Handbook Of: Surface AND Colloid Chemistry
Handbook Of: Surface AND Colloid Chemistry
Handbook Of: Surface AND Colloid Chemistry
SURFACE
AND
COLLOID
CHEMISTRY
FOURTH EDITION
SURFACE
AND
COLLOID
CHEMISTRY
FOURTH EDITION
HANDBOOK OF
SURFACE
AND
COLLOID
CHEMISTRY
FOURTH EDITION
This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid-
ity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright
holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this
form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may
rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or uti-
lized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopy-
ing, microfilming, and recording, or in any information storage or retrieval system, without written permission from the
publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923,
978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com
To Leon, Esma, and David.
Contents
Preface...............................................................................................................................................ix
Editor.................................................................................................................................................xi
Contributors.................................................................................................................................... xiii
Chapter 1 Introduction to Surface and Colloid Chemistry: Recent Advances and General
Remarks.........................................................................................................................1
K.S. Birdi
vii
viii Contents
ix
x Preface
polymer solutions is reviewed. Polymer–surfactant systems are described. The colloidal structures
and their stability have been found to be of much interest as described extensively in this handbook.
Finally, with great pleasure, I thank the staff of CRC Press for their patience and endurance in
helping me through at every stage in this task as an editor.
I also thank my family for providing the right kind of consideration while working through the
material for this handbook.
Editor
Professor K.S. Birdi received his BSc (Hons) from Delhi University, Delhi, India, in 1952. He
then traveled to the United States for further studies, majoring in chemistry at the University of
California at Berkeley. After graduation in 1957, he joined Standard Oil of California, Richmond.
In 1959, he moved to Copenhagen, Denmark, where he joined Lever Brothers as chief chemist in the
Development Laboratory. During this period, he became interested in surface chemistry and joined,
as assistant professor, the Institute of Physical Chemistry (founded by Professor J. Brønsted), Danish
Technical University, Lyngby, Denmark, in 1966. He initially did research on surface s cience aspects
(e.g., detergents, micelle formation and emulsion technology, adsorption and adhesion, and biophys-
ics). During the early exploration and discovery stages of oil and gas in the North Sea, Professor
Birdi got involved in Research Science Foundation programs, with other research institutes around
Copenhagen, in the oil recovery phenomena and surface science (enhanced oil recovery [EOR]).
Later, research grants on the same subject were awarded from the European Union projects. These
projects also involved extensive visits to other universities and an exchange of guests from all over
the world. Professor Birdi was appointed research professor in 1985 (Nordic Science Foundation)
and was then appointed, in 1990 (retired in 1999), to the School of Pharmacy, Copenhagen, as pro-
fessor in physical chemistry.
There was continuous involvement with various industrial contract research programs through-
out these years. These projects have actually been a very important source of information in
keeping up with real problems and helped in the guidance of research planning at all levels.
Professor Birdi has been a consultant to various national and international industries. He has
been a member of various chemical societies and a member of organizing committees of national
and international meetings related to surface science. He has been a member of selection commit-
tees for assistant professors and professors and was an advisory member (1985–1987) of the ACS
journal Langmuir.
Professor Birdi has been an advisor for about 90 advanced student projects and various PhD
projects. He is the author of about 100 papers and articles (and a few hundred citations).
In order to describe these research observations and data, he realized that it was essential to
delineate these in the form of books related to surface and colloid chemistry. His first book on
surface science was published in 1984: Adsorption and the Gibbs Surface Excess, Chattoraj, D.K.
and Birdi, K.S., Plenum Press, New York. His further publications include Lipid and Biopolymer
Monolayers at Liquid Interfaces, K. S. Birdi, Plenum Press, New York, 1989; Fractals, in Chemistry,
Geochemistry and Biophysics, K. S. Birdi, Plenum Press, New York, 1994; Handbook of Surface
and Colloid Chemistry, K. S. Birdi (editor) (first edition, 1997; second edition, 2003; third edition,
2009; CD-ROM, 1999), CRC Press, Boca Raton, FL; Self-Assembly Monolayer, Plenum Press,
New York, 1999; Scanning Probe Microscopes, CRC Press, Boca Raton, FL, 2002; Surface and
Colloid Chemistry, CRC Press, Boca Raton, FL, 2010 (translated to Kazakh, 2013); Introduction to
Electrical Interfacial Phenomena, K. S. Birdi (editor), 2010, CRC Press, Boca Raton, FL, 2013; and
Surface Chemistry Essentials, CRC Press, Boca Raton, FL, 2014. Surface chemistry (theory and
applications) has remained his major area of research interest throughout these years.
xi
Contributors
Waisudin Badri Peter A. Kralchevsky
University of Lyon Faculty of Chemistry and Pharmacy
Lyon, France Department of Chemical and Pharmaceutical
Engineering
and
Sofia University
University Lyon-1 Sofia, Bulgaria
Villeurbanne, France
Qiuli Lu
Sanjib Bagchi Department of Biomedical, Chemical and
Department of Chemical Sciences Environmental Engineering
Indian Institute of Science Education and University of Cincinnati
Research, Kolkata Cincinnati, Ohio
Kolkata, India
xiii
xiv Contributors
CONTENTS
1.1 Introduction............................................................................................................................... 3
1.2 Capillarity and Surface Forces (in Liquids) (Curved Surfaces).............................................. 14
1.2.1 Origin of Surface Forces (in Liquids).......................................................................... 15
1.2.1.1 Surface Energy.............................................................................................. 16
1.2.2 Capillary Forces: Laplace Equation (Liquid Curvature and Pressure)
(Mechanical Definition)............................................................................................... 18
1.2.3 Capillary Rise (or Fall ) of Liquids (Capillary Forces)............................................... 23
1.2.4 Soap Bubbles (Formation and Stability)......................................................................25
1.2.5 Measurement of Surface Tension of Liquids............................................................... 27
1.2.5.1 Shape and Weight of Liquid Drop................................................................28
1.2.5.2 Ring Method (Detachment).......................................................................... 30
1.2.5.3 Plate (Wilhelmy) Method............................................................................. 31
1.3 Typical Surface Tension Data of Liquids................................................................................. 32
1.3.1 Effect of Temperature and Pressure on Surface Tension of Liquids........................... 35
1.3.2 Heat of Liquid Surface Formation and Evaporation.................................................... 37
1.3.3 Other Surface Properties of Liquids............................................................................ 38
1.3.4 Interfacial Tension of Liquid1–Liquid2........................................................................ 39
1.3.5 Thermodynamics of Liquid Surfaces (Corresponding States Theory of Liquids)...... 41
1.4 Surfactants (Soaps and Detergents) Aqueous Solutions (Surface-Active Substances)............44
1.4.1 Surface Tension Properties of Aqueous Surfactant Solutions.....................................46
1.4.2 Surface-Active Substances (Amphiphiles)................................................................... 47
1.4.3 Aqueous Solution of Surfactants................................................................................. 48
1.4.4 Solubility Characteristics of Surfactants in Water...................................................... 48
1.4.4.1 Ionic Surfactants........................................................................................... 49
1.4.4.2 Nonionic Surfactants.................................................................................... 50
1.4.5 Micelle Formation of Surfactants (in Aqueous Media) (Critical Micelle
Concentration)............................................................................................................. 51
1.4.5.1 Analyses of CMC of Surfactants.................................................................. 54
1.4.6 Gibbs Adsorption Equation in Solutions..................................................................... 56
1
2 Handbook of Surface and Colloid Chemistry
1.1 INTRODUCTION
Surface and colloid chemistry is now recognized to play a very important role in everyday life.
Accordingly, the number of research publications related to this subject is also very extensive.
Therefore, it is useful to delineate in this chapter some recent advances and general remarks, as
related to the fundamental principles of the subject. All natural phenomena are defined by scientific
observations of the matter surrounding the earth and heaven. It has been recognized that the science
needs information on the structures of the matter (defined as solids, liquids, gases) (Figure 1.1).
Research conducted during the last decades has shown that one needs a much more detailed pic-
ture of these structures in all kinds of processes (chemical industry and technology and natural
biological phenomena). Application of this research has been relevant to many modern industrial
innovations (as regards future challenges: drinking water, energy resources [conventional and non-
conventional oil/gas reservoirs], food, clean air, transportation, pollution control, housing, health
and medicine, paper, printing, etc.).
Matter exists as
Gas
Liquid
Solid
4 Handbook of Surface and Colloid Chemistry
(a)
(b)
(c)
FIGURE 1.1 Molecular structure (schematic) of solid (a), liquid (b), and gas (c).
This has been recognized by classical science (as depicted in the following text):
Several decades of research has shown that the molecules that are situated at the interfaces (e.g.,
between gas–liquid, gas–solid, liquid–solid, liquid1–liquid2, solid1–solid2) are known to behave dif-
ferently (Figure 1.2) than those in the bulk phase (Adam, 1930; Adamson and Gast, 1997; Aveyard
and Hayden, 1973; Bakker, 1928; Bancroft, 1932; Barnes, 2011; Birdi, 2003a, 2003b, 2009, 2010a,
2014; Birdi and Nikolov, 1979; Biresaw and Mittal, 2008; Butt, 2010; Butt et al., 2013; Castner
and Ratner, 2002; Chattoraj and Birdi, 1984; Cini et al., 1972; David and Neumann, 2014; Davies
and Rideal, 1963; De Gennes, 2003; Defay et al., 1966; Fanum, 2008; Fendler and Fendler, 1975;
Gaines, 1966; Harkins, 1952; Hiemenz and Rajagopalan, 1997; Holmberg, 2002; Israelachvili,
2011; Janaiaud et al., 2014; Kelkar et al., 2014; Kolasinski, 2008; Krungleviclute et al., 2012;
Lara et al., 1998; MacDowell et al., 2014; Masel, 1996; Matijevic, 1969; McCash, 2001; Merk
et al., 2014; Miller and Neogi, 2008; Mousny et al., 2008; Mun et al., 2014; Nicot and Scanlon,
2012; Oura, 2003; Packham, 2005; Papachristodoulou and Trass, 1987; Partington, 1951; Rao and
Geckeler, 2011; Rosen, 2004; Schramm, 2005; Sinfelt, 2002; Soltis et al., 2004; Somasundaran,
2006; Somorjai and Li, 2010, 2011; Taylor, 1984; Yates and Cambell, 2011; Zana, 2008.).
Surface
FIGURE 1.2 Molecules in the bulk phase (liquid or solid) and in surface phase (molecular dimension:
schematic).
Introduction to Surface and Colloid Chemistry 5
Liquid surfaces
Surfaces of oceans, lakes, and rivers
Lung surface, biological cell surfaces
Solid surfaces
Adhesion, glues, tapes
Cement (and building) industry
Paper and printing industry
Construction industry (dams, tunnels, etc.)
Catalysis
Road surfaces (car tire, etc.)
Paint industry
Liquid–solid interfaces
Washing and cleaning (dry cleaning)
Wastewater treatment
Air pollution
Power plants
Liquid–liquid interfaces (oil–water systems)
Emulsions (cosmetics; pharmaceutical products)
Diverse industrial applications:
Oil and gas (conventional reservoirs and extraction of shale oil and gas fracking technology,
methane hydrates)
Food and milk products
Cleaning and washing
Medical applications
Electronics and related industries
Modern surface and colloid chemistry science is based on molecular-level understanding and prin-
ciples. These scientific developments have been useful in providing the information (at molecular
level) on various phenomena. Further, in some instances, such as oil spills, one can easily realize the
importance of the role of surface of oceans (Figure 1.3). It is found that part of oil evaporates, while
Evaporation
Oil spill
Ocean
water
Dissolution
FIGURE 1.3 Ocean surface and oil spill (evaporation; solution; sinking; floating states).
6 Handbook of Surface and Colloid Chemistry
some sinks to the bottom, and main part remains floating on the surface of water. This process is in
fact one of the major areas of surface chemistry applications.
Another observation of great importance arises from the fact that oceans cover some 75% of the
surface of the earth. Obviously, in this case, the reactions on the surface of oceans will have much
consequence for life on the earth. For example, carbon dioxide as found in air is distributed in vari-
ous systems, such as
In air
In plants
In oceans, rivers, and lakes (CO2 is soluble in water)
Thus, CO2 (as present in air [concentration around 350 ppm]) is in equilibrium in these different
systems, which will have consequences on the pollution control. The adsorption of CO2 in oceans
(lakes, rivers: about 25% of CO2 in air) thus is primarily a surface phenomenon. In other words, it is
not enough to monitor CO2 in air alone; one must have the knowledge of CO2 concentrations in the
other two states (i.e., plants and oceans). The chemical potentials of CO2 (CO2) in different phases
are at equilibrium:
At the same time, it is important to mention that CO2 being the main component of photosynthesis
in all plant (i.e., food) growth on the earth, it is absolutely the only necessary element for the exis-
tence of life on the earth! The global CO2 equilibrium is thus a complex quantity and an important
challenge for scientists.
Experiments have shown that the molecules situated near or at the interface (i.e., liquid–gas)
will be interacting differently with respect to each other than the molecules in the bulk phase. The
intramolecular forces acting would thus be different in these two cases. In other words, all processes
occurring near any interface will be dependent on these molecular orientations and interactions.
Furthermore, it has been pointed out that, for a dense fluid, the repulsive forces dominate the fluid
structure and are of primary importance. The main effect of the repulsive forces is to provide a uni-
form background potential in which the molecules move as hard spheres. These considerations have
shown that the molecules at the interface would be under an asymmetrical force field, which gives
rise to the so-called surface tension or interfacial tension (IFT) (Adamson and Gast, 1997; Birdi,
1989, 1997, 1999, 2003a; Chattoraj and Birdi, 1984). At a molecular level, when one moves from
one phase to another, that is, across an interface, this leads to the adhesion forces between liquids
and solids, which is a major application area of surface and colloid science. The resultant force
on molecules will vary with time because of the movement of the molecules; the molecules at the
surface will be pointed downward into the bulk phase. The nearer the molecule is to the surface, the
greater is the magnitude of the force due to asymmetry. The region of asymmetry near the surface
plays a very important role. Thus, when the surface area of a liquid is increased, some molecules
must move from the interior of the continuous phase to the interface. Surface tension of a liquid is
the force acting normal to the surface per unit length of the interface, thus tending to decrease the
surface area. Neighboring molecules surround the molecules in the liquid phase, and these interact
with each other in a symmetrical way. Further, molecules at the surface in most cases will also be
oriented in a different way than in the bulk phase (as found from experiments).
In the gas phase, where the density is 1000 times lesser than that in the liquid phase, the inter-
actions between molecules are very weak as compared to those in the dense liquid phase. Thus,
when one crosses the line from the liquid phase to the gas phase, there is an abrupt change in
the density of factor 1000. This means that while in liquid phase a molecule occupies a volume,
which is 1000 times smaller than when in the gas phase. The interfacial region is found to be of
Introduction to Surface and Colloid Chemistry 7
Insect on
water surface
Water
FIGURE 1.4 Insect (many different kinds of insects, such as mosquitoes, etc.) strides on surface of water
(such as lakes, rivers, oceans) (schematic).
Iron needle
Surface of water (surface tension)
The reason a heavy object can float on water is due to the fact that in order for the latter to sink, it
must overcome the surface forces. Of course, if one merely drops the metal object, it will overcome
the surface tension force and sink, which one generally observes. This clearly shows that at any
liquid surface, there exists a tension (surface tension), which needs to be broken when any contact
is made between the liquid surface and the material (here the metal needle). One notices ample of
examples on the surfaces of rivers and lakes, where stuff is seen floating about. Based on the same
8 Handbook of Surface and Colloid Chemistry
principles, it is found that the smooth hull of a ship exerts less resistance to sailing than a rough
bottom, thus saving in energy.
Definition of liquid interfaces:
Surface properties of solids: For example, in the case when a solid sample is crushed and surface
area increases per unit gram (Figure 1.5), the surface area per gram of the solid increases. For exam-
ple, finely divided talcum powder has a surface area of 10 m2/g. Active charcoal exhibits surface
areas corresponding to over 1000 m2/g. In fact, in cement, the energy input in producing very finely
divided particles is much greater than the rest of the production costs. This property has been found
to be an appreciable quantity and has important consequences. Qualitatively, one must notice that
work has to be put into the system when one increases the surface area (both for liquids and solids
or any other interface). Cement is mainly based on the mechanical energy used to make the particles
as small as possible, such that cost is dependent on this process. The solid surface properties change
drastically when one increases the surface area per unit weight.
The surface chemistry of small particles is an important part of everyday life (such as dust, tal-
cum powder, sand, rain drops, emissions, etc.). Thomas Graham already defined colloid chemistry
and its relation to surface chemistry a century ago. A particle having dimensions in the range of
10 −9 m (= 1 nm = 10 Å) and 10 −6 m (1 μm) was designated to be a colloidal (Birdi, 2003a, 2010a;
Gitis and Sivamani, 2004; Scheludko, 1966). Colloids are an important class of materials, interme-
diate between bulk and molecularly dispersed systems. Colloidal particles may be spherical, but in
some cases, one dimension can be much larger than the other two (as in a needle-like shape). The
size of particles also determines whether one can see them by the naked eye. Colloids are not visible
to the naked eye or under an ordinary optical microscope. The scattering of light can be suitably
used to see such colloidal particles (such as dust particles, etc.). The size of colloidal particles may
range from 10 −4 to 10 −7 cm. The units used are as follows:
1 μm = 10 −6 m
1 nm = 10 −9 m
1 Å (Angstrom) = 10 −8 cm = 0.1 nm = 10 −10 m (nm = nano-meter)
FIGURE 1.5 Formation of fine (colloidal) particles (such as talcum powder, active charcoal, cement)
(schematic: size less than μm).
Introduction to Surface and Colloid Chemistry 9
Considering the following simple examples can provide some information about the nanodimension
scale:
In surface chemistry, there is a great range of dimensions as needed to describe a variety of systems.
