Again, one must be careful in drawing definite sweep experiment, whereby the stress is gradu-
conclusions without other independent tech- ally increased (within a predetermined time
niques (e.g., microscopy). period to ensure that one is not too far from
S 1380 Suspensions
reaching the steady state) and plots of Zs versus γo σo

s are established. These experiments are carried
out at various storage times (say every 2 weeks)
and temperatures. From the change of Z(o) and strain Δt
stress
scr with storage time and temperature, one may
obtain information on the degree and the rate of Suspensions, Fig. 33 Stress–strain relationship for
a viscoelastic system
flocculation of the system. Clearly, interpreta-
tion of the rheological results requires expert
knowledge of rheology and measurement of From the amplitudes of stress and strain and
the particle size distribution as a function of the phase angle shift, one can obtain the various
time. One main problem in carrying these viscoelastic parameters: The complex modulus
experiments is sample preparation. When G*, the storage modulus (the elastic component
a flocculated emulsion is removed from the con- of the complex modulus) G0 , the loss modulus
tainer, care should be taken not to cause much (the viscous component of the complex modulus)
disturbance to that structure (minimum shear G00 , tan d, and the dynamic viscosity Z0 :
should be applied on transferring the suspension
so
to the rheometer). It is also advisable to use Complex modulus jG
j ¼ (104)
separate containers for assessment of the floc- go
culation. A relatively large sample is prepared,
Storage modulus G0 ¼ jG
jcosd (105)
and this is then transferred to a number of sep-
arate containers.
Loss modulus G00 ¼ jG
j sin d (106)
Dynamic (oscillatory) measurements are by
far the most useful methods to obtain information
G00
on the flocculation of a suspension (Tadros 2010). tand ¼ (107)
G0
A strain is applied in a sinusoidal manner, with an
amplitude go and a frequency v (cycles/s or Hz) G00
or o (rad s1). This is usually carried out by Dynamic viscosity Z0 ¼ (108)
o
moving one of the platens say the cup
(in a concentric cylinder geometry) or the plate G0 is a measure of the energy stored in a cycle
(in a cone-and-plate geometry) back and forth in of oscillation. G00 is a measure of the energy
a sinusoidal manner. The stress on the other dissipated as viscous flow in a cycle of oscilla-
platen, the bob, or the cone is simultaneously tion. Tan d is a measure of the relative magni-
measured. These platens are usually connected tudes of the viscous and elastic components.
to interchangeable torque bars, whereby the Clearly, the smaller the value of tan d, the more
stress can be directly measured. The stress ampli- elastic the system is and vice versa. Z0 , the
tude so is simultaneously measured. In dynamic viscosity, shows a decrease with
a viscoelastic system (such as the case with increase of frequency o. Z0 reaches a limiting
a flocculated suspension), the stress oscillates value as o ! 0 – the value of Z0 in this limit is
with the same frequency, but out of phase from identical to the residual (or zero shear) viscosity
the strain. From measurement of the time shift Z(o). This is referred to as the Cox–Merz rule.
between strain and stress amplitudes (Dt), one In oscillatory measurements, one carries out
can obtain the phase angle shift d: two sets of experiments:
(a) Strain sweep measurements. In this case,
d ¼ Dto (103) the oscillation is fixed (say at 1 Hz or 6.28
rads1), and the viscoelastic parameters are mea-
A schematic representation of the variation of sured as a function of strain amplitude. This
strain and stress with frequency is shown in allows one to obtain the linear viscoelastic
Fig. 33. region. In this region, all moduli are independent
Suspensions 1381 S
gcr primary or secondary minimum. Flocculation in
G* G* the primary minimum is associated with a large
G′ G′ attractive energy, and this leads to higher values of
G* linear Ec when compared with the values obtained for
viscoelastic secondary minimum flocculation. For a weakly
G*
flocculated suspension, such as the case with
secondary minimum flocculation of an electrostat-
go ically stabilized suspension, the deeper the second-
ary minimum, the higher the value of Ec (at any
Suspensions, Fig. 34 Strain sweep results given volume fraction and particle size distribution
of the suspension). With a sterically stabilized
of the applied strain amplitude and become only suspension, weak flocculation can also occur
a function of time or frequency. This is illustrated when the thickness of the adsorbed layer decreases.
in Fig. 34, which shows a schematic representa- Again, the value of Ec can be used as a measure of
tion of the variation of G*, G0 , and G00 with strain the flocculation; the higher the value of Ec, the
amplitude (at a fixed frequency). stronger the flocculation. If incipient flocculation
It can be seen from Fig. 34, that G*, G0 , and G00 occurs (on reducing the solvency of the medium
remain virtually constant up to a critical strain for the change to worse than y condition), a much
value, gcr. This region is the linear viscoelastic deeper minimum is observed, and this is accompa-
region. Above gcr, G* and G0 start to fall, whereas nied by a much larger increase in Ec.
G00 starts to increase. This is the nonlinear region. To apply the above analysis, one must have an
The value of gcr may be identified with the min- independent method for assessing the nature of
imum strain above which the “structure” of the the flocculation. Rheology is a bulk property that
emulsion starts to break down (e.g., breakdown of can give information on the interparticle interac-
flocs into smaller units and/or breakdown of tion (whether repulsive or attractive) (Tadros
a “structuring” agent). From gcr and G0 , one can 1993, 1996), and to apply it in a quantitative
obtain the cohesive energy Ec (Jm3) of the floc- manner, one must know the nature of these inter-
culated structure (Tadros 2010): action forces. However, rheology can be used in
a qualitative manner to follow the change of the
gðcr gðcr suspension on storage. Providing the system does
1
Ec ¼ sdg ¼ G0 gdg ¼ G0 g2cr (109) not undergo any Ostwald ripening,the change of
2 the moduli with time and in particular the change
0 0
of the linear viscoelastic region may be used as an
Ec may be used in a quantitative manner as indication of flocculation. Strong flocculation is
a measure of the extent and strength of the floccu- usually accompanied by a rapid increase in G0 , S
lated structure in a suspension. The higher the and this may be accompanied by a decrease in the
value of Ec, the more flocculated the structure is. critical strain above which the “structure” breaks
Clearly, Ec depends on the volume fraction of the down. This may be used as an indication of for-
suspension as well as the particle size distribution mation of “irregular” flocs which become sensi-
(which determines the number of contact points in tive to the applied strain. The floc structure will
a floc). Therefore, for quantitative comparison entrap a large amount of the continuous phase,
between various systems, one has to make sure and this leads to an apparent increase in the vol-
that the volume fraction of the disperse particles ume fraction of the suspension and hence an
is the same and the suspensions have very similar increase in G0 .
particle size distribution. Ec also depends on the (b) Oscillatory sweep: In this case, the strain
strength of the flocculated structure, that is, the amplitude is kept constant in the linear viscoelas-
energy of attraction between the particles. This tic region (one usually takes a point far from gcr
depends on whether the flocculation is in the but not too low, i.e., in the midpoint of the linear
S 1382 Suspensions
G* stable suspension (at a given particle size distri-

G′ bution), t* increases with increase of the volume
G* fraction of the oil phase, j. In other words, the
G′ crossover point shifts to lower frequency with
G*
increase in j. For a given suspension, t* increases
G*
with increase in flocculation providing the parti-
cle size distribution remains the same (i.e., no
w* Ostwald ripening). The value of G0 also increases
w with increase in flocculation since aggregation of
droplets usually results in liquid entrapment, and
Suspensions, Fig. 35 Oscillatory sweep results the effective volume fraction of the suspension
shows an apparent increase. With flocculation,
viscoelastic region), and measurements are car- the net attraction between the droplets also
ried out as a function of frequency. This is sche- increases, and this results in an increase in G0 .
matically represented in Fig. 35 for a viscoelastic G0 is determined by the number of contacts
liquid system. Both G* and G0 increase with between the particles and the strength of each
increase in frequency, and ultimately above contact (which is determined by the attractive
a certain frequency, they reach a limiting value energy).
and show little dependence on frequency. G00 is It should be mentioned that in practice, one
higher than G0 in the low frequency regime; it may not obtain the full curve, due to the fre-
also increases with increase in frequency, and at quency limit of the instrument, and also, mea-
a certain characteristic frequency o* (that surement at low frequency is time consuming.
depends on the system), it becomes equal to G0 Usually, one obtains part of the frequency depen-
(usually referred to as the crossover point), after dence of G0 and G00 . In most cases, one has a more
which it reaches a maximum and then shows elastic than viscous system. Most suspension sys-
a reduction with further increase in frequency. tems used in practice are weakly flocculated, and
In the low frequency regime, that is, below o*, they also contain “thickeners” or “structuring”
G00 > G0 ; this regime corresponds to longer times agents to reduce sedimentation and to acquire
(remember that the time is reciprocal of fre- the right rheological characteristics for applica-
quency), and under these conditions, the response tion. The exact values of G0 and G00 required
is more viscous than elastic. In the high- depend on the system and its application. In
frequency regime, that is, above o*, G0 > G00 ; most cases, a compromise has to be made
this regime corresponds to short times, and between acquiring the right rheological charac-
under these conditions, the response is more elas- teristics for application and the optimum rheolog-
tic than viscous. At sufficiently high frequency, ical parameters for long-term physical stability.
G00 approaches zero, and G0 becomes nearly equal Application of rheological measurements to
to G*; this corresponds to very short timescales achieve these conditions requires a great deal of
whereby the system behaves as a near elastic skill and understanding of the factors that affect
solid. Very little energy dissipation occurs at rheology.
such high frequency.
The characteristic frequency o* can be used to
calculate the relaxation time of the system t*: Sedimentation of Suspensions and
Prevention of Formation of Dilatant
1 Sediments (Clays)
t
¼ (110)
o
As discussed before, most suspensions undergo