As seen here, the range of dimensions is many folded! Accordingly, a unit Angstrom (Å = 10 −8 cm)
was used for systems of molecular dimension (famous Swedish scientist). Most common unit is
though nanometer (= 10 Å = 10 −9 m), which is mainly used for molecular scale features. In recent
years, nanosize (nanometer range) particles are of much interest in different applied science systems
(Nano = from Greek and means dwarf) (Lin et al., 2005). Nanotechnology is actually strongly
getting a boost from the last decade of innovation as reported by the surface and colloid literature
(Mun et al., 2014; Rao and Geckeler, 2011). In fact, light-scattering is generally used to study the
size and size distribution of such systems. Since colloidal systems consist of two or more phases and
components, the interfacial area-to-volume ratio becomes very significant. Colloidal particles have
a high surface area-to-volume ratio compared with ordinary bulk materials. A significant propor-
tion of colloidal molecules lie within, or close to, the interfacial region. Hence, the interfacial region
has significant control over the properties of colloids. To understand why colloidal dispersions can
be either stable or unstable, we need to consider
1. The effect of the large surface area-to-volume ratio (e.g., 1000 m2 surface area per gram of
solid [active charcoal])
2. The forces operating between the colloidal particles
3. Surface charges (varying from positive or negative or neutral surface charge) are very
important characteristics of such systems
There are some very special characteristics, which must be considered as regards colloidal particle
behavior: size and shape, surface area, and surface charge density.
10 Handbook of Surface and Colloid Chemistry
FIGURE 1.7 Brownian motion (schematic movement of small [dust] particles or large molecules [polymers,
etc.]) (in air or liquid media).
The Brownian motion of particles is a much studied field (Figure 1.7). The motion of col-
loidal particles is the basis of the physical characteristics of the system. The fractal nature
of surface roughness has recently been shown to be of importance (Avnir, 1989; Birdi, 1993;
Feder, 1988). Recent applications have been reported where nanocolloids have been employed.
It is thus found that some relevant terms are needed to be defined at this stage. The definitions
generally employed are as follows. Surface is a term used when one considers the dividing
phase between
Gas–Liquid
Gas–Solid
Interface (Figure 1.8) is the term used when one considers the dividing phase:
Solid–Liquid (colloids)
Liquid1–Liquid2 (oil–water emulsion)
Solid1–Solid2 (adhesion: glue, cement)
It is thus found that surface tension may be considered to arise due to a certain well-defined
degree of unsaturation of bonds that occurs when a molecule resides at the surface and not in the
bulk. The term surface tension is used for solid–vapor or liquid–vapor interfaces. The term IFT
is more generally used for the interface between two liquids (oil–water), two solids, or a liquid
and solid. It is, of course, obvious that in a one-component system, the fluid is uniform from the
bulk phase to the surface. However, the orientation of the surface molecules will be different
from those molecules in the bulk phase in all systems. For instance, in the case of water, the
orientation of molecules inside the bulk phase will be different than those at the interface. The
hydrogen bonding will orient the oxygen atom toward the interface. The question one may ask,
then, is how sharply does the density change from that of being fluid to that of gas (a change by
a factor of 1000). Is this a transition region a monolayer deep or many layers deep? The Gibbs
adsorption theory (Birdi, 1989, 1999, 2003a, 2009, 2010a; Chattoraj and Birdi, 1984; Defay
et al., 1966) considers surface of liquids to be monolayer. The surface tension of water decreases
appreciably on the addition of very small quantities of soaps and detergents. The Gibbs adsorp-
tion theory relates the change in surface tension to the change in soap concentration (chemical
potential). The experiments, which analyze the spread monolayers, are also based on one molec-
ular layer. The latter data conclusively indeed verify the Gibbs assumption. Detergents and
other (soaps, etc.) similar kind of molecules are found to exhibit self-assembly characteristics
Introduction to Surface and Colloid Chemistry 11
Air Air
Surface Surface
Solid Liquid
(a) (b)
Solid1
Liquid
Surface Surface
Solid Solid2
(c) (d)
FIGURE 1.8 Characterization of different interfaces: (a) solid–gas (air); (b) liquid–gas (air); (c) solid–liquid;
and (d) solid1–solid2.
self-assembly monolayer (SAM) (Birdi, 1999). However, there exists no procedure, which can
provide information by a direct measurement.
Colloids (Greek word for glue-like) are a wide variety of systems consisting of finely divided
particles (or macromolecules [such as glue, gelatine, proteins, etc.]), which are found in everyday
life (Table 1.1). In general, one classifies colloidal systems into three distinct types (Adamson and
Gast, 1997; Birdi, 2003a, 2009; Dukhin and Goetz, 2002; Lyklema, 2000):
1. In simple colloids, clear distinction can be made between the disperse phase and the dis-
perse medium, for example, simple emulsions of oil-in-water (o/w) or water-in-oil (w/o).
2. Multiple colloids involve the coexistence of three phases of which two are finely divided,
for example, multiple emulsions (mayonnaise, milk) of water-in-oil-in-water (w/o/w) or oil-
in-water-in-oil (o/w/o).
3. Network colloids have two phases forming an interpenetrating network, for example,
polymer matrix.
The stability of colloidal systems (in the form as solids or liquid drops) is determined by the free
energy (surface free energy or the interfacial free energy) of the system. The main parameter of
interest is the large surface area exposed between the dispersed phase and the continuous phase.
Since colloidal particles move about constantly, the Brownian motion determines their disper-
sion energy. The energy imparted by collisions with the surrounding molecules at temperature
T = 300 K is 3/2kBT = (3/2)(1.38 × 10 −23)(300) = 0.6 × 10 −20 J (where k B is the Boltzmann constant).
12 Handbook of Surface and Colloid Chemistry
TABLE 1.1
Typical Colloidal Systems
Phases
Dispersed Continuous System Name
Liquid Gas Aerosol fog, spray
Gas Liquid Foam, thin films, froth
Fire extinguisher foam
Liquid Liquid Emulsion (milk)
Mayonnaise, butter
Solid Liquid Sols, AgI, photography films
Suspension wastewater
Cement
Oil recovery (shale oil)
Coal slurry
Bio colloids
Corpuscles Serum Blood
Blood-coagulants
Hydroxyapatite Collagen Bone–teeth
Liquid Solid Solid emulsion (toothpaste; paints)
Solid Gas Solid aerosol (dust)
Gas Solid Solid foam (polystyrene)
Insulating foam
Solid Solid Solid suspension/solids in plastics
This energy and intermolecular forces would thus determine the colloidal stability. In the case
of colloid systems (particles or droplets), the kinetic energy transferred on collision will be thus
kBT = 10 −20 J. However, at a given moment, there is a high probability that a particle may have a
larger or smaller energy in comparison to kBT. Further, the probability of total energy several times
kBT (over 10 times kBT) thus becomes very small. The instability will be observed if the ratio of the
barrier height to kBT is around 1–2 units.
The idea that two species (solid1–solid2) should interact with each other, so that their mutual
potential energy can be represented by some function of the distance between them, has been
described in the literature (Israelachvili, 2011). Furthermore, colloidal particles frequently adsorb
(and even absorb) ions from their dispersing medium (such as in ground water treatment and purifi-
cation). Sorption that is much stronger than what would be expected from dispersion forces is called
chemisorption, a process that is of both chemical and physical interest.
Emulsions: As one knows from experience: oil and water do not mix, which suggests that these
systems are dependent on the oil–water interface (Figure 1.9). The liquid1–liquid2 (oil–water)
interface is found in many systems, which is most important in the world of emulsions.
The trick in using emulsions is based on the fact that one can apply both water and oil (latter is
insoluble in water) simultaneously. Further, one can then include other molecules, which may be
soluble in either phase (water or oil). This obviously leads to the common observation where we
find thousands of applications of emulsions. It is very important to mention here that actually nature
uses this trick in most of the major biological fluids. The most striking example is milk. The emul-
sion chemistry of milk (and other food emulsion systems) has been found to be the most complex
(Kristensen, 1997). Paint consists of polymer molecules dispersed in water phase. After application,
water evaporates, leaving behind a glossy layer of paint.
Introduction to Surface and Colloid Chemistry 13
Oil
Water
Oil in water
Oil
Water
In fact, the state of mixing oil and water is an important example of interfacial behavior at
l iquid1–liquid2. Emulsions of oil–water systems are useful in many aspects of daily life: milk, foods,
paint, oil recovery, pharmaceutical, and cosmetics.
If one mixes olive oil with water and on shaking, one gets the following (Figure 1.9):
However, if one adds suitable substances, which change surface forces, then the olive oil drops
formed can be very small (micrometer [μm] range). The latter leads to a stable emulsion.
The degree of emulsion stability is basically dependent on the size of the oil drops (as dispersed
in water), besides other factors:
In addition, one finds that these considerations are important in regard to the different systems as
follows:
Paints, cements, adhesives, photographic products, water purification, sewage disposal, emul-
sions, chromatography, oil recovery, paper and printing industry, microelectronics, soap and deter-
gents, catalysts, and biology (cell, virus). In some oil–surfactant–water–diverse components, liquid
crystal (LC) phases (lyotropic LC) are observed. These lyotropic LCs are indeed the basic building
blocks in many applications of emulsions in technology. LC structures can be compared with a
layer-cake where each layer is molecular thick. It is thus seen that surface science pertains to inves-
tigations, which take place between two different phases.
Further, the degree of curvature is found to impart different characteristics to liquid surfaces.
The state of molecules in different phases can be described as follows:
1.
Molecules inside the bulk phase: molecules are surrounded symmetrically in all directions.
2.
Molecules at the surface: molecules are interacting with molecules in the bulk phase (same
as under I), but towards the gas phase, the interaction is weaker due to larger distances
between molecules (Figure 1.1).
The molecular forces, which are present between the surface molecules, are found to be differ-
ent from the forces acting on molecules in the bulk phase or the gas phase. Accordingly, these
forces are called surface forces. Surface forces make the liquid surface behave like a stretched
elastic membrane in that it tends to contract. One cannot see this phenomenon directly, but it
is observed through indirect experimental observations (both qualitatively and quantitatively!).
The latter arises from the observation that when one empties a beaker with a liquid, the liq-
uid breaks up into spherical drops. This indicates that drops are being created under some
forces, which must be present at the surface of the newly formed interface. These surface forces
become even more important when a liquid is in contact with a solid (such as ground water,
oil reservoir). The flow of liquid (e.g., water or oil) through small pores in the underground is
mainly governed by the capillary forces. One can also accept that recovering oil from large
depths (10 km!) in underground needs appreciable pressure if pores are very small. It is found
that capillary forces play a very dominant role in many of these systems. Thus, the interac-
tion between liquid and any solid (with pores) will form curved surfaces, which being dif-
ferent from planar fluid surface gives rise to capillary forces. Another essential aspect is the
mechanical surface tension present at curved surfaces. The curved liquid surfaces, such as
in drops, or small capillaries, exhibit very special properties than flat liquid surfaces. This is
essential, since these principles are the building blocks for the understanding of the subject.
These surface forces interact both at liquid1–liquid 2 (such as oil–water) and liquid–solid (such
as cement–water) interfaces.
Example: Water
Volume per mole liquid water = Vliquid = 18 mL/mol
Volume per mole water in gas state (at STP) (Vgas) = 22 L/mol
Ratio Vgas/Vliquid ≅ 1000
This illustrates that the approximate ratio of distance between molecules in gas phase and liquid
phase will be about 10 ≅ (1000)1/3 (from simple geometrical considerations of volume [proportional
to length3] and length). In other words, the surface chemistry is related to those molecules, which
are situated in this transition region. Experiments have clearly shown that this transition region is
of molecular dimension. The same is true for all liquids, with only minor differences.
In other words, the density of water changes 1000 times as the surface is crossed from liquid
phase to the gas phase (air). This means that in gas phase, each molecule occupies 1000 more vol-
ume than in the liquid phase. Thus, the molecules in the gas phase move larger distances before
interacting with another molecule. This large change means that the surface molecules must be
under different environment than in the liquid phase or in the gas phase. Surface phase is the
transition region. Distance between gas molecules is approximately 10 times larger than in a liquid.
Hence, the forces (all forces increase when distances between molecules decrease) between gas
molecules are much weaker than in the case of liquid phase. All interaction forces between mol-
ecules (solid phase, liquid phase, and gas phase) are related to the distance between molecules. It is
also obvious that the surface of a liquid is a very busy place, since molecules are evaporating into
the air and some are also returning into the liquid phase. This is a snapshot at a molecular level, so
one does not see it by the naked eye.
16 Handbook of Surface and Colloid Chemistry
In the case of water, it is cohesive forces that maintain water, for example, in liquid state at room
temperature and pressure (Franks, 1975; Fraxedas, 2014). It is useful as an example to compare
cohesive forces in two different molecules, such as H2O and H2S. At room temperature and pres-
sure, H2O is liquid while H2S is a gas. This means that H2O molecules interact with different forces,
which are stronger and thus form a liquid phase. On the other hand, H2S molecules exhibit much
lower interactions and thus are in a gas phase at room temperature and pressure. In other words,
hydrogen bonds (between H and O) in water are stronger than hydrogen–sulfur bonds.
or
⎛ dx ⎞ f
γ = f⎜ ⎟ = (1.2)
⎝ dA ⎠ 2l
where
dx is the change in displacement
l is the length of the thin film
The quantity γ represents the force per unit length of the surface (mN/m = dyn/cm), and this force is
defined as surface tension or IFT. Surface tension, γ, is the differential change of free energy with
the change of surface area at constant temperature, pressure, and composition.
One may consider another example to describe the surface energy. Let us imagine that a liq-
uid fills a container of the shape of a funnel. In the funnel, if one moves the liquid upward, then
there will be an increase in surface area. This requires that some molecules from bulk phase have
to move into the surface area and create extra surface AS. The work required to do so will be
dx
FIGURE 1.11 Surface film of a liquid (schematic) (see text for details).
Introduction to Surface and Colloid Chemistry 17
(force times area) γAS. This is reversible work at constant temperature and pressure; thus, it gives
the increase in free energy of the system:
dG = −γA S (1.3)
Thus, the tension per unit length in a single surface, or surface tension, γ, is numerically equal to the
surface energy per unit area. Then GS, the surface free energy per unit area, is
⎛ dG ⎞
GS = γ = ⎜ ⎟ (1.4)
⎝ dA ⎠
Under reversible conditions, the heat (q) associated with it gives the surface entropy, SS
dq = TdSS (1.5)
dγ
= −SS (1.6)
dT
H S = G S + TS (1.7)
E S = G S + TSS (1.8)
⎛ dγ ⎞
ES = γ − T ⎜ ⎟ (1.9)
⎝ dT ⎠
The quantity ES has been found to provide more useful information on surface phenomena than any
other quantities (Birdi, 2009).
Here we find that the term SS is the surface entropy per square centimeter of surface. This shows
that to change the surface area of a liquid (or solid: as described later), there exists a surface energy
(γ: surface tension), which one needs to consider.
The quantity γ means that to create a fresh surface of 1 m2 (=1020 Å2) of new surface of water,
one will need to use 72 mJ energy. To transfer a molecule of water from the bulk phase (where it is
surrounded by about 10 near neighbors by about 7 kBT) (kBT = 4.12 × 10 −21 J) to the surface, one
need to break about half of these hydrogen bonds (i.e., 7/2kBT = 3.5kBT). The free energy of trans-
fer of one molecule of water (with area of 12 Å2) will be thus about 10 −20 J (or about 3kBT). This
is a reasonable quantity under these simple geometrical assumptions (Adamson and Gast, 1997;
Birdi, 2009).
In the case of solid systems, similar consideration is needed if one increases the surface area of
a solid (e.g., by crushing). In the latter case, one needs to measure and analyze the surface tension
of the solid. It is found that energy needed to crush a solid is related to the surface forces (i.e., solid
surface tension).
18 Handbook of Surface and Colloid Chemistry
Liquids show characteristics, which are specific due to the fact that molecules in liquids are able to
move inside a container, while a solid cannot exhibit this property. It is found that this property of
liquid gives some specific properties, such as curved surfaces in narrow tubings (Adamson and Gast,
1997; Birdi, 2009; Goodrich et al., 1981). It is interesting to consider aspects in the field of the wet-
tability. Surely everybody has noticed that water tends to rise near the walls of a glass container. This
happens because the molecules of this liquid have a strong tendency to adhere to the glass. Liquids
that wet the walls make concave surfaces (e.g., water/glass), and those that do not wet them make
convex surfaces (e.g., mercury/glass). Inside tubes with internal diameter smaller than 2 mm, called
capillary tubes, a wettable liquid forms a concave meniscus in its upper surface and tends to go up
along the tube. On the contrary, a nonwettable liquid forms a convex meniscus and its level tends
to go down. The amount of liquid attracted by the capillary rises until the forces, which attract it,
balance the weight of the fluid column. The rising or the lowering of the level of the liquids into thin
tubes is named capillarity (capillary force). Also, the capillarity is driven by the forces of cohesion
and adhesion we have already mentioned.
Any liquid inside a large beaker has almost a flat surface. However, the same liquid inside a
fine tubing will be found to be curved at the surface (Figure 1.12). In other words, the curved
liquid surface gives rise to some very characteristic properties, which one finds in everyday
life (e.g., water flow in earth, oil recovery from reservoirs, blood flow in arteries) (Murrant and
Sarelius, 2000).
The surface tension, γ, and the mechanical equilibrium at interfaces have been described in the
literature in detail (Adamson and Gast, 1997; Birdi, 1989, 2003a, 2009, 2010a; Chattoraj and Birdi,
1984). The surface has been considered as a hypothetical stretched membrane; this is termed as the
surface tension. In a real system undergoing an infinitesimal process, it can be written as
where
dW is the work done by the systems when a change in volume, dV and dV′, occurs
p and p′ are pressures in the two phases α and ß, respectively, at equilibrium
dA is the change in interfacial area
Tube
Beaker
FIGURE 1.12 Surface of water inside a large beaker and in a narrow tubing. The rise of water in the tubing
is due to the capillary force.
Introduction to Surface and Colloid Chemistry 19
The sign of the interfacial work is designated negative by convention (Adamson and Gast, 1997;
Chattoraj and Birdi, 1984).
The fundamental property of liquid surfaces is that they tend to contract to the smallest possible
area. This is not observed in the case of solids. This property is observed in the spherical form of
small drops of liquid, in the tension exerted by soap films as they tend to become less extended, and
in many other properties of liquid surfaces. In the absence of gravity effects, these curved surfaces
are described by the Laplace equation, which relates the mechanical forces as (Adamson and Gast,
1997; Birdi, 1997; Chattoraj and Birdi, 1984)
⎛1 1⎞
p − pʹ = γ ⎜ + ⎟ (1.11)
⎝ r1 r2 ⎠
⎛γ⎞
p − pʹ = 2 ⎜ ⎟ (1.12)
⎝r⎠
where r1 and r2 are the radii of curvature (in the case of an ellipse), while r is the radius of curvature
for a spherical-shaped interface. It is a geometric fact that surfaces for which Equation 1.12 holds
are surfaces of minimum area. These equations thus give
dW = pd ( V + Vʹ ) − γdA
(1.13)
where
p = p′ for plane surface
Vt is the total volume of the system
It is found that there exists a pressure difference across the curved interfaces of liquids (such as
drops or bubbles). For example, if one dips a tube into water (or any fluid) and applies a suitable
pressure, then a bubble is formed (Figure 1.13). This means that the pressure inside the bubble is
greater than the atmosphere pressure. It thus becomes apparent that curved liquid surfaces induce
effects, which need special physicochemical analyses in comparison to flat liquid surfaces. It must
be noticed that in this system a mechanical force has induced a change on the surface of a liquid.
This phenomenon is also called capillary forces. Then one may ask, does this also require similar
consideration in the case of solids? The answer is yes, and will be discussed later in detail. For
example, in order to remove liquid, which is inside a porous media such as a sponge, one would need
force equivalent to these capillary forces. Man has been fascinated with bubbles for many centu-
ries. As seen in Figure 1.13, the bubble is produced by applying a suitable pressure, ΔP, to obtain a
bubble of radius R, where the surface tension of the liquid is γ.
It is useful to consider the phenomena where one expands the bubble by applying pressure, Pinside.
In this process, one has two processes to consider:
In other words, there are two opposing actions: expansion of volume and increase of surface area.
20 Handbook of Surface and Colloid Chemistry
Liquid
Bubble
FIGURE 1.13 An air bubble in a liquid. The pressure inside the bubble is greater than the pressure outside
the bubble.
The work done can be expressed in terms of that done against the forces of surface tension and
that done in increasing the volume. At equilibrium, there will exist following condition between
these two kinds of work:
and
2γ (1.17)
ΔP =
R
where ΔP = (Pg − Pliquid). Since the free energy of the system at equilibrium is constant, dG = 0, then
these two changes in system are equal. If the same consideration is applied to the soap bubble, then
the expression for ΔP bubble will be
4γ
ΔPbubble = p inside − poutside = (1.18)
R
because now there exist two surfaces and the factor 2 is needed to consider this state.
The pressure applied gives rise to work on the system, and the creation of the bubble gives rise
to the creation of surface area increase in the fluid. The Laplace equation relates the pressure differ-
ence across any curved fluid surface to the curvature, 1/radius, and its surface tension, γ. In those
cases where nonspherical curvatures are present (Birdi et al., 1988; Kendoush, 2007), one obtains
the more universal equation:
⎛ 1 1 ⎞
ΔP = γ ⎜ + ⎟ (1.19)
⎝ R 1 R 2 ⎠
Introduction to Surface and Colloid Chemistry 21
It is also seen that in the case of spherical bubbles, since R1 = R2, this equation becomes identical
to Equation 1.18. It is thus seen that in the case of a liquid drop in air (or gas phase), the Laplace
pressure would be the difference between the pressure inside the drop, pL, and the gas pressure, pG:
p L − pG = ΔP (1.20)
2γ
p L − pG = (1.21)
Radius
In the case of a water drop with radius 2μ, there will be ΔP of magnitude:
2(72 mN/m )
ΔP = = 72,000 N/m 2 = 0.72 atm (1.22)
2 × 10 −6 m
The magnitude of vapor pressure is known to be dependent on the pressure. Thus, ΔP will effect the
vapor pressure and lead to many consequences in different systems. In fact, the capillary (Laplace)
pressure determines many industrial and biological systems. The lung alveoli are dependent on the
radii during the inhale–exhale process, and the change in the surface tension of the fluid lining the
lungs alveoli. In fact, many lung diseases are related to the lack of surface pressure and capillary
pressure balance. Blood flow through arteries of different diameter throughout the body is another
system where Laplace pressure is of much interest for analytical methods (such as heart function
and control).
The Laplace equation is useful for analyses in a variety of systems as described in the following:
Open
Closed
(a) (b)
FIGURE 1.15 (a, b) Equilibrium state of two bubbles of different radii (see text).
It has been observed that a system with varying sized bubbles collapses faster than bubbles
(or liquid drops) that are of exactly the same size. In other words, in systems with bubbles of similar
diameter, there will be a slow coalescence than in systems with varying sized bubbles. Another
major consequence is observed in the oil recovery phenomena (see later). Oil production takes place
(in general) by applying gas or water injection. When gas or water injection is applied, where there
are small pores, the pressure needed will be higher than that in the large pore zone. Thus, the gas
or water will bypass the small pore zone and leave the oil behind (at present more than 30%–50%
of oil in place is not recovered under normal production methods). This is obviously a great chal-
lenge to the surface and colloid chemists in the future. Enhanced oil recovery (EOR) technology is
of enormous interest at this stage in the literature (Birdi, 2003a).
Another aspect is that vapor pressure over a curved liquid surface, pcurve, will be larger than on a
flat surface, pflat. A relation between pressure over curved and flat liquid surface was derived (Kelvin
equation):
⎛p ⎞ ⎛ v L ⎞ ⎛ 2γ ⎞
ln ⎜ curve ⎟=⎜ ⎟⎜ ⎟ (1.23)
⎝ pflat ⎠ ⎝ RT ⎠ ⎝ R curve ⎠
where
pcurve and pflat are the vapor pressures over curved and flat surfaces, respectively
Rcurve is the radius of curvature
vL is the molar volume
Kelvin equation thus suggests that if liquid is present in a porous material, such as cement, then the
difference in vapor pressure exists between two pores of different radii. Infact, in cement industry, one
reduces the value of γ, such that the vapor pressure difference between pores of different radii is reduced.
Similar consequence of vapor pressure exists when two solid crystals of different size are concerned.
The smaller sized crystal will exhibit higher vapor pressure and will also result in faster solubility rate.
Rain drop formation from clouds: In the clouds (water is in a gas phase) where the distance
between water molecules is roughly the same as in the gas phase, the transition from water vapor
in clouds to rain drops (liquid state) is not as straightforward process as might seem. The forma-
tion of a large liquid rain drop requires that a certain number of water molecules in the clouds (as
gas phase) have formed a nuclei (which is the formation of first liquid drop). This is observed in
some cases where there are many clouds, but one does not get any rain as needed for irrigation
(this happens if nuclei are absent). The nuclei or embryo will grow, and Kelvin relation will be the
determining factor. Artificial rain has been attempted by using fine particles (finely divided silver
particles or similar), which leads to nuclei formation and assisting in rain drop formation (Adamson
and Gast, 1997). Pollution (such as dust particles) will also give rise to abnormal rain fall due to the
latter effect on nucleation.
Introduction to Surface and Colloid Chemistry 23
1. The range of these applications is from the blood flow in the veins to the oil recovery in the
reservoir.
2. Fabric’s properties are also governed by capillary forces (i.e., wetting, etc.).
3. The sponge absorbs water or other fluids where the capillary forces push the fluid into
many pores. This is also called wicking process (as in the candle wicks).
The curvature in a system where a narrow capillary circular tube is dipped into a liquid exhibits
properties, which are not observed in a large beaker. The liquid is found to rise in the capillary,
when the fluid wets the capillary (like water and glass or water and metal). The curvature of the
liquid inside the capillary will lead to pressure difference between this state and the relatively flat
surface outside the capillary (Figure 1.16).
The rise or fall of a liquid in a capillary (arising from the capillary force) is dependent on the wet-
ting characteristics. Inside the capillary, the liquid (with surface tension, γ) attains an equilibrium
of capillary forces. However, if the fluid is nonwetting (such as Hg in glass), then one finds that the
fluid falls. This arises from the fact that Hg does not wet the tube. Capillary forces arise from the
difference in attraction of the liquid molecules to each other and the attraction of the liquid molecule
to those of the capillary tube. The fluid rises inside the narrow tube to a height, h, until the surface
tension forces balance the weight of the fluid. This equilibrium gives following relation:
where
γ is the surface tension of the liquid
R is the radius of curvature
Air Air
Liquid Hg
(a) (b)
FIGURE 1.16 Rise or fall of a liquid in a glass capillary: (a) rise of water and (b) fall of Hg in a capillary.
24 Handbook of Surface and Colloid Chemistry
In the case of narrow capillary tubes, less than 0.5 mm, the curvature can be safely set equal to the
radius of the capillary tubing. The fluid will rise inside the tube to compensate for surface tension
force; thus, at equilibrium, we get
γ = 2RρL gh (1.26)
where
ρL is the density of the fluid
g is the acceleration of gravity
h is the rise in the tube
Example
γ 72.8
h= =
2RρL g 2R(1000)9.8
The magnitudes for h for different radii of tubings are the following:
One can measure the magnitude of h very accurately and thus allows one to measure the value of
γ (Adamson and Gast, 1997; Birdi, 2009).
It is found that these assumptions are only precise when the capillary tubing is rather small. In the
case of larger sized capillaries, correction tables are found in the literature (Adamson and Gast,
1997). However, in some particular case where the contact angle, θ, is not zero, one will need a cor-
rection, and equation will become
⎛ 1 ⎞
γ = 2RρL gh ⎜ ⎟ (1.27)
⎝ cos(θ) ⎠
It is seen that when the liquid wets the capillary wall, the magnitude of θ is 0, and cos(0) = 1.
In the case of Hg, the contact angle is 180°, since it is nonwetting fluid (see Figure 1.16b). Since
cos(180) = −1, then the sign of h in equation will be negative. This means that Hg will show a drop
in height in glass tubing. Hence, the rise or fall of a liquid in a tubing will be governed by the sign
of cos(θ).
Thus, capillary forces will play an important factor in all systems where liquids are present
in porous environment. Similar result can also be derived by using the Laplace equation (1.28)
(1/radius = 1/R):
2γ
ΔP = (1.28)
R
Introduction to Surface and Colloid Chemistry 25
Air
Thin liquid
film
Air
The liquid rises to a height h, and the systems achieve equilibrium, and the following relation is found:
2γ
= hggρL (1.29)
R
γ = 2RρL gg h (1.30)
It is found that the various surface forces are responsible for the capillary rise. The lower the surface
tension, the lower is the height of column in the capillary. The magnitude of γ is determined from
the measured value of h for a fluid with known ρL. The magnitude of h can be measured directly by
using a suitable device (e.g., photograph image).
Further, it is known that in real world, capillaries or pores are not always circular shaped. In
fact, one considers that in oil reservoirs (or water seepage), pores are more like triangular or square
shaped than circular. In this case, one can measure the rise in other kinds of shaped capillaries,
such as rectangular or triangular (Birdi, 1997, 2003a; Birdi et al., 1988). These studies have much
importance for oil recovery or water treatment systems. Especially in shale oil/gas technology, one
uses large amounts of water to transport chemicals through pores where curved liquid surfaces are
present (Nicot and Scanlon, 2012). In any system where fluid flows through porous material (e.g.,
seepage of water), one would expect that capillary forces will be one of the most dominant factors.
Further, it is known that the vegetable world is dependent on capillary pressure (and osmotic pres-
sure) to bring water up to the higher parts of plants. In this way, some trees succeed in bringing this
essential liquid (water) up to 120 m above the ground.
One also knows that soap bubbles are extremely thin (1000 times thinner than the diameter of
a hair!) and unstable (Birdi, 1997; Boys, 1959; Li et al., 2014; Lovett, 1994; Scheludko, 1966;
26 Handbook of Surface and Colloid Chemistry
Taylor, 2011). In spite of the latter, under special conditions, one can keep soap bubbles for
long lengths of time, which thus allows one to study its physical properties (such as thickness,
composition, conductivity, spectral reflection, etc.). The thickness of a bubble is in most cases
over hundreds of micrometers in the initial state. The thin liquid film (TLF) consists of bilayer
of detergent that contains the solution. The film thickness decreases with time due to following
reasons:
Therefore, the stability and lifetime of such thin films will be dependent on these different char-
acteristics. This is found from the fact as an air bubble is blown under the surface of a soap or
detergent solution, air bubble will rise up to the surface. It may remain at the surface, if the speed
is slow, or it may escape into the air as a soap bubble. Experiments show that a soap bubble consists
of a very TLF with an iridescent surface. But as the fluid drains away and the thickness decreases,
the latter approaches to the equivalent of barely two surfactant molecules plus a few molecules
of water. It is worth noting that the limiting thickness is on the order of two or more surfactant
molecules. This means that one can see with the naked eye molecular size structures of TLFs. As
the air bubble enters the surface region, the soap molecules along with water molecules are pushed
up and as the bubble is detached, it leaves as a TLF with following characteristics (as found from
various measurements):
A bilayer of soap (approximately 200 Å thick) on the outer region contains the aqueous phase
(Figure 1.17).
The thickness of the initial soap bubble is some micrometers.
The thickness decreases with time and one starts to observe rainbow colors, as the reflected
light is of the same wavelength as the thickness of the bubble (few hundreds of Angstroms).
The thinnest liquid film consists mainly of the bilayer of surface-active substance (SAS)
(such as soap = 50 Å) and some layers of water. The light interference and reflection studies
show many aspects of these TLFs.
The bilayer soap film may be depicted as arrays of the soap molecule (with a few layers of water)
(-----O: length of the soap molecule is about 15 Å):
Thick bubble film (micrometer or more) (shows no colors):
-----OWWWWWO-----
-----OWWWWWO-----
-----OWWWWWO-----
-----OWO-----
-----OWO-----
-----OWO-----
The thickness of the bilayer is thus about 30–50 Å (almost twice the length of the soap molecule).
The thickness is approximately the size of wavelength of light.
The iridescent colors of the soap bubble arise from the interference of reflected light waves.
The reflected light from the outer surface and the inner layer gives rise to this interference effect
(Figure 1.17). The rainbow colors are observed as the bubble thickness decreases (and reaches
magnitudes corresponding to that in the light waves) due to the evaporation of water.
Introduction to Surface and Colloid Chemistry 27
Rainbow colors (violet, indigo, blue, green, yellow, orange, red: VIBGYOR) are observed when
the thickness is on the order of wavelength of light (Adamson and Gast, 1997; Birdi, 2009). The
wavelength increases from violet to red color. One observes rainbow colors in all sorts of situations
where interference of waves of light takes place. Another interesting observation is that in some
natural phenomena, colors are produced by simply structures made similar to the bubble; that is,
thickness (of lipid-like molecules) varies in the range of wavelength of light (Orna, 2013). This is,
for instance, the case in the colors observed in peacock feathers (and many other colored objects in
nature and otherwise).
Soap films that are rather thick (thickness is mainly due to the water) reflect red light and one
observes blue–green colors. Lesser thin films cancel out yellow wavelength and blue color is
observed. As the thickness approaches the wavelength of light, all colors are cancelled out and a
black (or gray) film is observed. This corresponds to 25 nm (250 Å).
The transmitted light, Itr, is related to the incident, Iin, and the reflected intensity, Ire:
I tr = I in − I re (1.31)
As one finds from the analyses of the daylight, it consists of different wavelengths of colors (violet,
indigo, blue, green, yellow, orange, red [VIBGYOR: easy to remember!]):
Red 680 nm
Orange 590 nm
Yellow 580 nm
Green 530 nm
Blue 470 nm
Violet 405 nm
Slightly thicker soap films (ca. 1500 nm) sometimes look golden. In the thinning process, the differ-
ent colors get cut off. Thus, if the blue color gets cut off, the film looks amber to magenta.
The bubble film’s (which consists of SAS + water + salts, etc.) stability can be described as
follows:
Bubble film:
It is thus obvious that the rate of water evaporation plays an important role. The evaporation can
be reduced by containing the bubble in a closed bottle. One also finds that in such closed system,
the bubbles remain stable for a very long time. The drainage of water away from the film is depen-
dent on the viscosity of the fluid. Therefore, additives such as glycerin (or other thickening agents
[polymers]) assist in the stability of the bubble films.
m m g = 2πRγ (1.32)
where
mm is the weight of the detached drop
R is the radius of the tubing
A simple method is to count the number of drops (e.g., 10 or more) and measure the weight (with a
very high accuracy: less than microgram).
One may also use a more convenient method where a fluid is pumped and the drops are collected
and weighed. Since in some systems (solutions) there may be kinetic effects, one must be careful to
keep the flow as slow as possible. This system is very useful in studying systems, which one finds in
daily life phenomena: oil flow in reservoirs, blood cells flow through arteries. In those cases where
the volume of fluid available is limited, one may use this method with advantage. By decreasing
the diameters of the tubing, one can work with less than 1 μL fluids. This may be useful in systems
such as eye fluids, etc.
Liquid
Drop
In the process from initial to the final stage, the liquid molecules that were near the solid surface
have been moved away and are now under the influence of their own molecules. This requires energy
(interfacial energy), and the force required to make this happen is proportional to the surface area of
contact and to the surface tension of the liquid. The methods of determining the surface tension by
measuring the force required to detach a body from a liquid are therefore similar to the stalagmom-
eter method described earlier. However, their advantage over the latter method is that is, possible to
choose the most convenient form and size of the body (platinum rod, ring or plate) so as to enable
the measurement to be carried out rapidly, but without any detriment to its accuracy. The detach-
ment method has found an application in the case of liquid whose surface tension changes with time.
ds
Ss = (1.33)
de
where
de is the equatorial diameter
ds is the diameter at a distance de from the tip of the drop, Figure 1.19
ds
de
de
ρL gd 2e
γ= (1.34)
H
where ρL is the density of the liquid and H is related to Ss, but the values of 1/H for varying Ss were
obtained from experimental data. For example, when Ss = 0.3, 1/H = 7.09837, while when Ss = 0.6,
1/H = 1.20399. The magnitude of γ is estimated from accurate mathematical functions used to
estimate 1/H for a given de value (Adamson and Gast, 1997; Birdi, 2003a). The accuracy (0.1%) of
γ is satisfactory for most of the systems especially when experiments are carried out under extreme
conditions (such as high temperatures and pressures).