The relaxation time may be used as a guide for separation on standing as a result of the density
the state of the suspension. For a colloidally difference between the particles and the medium,
Suspensions 1383 S
a b c
h h c
Suspensions, Fig. 36 Schematic representation of particle sedimentation: (a) submicron particles, Brownian diffu-
sion > gravity; (b) coarse particles (> 1 mm) with uniform size; (c) coarse particles with size distribution
unless the particles are small enough for 4.4 109, 4.4 107, and 4.4 105 ms1,
Brownian motion to overcome gravity. This is respectively. The time needed for complete sed-
illustrated in Fig. 36 for three cases of imentation in a 0.1 m container is 250 days, 60 h,
suspensions. and 40 min, respectively.
The most practical situation is that represented For moderately concentrated suspensions,
by (C), whereby a concentration gradient of the 0.2 > j > 0.01, sedimentation is reduced as
particles occurs across the container. The concen- a result of hydrodynamic interaction between
tration of particles C can be related to that before the particles, which no longer sediment indepen-
any settling Co by the following equation: dently of each other. The sedimentation velocity,
v, can be related to the Stokes’ velocity vo by the
mgh following equation:
C ¼ Co exp (111)
kT
v ¼ vo ð1 6:55fÞ (115)
where m is the mass of the particles that is given
by (4/3) p R3 Dr (R is the particle radius, and Dr This means that for a suspension with
is the density difference between particle and j ¼ 0.1, v ¼ 0.345 vo, that is, the rate is reduced
medium), g is the acceleration due to gravity, by a factor of 3.
and h is the height of the container. For more concentrated suspensions (j > 0.2),
For a very dilute suspension of rigid the sedimentation velocity becomes a complex
noninteracting particles, the rate of sedimentation function of j. At j > 0.4, one usually enters
vo can be calculated by application of Stokes’ the hindered settling regime, whereby all the par-
law, whereby the hydrodynamic force is balanced ticles sediment at the same rate (independent of S
by the gravitational force: size). A schematic representation for the varia-
tion of v with j is shown in Fig. 37, which also
Hydrodynamic force ¼ 6pZRvo (112) shows the variation of relative viscosity with j. It
can be seen that v decreases exponentially with
Gravity force ¼ ð4=3ÞpR3 Drg (113) increase in j, and ultimately, it approaches zero
when j approaches a critical value jp (the max-
2 R2 Drg imum packing fraction). The relative viscosity
vo ¼ (114)
9 Z shows a gradual increase with increase in j, and
when j ¼ jp, the relative viscosity approaches
where Z is the viscosity of the medium (water). infinity.
vo was calculated for three particle sizes The maximum packing fraction jp can be
(0.1, 1, and 10 mm) for a suspension with density easily calculated for monodisperse rigid spheres.
difference Dr ¼ 0.2. The values of vo are For hexagonal packing jp ¼ 0.74, whereas for
S 1384 Suspensions
Suspensions, h f –[h]fp
Fig. 37 Variation of = 1–
v0 h0 fp
sedimentation velocity and
relative viscosity with j
v
h
h0
v 0 Maximum
fp packing
[h] = 2.5 fraction
f f fp
random packing, jp ¼ 0.64. The maximum v /v0

packing fraction increases with polydisperse 1.0
suspensions. For example, for a bimodal particle
size distribution (with a ratio of 10:1), o Experimental data
0.8
jp > 0.8. Line drawn with

fp = 0.58 and 0.6
It is possible to relate the relative sedimenta-
k = 5.4
tion rate (v/vo) to the relative viscosity Z/Zo:
0.4