The pendant drop method is very useful under specific conditions:
1. Technically, only a drop (a few microliters) is required. For example, eye fluid can be stud-
ied, since only a drop of microliter is needed.
2. It can be used under very extreme conditions (very high temperature or corrosive fluids).
3. Under very high pressure and temperatures. Oil reservoirs are found typically at 100°C and
300 atm pressure. Surface tension of such systems can be conveniently studied by using
high pressure and temperature cells with optical clear windows (sapphire windows:1 cm
thick; up to 2000 atm). For example, γ of inorganic salts at high temperatures (ca. 1000°C)
can be measured by using this method. The variation of surface tension can be studied as
a function of various parameters (temperature and pressure, additives [gas, etc.]).
where
Wtotal is the total weight of the ring
Wring is the weight of the ring in air
R ring is the radius of the ring
The circumference is 2πR ring, and factor 2 is because of the two sides of contact.
This relation assumes that the contact between the fluid and the ring is geometrically simple. It is
also found that this relation is fairly correct (better than 1%) for most working situations. However,
it was observed that Equation 1.36 needed correction factor, in much the same way as was done for
the drop-weight method. Here, however, there is one additional variable so that the correction factor
f now depends on two dimensionless ratios (Adamson and Gast, 1997).
Introduction to Surface and Colloid Chemistry 31
Perimeter of a plate is the sum of twice the length + breadth. The surface force will act along the
perimeter of the plate (i.e., length [Lp] + width [Wp]). The plate is often very thin (less than 0.1 mm)
and made of platinum, but even plates made of glass, quartz, mica, and filter paper can be used. The
forces acting on the plate consist of the gravity and surface tension force (downward), and buoyancy
due to displaced water upward. For a rectangular plate of dimensions Lp and Wp and of material
density ρ, immersed to a depth hp in a liquid of density ρL, the net downward force, F, is given by the
following equation (i.e., weight of plate + surface force (γ) × perimeter of the plate–upward drift):
where
γ is the liquid surface tension
θ is the contact angle of the liquid on the solid plate
g is the gravitational constant
The weight of plate is constant and can be tared. If the plate used is very thin (i.e., tp ≪ Wp) and the
up-drift is negligible (i.e., hp is almost zero), then γ is found from
F
γ= (1.39)
2Wp
Plate
Liquid
FIGURE 1.20 Wilhelmy plate in a liquid (plate with dimensions: length = Lp, width = Wp).
32 Handbook of Surface and Colloid Chemistry
The sensitivity of γ by using these procedures has been found to be very high (±0.001 dyn/cm
[mN/m]) (Birdi, 2009). The change in surface tension (surface pressure = Π) is then determined by
measuring the change in F for a stationary plate between a clean surface and the same surface with
a monolayer present. If the plate is completely wetted by the liquid (i.e., cos(θ) = cos(0) = 1), the
surface pressure is then obtained from the following equation:
⎡ ΔF ⎤ −ΔF
Π = −⎢ ⎥= , if Wp t p (1.40)
⎢⎣ 2 ( t p + Wp ) ⎥⎦ 2Wp
In general, by using very thin plates with thickness 0.1–0.002 mm, one can measure surface ten-
sion with very high sensitivity. The apparatus is calibrated using pure liquids, such as water and
ethanol. The buoyancy correction is made very small (and negligible) by using a very thin plate
and dipping the plate as little as possible. The wetting of water on platinum plate is achieved by
using commercially available platinum plates that are roughened to increase wettability. The latter
property gives rise to almost complete wetting, that is, θ = 0. The force is in this way determined by
measuring the changes in the mass of the plate, which is directly coupled to a sensitive electrobal-
ance or some other suitable device (such as pressure transducer, etc.). It must be noticed that in all
systems where an object is in contact with a liquid there will exist a force toward the liquid (related
to surface tension).
n-Heptane 20.14
n-Hexadecane 27.47
n-Hexane 18.43
n-Octane 21.62
Alcohols: The magnitude of γ changes by 23.7 − 22.1 = 1.6 mN/m per –CH2– group. This is
based upon the γ data of ethanol (22.1 mN/m) and propanol (23.7 mN/m).
These observations indicate the molecular correlation between bulk forces and surface forces
(tension) (γ) for homologous series of substances.
Introduction to Surface and Colloid Chemistry 33
TABLE 1.2
Surface Tension Values of Some Common Liquids
Liquid Surface Tension (20°C; mN/m [dyn/cm])
1,2-Dichloro-ethane 33.3
1,2,3-Tribromo propane 45.4
1,3,5-Trimethylbenzene (mesitylene) 28.8
1,4-Dioxane 33.0
1,5-Pentanediol 43.3
1-Chlorobutane 23.1
1-Decanol 28.5
1-Nitro propane 29.4
1-Octanol 27.6
Acetone (2-propanone) 25.2
Aniline 43.4
2-Aminoethanol 48.9
Anthranilic acid ethylester 39.3
Anthranilic acid methylester 43.7
Benzene 28.9
Benzylalcohol 39.0
Benzylbenzoate (BNBZ) 45.9
Bromobenzene 36.5
Bromoform 41.5
Butyronitrile 28.1
Carbon disulfide 32.3
Quinoline 43.1
Chlorobenzene 33.6
Chloroform 27.5
Cyclohexane 24.9
Cyclohexanol 34.4
Cyclopentanol 32.7
p-Cymene 28.1
Decalin 31.5
Dichloromethane 26.5
Diiodomethane (DI) 50.8
1,3-Diiodopropane 46.5
Diethylene glycol 44.8
Dipropylene glycol 33.9
Dipropylene glycol monomethylether 28.4
Dodecyl benzene 30.7
Ethanol 22.1
Ethylbenzene 29.2
Ethylbromide 24.2
Ethylene glycol 47.7
Formamide 58.2
Fumaric acid diethylester 31.4
Furfural (2-furaldehyde) 41.9
Glycerol 64.0
Ethylene glycol monoethyl ether (ethyl cellosolve) 28.6
Hexachlorobutadiene 36.0
Iodobenzene 39.7
(Continued)
34 Handbook of Surface and Colloid Chemistry
dG = γdA (1.41)
⎛ dG ⎞
GS = γ = ⎜ ⎟ (1.42)
⎝ dA ⎠T,P
⎛ dG ⎞
SS = − ⎜ S ⎟ (1.43)
⎝ dT ⎠P
⎛ dγ ⎞
= −⎜ ⎟ (1.44)
⎝ dT ⎠
36 Handbook of Surface and Colloid Chemistry
H S = G S + TSS (1.45)
All natural processes are dependent on the effect of temperature and pressure. For instance, oil res-
ervoirs are found under high temperatures (ca. 80°C) and pressure (around 100–400 atm [depend-
ing on the depth]). The thermodynamics of the system is defined by the fundamental equation for
the free energy, G, and to the enthalpy, H, and entropy, S, of the system:
G = H − TS (1.46)
It is thus important that in all practical analyses, one should be aware of the effects of temperature
and pressure. The molecular forces that stabilize liquids will be expected to decrease as temperature
increases. Experiments also show that in all cases, surface tension decreases with increasing tem-
perature. Surface entropy of liquids is given by (−dγ/dT). This means that the entropy is positive at
higher temperatures (because the magnitude of γ always decreases with temperature for all liquids).
The rate of decrease of surface tension with temperature is found to be different for different liquids,
which supports the aforementioned description of liquids (Jasper, 1972).
For example, the surface tension of water data is given as
At 5°C, γ = 75 mN/m
At 25°C, γ = 72 mN/m
At 90°C, γ = 60 mN/m
Extensive γ data for water were fitted to the following equation (Birdi, 2003a; Cini et al., 1972):
where tC is in °C. This equation gives the value of γ at 0°C as 75.69 mN/m. The value of γ at 50°C is
found to be (75.69 − 0.1413 × 50 − 0.0002985 × 502) = 67.88 mN/m, and at 25°C, it is 71.97 mN/m.
In the literature, one finds such relationships for other liquids, which allows one to calculate the
magnitude of γ at different temperature. This kind of application is important in oil and other
industrial phenomena.
Further, these data show that γ of water decreases with temperature from 25°C to 60°C,
(72 − 60)/(90 − 25) = 0.19 mN/m °C. This is as one should expect from physicochemical theory.
The surface tension change with temperature analyses thus provides a very sensitive information
as regards the molecular forces in any liquid (surface entropy). In all systems, as the temperature
increases, the energy between molecules gets weaker and thus the surface energy, that is, surface
tension, will decrease. The difference in the surface entropy gives information on the structures of
different liquids. It is also observed that the effect of temperature on γ will be lower on liquids with
higher boiling point (such as Hg) than for low boiling liquids (such as n-hexane). Actually, there
exists a correlation between γ and heat of vaporization (or boiling point). In fact, many systems even
show big differences (due to effects of temperature on γ) when comparing winter or summer months
(such as rain drops, sea waves, foaming in natural environments). Different thermodynamic rela-
tions have been derived, which can be used to estimate the surface tension at different temperatures.
Especially, straight chain alkanes have been extensively analyzed. The data show that a simple cor-
relation between surface tension, temperature, and nC exists. This allows one to estimate the value of
γ at any temperature of a given alkane. This observation has many aspects in the applied industry.
It allows one to estimate the magnitude of surface tension of an alkane at the required temperature.
Further, one has a fairly good quantitative analyses about how surface tension will change for a
given alkane under a given experimental condition (e.g., in oil industry).
Introduction to Surface and Colloid Chemistry 37
TABLE 1.3
Ratio of Enthalpy of Surface Formation, hs, and the
Enthalpies of Vaporization, hvap (See Text for Details)
Molecules (Liquid) hs/hvap
Hg 0.64
N2 0.51
O2 0.5
CCl4 0.45
C6H6 0.44
Diethyl ether 0.42
Cl–C6H5 0.42
Methyl formate 0.40
Ethyl acetate 0.4
Acetic acid 0.34
H2O 0.28
C2H5OH 0.19
CH3OH 0.16
38 Handbook of Surface and Colloid Chemistry
Stefans law is an indication that the surface molecules are oriented differently than in the bulk
phase. This observation is useful in order to understand the surface phenomena.
As an example, one may proceed with this theory and estimate the surface tension of a liquid
with data on its heat of evaporation. The number of near neighbors of a surface molecule will be
about half (6 = 12 [near neighbors in the bulk phase]/2) than those in the bulk phase (12 neighbors).
It is now possible to estimate the ratio of the attractive energies in the bulk and in the surface, per
molecule. We have the following data for a liquid such as CCl4:
31,522 J mol −1
Energy change per molecule = = 5.23 × 10 −2 0 J (1.50)
6.023 × 1023 mol −1
If we assume that about half of energy is gained when a molecule is transferred to the surface, then
we get
Energy per molecule at surface = 5.23 × 10 −20 (2) J = 2.6 × 10 −20 J (1.51)
The molecules at the surface of CCl4 occupy a certain value of area, which can be estimated only
roughly as follows.
The measured value of γ of CCl4 is 27 mN/m (Table 1.2). The large difference can be ascribed to
the assumption that a Stefan ratio of 2 was used in this example. As expected, the simple ratio with
factor 2 may vary for nonspherical molecules (as in the case of CCl4). Under these assumptions, one
may conclude that the estimated value of γ is an acceptable description of the surface molecules.
This example is useful for basic considerations about the molecular interactions at the surfaces.
energy being transposed to waves. Hence, mankind has tried to convert wave energy to other useful
forms of energy source. Both transverse capillary waves and longitudinal waves can deliver infor-
mation about the elasticity and viscosity of surfaces, albeit on very different timescales. Rates of
adsorption and desorption can also be deduced. Transverse capillary waves are usually generated
with frequencies between 100 and 300 Hz. The generator is a hydrophobic knife-edge situated in the
surface and oscillating vertically, while the usual detector is a lightweight hydrophobic wire lying
in the surface parallel to the generator edge. The generator and detector are usually close together
(15–20 mm), so reflections set up a pattern of standing transverse wave. Optical detection, on the
other hand, causes no interference to the generated pattern. The damping of capillary ripples arises
primarily from the compression and expansion of the surface and the interaction between surface
film and water phase (Adamson and Gast, 1997; Birdi, 1997). This leads to compression and expan-
sion of the surface. If a surface film is present, compression tends to lower the surface tension, while
expansion raises it. This generates a Marangoni flow, which opposes the wave motion and dampens
it (Birdi, 2003a, 2007, 2010a). Furthermore, if the material is soluble, the compression–expansion
cycle will be accompanied with the hydrodynamic characteristics of capillary ripples.
γ AB = γ A ( B) − γ B( A ) (1.52)
The prediction of γAB from this rule is approximate but found to be useful in a large number of
systems (such as alkanes: water), with some exceptions (such as water: butanol) (Tables 1.3 and 1.4).
For example,
However, for general considerations, one may only use it as a reliable guideline and when exact data
are not available. The Antonow rule can be understood in terms of a simple physical picture. There
should be an adsorbed film or Gibbs monolayer of substance B (the one of lower surface tension) on
the surface of liquid A. If we regard this film as having the properties of bulk liquid B, then γA(B) is
effectively the IFT of a duplex surface and would be equal to [γA(B) + γB(A)].
Measurement of IFT (between two immiscible liquids):
IFT can be measured by different methods, depending on the characteristics of the system.
Following methods can be applied:
TABLE 1.4
Antonow’s Rule and Interfacial Tension Data (mN/m) (See Text for Details)
Oil Phase w(o) o(w) o/w w(o)–o
Benzene 62 28 34 34
Chloroform 52 27 23 24
Ether 27 17 8 9
Toluene 64 28 36 36
n-Propylbenzene 68 29 39 40
n-Butylbenzene 69 29 41 40
Nitrobenzene 68 43 25 25
i-Pentanol 28 25 5 3
n-Heptanol 29 27 8 2
CS2 72 52 41 20
Methylene iodide 72 51 46 22
TABLE 1.5
IFT between Water and Organic Liquids (20°C)
Water/Organic Liquid IFT (mN/m)
n-Hexane 51.0
n-Octane 50.8
CS2 48.0
CCl4 45.1
Br–C6H5 38.1
C6H6 35.0
NO2–C6H5 26.0
Ethyl ether 10.7
n-Decanol 10
n-Octanol 8.5
n-Hexanol 6.8
Aniline 5.9
n-Pentanol 4.4
Ethyl acetate 2.9
Isobutanol 2.1
n-Butanol 1.6
The Wilhelmy plate is placed at the surface of water, and oil phase is added until the latter covers
the whole plate. The apparatus must be calibrated with known IFT data, such as water–hexadecane
(52 mN/m; 25°C) (Table 1.5).
The drop-weight method is carried out by using a pump (or a syringe) to deliver liquid phase into
the oil phase (or vice versa, as one finds suitable). In the case of water, water drops sink to the bot-
tom of the oil phase. The weight of the drops is measured (by using an electrobalance), and IFT can
be calculated. The accuracy can be very high by choosing the right kind of setup. The drop shape
(pendant drop) is most convenient if small amounts of fluids are available and if extreme tempera-
ture and pressures are involved. Modern digital image analyses also make this method very easy to
apply in extreme situations.
Introduction to Surface and Colloid Chemistry 41
γ t = k o (1 − k1t ) (1.54)
where ko is a constant. It was found that coefficient k1 is approximately equal to the rate of decrease
of density, (ρ), of liquids with the rise of temperature:
ρt = ρo (1 − k1t ) (1.55)
where ρo is the value of density at t = 0°C, and values of constant k1 were found to be different for
different liquids. The magnitude of γ decreases with temperature and vanishes at the critical point
(Tc and Pc). Experiments show that if one heats a liquid in a losed tuning, the liquid expands with
higher T and P, until at the critical point, the phase boundary between the liquid and the gas disap-
pears. At the critical point, the liquid and the gas have the same density. The following equation
relates surface tension of a liquid to the density of liquid, ρl, and vapor, ρv (Birdi, 1989, 2010a,b;
Partington, 1951):
γ
= C (1.56)
(ρl − ρv )4
where the value of constant C is nonvariable only for organic liquids, while it is not constant for liq-
uid metals. At the critical temperature, Tc, and critical pressure, Pc, a liquid and its vapor are identi-
cal, and the surface tension, γ, and total surface energy, like the energy of vaporization, must be zero
(Birdi, 1997). At temperatures below the boiling point, which is 2/3Tc, the total surface energy and
the energy of evaporation are nearly constant. The variation in surface tension, γ, with temperature
is given in Figure 1.21 for different liquids.
These data clearly show that the variation of γ with temperature is a very characteristic physical
property. This observation becomes even more important when it is considered that the sensitivity
of γ measurements can be as high as ±0.001 dyn/cm (=mN/m).
The change in γ with temperature in the case of mixtures would thus be dependent on the
composition. For example, the variation of γ of the system: CH4 + hexane is given as follows:
This relation shows that addition of hexane to CH4 increases the magnitude of γ of the mixture. In
fact, one can estimate the concentration of hexane in CH4 by using Equation 1.57. This has much
interest in oil reservoir engineering operations where one finds CH4 in the crude oil. Experiments
show that addition of a gas to a liquid will always decrease the value of γ of the mixture.
42 Handbook of Surface and Colloid Chemistry
Surface tension
20
C10
C5
10
50 100
Temperature (°C)
FIGURE 1.21 Variation of γ with temperature of different alkanes (C5: n-pentane; C10: n-decane).
It is well known that the corresponding states theory can provide much useful information
about the thermodynamics and transport properties of fluids. For example, the most useful two-
parameter empirical expression, which relates the surface tension, γ, to the critical temperature,
Tc, is given as
k1
⎛ 1− T ⎞
γ = ko ⎜ ⎟ (1.58)
⎝ Tc ⎠
where ko and k1 are constants. The magnitude of constant k1 has been reported as follows:
=3/2, although experiments indicated that k1 = 1.23 or 11/9 (Birdi, 1997, 2003a). Data for some
liquids have also reported the value of k1 to be between 6/5 and 5/4.
It was found that constant ko was proportional to Tc1/ 3Pc2 / 3. The relation in Equation 1.58 when
fitted to the surface tension, γ, data of liquid CH4, has been found to give the following correlation:
1.287 1.287
⎛ 1− T ⎞ ⎛ 1− T ⎞
γ CH4 = 40.52 ⎜ ⎟ = 40.52 ⎜ ⎟ (1.59)
⎝ Tc ⎠ ⎝ 190.55 ⎠
where Tc = 190.55 K. This equation has been found to fit the γ data for liquid methane from 91
to −190 K, with an accuracy of ±0.5 mN/m. In a recent study, the γ versus T data on n-alkanes,
from n–pentane to n-hexadecane, were analyzed (Birdi, 1997, 2010a). The constants ko (between 52
and 58) and k1 (between 1.2 and 1.5) were found to be dependent on the number of carbon atoms,
nC, and since Tc is also found to be dependent on nC (Birdi, 1997). The estimated values of differ-
ent n-alkanes were found to agree with the measured data within a few percent: γ for n-C18H38, at
100°C, was 21.6 mN/m, from both measured and calculated values. This agreement shows that the
surface tension data on n-alkanes fits the corresponding state equation very satisfactorily. It is worth
mentioning that the equation for the data on γ versus T, for polar (and associating) molecules like
water and alcohols, gives magnitudes of ko and k1 that are significantly different than those found
for nonpolar molecules such as alkanes, etc.
In the following, calculated values of γ for different alkanes are given based upon the analyses
using these equations (Birdi, 1997).
Introduction to Surface and Colloid Chemistry 43
Comparison of calculated and measured values of surface tension (γ) (Birdi, 1997)
−dγ
SS = (1.60)
dT
k1
⎛ 1− T ⎞ 1
= k 1k o ⎜ ⎟ − T (1.61)
T
⎝ c ⎠ c
k1 −1
⎛ dγ ⎞ ⎛ 1− T ⎞ ⎛ 1 + ( k1 − 1)T ⎞
H s = G s − TSs = −T ⎜ ⎟ = ko ⎜ ⎟ ⎜ ⎟ (1.62)
⎝ dT ⎠ ⎝ Tc ⎠ ⎝ Tc ⎠
The reason heat is absorbed on the expansion of a surface is that the molecules must be transferred
from the interior against the inward attractive force to form the new surface. In this process, the
motion of the molecules is retarded by this inward attraction so that temperature of the surface lay-
ers is lower than that of the interior, unless heat is supplied from outside.
The following relationship relates γ to density (Birdi, 1997, 2003a):
2 /3
⎛M⎞
γ⎜ ⎟ = k ( Tc − T − 6 ) (1.63)
⎝ ρ⎠
where
M is the molecular weight
ρ is the density (M/ρ = molar volume)
The quantity (γ(M/ρ)2/3) is called the molecular surface energy. It is important to notice the cor-
rection term 6 on the right-hand side. This is the same as found for n-alkanes and n-alkenes in the
estimation of Tc from γ versus temperature data (Birdi, 1997; Jasper, 1972). Surface tension and
temperature relationships of alkanes are useful in oil industry.
44 Handbook of Surface and Colloid Chemistry
KCl gives a larger increase in γ than HCl (per mole of concentration). This indicates that the degree
of adsorption of KCl at the surface of water is larger than that of HCl. Further, it is found from
experiments that the aqueous solutions of salts exhibit surface charges (Birdi, 2010a; Chattoraj and
Birdi, 1984). This arises from the fact that the number of positive ions and negative ions is not equal
(contrary to the bulk phase!). In other words, it is found that the surface potential of KCl solution is
lesser than in the case of HCl. From this, it is concluded that K+ and Cl− ions adsorb at the surface
almost to the same degree.
Surface tension
5.0
80
NaCl
70
Soap/detergent
0
0 0.1
Mol/L
FIGURE 1.22 Change in surface tension of water on addition of inorganic salt (e.g., NaCl or KCl) or a soap.
Introduction to Surface and Colloid Chemistry 45
It is thus seen that even very small differences in surface tension of these solutions indicate some
significant characteristics that are unique for the system. These phenomena have very significant
consequences in systems where these ions are present in connection with surfaces. One example
is that in biological cells, the role of Na+ and K+ (and other cations) is significantly different (Birdi,
2010a,b, 2014).
Change in γ with the addition of solute (equal molar concentration):
Some typical surface tension data of different solutions are given in the following:
This shows that to reduce the value of γ of water from 72 to 30 mN/m, one would need 0.005
mol of SDS or 40% ethanol. There are special substances, which are called soaps or detergents
or surfactants, which exhibit unique physicochemical properties (Birdi, 2009, 2010a; Romsted,
2014; Rosen, 2004). The most significant structure of these molecules is due to the presence of a
hydrophobic (alkyl group) and a hydrophilic (polar group such as –OH; –CH2CH2O −; –COONa;
–SO3Na; –SO4Na; –CH33N–; etc.]).
The different polar groups are as follows:
Ionic groups
–COONa, SO3Na, –SO4Na (negatively charged—anionic)
–(N)(CH3)4Br (positively charged—cationic)
–(N)(CH3)2–CH2–COONa (amphoteric)
Nonionic groups
–CH2CH2OCH2CH2OCH2CH2OH
–(CH2CH2OCH2CH2O)x(CH2CH2CH2O)yOH
Accordingly, one also calls these substances amphiphile (meaning: two kinds) (i.e., alkyl part and
the polar group):
CCCCCCCCCCCCCCCCCCCC–O
Alkyl group (CCCCCCC–)–Polar group (–O)=
Amphiphile molecule
(CH3CH2CH2CH2CH2CH2CH2CH2CH2)–Polar
For instance, surfactants dissolve in water and give rise to low surface tension (even at very low
concentrations [few grams per liter or 1–100 mmol/L]) of the solution; therefore, these substances
are called surface-active molecules (surface-active agents or substances). On the other hand, most
inorganic salts increase the surface tension of water. All surfactant molecules are amphiphilic,
which means these molecules exhibit hydrophilic and hydrophobic properties. Ethanol reduces sur-
face tension of water, but one will need over few moles per liter to obtain the same reduction in γ as
when using a few millimoles of surface-active agents.
46 Handbook of Surface and Colloid Chemistry
As expected, if one adds ethanol (with γ of 22 mN/m) to water (with γ of 72 mN/m), then the
magnitude of γ in the mixed solution will decrease. This is always the case in mixtures. Further,
if one dissolves a gas in a liquid, then the magnitude of γ of the mixture will always decrease.
In fact, one can estimate the amount of gas dissolved by measuring the decrease in γ in such
mixtures.
However, the value of γ of surfactant solutions decreases to 30 mN/m with surfactant concentra-
tion around mmol/L (range of 1–10 g/L). Soaps have been used by mankind for many centuries. In
biology, one finds a whole range of natural amphiphile molecules (bile salts, fatty acids, cholesterol
and other related molecules, phospholipids). In fact, many important biological structures and func-
tions are based on amphiphile molecules.
Moreover, one finds that many surfactants as found in nature (such as bile acids in the stom-
ach) behave exactly the same way as the man-made surface-active agent. Proteins (which are large
molecules [with molecular weights varying from six thousand to millions]) also decrease γ when
dissolved in water (Chattoraj and Birdi, 1984; Tanford, 1980). It is found that aqueous protein solu-
tions exhibit low surface tension values. The decrease in surface tension is related to the amino acid
composition of the protein molecule (Chattoraj and Birdi, 1984; Tanford, 1980).
Soaps and surfactants are molecules, which are characterized as amphiphiles. Amphiphile is
a Greek word, which means likes both kinds. A part of the amphiphile likes oil or hydrophobic
(lipophilic = likes fat) (Tanford, 1980), while the other part likes water or hydrophilic (also called
lipophobic). The balance between these two parts, hydrophilic–lipophilic, is called hydrophilic–
lipophilic balance (HLB). The latter quantity can be estimated by experimental means, and theo-
retical analyses allow one to estimate its value (Adamson and Gast, 1997; Birdi, 2009). HLB values
are applied in the emulsion industry (Birdi, 2009; Hansen, 2007).
Soap molecules are made by reacting fats with strong alkaline solutions (this process is called
saponification). In water solution, the soap molecule, CnH2n+1COONa (with n greater than 12–22),
dissociates at high pH into RCOO − and Na+ ions. It is found that the magnitude of n must be 12 or
more for effective results.
A great variety of surfactants were synthesized from oil by-products, especially, C12H25C6H4SO3Na,
sodium dodecyl benzene sulfonates, were used in detergents. Later, these were replaced by sodium
dodecyl sulfates (SDS) or sulfonates, because sodium dodecyl benzene was found to be not biode-
gradable (Rosen, 2004). This means that bacteria in the sewage plants were not able to degrade the
alkyl group. The alkyl sulfonates were degraded to alkyl hydroxyls and alkyl aldehydes and later
to CO2, etc.
In many applications, one found it necessary to employ surfactants that were nonionic. For
example, nonionic detergents as used in washing clothes are much different in structure and proper-
ties than those used in dish-washing machines. In washing machines, foam is crucial as it helps in
keeping the dirt away from clothes once it has been removed. On the other hand, in machine dish-
washing, one does not need any foam. This arises from the fact that foam will hinder mechanical
effect of dish-washing process.
o bvious that in all those systems where surface tension plays an important role, the additives will
play an important role in these systems.
The magnitude of γ changes slowly in the case of methanol as compared to detergent solutions.
The methanol–water mixtures gave following γ data (20°C):
It is important to have an understanding about the change in the surface tension, γ, of water as
a function of molecular structure of solute. The surface tension data in the case of homologous
series of alcohols and acids show some simple relation to the alkyl chain length. It is found
that each addition of –CH2 – group gives a value of concentration and surface tension such that
the value of concentration is lower by about factor 3. However, it must be mentioned that such
dependence in the case of nonlinear alkyl chains will be different. The effective –CH 2 – increase
in the case of nonlinear chain will be lesser (ca. 50%) than in the case of a linear alkyl chain.
The tertiary –CH 2 – group effect would be even lesser. In general, though, one will expect that
the change in γ per mole substance will increase with any increase in the hydrocarbon group
of the amphiphile. The effect of chain length on surface tension arises from the fact that as the
hydrophobicity increases with each –CH 2 – group, the amphiphile molecule adsorbs more at the
surface. This will thus be a general trend also in more complicated molecules, such as in proteins
and other polymers.
In the case of proteins, the amphiphilic property arises from the different kinds of amino acids
(25 different amino acids) (Chattoraj and Birdi, 1984; Tanford, 1980). Some amino acids have lipo-
philic groups (such as phenylalanine, valine, leucine, etc.), while others have hydrophilic groups
(such as glycine, aspartic acid, etc.).
In fact, one finds from surface tension measurements that some proteins are considerably more
hydrophobic (such as hemoglobin) than others (such as bovine serum albumin, ovalbumin). These
properties of proteins have been extensively investigated, and these data have been found to be
related to biological functions (Birdi, 1999; Chattoraj and Birdi, 1984; Tanford, 1980).
Air–water
Oil–water
Solid–water
The magnitude of surface tension is reduced, since the hydrophobic (alkyl chain or group) is ener-
getically more attracted to the surface than being surrounded by water molecules inside the bulk
aqueous phase. Figure 1.23 shows the monolayer formation of the SAS at high bulk concentration.
Since the close-packed SAS at the surface looks like alkane, it would be thus expected that the
48 Handbook of Surface and Colloid Chemistry
Surface
Solution
FIGURE 1.23 Orientation of soap (SAS) at the surface of water (alkyl group: , polar group: ).
surface tension of SAS solution would decrease from 72 mN/m (surface tension of pure water) to
alkane-like surface tension (close to 25 mN/m).
The orientation of surface molecule at the interface will be dependent on the system. This is
shown as follows:
Air–water: Polar part toward water and hydrocarbon part toward air
Oil–water: Polar part toward water and hydrocarbon part toward oil
Solid–water: Polar part toward water and hydrocarbon part toward solid (in general)
Solubility of SDS
Solubility
curve
CMC
curve
Krafft
point
Temperature (°C)
FIGURE 1.24 Solubility (KP) of ionic (anionic or cationic) surfactants in water (as a function of temperature).
Temperature
50
10
Sodium alkyl
sulfates (number of C atoms)
anionic detergent. Therefore, the effect of various parameters on the KP needs to be considered in
the case of ionic surfactants. Some of these are given as follows:
Temperature
Two
phase
Cloud
point
LC
0 1
Water-nonionic
(weight fraction)
FIGURE 1.26 Solubility of a nonionic surfactant in water (CP) (schematic) (dependent on temperature).
Introduction to Surface and Colloid Chemistry 51
surfactants strongly adsorb at the surface and leave a smooth surface. These detergents are used in
the so-called hair conditioner formulations. This is because the (positive) charged end is oriented
toward the hair (negative) surface and the alkyl group is pointing away (as depicted in the following).
Cationic detergent (positive charge) + hair (negative charge):
This conditioner process is based on the interaction of a positively charged molecule (cationic deter-
gent) with a negatively charged substance (hair). This imparts hydrophobicity to hair and feels
soft and looks fluffy. Fabric softener is based on the same principle (cotton is generally negatively
charged).
According to the principles of physical chemistry science, the micelle formation is a very intriguing
system. One finds a great number of publication in the literature that describes the accurate state of
phases in such system (Bai and Lodge, 2010; Birdi, 2003a, 2010a; Birdi and Ben-Naim, 1981; Chen
and Ruckenstein, 2014; Danov et al., 2014; Golub and de Keizer, 2004; Raju et al., 2001; Rathman
and Scamehorn, 1988; Tanford, 1980; Zana, 1995). Below CMC, detergent molecules are present
as single monomers. Above CMC, one will have that monomers, Cmono, are in equilibrium with
micelles, Cmice. The physical chemistry of such an equilibrium is found to be of great interest in the
literature. The micelle with aggregation number, Nag, is formed from monomers:
Nag monomers which were surrounded by water aggregate together, above CMC, and form a
micelle. In this process, the alkyl chains have transferred from water phase to a alkane-like micelle
interior. This occurs because the alkyl part is at a lower energy in micelle than in the water phase.
The aggregation process is stepwise (i.e., monomer to dimer to trimer to tetramer and so on). This
is based on the fact that some surfactants, such as cholates, form micelles with few aggregation
numbers (between 5 and 20), while SDS can form micelles varying from 100 to 1000 aggregation
numbers. The micelle formation takes place when the alkyl chain of the molecule as surrounded
by water (above the CMC) is transferred to a micelle phase (where alkyl chain is in contact with
52 Handbook of Surface and Colloid Chemistry
neighboring alkyl chains). Thus, in the latter case, the repulsion between alkyl chain and water has
been removed. Instead, the alkyl chain–alkyl chain attraction (van der Waals forces) is the driving
force for the micelle formation. The surfactant molecule forms a micellar aggregate at a concentra-
tion higher than CMC, because it moves from water phase to micelle phase (lower energy). The
micelle reaches an equilibrium after a certain number of monomers have formed a micelle. This
means that there are both attractive and opposing forces involved in this process. Otherwise, one
would expect very large aggregates if there was only attractive force involved. This would mean
phase separation, that is, two phases: one water-rich phase and another surfactant-rich phase. Thus,
aggregation is a specific property where instead of phase separation molecules are able to form
small aggregates, micelles (not visible to naked eye), and very stable micellar solutions.
Thus, one can write the standard free energy of a micelle formation, G omice , as follows:
If there are only attractive forces present (as in alkanes), then one will not observe any significant
solubility in water. The attractive forces are associated with the hydrophobic interactions between
the alkyl part (alkyl–alkyl chain attraction) of the surfactant molecule, G ohydrophobic . The opposing
forces arise from the polar part (charge–charge repulsion, polar group–hydration), G opolar . These
forces are of opposite signs. The attractive forces would lead to larger aggregates. The opposing
forces would hinder the aggregation. A micelle with a definite aggregation number is where the
value of G omice is zero (Figure 1.27). Hence, we can write for G omice:
⎛C ⎞
ΔG omice = μomice − μomono = RT ln ⎜ mice ⎟ (1.67)
C
⎝ mono ⎠
Minimum at CMC
FIGURE 1.27 Attraction forces between alkyl chains and repulsion forces between polar groups give a
minimum energy in the system at CMC.
Introduction to Surface and Colloid Chemistry 53
This relation is valid for nonionic surfactants. In the case of ionic surfactants, one needs to modify
this equation. This equilibrium shows that if we dilute the system, then micelles will break down
to monomers to achieve equilibrium. This is a simple equilibrium for a nonionic surfactant. In the
case of ionic surfactants, there will be different charged species present in the solution. In the case
of ionic surfactant, such as SDS (with sodium S+, and alkyl sulfate, SD −, ions), the micelle with
aggregation number, N SD−, will consist of counter-ions (sodium ions), CS+:
NSD− ionic surfactant monomers + CS+ counter = micelle with charge (NSD− − CS+) (1.69)
Since N SD− is found to be larger than CS+ , all anionic surfactants are negatively charged. Similarly,
cationic micelles will be positively charged. For instance, CTAB, we have following equilibria in
micellar solutions:
CTAB dissociates into CTA+ and Br− ions.
The micelle with N CTA+ monomers will have CBr − counter-ions. The positive charge of the micelle
will be the sum of positive and negative ions (N CTA+ − CBr −). The actual concentration will vary for
each species with the total detergent concentration (e.g., SDS, Figure 1.28). The change in surface
tension, γ, versus concentration is given in Figure 1.29b.
The surface tension curve, Figure 1.29b, is typical for all kinds of detergents. Below CMC, the
SDS molecules in water are found to dissociate into SD− and Na+ ions. Conductivity measurements
show the following data:
1. SDS behaves as a strong salt, and SD− and Na+ ions are formed (same as one observes for
NaCl).