v Z
¼a o (116) 0.2
vo Z
0
The relative viscosity is related to the volume 0.1 0.2 0.3 0.4 0.5
φ
fraction j by the Dougherty–Krieger
Equation (21) for hard spheres: Suspensions, Fig. 38 Variation of v/vo with j for poly-
styrene latex suspensions (R ¼ 1.55 mm)
!½Z fp
Z f
¼ 1 (117)
Zo fp
The sedimentation of particles in non-
where [Z] is the intrinsic viscosity (2.5 for hard Newtonian fluids, such as aqueous solutions
spheres). containing high molecular weight compounds
Combining Eqs. 116 and 117, (e.g., hydroxyethyl cellulose or Xanthan gum),
is not simple since these non-Newtonian solu-
!a½Z fp !k fp tions are shear thinning with the viscosity
v f f
¼ 1 ¼ 1 (118) decreasing with increase in shear rate. As
vo fp fp discussed above, these solutions show
a Newtonian region at low shear rates or shear
The above empirical relationship was tested stresses, usually referred to as the residual or
for sedimentation of polystyrene latex suspen- zero-shear viscosity Z(o). As discussed above,
sions with R ¼ 1.55 mm in 103 mol dm3 the stress exerted by the particles is very small,
NaCl. The results are shown in Fig. 38. in the region of 103–101 Pa depending on
The circles are the experimental points, the particle size and the density of the particles.
whereas the solid line is calculated using Clearly, to predict sedimentation, one needs
Eq. 118 with jp and k ¼ 5.4 (Buscall et al. 1982). to measure the viscosity at this low stresses.
Suspensions 1385 S
v/R 2 Polystyrene latex
1.48%
m–1 s–1 R = 1.5 μm
in EHEC solutions
1.0 1.30%
h 1.08%
Pa s
Ethyl hydroxyethyl
cellulose (EHEC)
0.86%
0.1 10–3 10–2 10–1 100 101
h(0) / Pa s
0.65%
Suspensions, Fig. 40 Variation of v/R2 with Z(o) for

0.43%
polystyrene latex suspensions (R ¼ 1.55)
0.01
0.22% exerted by the gel network with the bulk “elastic”
modulus of the suspension. The latter is related to
0.01 0.1 1.0 10.0 the high-frequency modulus G0 .
σ/Pa
Suspensions, Fig. 39 Variation of viscosity with

Prevention of Sedimentation and
applied stress for solutions of EHEC at various
concentrations Formation of Dilatant Sediments
Several methods may be applied to prevent sed-

This is illustrated for solutions of ethyl imentation and formation of clays or cakes in
hydroxyethyl cellulose (EHEC) in Fig. 39. a suspension:
Good correlation is found between the rate of (i) Balance of the density of the disperse phase
sedimentation for polystyrene latex and Z(o) and medium. It is clear from Stokes’ law that if
(Buscall et al. 1982) as shown in Fig. 40. Dr ¼ 0, vo ¼ o. This method can be applied
For the above suspension, no sedimentation only when the density of the particles is not
occurred when Z(o) was greater than 10 Pas. much larger than that of the medium (e.g.,
The situation with more practical dispersions is Dr 0.1). By dissolving an inert substance in
more complex due to the interaction between the the continuous phase, one may achieve density
thickener and the particles. Most practical sus- matching. However, apart from its limitation to S
pensions show some weak flocculation, and the particles with density not much larger than the
“gel” produced between the particles and thick- medium, the method is not very practical since
ener may undergo some contraction as a result of density matching can only occur at one
the gravity force exerted on the whole network. temperature.
A useful method to describe separation in these (ii) Reduction of particle size: As mentioned
concentrated suspensions is to follow the relative above, if R is significantly reduced (to values
sediment volume Vt/Vo or relative sediment below 0.1 mm), the Brownian diffusion can over-
height ht/ho (where the subscripts t and o refer come the gravity force, and no sedimentation
to time t and zero time, respectively) with storage occurs. This is the principle of formation of
time. For good physical stability, the values of nano-suspensions.
Vt/Vo or ht/ho should be as close as possible to (iii) Use of high molecular weight thickeners:
unity (i.e., minimum separation). This can be As discussed above, high molecular weight mate-
achieved by balancing the gravitational force rials such as hydroxyethyl cellulose or Xanthan
S 1386 Suspensions
gum when added above a critical concentration Mw =