2. A break in the plot is observed at SDS concentration equal to the CMC. This clearly shows
that the number of ions decreases with concentration. The latter indicates that some ions
(in the present case cations, Na+) are partially bound to the SDS-micelles, which results in
change in the slope of the conductivity of the solution. Same behavior is observed for other
ionic detergents, such as cationic (CTAB) surfactants. The change in surface tension also
shows a break at CMC.
At CMC, micelles (aggregates of SD − with some counter-ions, Na+) are formed and some Na+ ions
are bound to these, which is also observed from conductivity data. In fact, an analysis of these
data has shown that approximately 70% Na+ ions are bound to SD − ions in the micelle. The surface
Concentration of each
species
(in aqueous solution of SDS) SDS
SD
Na
Micelle
CMC
Total SDS
FIGURE 1.28 Variation of concentration of different ionic species for SDS solutions (Na+, SD −, SDSmicelle)
and change in surface tension of a detergent solution with concentration.
54 Handbook of Surface and Colloid Chemistry
(a)
(b)
(c)
FIGURE 1.29 Different types of micellar aggregates: (a) spherical, (b) disc like, and (c) cylindrical (schematic).
charge (negative charge) was estimated from conductivity measurements (Birdi, 2003a). Therefore,
the concentration of Na+ will be higher than SD − ions after CMC. The same is true in the case of
cationic surfactants. In the case of CTAB solutions, one thus has CTA+ and Br− ions below CMC.
Above CMC, there are additionally CTAB+ micelles. In these systems, the counter-ion is Br−. It is
important to notice that due to these differences the two systems are completely divergent from
each other, as regards the areas of application (such as one cannot use CTAB for washing clothes!).
where
k1 and k2 are constants
Calkyl is the number of carbon atoms in the alkyl chain
The magnitude of CMC will change if the additive has an effect on the monomer–micelle equilib-
rium. It will also change if the additive changes the detergent solubility. The CMC of all ionic sur-
factants will decrease if co-ions are added. However, nonionic surfactants show very little change
in CMC on the addition of salts. This is as one should expect from theoretical considerations.
It is important to notice how merely the addition of 0.01 mol/L of NaCl changes the CMC by 65%.
This shows that the charge–charge repulsion is very significant and is reduced appreciably by the
addition of counter-ions (in this case Na+).
Introduction to Surface and Colloid Chemistry 55
The change of CMC with NaCl for SDS is as follows (at 25°C):
The radius of the spherical micelle is reported as 20 Å, which increases to 23 Å (for the nonspherical).
Experiments have shown that in most cases, such as for SDS, the initial spherical-shaped
micelles may grow under some influence into larger aggregates (disk-like, cylindrical, lamellar
vesicle) (Figure 1.29). The spherical micelle has a radius of 17 Å. The extended length of the SDS
molecule is about 17 Å. However, larger micelles (as found in 0.6 mol/L NaCl solution) have dimen-
sions of 17 and 25 Å, radii of an ellipse.
It is important to notice how CMC changes with even very small addition of NaCl (Birdi et al.,
1980; Mukerjee and Mysels, 1971). It has been found in general the change in CMC (in the case of
ionic surfactants only) with the addition of ions follows the relation:
The magnitude of Constant2 was found to be related to the degree of micelle charge. Its magnitude
varied from 0.6 to 0.7, which means that micelles have 30% (effective) charge. This indicates that if
there are 100 monomers per micelle, then ca. 70% counter-ions are bound.
Data were reported for the CMC of cationic surfactants that decreased on the addition of KBr
as follows:
Dodecyltrimethylammonium bromide (DTAB) and TrTAB:
TrTAB:
TTAB:
It is noticed that the magnitude of Constant1 increases with the increase in alkyl chain length.
Similar relationship has been reported for Na-alkyl sulfate homologous series (Birdi et al., 1980).
The alkyl chain length has a very significant effect (decrease with increase in nC) on the CMC.
The CMC data for soaps are found to give the following dependence on the alkyl chain length:
These data show that CMC decreases by a factor of 4 for each increase in chain length by –CH2CH2–.
This again indicates that due to lower solubility in water with increasing chain length, CMC is
related to the latter characteristic of the molecule. Further, this effect will be valid for all kinds of
detergent molecules (both with and without charges) (Mukerjee and Mysels, 1971; Tanford, 1980).
This may be compared to a sandwich kind of structure. It is important to mention that these bubble
structures are of nanometer dimensions, but still easily visible to the naked eye. It also indicates the
self-assembly properties of such SASs.
The thermodynamics of surface adsorption has been extensively described by Gibbs adsorption
theory (Chattoraj and Birdi, 1984; Spaull, 2004; Zhou, 1989). Further, Gibbs adsorption theory has
also been applied in the analyses of diverse phenomena (such as solid–liquid or liquid1–liquid2,
adsorption of solute on polymers, etc.). In fact, in any system where adsorption takes place at an
interface the Gibbs theory will be applicable (such as solid–liquid; protein molecule–solution with
solutes that may adsorb).
Gas Gas
Surface Surface
Liquid Liquid
(a) (b)
FIGURE 1.30 Surface composition (a) of pure water and (b) of an ethanol–water solution (shaded = ethanol).
Gas
Surface
Liquid
FIGURE 1.31 Concentration of detergent (shaded with tail) in solution and at surface. The shaded area at
surface is the excess concentration due to accumulation.
The schematic concentration profile of detergent molecules is such that the concentration is homog-
enous up to the surface. At the surface, there is almost only detergent molecule plus the necessary
number of water molecules (which are in a bound state to the detergent molecule). The solution
thus shows very low surface tension, ca. 30 mN/m. The surface concentration profile of detergent
is not easily determined by any direct method. Here it is shown as a rectangle for convenience, but
one may also imagine other forms of profiles, such as curved. One observes that surface tension
decreases due to ethanol. This suggests that there are more ethanol molecules at the surface than in
the bulk. This is also seen in a cognac glass. The ethanol vapors are observed to condense on the
edge of the glass. This shows that the concentration of ethanol in the surface of the solution is very
high. To analyze these data, one has to use the well-known Gibbs adsorption equation (Birdi, 1989,
2009; Chattoraj and Birdi, 1984).
n xi = n ti − n i α − n iβ (1.72)
where
n ti is the total moles of the ith component
niα and niβ moles in the neighboring phases, in the real system
In an exactly similar manner, one can define the respective surface excess internal energy, Ex, and
entropy, Sx by the following mathematical relationships (Birdi, 1989; Chattoraj and Birdi, 1984):
E x = E t − E α − E ß (1.73)
Sx = St − Sα − Sfl (1.74)
Here Et and St are the total energy and entropy, respectively, of the system as a whole for the actual
liquid system. The energy and entropy terms for α and ß phases are denoted by the respective super-
scripts. The excess (x) quantities thus refer to the surface molecules in an adsorbed state. At constant
T and p, for a two-component system (say water(1) + alcohol(2)), the classical Gibbs adsorption
equation has been derived (Adamson and Gast, 1997; Chattoraj and Birdi, 1984):
⎛ dγ ⎞
Γ2 = − ⎜ ⎟ (1.75)
⎝ dμ 2 ⎠T,p
μ 2 = μ 2o + RT ln(a 2 ) (1.76)
μ 2 = μ 2o + RT ln(C2 ) (1.77)
−1 ⎛ dγ ⎞ −C2 ⎛ dγ ⎞
Γ2 = ⎜ ⎟=
RT ⎝ d ln(C2 ) ⎠ RT ⎜⎝ dC2 ⎟ (1.78)
⎠
This shows that the surface excess quantity on the left-hand side is proportional to the change in sur-
face tension with the concentration of the solute (dγ/dln(Csurfaceactivesubstance)). A plot of ln(C2) versus
γ gives a slope equal to
Γ 2 ( RT )
From this, one can estimate the value of Γ2 (mol/area). This indicates that all SASs, will always have
a higher concentration at the surface than in the bulk of the solution. This relation has been verified
Introduction to Surface and Colloid Chemistry 59
by using radioactive tracers. Further, as will be shown later under spread monolayers, one finds a
very convincing support to this relation and the magnitudes of Γ for various systems. The surface
tension of water (72 mN/m, at 25°C) decreases to 63 mN/m in a solution of SDS of concentration
1.7 mmol/L. The large decrease in surface tension suggests that SDS molecules are concentrated at
the surface, as otherwise there should be very little change in surface tension. This means that the
concentration of SDS at the surface is much higher than in the bulk. The molar ratio of SDS:water
in the bulk is 0.002:55.5. At the surface, the ratio will be expected to be of a completely different
value, as found from the value of Γ (ratio is 1000:1). This is also obvious when considering that foam
bubbles form on solutions with very low surface-active agent concentrations. In fact, it is easy to
consider the state of surfactant solutions in terms of molecular ratios.
SDS is a strong electrolyte and in water it dissociates almost completely (at concentrations below
CMC):
C12H25SO4Na ≡ C12H25SO4 − + Na +
(1.79)
SDS ≡ DS− + S+
where surface excess, Γi, terms for each species in the solution, for example, DS− and S+, are
included.
This equation can be simplified, assuming electrical neutrality is maintained in the interface:
and
⎛ 2RT ⎞
=⎜ ⎟ ΓSDSdCSDS (1.84)
⎝ CSDS ⎠
−1 ⎛ dγ ⎞
ΓSDS = ⎜ ⎟ (1.85)
2(RT) ⎝ (d ln CSDS ) ⎠
In the case where the ion strength is kept constant, that is in the presence of added NaCl, then
the equation becomes
−1 ⎛ dγ ⎞
ΓSDS = ⎜ ⎟ (1.86)
(RT) ⎝ (d ln CSDS ) ⎠
60 Handbook of Surface and Colloid Chemistry
Comparing Equation 1.85 with 1.86, it will be seen that they differ by a factor of 2 and that the
appropriate form will need to be used in the experimental test of Gibbs equation (Adamson and
Gast, 1997; Chattoraj and Birdi, 1984). It is also quite clear that any partial of ionization would
lead to considerable difficulty in applying the Gibbs equation. Further, if SDS were investigated in
a solution using a large excess of sodium ions, produced by the addition of, say, sodium chloride,
then the sodium ion term in Equation 1.85 will vanish and will arrive back at an equation equiva-
lent to Equation 1.86.
In the plots of γ versus concentration, the slope is related to the surface excess, ΓSalkylsulfate.
The Gibbs adsorption equation for a simple 1:1 ionic surfactant, such as NaSDS (SD, Na).
In the absence of any other additives, one gets the following relationship between change in and
the change in NaSDS concentration (at a given T):
−1 ⎛ dγ ⎞
ΓSDS = ⎜⎜ ⎟⎟
4.606RT ⎝ ( d log CSDS ) ⎠
The magnitude of Γ thus obtained at the interface provides information about the orientation
and packing of the NaSDS molecule. This information is otherwise not available by any other
means.
The area/molecule values indicate that the molecules are aligned vertically on the surface, irre-
spective of the alkyl chain length. If the molecules were oriented flat, then the value of area/mol
would be much larger (greater than 100 Å2). Further, since the alkyl chain length has no effect
on the area also proves this assumption is acceptable. These conclusions have been verified from
spread monolayer studies. Further, one also finds that the polar group, that is, –SO4−, would occupy
something like 50 Å2.
Gibbs adsorption equation is a relation about the solvent and a solute (or many solutes). The
solute is present either as excess (if there is an excess surface concentration) if the solute decreases
the γ or a deficient solute concentration (if surface tension is increased by the addition of the solute)
(Chattoraj and Birdi, 1984).
Let us consider the system: a solution of water with a surfactant (soap, etc.), such as SDS. In an
aqueous solution, SDS molecule dissociates into SD − and S+ ions. The composition of surface of
water in these systems will be as depicted in the following:
Pure water ([w] bulk and surface [w] phase):
Water plus SDS ([S] in bulk phase and [S] in surface phase) (2 g/L):
It is found that the surface tension of pure water of 72 mN/m decreases to 30 mN/m by the addition
of 2.3 g/L (8 mmol/L) of SDS. Thus, the surface of SDS solution is mostly a monolayer of SDS plus
some bound water. The ratio of water:SDS in the system (1 L solution) is roughly as follows:
These data show that it would require 100 million bubbles to remove 1 g of SDS from the solution!
However, pollutants generally are found in concentrations less than 1 mg/L. Thus, one would need
about 100,000 bubbles to remove 1 mg of SDS/L of water solution. This is a rather low number of
bubbles. This seems to be a very large number. Since bubbles can be easily produced at very fast
rates (ca. 100–1000 bubbles/min), this is not a big hindrance. Consequently, any kind of other SAS
(such as pollutants in industry) can be thus removed by foaming. For example, in the recycle pro-
cess in paper industry, ink pollutant molecules are removed by bubble foam technology. It is also
important to consider that if an impurity in water was surface-active molecule, then this procedure
can be used to purify water.
During the past decades, a few experiments have been reported where verification of Gibbs adsorp-
tion has been reported. One of these methods has been carried out by removing using a microtone
62 Handbook of Surface and Colloid Chemistry
blade the thin layer of surface of a surfactant solution. Actually, this is almost the same as the proce-
dure of bubble extraction or merely by a careful suction of the surface layer of solution. The surface
excess data for a solution of SDS were found to be acceptable. The experimental data were 1.57 ×
10 −18 mol/cm2, while from Gibbs adsorption equation one expected it to be 1.44 × 10 −18 mol/cm2.
dγ
=
( 47 − 39) =
8
= 17
d log(CCTAB ) (log(0.6) − log(0.96)) −0.47
From these data, the area/molecule for CTAB is found to be 90 Å 2, which is reasonable.
The Gibbs adsorption equation thus shows that near the CMC the surfactant molecules are oriented
horizontally with the alkyl chains pointing up while the polar groups are interacting with the water.
Accordingly, if one analyzes data of systems with varying alkyl chain lengths, say, C8SO4Na and
C12SO4Na, then the area per molecule should be the same. This is indeed the case as found from
experimental data.
2
dΓ ⎛ D ⎞
= Cbulk t −2 (1.87)
dt ⎜⎝ π ⎟⎠
2
⎛ Dt ⎞
Γ = 2Cbulk ⎜ ⎟ (1.88)
⎝ π ⎠
where
D is the diffusion constant coefficient
Cbulk is the bulk concentration of the solute
The following procedure has been used to investigate the surface adsorption kinetics.
In the literature, studies using suction at the surface have been used to investigate the surface
adsorption kinetics using high-speed measurement techniques (Birdi, 1989).
The magnitude of γ increases right after suction, corresponding to pure water (i.e., 72 mN/m)
and decreases with time as Γ increases (from initial value of zero). This experiment actually verifies
the various assumptions as made in the Gibbs adsorption equation. Experimental data show good
correlation to this equation when the magnitude of t is very small. It is also obvious that different
detergents will exhibit different adsorption rates. This property will effect the functional properties
of the detergent solution.
Polar Polar
outer outer
core core Water-insoluble
organic molecule
Polar
outer
core Water
soluble ion
FIGURE 1.32 Micelle structure: (a) inner part = liquid paraffin-like; outer polar part; (b) solubilization of
apolar molecule; and (c) binding of counter-ion to the polar part (schematic).
to electrical double layer (EDL)). These reactions would otherwise have been impossible (Birdi,
2009, 2010a). The different characteristics of micelle can be delineated as shown in Figure 1.32.
Alkyl chains attract each other (van der Waals forces); thus, the inner part consists of alkyl
groups that are closely packed. It is known (from experimental data) that these clusters behave as
liquid paraffin (CnH2n+2). This was concluded from the fact that micelles of different alkyl chain
lengths are miscible (Tanford, 1980). The alkyl chains are thus not fully extended. Hence, one
would expect that this inner hydrophobic part of micelle should exhibit properties that are common
for alkanes, such as ability to solubilize all kinds of water-insoluble organic compounds (as also
found from experimental data). The solute enters the alkyl core of the micelle and it swells. The
equilibrium is reached when the ratio between moles solute: moles detergent is reached correspond-
ing to the (free energy) thermodynamic value. Size analyses of micelles (by using light-scattering)
of some spherical micelles of SDS have indeed shown that the radius of the micelle is almost the
same as the length of the SDS molecule. However, if the solute interferes (that is forms aggregates)
with the outer polar part of micelle, then the micelle free energy may change (i.e., the CMC and
other properties change). This is observed in the case of n-dodecanol addition to SDS solutions. In
general, the addition of very small amounts of solutes show very little effect on CMC (i.e., the free
energy of the micelle formation). However, if a solute makes mixed-micelles, then the free energy
of micelle system changes and gives a change in CMC (Birdi, 2010a; Tanford, 1980). The data in
Introduction to Surface and Colloid Chemistry 65
Solubility of
naphthalene
Slope = 14
Micellar
phase
CMC
SDS (mol/L)
Figure 1.33 show the change in the solubility of naphthalene (a water-insoluble organic molecule)
in SDS aqueous solutions. Similar data have been reported for many other water-insoluble organic
compounds (such as anthracene, phenanthrene, azobenzene, etc.).
Below CMC, the amount dissolved remains constant, which corresponds to its solubility in pure
water. The slope of the plot above CMC corresponds to 14 mol SDS:1 mol naphthalene. It is seen
that at the CMC the solubility of naphthalene abruptly increases. This is due to the fact that micelles
can solubilize water-insoluble organic compounds. A more useful analysis can be carried out by
considering the thermodynamics of this solubilization process.
At equilibrium, the chemical potential of a solute (naphthalene; etc.) will be given as
μss = μaq M
s = μ s (1.89)
where ss, saq, and sM are the chemical potentials of the solute in the solid state, aqueous phase,
and micellar phase, respectively. This equilibrium state is common in all kinds of physic/chemical
systems. It must be noted that in these micellar solutions we will describe the system in terms of
aqueous phase and micellar phase. Before CMC, the solute will only be present in the water phase
o
(as if no detergent is added). The standard free energy change involved in the solubilization, G so ,
is given as follows:
⎛C ⎞
o
ΔG so = −RT ln ⎜ s,M ⎟ (1.90)
⎝ Cs,aq ⎠
where Cs,aq and Cs,M are the concentrations of the solute in the aqueous phase and in the micellar
phase, respectively. The free energy change is the difference between the energy when solute is
transferred from solid state to the micelle interior. It has been found from many systematic s tudies
that ΔGso is dependent on the chain length of the alkyl group of the surfactant. The magnitude
of ΔGso changes by −837 J (−200 cal)/mol, with the addition of –CH2– group. In most cases, the
addition of electrolytes to the solution has no effect (Birdi, 1982, 1999, 2003a). The kinetics of solu-
bilization has an effect on its applications (Birdi, 2003a; Yoshida et al., 2002).
Another important aspect is that the slope in Figure 3.20 corresponds to (1/Cs,M). This allows
one to determine moles SDS required to solubilize one mole of solute. This magnitude is useful in
66 Handbook of Surface and Colloid Chemistry
understanding the mechanism of solubilization in micellar systems (Birdi, 2009; Chaibundit et al.,
2002; Yan et al., 2012).
Analyses of various solutes in SDS micellar systems showed that
The ratio of detergent:solute (in the case of naphthalene, etc.) decreases as the chain length of
the detergent molecule increases. This kind of study thus allows one to determine (quantitatively)
the range of solubilization in any such application. These systems when used to solubilize water-
insoluble organic compounds would require such information (in systems such as pharmaceutical,
agriculture sprays, paints, etc.). Dosage of any substance is based on the amount of material per
volume of a solution. Thus, this also shows that wherever detergents are employed, the major role
(besides lower surface tension) would be the solubilization of any water-insoluble organic com-
pounds (e.g., in pharmaceutical products). This process would then assist in the cleaning or washing
or any other effect. In some cases, such as bile salts, the solubilization of lipids (especially lecithins)
gives rise to some complicated micellar structures (Birdi, 2010a; Tanford, 1980). Due to the forma-
tion of mixed lipid–bile salt micelles, one observes changes in CMC and aggregation number. This
has major consequences in the bile salts in biology. Dietary fat consists essentially of mixed triglyc-
erides. These fatty lipids pass through the stomach into the small intestine, without much change
in structure. In the small intestine, the triglycerides are partly hydrolyzed by an enzyme (lipase),
which leads to the formation of oil–water emulsion. This shows the importance of surface agents in
biological systems, such as stomach.
important requirement is that the ingredients in the shampoo should not damage or irritate the eye
or the skin with which it may come in contact. In fact, all shampoos are tested for eye irritation
and skin irritation before marketing. In some cases by merely increasing the viscosity, one achieves
a great deal of protection. For example, a surfactant solution (ca. 20%), alkyl sulfate with two EO
(ethylene oxide groups), gives very high viscosity if a small amount of salt is added. Shampoo indus-
try is highly specialized, and a large industrial state-of-the art research is applied in this product.
Pure water
WSWSWSWSWSWSWSWSWSWSWSWSWSW
FIGURE 1.34 Pure water (W) and a foreign solute (organic) molecule (S) is dissolved in water. The cavity
is created for S (see text).
68 Handbook of Surface and Colloid Chemistry
This indicates that some kind of structure will be present in liquid water surrounding CH4 or alkane
molecule, which would be indicative of hydrate structure (Tanford, 1980). These data are of cur-
rent interest in the future (from nonconventional reservoirs) gas recovery from methane hydrate
reservoirs.
Inorganic salts, such as NaCl, in aqueous solutions dissociate into Na+ and Cl− ions and interact
with water through hydrogen bonds. Hexane molecule merely dissolves (although very low solu-
bility) in water by placing inside water structure without taking part in the water bonds (hydrogen
bonds). Since water structure is stabilized mainly by hydrogen bonds, hexane molecule will give rise
to some rearrangements of these bonds but without breaking (since no heat is evolved in this process).
If a salt exhibits a maximum solubility (called saturation solubility: at a specific temperature and
pressure) of 10 mol/L in water, then it corresponds to ca. 10 mol of salt:55 mol of water (ratio of
1:5.5). On the other hand, an alkane molecule may show a maximum solubility of 0.0001 mol/L in
water (0.0001:55 mol), or a ratio of 1:550.000. Many decades ago, this model was found to be able
to predict the solubilities of both simple organic molecules, and for the case of more complicated.
In most simple case, the solubility of heptane is lower than that of hexane, due to the addition of
one –CH2– group. In the case of alkane molecules, a linear relation between the solubility and the
number of –CH2– groups (Acree, 2004; Birdi, 1997; Kyte, 2003; Tanford, 1980).
This model is thus based on the following assumptions when alkane molecule is placed in water:
Alkane (CCCCCCCCC) is placed in a cavity in water (ww).
wwwwwwwwwwwwwwwwwwwwwwwwwwwwwww
wwwwwwwwwwwCCCCCCCCCwwwwwwwwwwww
wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww
The alkane molecule merely makes a cavity in water without breaking hydrogen bonds of water
structure. It has been argued that the structure of water exhibits voids where a solute (nonpolar) can
reside (Birdi, 2009; Tanford, 1980). Some analogy has been related to the data of methane hydrate
structures (Tanford, 1980). In the ice structure, it is found that methane (and other gases, such as
chlorine, Cl2) does fit in voids in ice and is trapped. In water, the ice-like structure persists, and
thus, alkane molecules can dissolve in the cavity spaces. The energy needed to create a surface area
of the cavity will be proportional to the degree of solubility of the alkane. Thus, the free energy of
solubility of any alkane molecule will be given as follows:
By analyzing the solubility data of a whole range of alkane molecules in water, the following rela-
tion was found to fit the experimental data:
For the solubility of alkanes in water, the total surface area (TSA) gives the solubility:
where solubility (sol) is in molar units and TSA in A2. For example,
The surface areas of each group in n-nonanol (C9H19OH) were estimated by different methods.
These data are as follows:
The magnitude of the surface area of the CH3 group is obviously expected to be larger than the CH2
groups. One can estimate the magnitude of TSA of n-decanol: it will be (TSA of nonanol + TSA of
CH2) 431 + 32 = 463 Å2. From this value, one can estimate its solubility.
The data for solubility of homologous series of n-alcohols in water (at 25°C) are of interest, as
shown in the following.
Accordingly, this algorithm allows one to estimate the solubility of water of any organic substance.
The estimated solubility of cholesterol (add) was almost in accord with the experimental data (Birdi,
1997, 2009). The solubility analysis of cholesterol is of much interest in the case of biological phe-
nomena (Tanford, 1980). It is seen that log(S) is a linear function of the number of carbon atoms in
the alcohol. Each –CH2– group reduces log(S) with 0.06 unit.
Water
Teflon barrier
Clean
Monolayer
surface
FIGURE 1.36 Monolayer film balance: barrier and lipid film and surface pressure.
(Figure 1.35) on water surface (Adamson and Gast, 1997; Birdi, 1989, 1999, 2003a; Bouvrais et al.,
2014; Chattoraj and Birdi, 1984; Gagnon and Meli, 2014; Gaines 1966; Imae, 2007).
A few decades ago, experiments showed that monomolecular films of lipids could be studied by
using rather simple experimental methods, Figure 1.36. It is amazing to find that even a monolayer
of lipid film (thickness ca. 20 Å) can easily be investigated with a very high precision. In fact,
monolayer method is the only technique that can provide detailed information about adsorbed state
of molecules in such systems. Furthermore, Langmuir was awarded Nobel Prize (in 1920) for his
pioneer monolayer film studies (Birdi, 1999, 2003a; Gaines, 1966).
It was already explained that when a surface-active agent (such as surfactant or a soap) is dis-
solved in water, it adsorbs preferentially at the surface (surface excess, Γ). This means that the
concentration of a surface-active agent at the surface may be as high as 1000 times more than in
the bulk. The decrease in surface tension indicates this and also suggests that only a monolayer is
present at the surface. For example, in a solution of SDS of concentration 0.008 mol/L, the surface is
completely covered with SDS molecules. Let us consider systems where the lipid (almost insoluble
in water) is present as a monolayer on the surface of water. In these systems, almost all the substance
applied to the surface (in the range of few micrograms) is supposed to be present at the interface.
This means that one knows (quantitatively) the magnitude of surface concentration (same as the
surface excess, Γ).
If one places a very small amount of a lipid on the surface of water, it may affect the surface ten-
sion in different ways. It may not show any effect (such as cholesterol). It may also show a drastic
decrease in surface tension (such as stearic acid or tetra-decanol). An amphiphile molecule will
adsorb at the air–water or oil–water interface, with its alkyl group pointing away from the water
phase. The alkyl group is at a lower energy state when pointing toward air that being surrounded
by water molecules.
The system can be easily considered as a two-dimensional analog to the classical three-
dimensional systems (such as the gas theory).
It is thus seen that Π of a monolayer is the lowering of surface tension due to the presence of
monomolecular film. This arises from the orientation of the amphiphile molecules at the air–water
or oil–water interface, where the polar group would be oriented toward the water phase, while the
Introduction to Surface and Colloid Chemistry 71
nonpolar part (hydrocarbon) would be oriented away from the aqueous phase. Later it was observed
that if a clean smooth solid is dipped through monolayer, then in most cases a single layer of lipid
will be deposited on the solid. This film was called Langmuir–Blodgett (LB) film (Figure 1.37).
Scanning probe microscopes (Birdi, 2003b) have been used to verify these structures in a very high
detail. Further if one repeated this process, one could deposit multilayers. This LB film technique
has found much application in the electronic industry. Further, the presence of only one layer of
lipid changes the surface properties (such as contact angle-wetting, friction, light reflection, charge,
adhesion, etc.) of the solid.
Constant area monolayer film method: One can also study Π versus surface concentration
(Cs = area/molecule) isotherms by keeping the surface area constant. The surface concentration, Cs,
is changed by adding small amounts of a substance to the surface (by using a microliter syring—1
or 5 μL). In general, one obtains a good correlation with the Π versus area isotherms (Adamson and
Gast, 1997; Birdi, 1989, 1999; Gaines, 1966).
In the case of alkanes, the distance between the molecules in the solid state phase is ca. 5 Å,
while it is 5.5–6 Å in the case of liquid state. The distance between molecules in the gas phase, in
general, is ca. 10001/3 = 10 times larger than in the liquid state (water:volume of 1 mol water = 18 cc;
volume of 1 mol gas = 22.4 L). In the three-dimensional structural buildup, the molecules are in
contact with near neighbors and in contact with molecules that may be 5–10 molecular dimensions
apart (as found from x-ray diffraction). This is apparent from the fact that in liquids there is a long
range order up to 5–10 molecular dimensions. On the other hand, in the two-dimensional films the
state is much different. The amphiphile molecules are oriented at the interface such that the polar
groups are pointed toward water (subphase), while the alkyl groups are oriented away from the
subphase. This orientation gives the minimum surface energy. The structure is stabilized through
lateral interaction between
The alkyl–alkyl groups attraction arises from the van der Waals forces. The magnitude of van der
Waals forces increases with
Experiments show that the alkyl chain length increases the magnitude of Π of films and thus
becomes more stable, thus giving higher collapse pressure, Πco. The stable films are thus formed
when the attraction forces are stronger than the repulsive forces.
The most convincing results were those as obtained with the normal fatty alcohols and acids.
Their monomolecular films were stable and exhibit very high surface pressures (Birdi, 1989, 1999).
A steep rise in Π is observed around 20.5 Å, regardless of the number of carbon atoms in the
chains. The volume of a (–CH2–) group is 29.4 Å3. This gives the length of each (–CH2–) group
perpendicular to the surface, or the vertical height of each group as ca. 1.42 Å. This compares very
satisfactorily with the x-ray data with this value of 1.5 Å. This means that such straight chain lipids
are oriented in this compressed state in vertical orientation.
As high pressures lead to transitions from gas to liquid to solid phases in the three-dimensional
systems, similar state of affairs would be expected in the two-dimensional film compression Π
versus A isotherms, Figure 1.38, and as described in the following.
Surface pressure
50
Collapse
Solid
LEX
LCO
GAS
0
Area per molecule
FIGURE 1.38 Lipid monolayer phases (two-dimensional) (monolayer phases are dependent on both molecu-
lar structure and temperature).
74 Handbook of Surface and Colloid Chemistry
In general, ideal gas behavior is only observed when distances between the amphiphiles are very
large, and thus the value of Π is very small, that is, <0.1 mN/m. It is also noticed that from such
sensitive data one can estimate the molecular weight of the molecule in the monolayer. This has
been extensively reported for protein monolayers (Adamson and Gast, 1997; Birdi, 1989, 1999). The
latter observation requires an instrument with very high sensitivity, ±0.001 mN/m. The Π versus A
isotherms of n-tetradecanol, pentadecanol, pentadecyclic acid and palmitic acid in the low Π region
showed data that agreed with the ideal film. Similar data for isotherms were reported for other lipid
monolayers by other workers. The various forces that are known to stabilize the monolayers are
mentioned as follows:
where
Πkin arises from kinetic forces
ΠvdW is related to the van der Waals forces acting between the alkyl chains (or groups)
Πelectro is related to polar group interactions (polar group–water interaction; polar group–polar
group repulsion; charge–charge repulsion)
When the magnitude of A is very large, the distance between molecules is large. If there are no
van der Waals or electrostatic interactions, then the film obeys the ideal equation. As the area per
molecule is decreased, other interactions become significant. The Π versus A isotherm can be used
to estimate these different interaction forces (Adamson and Gast, 1997; Birdi, 1989, 1999; Bouvrais
et al., 2014; Gaines, 1966). The ideal equation has been modified to fit Π versus A data, in those
films where co-area, Ao, correction is needed (Birdi, 1989) (Table 1.6):
Π( A − Ao ) = k BT (1.99)
In the case of straight chain alcohols or fatty acids, Ao is almost 20 Å2, which is the same as found
from the x-ray diffraction data of the packing area per molecule of solid alkanes.
Introduction to Surface and Colloid Chemistry 75
TABLE 1.6
Magnitudes of Ao for Different Film-Forming Molecules on the Surface of Water
Compound Ao (Å2)
Straight chain acid 20.5
Straight chain acid (on dilute HCl) 25.1
N-fatty alcohols 21.6
Cholesterol 40
Lecithins ca. 50
Proteins ca. 1 m2/mg
Diverse synthetic polymers (poly-amino acids, etc.) ca. 1 m2/mg
Source: Birdi, K.S., Lipid and Biopolymer Monolayers at Liquid Interfaces, Plenum Press, New York, 1989.
This equation (Equation 1.99) is thus valid when A ≫ Ao. The magnitude of Π is 0.2 mN/m for
A = 2000 Å2, for ideal film. However, Π will be about 0.2 mN/m for A = 20 Å2 for a solid-like film
of a straight chain alcohol.
Π versus (A) for a monolayer of valinomycin (a dodecacyclic peptide) shows that the relation as
given in Equation 1.99 is valid. In this equation, it is assumed that the amphiphiles are present as
monomers. However, if any association takes place, then the measured values of (ΠA) would be less
than kBT < 411, as has also been found (Birdi, 1989, 1999). The magnitude of kBT = 411 dyn/cm =
4 × 10 −21 J, at 25°C. In the case of nonideal films, one will find that the versus data do not fit the
relation in equation. This deviation requires that one uses other modified equations-of-state. This
procedure is the same as one uses in the case of three-dimensional gas systems.
This means that such deviations thus allow one to estimate these interactions.
1.
Liquid expanded films (Lexp): In general, there are two distinguishable types of liquid films.
The first state is called the liquid expanded (Lexp) (Adamson and Gast, 1997; Chattoraj and
Birdi, 1984; Gaines, 1966). If one extrapolates the Π–A isotherm to zero Π, the value of A
obtained is much larger than that obtained for close packed films. This shows that the dis-
tance between the molecules is much larger than one will find in the solid film, as will be
discussed later. These films exhibit very characteristic elasticity, which will be described
further in the following.
76 Handbook of Surface and Colloid Chemistry
2.
Liquid condensed films (Lco): As the area per molecule (or the distance between molecules)
is further decreased, there is observed a transition to a so-called liquid condensed (Lco)
state. These states have also been called “solid expanded” films (Adam, 1930; Adamson
and Gast, 1997; Birdi, 1989, 1999; Gaines, 1966), which will be later discussed in further
detail. The Π versus A isotherms of n-pentadecylic acid (amphiphile with a single alkyl
chain) have been studied, as a function of temperature. Π–A isotherms for two chain alkyl
groups, as lecithins, also showed a similar behavior.
Surface pressure
45
Cholesterol
Collapse
0
Area/molecule 40
FIGURE 1.39 Surface pressure (Π) versus area/molecule (A) isotherm of cholesterol monolayer on water
(at 25°C).
Introduction to Surface and Colloid Chemistry 77
rise in Π is observed and a distinct break in the isotherm is found at the collapse. This occurs at
Π = 40 mN/m and A = 40 Å 2.
This value of Aco corresponds to the cholesterol molecule oriented with the hydroxyl group pointing
toward the water phase. AFM studies cholesterol as LB films have shown that there exist domain struc-
tures. This has been found for different collapse lipid monolayers (Birdi, 2003b; Kuo and Chang, 2014).
It should be mentioned that monolayer studies are the only procedure, which allows one to estimate
the area per molecule of any molecule as situated at the water surface. In general, the collapse pres-
sure, Πcol, is the highest surface pressure to which a monolayer can be compressed without a detectable
movement of the molecules in the films to form a new phase, Figure 1.39. In other words, this surface
pressure will be related to the nature of the substance and the interaction between the subphase and the
polar part of the lipid or the polymer molecule. The monolayer collapse pressure has been shown to
provide much information also in the case of protein monolayers. It was found that a-helical polypep-
tides exhibited shaper collapse state than b-shaped polypeptides (Birdi and Fasman, 1973).