(at which polymer coil overlap occurs) will pro- 223,000
100
duce very high viscosity at low stresses or shear
rates (usually in excess of several hundred Pas),
G*
and this will prevent sedimentation of the (Pa) 124,000
particles.
(iv) Use of “inert” fine particles: Several fine 50
particulate inorganic materials produce “gels”
70,000
when dispersed in aqueous media, for example,
sodium montmorillonite or silica. These particu-
late materials produce three-dimensional struc-
tures in the continuous phase as a result of 0.002 0.004 0.006 0.008 0.010
φp
interparticle interaction. For example, sodium
montmorillonite (referred to as swellable clays) Suspensions, Fig. 41 Variation of G* with jp for HEC
forms gels at low electrolyte concentrations by with various molecular weights
simple double layer interaction. At intermediate
electrolyte concentrations, the clay particles pro-
duce gels by “face-to-edge” association since the oxide side chains) to which hydroxyethyl
faces of the platelets are negatively charged, cellulose with various molecular weights was
whereas the edges are positively charged. At suf- added to the suspension. The weak flocculation
ficient particle concentration, the T-junctions was studied using oscillatory measurements.
produce a continuous gel network in the continu- Figure 41 shows the variation of the complex
ous phase, preventing the sedimentation of the modulus G* with jp.
coarse suspension particles. Finely divided silica Above a critical jp value (that depends on the
such as Aerosil 200 (produced by Degussa) pro- molecular weight of HEC), G* increases very
duces gel structures by simple association (by van rapidly with further increase in jp. When jp
der Waals attraction) of the particles into chains reaches an optimum concentration, sedimenta-
and cross chains. When incorporated in the con- tion is prevented. This is illustrated in Fig. 42
tinuous phase of a suspension, these gels prevent which shows the sediment volume in 10 cm cyl-
sedimentation. inders as a function jp for various volume frac-
(v) Use of mixtures of polymers and finely tions of the suspension js.
divided particulate solids: By combining the At sufficiently high volume fraction of the
thickeners such as hydroxyethyl cellulose or suspensions js and high volume fraction of free
Xanthan gum with particulate solids such as polymer jp, a 100 % sediment volume is reached,
sodium montmorillonite, a more robust gel struc- and this is effective in eliminating sedimentation
ture could be produced. This gel structure may be and formation of dilatant sediments.
less temperature dependent and could be opti- (vii) Use of liquid crystalline phases: Surfac-
mized by controlling the ratio of the polymer tants produce liquid crystalline phases at high
and the particles. concentrations. Three main types of liquid crys-
(vi) Depletion flocculation: As discussed tals can be identified: Hexagonal phase (some-
before, addition of free non-adsorbing polymer times referred to as middle phase), cubic phase,
can produce weak flocculation above a critical and lamellar (neat phase). All these structures
volume fraction of the free polymer, jp. This are highly viscous, and they also show elastic
weak flocculation produces a “gel” structure response. If produced in the continuous phase of
that reduces sedimentation. As an illustration, suspensions, they can eliminate sedimentation
results were obtained for a sterically stabilized of the particles. These liquid crystalline phases
suspension (using a graft copolymer of are particularly useful for application in liquid
polymethylmethacrylate with polyethylene detergents which contain high surfactant
Suspensions 1387 S
Turbid Flocculation starts fs = 0.3 Cohen-Stuart MA, Mulder JW (1985) Colloid Surf 15:49
10 Conner P, Ottewill RH (1971) J Colloid Interface Sci
fs = 0.2 37:642
8 Cosgrove T, Crowley TL, Ryan T (1987) Macromolecules
Sed.vol. (cm3)
20:2879
6 Day RE, Greenwood FG, Parfitt GD (1967) In Proceed-
fs = 0.1
ings of the 4th International congress of surface active
4 substances, vol 18, p 1005
High sediment volume Deryaguin BV, Landau L (1941) Acta Physicochem
2 USSR 14:633
with no claying
Einstein A (1906) Investigations on the theory of the
0.005 0.01 0.015 0.02 0.025 brownian movement. Dover, New York
φp Feeney PJ, Napper DH, Gilbert RG (1984) Macromole-
cules 17:2520
Ferry JD (1980) Viscoelastic properties of polymers.
Suspensions, Fig. 42 Variation of sediment volume
Wiley, New York
with jp
Fischer EW (1958) Kolloid Z 160:120
Flory PJ, Krigbaum WR (1950) J Chem Phys 18:1086
concentrations. Their presence reduces sedi- Fleer GJ, Cohen-Stuart MA, Scheutjens JMHM, Cosgrove
mentation of the coarse builder particles (phos- T, Vincent B (1993) Polymers at interfaces. Chapman
phates and silicates). and Hall, London
Fontana BJ, Thomas JR (1961) J Phys Chem 65:480
Fuerestenau DW (1971) In: Hair ML (ed) The chemistry
of biosurfaces. Marcel Dekker, New York, pp 91
Cross-References Fuerstenau DW, Healy TW (1972) In: Lemlich R (ed)
Adsorptive bubble separation technique. Academic
▶ Food Systems Press, London, pp 91
Garvey MJ, Tadros TF, Vincent B (1974) J Colloid Inter-
face Sci 49:57
Garvey MJ, Tadros TF, Vincent B (1976) J Colloid Inter-
References face Sci 55:440
Gaudin AM, Fuertsenau DW (1955) Trans AIME 202:958
Abeles F (1964) In: Passaglia E, Stromberg RR, Kruger Gibbs JW (1906) Scientific papers, vol 1. Longman Green,
J (eds) Ellipsometry in the measurement of surfaces London
and thin films, vol 256. National Bureau Standards Goodwin JW, Hughes R (2000) Rheology for chemists.
Miscellaneous Publications, Washington, DC, p 41 Royal Society of Chemistry Publication, Cambridge
Asakura A, Oosawa F (1954) J Chem Phys 22:1235 Greenwood FG, Parfitt GD, Picton NH, Wharton DG
Asakura A, Oosawa F (1958) J Chem Phys 93:183 (1968) Adv Chem Ser 79:135
Barrett KEJ (1975) Dispersion polymerisation in non- Healy TW (1974) J Macromol Sci Chem 118:603
aqueous media. Wiley, London Henry DC (1948) Proc R Soc London A 133:106
Barnett KG, Cosgrove T, Vincent B, Burgess A, Crowley Herschel WH, Buckley R (1926a) Measurement of con-
TL, Kims J, Turner JD, Tadros TF (1981) Polymer sistency as applied to rubber benzene solutions. Proc
Comm 22:283 Amer Soc Test Materials 26:621
Bijesterbosch BH (1987) Stability of solid/liquid disper- Herschel WH, Buckley R (1926b) Konsistenzmessungen S
sions. In: Tadros TF (ed) Solid/liquid dispersions. von Gummi- Benzollösungen. Kolloid Z 39:291–300.
Academic Press, London doi:10.1007/BF01432034
Blake T (1984) Wetting. In: Tadros TF (ed) Surfactants. Hesselink FT, Vrij A, Overbeek JTG (1971) J Phys Chem
Academic Press, London 75:2094
Blakely D (1975) Emulsion polymerisation. Applied Hirst EL, McGilvary DI and Percival EG (1950 ) Studies
Science Publication, London on fructosans. Part I. Inulin from dahlia tubers. J Chem
Buscall R, Goodwin JW, Ottewill RH, Tadros TF Soc 1297–1302
(1982) J Colloid Interface Sci 85:78 Hoeve CA (1970) On the general theory of polymer
Casson N (1959) In: Mill CC (ed) Rheology of Disperse adsorption at a surface. J Polym Sci 30:361
Systems. Pergamon Press, London, pp 84–104 Hoeve CA (1971) Theory of polymer adsorption at inter-
Clunie JS, Ingram BT (1983) In: Parfitt GD, Rochester CH faces. J Polym Sci 34:1
(eds) Adsorption from solution at the solid/liquid inter- Hough DB, Randall HM (1983) In: Parfitt GD, Rochester
face. Academic Press, London, pp 105 CH (eds) Adsorption from solution at the solid/liquid
Cohen-Stuart MA, Fleer GJ, Bijesterbosch J (1982) interface. Academic Press, London, pp 247
J Colloid Interface Sci 90:321 Huckel E (1924) Phys Z 25:204
S 1388 Suspensions (Solid/Liquid Dispersions)
Hunter RJ (1981) Zeta potential in colloid science: prin- Somasundaran P, Hannah HS (1979) In: Shah DO,
ciples and applications. Academic, London Schechter RS (eds) Improved oil recovery by surfac-
Killmann E, Eisenlauer E, Korn MJ (1977) Polymer Sci tant and polymer flooding. Academic Press, London,
Polymer Symposium 61:413 pp 205
Klimenko NA (1974) Tryasorukova and permilouskayan. Stevens CV, Meriggi A, Peristerpoulou M, Christov PP,
Kolloid Zh 36:678 Booten K, Levecke B, Vandamme A, Pittevils N,
Kleminko NA (1978) Kolloid Zh 40:1105 Tadros TF (2001) Biomacromolecules 2:1256
Kleminko NA (1979) Kolloid Zh 41:78 Strobel G (2006) Physics of polymers, 3rd edn. Springer,
Lyklema J (1987) Structure of the solid/liquid interface Berlin
and the electrical double layer. In: Tadros TF (ed) Suzuki M (1993) In: Suzuki M, Chatterton NJ (eds) Sci-
Solid/liquid dispersions. Academic Press, London ence and technology of fructans. CRC Press, Boca
Litchi G, Gilbert RG, Napper DH (1983) J Polym Raton Fl, p 21
Sci 21:269 Tadros TF, Booten K, Levecke B, Vandamme A, Stevens
Mackor EL, van der Waals JH (1951) J Colloid Sci 7:535 CV (2004) Adv Colloid Interface Sci 108–109:207
Napper DH (1981) Polymeric stabilisation of colloidal Tadros TF (1985) Polymer Colloids. In: Buscall R, Corner
dispersions. Academic, London T, Stageman JF (eds) Applied sciences. Elsevier,
Obey TM, Griffiths PC (1999) In: Goddard ED, Gruber JV London, p 105
(eds) Principles of polymer science and technology in Tadros TF, Vincent B (1978) J Colloid Interface Sci
cosmetics and personal care. Marcel Dekker, New 72:505
York, Chapter 2 Tadros TF (1981) Polymer adsorption and dispersion sta-
Ottewill RH (1982) In: Goodwin JW (ed) Concentrated bility. In: Tadros TF (ed) The effect of polymers on
dispersions. Royal Society of Chemistry Publication, dispersion properties. Academic Press, London
London, N.43, Chapter 9 Tadros TF (1980) Adv Colloid Interface Sci 12:141
Ottewill RH (1983) In: Poehlein GW, Ottewill RH, Tadros TF (1983) In: Poehlein GW, Ottewill RH (eds)
Goodwin JW (eds) Science and technology of polymer Science and technology of polymer colloids, vol II.
colloids, vol II. Martinus Nijhof Publishing, Boston/ Martinus Nijhof Publishing, Boston/The Hague
The Hague, p 503 Tadros TF (ed) (1987) Solid/liquid dispersions. Academic
Ottewill RH (1987) Properties of concentrated suspen- Press, London
sions. In: Tadros TF (ed) Solid/liquid dispersions. Tadros T (2010) Rheology of dispersions. Wiley-VCH,
Academic Press, London Germany
Ottewill RH, Shaw JN (1972) J Electroanal Interfacial Tadros TF (1993) Adv Colloid Interface Sci 46:1
Electrochem 37:133 Tadros TF (1996) Adv Colloid Interface Sci 68:91
Parfitt GD (1973) Fundamental aspects of dispersions. In: van den Boomgaard T, King TA, Tadros TF, Tang H,
Parfitt GD (ed) Dispersion of powders in liquids. Vincent B (1978) J Colloid Interface Sci 61:68
Applied Science Publishers, London Verwey EJW, Overbeek JTG (1948) Theory of stability of
Parfitt GD (ed) (1977) Dispersion of powders in liquids. lyophobic colloids. Elsevier, Amsterdam
Applied Science Publishers, London Volmer M (1939) Kinetik der Phase Buildung. Stemkopf,
Piirma I (1992) Polymeric surfactants, vol 42, Surfactant Dresden
science series. Marcel Dekker, New York von Smoluchowski M (191) Handbuch der Electricitat und
Pusey PN (1973) In: Green JHS, Dietz R (eds) Industrial des Magnetismu, vol II. Barth, Leipzig
polymers: characterisation by molecular weights. Wakamatsu T, Fuerstenau DW (1968) Adv Chem Ser
Transcripta Books, London 71:161
Rideal EK (1922) Phil Mag 44:1152 Washburn ED (1921) Phys Rev 17:273
Robb ID, Smith R (1974) Eur Polym J 10:1005 Whorlow RW (1980) Rheological techniques. Ellis
Roe RJ (1974) J Chem Phys 60:4192 Horwood, Chister
Saleeb FZ and Kitchener JA (1965) The effect of graphi- Wiersema PH, Loeb AL, Overbeek JTG (1967) J Colloid
tization on the adsorption of surfactants by carbon Interface Sci 22:78
blacks. J Chem Soc 911–917 Zisman WA (1964) Contact angles, wettability and adhe-
Scheutjens JMHM, Fleer GJ (1979) J Phys Chem 83:1919 sion, vol 43, Advances in chemistry series. ACS,
Scheutjens JMHM, Fleer GJ (1980) J Phys Chem 84:178 Washington, p 1
Scheutjens JMHM, Fleer GJ (1982) Adv Colloid Interface
Sci 16:341
Silberberg A (1968) J Chem Phys 48:2835
Smith WV, Ewart RH (1948) J Chem Phys 16:592
Smolders CA (1961) Contact angles; wetting and de- Suspensions (Solid/Liquid
wetting of mercury: Part II. Theory of wetting. Recl
Trav Chim 80:650
Dispersions)
Somasundaran P, Goddard ED (1979) Modern Aspects
Electrochem 13:207 ▶ Suspensions
Swollen Micelle 1389 S
the O/W or W/O interface, resulting in its
Suspoemulsions partial or complete displacement by the sus-
pension dispersant which is not a good emul-
Tharwat Tadros sion stabilizer, and this results in
Wokingham, Berkshire, UK coalescence of the emulsion droplets with
ultimate separation of oil (for O/W) or
water (for W/O)).
Synonyms (3) Heteroflocculation between the oil droplets
and suspension particles (the latter may be
Mixtures of suspensions and emulsions partially wetted by the oil, and they may
reside at the O/W interface (this is particu-
larly the case if the oil droplets are much
Definition larger than the suspension particles).
Heteroflocculation can also occur with sus-
Suspoemulsions are mixtures of suspensions and pension particles dispersed in a W/O
emulsions that are applied in many agrochemi- emulsion).
cal formulations whereby two active ingredients (4) Phase transfer and crystallization (this hap-
are formulated with one as an aqueous suspen- pens when the suspension particles have
sion and the other as an oil/water emulsion. Two some solubility in the oil phase). The small
main types can be distinguished in the follow- suspension particles which have higher solu-
ing: (1) a system whereby the solid particles and bility than the larger ones (due to curvature
emulsion droplets remain as separate entities effects) may become dissolved in the oil
and (2) a system whereby the solid particles are phase, and they become recrystallized onto
dispersed in the oil droplets. The first system is the larger suspension particles (a form of an
the one that is commonly applied in agrochem- Ostwald ripening process). Large and some-
ical formulations. With suspoemulsions, two times needle-shaped crystals may be pro-
active ingredients are formulated together duced as a result of crystal habit
which offers convenience to the farmer and modification (that sometimes occurs with
also may result in synergism in biological effi- Ostwald ripening). One of the most useful
cacy. A wider spectrum of disease control may methods to study interaction in
be achieved particularly with many fungicides suspoemulsions is rheology, in particular vis-
and herbicides. With many suspoemulsions, an coelastic measurements.
adjuvant that enhances the biological efficacy is
added. The formulation of suspoemulsions is not
an easy task; one may produce a stable suspen- Cross-References S
sion and emulsion separately, but when these
mixed, they become unstable due to the follow- ▶ Colloid Stability
ing interactions: ▶ Crystal Growth
(1) Homoflocculation of the suspension particles ▶ Emulsions
(this can happen if the dispersing agent used ▶ Surfactants
for preparation of the suspension is not ▶ Suspensions
strongly adsorbed, and hence it becomes
displaced by the emulsifier which is more
strongly adsorbed but not a good stabilizer
for the suspension particles) Swollen Micelle
(2) Emulsion coalescence (this can happen if
the emulsifier is not strongly adsorbed at ▶ Microemulsions

Marques2013 PDF

Uploaded by

Copyright:

Available Formats

Marques2013 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Marques2013 PDF

Uploaded by

Copyright:

Available Formats

S

Sedimentation decreasing exponentially with increase of height

T. Tadros (ed.), Encyclopedia of Colloid and Interface Science,

structures. The hexagonal phase is built up of

Synonyms Shear Rate Given in Reciprocal

The shear rate or strain rate is a measure of the Short-Range Interactions

When g is very small, tan g g. The shear Definition

interface in order to be able to remove the parti- Definition

Specific Surface Interactions ▶ Deposition

Colloid forces; Colloid interactions; Dispersion

Dispersion interactions; electric double-layer Synonyms

flocculation (Smoluchowski rate) is diffusion

Stern-Langmuir Isotherm and

Tharwat Tadros Tharwat Tadros S

Double layer Stern-Langmuir theory for surfactant adsorption

coverage to surfactant adsorption and free energy Cross-References

colloidal stabilization in order to optimize this pre- Definition

Surface Forces and the Equilibrium Cross-References

Surfactant solution Adsorption of surfactants

The Gibbs Adsorption Isotherm dg ¼ Gs1;2 d m2 (5)

or dg ¼ Gs2;1 RT d ln aL2 (8) S

1 Just above the break point,

a hydrophobic plate is used so that when the plate

The Pendent Drop Method

If a drop of oil is allowed to hang from the end of

Basically one measures the force required

Prolate ellipsoid Elongated cyclinder

This method is particularly useful for the Adsortion of Surfactants at the

no Dx1 Adsorption of Ionic Surfactants on

carbon black (Saleeb and Kitchener 1965;

Graphon and polystyrene (0.4 nm2). As with 0

Aveyard R, Haydon DA (1973) An introduction to the

ity, from a surface or interface provides

Surfactant Adsorption at 1e−3

Adsorbed amount (x10−10 mol cm−2)

where bi, ni(z) are the scattering lengths and num-

The adsorption is cooperative in nature and

Surfactant Adsorption at 1e+1

Surfactant Adsorption Kinetics The adsorption of surfactants at liquid interfaces

the surface layer depend on its composition and

" "  n1 !#

Diffusion-Controlled Adsorption where Di is the diffusion coefficient of ith com-

where the time-scaling variable t ¼ 0 ½ f ðtÞ 2 dt;

55 experimental data. Indeed, it is one of the

35 System with the Surfactant Soluble in Both Liquid

the adsorption equilibrium is quickly attained

Fast Dynamic Adsorption for Gas/Liquid Interfaces

properties of liquid/liquid interfaces. Figure 19 Dilational Viscoelasticity of Surfactant

where for diffusion from outside of a drop or a bubble and

Experimental dependency becomes considerable. Noteworthy

model assuming intrinsic compressibility of solutions. Dodecanol is a nonionic surfactant

ber, Rai ¼ rgc0iniLDi @r

@ci is the Rayleigh number, ni is

density), or it can be constructed from well as oscillation mechanism is not completely

of spread monolayer. The results presented in

monooctyl ether (C8EO4), in contact with References

Surfactant Aggregation Behavior

b1 b2 Some typical neutron reflectivity data for a binary

Surfactant Mixtures, a 1e−3

" " n1 !#

where the time-scaling variable t ¼ 0 ½ f ðtÞ2 dt;