A monomolecular film is resistant to shear stress in the plane of the surface, as also is the case for
the bulk phase: A liquid is retarded in its flow by viscous forces. The viscosity of the monolayer
may indeed be measured in two dimensions by flow through a canal on a surface or by its drag on
a ring in the surface, corresponding to the Ostwald and Couttte instruments for the study of bulk
viscosities. The surface viscosity, ηs, is defined by the relation
{Tangential force per centimeter of surface} = ηs ◊ (rate of strain) (1.100)
and is thus expressed in units of (m/t) (surface poises), whereas bulk viscosity, η, is in units of poise
(L/(m t)). The relationship between these two kinds of viscosity is
η
ηs = (1.101)
d
WWWWWWWWWWWWWWWWWWWWWWWWWW: surface
While a system with lipid monolayer (L) gives a surface that is different:
LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLL: surface
WWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWW: first layer
WWWWWWWWWWWWWWWWWWWWWWWWWWWWW: bulk phase
78 Handbook of Surface and Colloid Chemistry
Thus, the lipid monolayer gives rise to hindrance to any flow or movement at the surface (the layers
with LWLW). This arises mainly from the fact that water molecules (first layer) bound to the polar
part of the lipid are subjected to high viscosity (surface viscosity).
It is easily realized that if a monolayer is moving along the surface under the influence of a gradi-
ent of surface pressure, it will carry some of the underlying water with it. In other words, there is no
slippage between the monolayer molecules and the adjacent water molecules. The thickness of such
regions has been reported to be on the order of 0.003 cm. It has also been asserted that the thickness
would be expected to increase as the magnitude of ηs increases. However, analogous to bulk phase,
the concept of free volume of fluids should be also considered in these films. As mentioned earlier,
the soap films are made of these films when air bubble forms at the liquid surface. Therefore, the
soap bubble characteristics are dependent on surface viscosity of the lipid monolayer.
The magnitude of ΔV is measured most conveniently by placing an air electrode (a radiation emit-
ter: e.g., Po210 [alpha-emitter]) near the surface (ca. mm in air) connected to a very high impedance
electrometer. This is required, since the resistance in air is very high, but it is appreciably reduced
by the radiation electrode.
Air electrode
Lipid monolayer IIIIIIIIIIIIIIIIIIIIIIIIIIIIII
Water phase electrode
The surface voltage is measured as a change in V for pure water surface to that of lipid monolayer
system.
Since these monolayers are found to be very useful biological cell membrane structures, it is thus
seen that such studies can provide information on many systems where ions are carried actively
through cell membranes (Birdi, 1989, 1999; Chattoraj and Birdi, 1984).
The transport of K+ ions through cell membranes by antibiotics (valinomycin) has been a very
important example. Addition of K+ ions to the subphase of a valinomycin monolayer showed that
the surface potential became positive. This clearly indicated the ion-specific binding of K+ to
valinomycin. In biological cells the concentrations of different ions are sometimes 20 times dif-
ferent than in the solution outside the cell. The membrane peptides, such as valinomycin, create
Introduction to Surface and Colloid Chemistry 79
a channel in the cell membrane for K+ ions. This leads to the effect that concentration of K+ ions
becomes the same both inside and outside the cell. This leads to the collapse of the cell. Further,
one finds that the theoretical basis of charged lipid monolayers is very well verified from such model
monolayer studies.
Further,
μ RNa = μ R + μ Na (1.104)
Γ NaCl = Γ Cl (1.106)
and
Γ RNa = Γ R (1.107)
It is also seen that the following equation will be valid for this system:
This is the form of the Gibbs equation for an aqueous solution containing three different ionic spe-
cies (e.g., R; Na; Cl). Thus, the more general form would for solutions containing i-number of ionic
species as
In the case of charged lipid film, the interface will acquire surface charge. The surface charge
may be positive or negative depending upon the cationic or anionic nature of the lipid or polymer
ions. This would lead to the corresponding surface potential, ψ, also having a positive or negative
potential (Attard, 1996; Birdi, 1989; Chattoraj and Birdi, 1984). The interfacial phase must be elec-
troneutral. This can only be possible if the inorganic counter-ions also are preferentially adsorbed
in the interfacial region. If a negatively charged lipid molecule, R−Na+, is adsorbed at the interface,
the latter will be negatively charged [air–water or oil–water]. According to the Helmholtz model for
80 Handbook of Surface and Colloid Chemistry
monolayer, Na+ ions on the interfacial phase will be arranged in a plane toward the aqueous phase.
The charge densities are equal in magnitude, but with opposite signs, Γ (charge per unit surface
area), in the adjacent planes. The (negative) charge density of the plane is related to the surface
potential [negative], ψo, at the Helmholtz charged plane:
(4πσδ)
ψo = (1.110)
D
where D is the dielectric constant of the medium (aqueous). According to Helmholtz double layer
model, the potential ψ decreases sharply from its maximum value, ψo, to zero as δ becomes zero
(Birdi, 1989, 1999, 2010a). The Helmholtz model was found not to be able to give a satisfactory
analyses of measured data. Later, another theory of the diffuse double layer was proposed by Gouy
and Chapman. Due to surface potential, the Na+ and Cl− ions will be distributed nonuniformly due
to the electrostatic forces. The concentrations of the ions near the surface can be estimated by the
Boltzmann distribution. The surface charge, σ, is derived from these assumptions (Adamson and
Gast, 1997; Birdi, 2010a; Chattoraj and Birdi, 1984):
1/2
⎡ 2DRTC ⎤ ⎡ ⎛ εψ o ⎞ ⎤
σ=⎢ ⎢sinh ⎜ ⎟ ⎥ (1.111)
⎣ 1000π ⎦ ⎝ 2kT ⎠ ⎥⎦
⎥
⎢⎣
At 25°C, for uni-univalent electrolytes one gets the following (Adamson and Gast, 1997; Birdi, 2009):
This relates the potential charge of a plane plate condenser to the thickness 1/k. The expression for
Πel is derived as
−1
⎛ ⎛ 134 ⎞ ⎞
Π el = 6.1C1/ 2 ⎜ cosh sinh −1 ⎜ 1/ 2 ⎟
⎟ (1.113)
⎜
⎝ ⎝ A elC ⎠ ⎟⎠
The quantity (kT) is approximately 4 × 10 −14 erg at ordinary room temperature (25°C), and
(kT/ε) = 25 mV. The magnitude of Πel can be estimated from monolayer studies at varying pH.
At the iso-electric pH, the magnitude of Πel will be zero (Birdi, 1989). These Π versus A isotherms
data at varying pH subphase have been used to estimate Πel in different monolayers.
At the iso-electric pH, the value of Πel was found to be zero, as expected.
The amount of loss of water from surface will be dependent on different parameters:
Lipid monolayers are found to have extensive effect on the evaporation rates of water.
In one example, one finds the following data:
The presence of lipid film thus has a significant effect on the evaporation rate of water.
At air–water interface, water molecules are constantly evaporating and condensing in a closed
container. In an open container, water molecules at the surface will desorb and diffuse into the gas
phase. It thus becomes important to determine the effect of a monomolecular film of amphiphile at
the interface. Measurement of the evaporation of water through monolayer films was found to be
of considerable interest in the study of methods for controlling evaporation from great lakes. Many
important atmospheric reactions involve interfacial interactions of gas molecules (oxygen and dif-
ferent pollutants) with aqueous droplets of clouds and fogs, as well as ocean surfaces. The presence
of lipid monolayer films would thus have an appreciable effect on such mass transfer reactions.
In the original procedure, the box containing the desiccant is placed over the water surface, and
the amount of water sorbed is determined by simply removing the box and weighing it (Adanson and
Gast, 1997; Birdi, 1989; Moroi et al., 2000). The results are generally expressed in terms of specific
evaporation resistance, r. The methods for calculating r from the water uptake values, together with
the assumptions involved, are described in detail in the aforementioned references. The rates of evap-
oration are measured both without (Rw) and with (Rf ) the film. The resistance r is given as follows:
where
A is the area of the dish
vw and vd are the water vapor concentrations for water and desiccant, respectively
The condensed monolayers gave a much higher r values than the expanded films, as expected. Many
decades ago, it was found that the evaporation rates of water were reduced by the presence of mono-
molecular lipid film (Adamson and Gast, 1997). This observation gave rise to important application
in reducing the loss of water from great lakes during summer in countries such as Australia and
Africa.
The most simple method to investigate is to measure the loss of weight of a water container,
with or without the presence of a monomolecular lipid film. La Mer investigated the effect of long
chain films (C14H29OH, C16H33OH, C18H37OH, C20H41OH, C22H45OH) and found that the resistance
to evaporation increased with the chain length (Adamson and Gast, 1997). For instance, the resis-
tance increased by a factor 40 for C22H45OH as compared to C14H29OH monolayer. This indicates
that evaporation takes place mainly through the alkyl chains films. Since the attraction between
alkyl chains increases with the number of carbon atoms (as also observed from the collapse pres-
sure, Πcol), the resistance to evaporation increases.
stable monolayers at water surface, if the HLB is of the right balance. Especially, almost all proteins
(hemoglobin, ovalbumin, bovine serum albumin, lactoglobulin, etc.) are reported to form stable
monolayers at the water surface (Birdi and Fasman, 1973; Birdi, 1999).
Micro-lithography
Solid-state polymerization
Light guiding, electron tunneling
Medical diagnostic chips
Photovoltaic effects
In the case of films such as Mg-stearate, if one dips a clean glass slide through the film, a monolayer
is adsorbed on the down-stroke. Another layer is adsorbed on the up-stroke. Under careful condi-
tions, one may make LB films with multilayers (varying from a few layers to over thousands). One
can monitor the adsorption by measuring the decrease in Π on each stroke. If no adsorption takes
place, then one observes no change in surface pressure. There are some lipids, such as cholesterol,
which do not form LB films. There are also other methods, such as
Light reflection
IR spectroscopy
STM (scanning tunneling microscope) and AFM (atomic force microscope)
Change in contact angle
which can provide detailed information on these LB films. The LB deposition is traditionally car-
ried out in the solid phase. The surface pressure, Π, is then high enough to ensure sufficient cohesion
in the monolayer, for example, the attraction between the molecules in the monolayer is high enough
so that the monolayer does not fall apart during transfer to the solid substrate. This also ensures the
buildup of homogeneous multilayers. The surface pressure value that gives the best results depends
on the nature of the monolayer and is usually established empirically. However, amphiphiles can
seldom be successfully deposited at surface pressures lower than 10 mN/m and at surface pres-
sures above 40 mN/m collapse and film rigidity often pose problems. When the solid substrate
Introduction to Surface and Colloid Chemistry 83
is hydrophilic (glass, SiO2, etc.), the first layer is deposited by raising the solid substrate from the
subphase through the monolayer, whereas if the solid substrate is hydrophobic (graphite, silanized
SiO2, etc.), the first layer is deposited by lowering the substrate into the subphase through the mono-
layer. One finds that in some LB systems, the magnitude of Π drops to/by about 1–2 mN/m as each
time a plate (with surface area of 5 cm2) is moved down or up through the monolayer. It is thus a
very sensitive method to study the LB deposition phenomena directly. There are several parameters
that affect what type of LB film is produced. Depending on the behavior of the molecule, the solid
substrate can be dipped through the film until the desired thickness of the film is achieved.
It is obvious that such processes involving monomolecular film transfers will easily be disturbed
by defects, arising from various sources. The structural analysis of the molecular ordering within a
single LB monolayer is important both to understand how the environment in the immediate vicin-
ity of the surface (i.e., solid) affects the structure of the molecular monolayer and to ascertain how
the structure of one layer forms a template for subsequent layers in a multilayer formation.
investigated. The contact angles and adhesional energy changes during the transfer of monolayers
from the air–water interface to solid (hydrophobic glass) supports have been analyzed (Birdi, 1989;
Gaines, 1966).
Interest in the dielectric studies of deposited LB films of fatty acids and their metal salts was one
of the parameters of main investigations in the early stages of research on LB films, for example,
capacitance, resistance, and dielectric constant. In early investigations, measurements on impedance
of films and related phenomena were carried on Cu-stearate, Ba-stearate, and Ca-stearate films.
Initially, Hg droplets were used for small area probe measurements and an AC bridge was used for
impedance measurements. The resistance of the films was found to be very low (<1 Ω) with high
signal voltages, whereas it was on the order of mega-ohms with signals of 1 or 2 V. In both types of
films, the capacity decreased with thickness, as can be expected from the following relation:
Aε
Capacitance of the deposited LB films = CC = (1.115)
4πNd
where
CC is the capacity
ε denotes the dielectric constant
N is the number of layers
d is the layer thickness
A is the area of contact between drop and film
On the other hand, the values of the resistance per layer showed a definite increase with the thick-
ness of the film. The specific resistance of the films thus determined from their values of the resis-
tance per layer was ca. 1013 Ω. This was based on the results of capacity measurements on some
75 samples. The capacitance measurements thus performed on stearate films (1–10 layers) led to ε
values between 2.1 and 4.2, with a bulk value of 2.5.
In many of the measurements reported in the literature, the organic film was sandwiched between
an evaporated aluminum electrodes. The fact that an oxide layer grows on the base of aluminum
electrode was present and its effect on the capacitance values of the device was neglected consid-
ering that the resistivity of oxide film is small compared with resistivity of the organic LB layers.
The presence of such thin oxide layer between metal electrodes and fatty acids can be analyzed.
The capacitance has been reported to be a linear function of [1/CC] with respect to the number of
transferred monolayers (Figure 4.12). LB films of Ba-salts of fatty acids deposited at Π = 50 mN/m
(Birdi, 1999, 2003b) gave following relation between [1/CC] and N:
Ba-stearate:
1
= 15.9N + 1.13 (106 F −1 ) (1.116)
CC
Ba-behenate:
1
= 17.2N + 8 (106 F −1 ) (1.117)
CC
The dielectric anisotropy of long-chain fatty acid monolayers was analyzed. These fatty acids were
considered as being oriented in a cylinder cavity with length [L] ≫ diameter [D]. Considering each
bond in these molecules as a polarization ellipsoid with axial symmetry about the –C–C– bonds, the
mean polarizability of the bonds was calculated.
Introduction to Surface and Colloid Chemistry 85
FIGURE 1.40 Different lipid self-assembly monolayer (SAM) structures: (a) micelle; (b) monolayer; (c) LB
film; and (d) vesicle.
is dispersed in water at 25°C, it forms self-assembly structure, which is called liposome or vesicle.
A liposome is a spherical vesicle with a membrane composed of a phospholipid and cholesterol (less
than 50%) bilayer. Liposomes can be composed of naturally derived phospholipids with mixed lipid
chains (such as egg phosphatidylethanolamine) or of pure surfactant components like dioleolylphos-
phatidylethanolamine (DOPE). Liposomes usually contain a core of aqueous solution. Multilayer
liposomes are called vesicles. However, one finds a range of mixtures of these structures in mixed
lipid systems.
The inherent lipid self-assembly characteristic is the main driving force in these structures.
Monolayer studies are the only source of data which provides direct estimation of the stabiliz-
ing forces. Hence, it is safe to conclude that many important natural systems are based upon this
molecular characteristics of lipids. Vesicles are unilamellar phospholipid liposome.
The word liposome comprises two words (from Greek-lipid [fat] and Soma [body]). The word
liposome does not in itself denote any size characteristics. Furthermore the term liposome does not
necessarily mean that it must contain lipophobic contents, such as water, although it usually does.
The vesicles may be conceived as microscopic (or nanosized) containers of carrying molecules
(drugs) from one place to another. The structures are suitable for both transporting water-soluble
or water-insoluble drugs. Since the lipids used are biocompatible molecules, this may also enhance
their adsorption and penetration into cells.
During the past decades, liposomes have been used for drug delivery due to their unique solu-
bilization characteristics for water-insoluble organic substances. A liposome encapsulates a region
on aqueous solution inside a hydrophobic membrane; dissolved hydrophilic solutes cannot readily
pass through the lipids. Hydrophobic chemicals can be dissolved into the membrane, and in this
way, liposome can carry both hydrophobic molecules and hydrophilic molecules. Liposomes can
also be designed to deliver drugs in different specific ways. Liposomes that contain low (or high)
pH can be constructed such that dissolved aqueous drugs will be charged in solution (i.e., the pH
is outside the drug’s pI range) (isoelectric pH = pI). As the pH naturally neutralizes within the
liposome (protons can pass through a membrane), the drug will also be neutralized, allowing it to
freely pass through a membrane. These liposomes work to deliver drug by diffusion rather than by
direct cell fusion.
Further, another important property of liposomes is their natural property to target cancer
cells. The endothelial walls of all healthy human blood vessels are encapsulated by endothelial
cells that are bound together by tight junctions. These tight junctions block large particle in the
blood from leaking out of the vessel. It is known that the tumor vessels do not contain the same
level of seal between cells and are diagnostically leaky. The size of liposomes can be used to play
a specific application. For example, liposomes of certain sizes, typically (i.e., than 400 nm), can
rapidly enter tumor sites from the blood but are kept in the bloodstream by the endothelial wall in
healthy tissue vasculature. Liposome-based anticancer drugs are now being used as drug delivery
systems.
Liposomes can be created by shaking or sonicating phospholipids (dissolved in alcohol) in water.
Low shear rates create multilamellar liposomes, which have many layers like an onion. Continued
Introduction to Surface and Colloid Chemistry 87
high-shear sonication tends to form smaller unilamellar liposomes. In this technique, the liposome
contents are the same as the contents of the aqueous phase. Sonication is generally considered a
“gross” method of preparation, and newer methods such as extrusion are employed to produce
materials for human use.
(a) (b)
FIGURE 1.41 Solid surface molecules: (a) perfect crystal and (b) surface with defects.
88 Handbook of Surface and Colloid Chemistry
GL
Air Liquid
Contact
angle
GS GLS
Solid
FIGURE 1.42 The surface force equilibrium between surface tensions of liquid (GL)–solid (GS)–liquid–
solid (GLS)–contact angle (CA).
as follows (Adamson and Gast, 1997; Birdi, 1997, 2003a; Chattoraj and Birdi, 1984; Coertjens et al.,
2014; Karapetsas et al., 2014; Nikolov and Wasan, 2014; Ramiasa et al., 2014):
γ S = γ L cos(θ) + γ SL (1.119)
γ L cos(θ) = γ S − γ SL (1.120)
where γ is the IFT at the various boundaries between solid, S, liquid, L, and air (or vapor) phases,
respectively. The relation of Young’s equation is easy to understand as it follows from simple
physics laws. At the equilibrium contact angle, all the relevant surface forces come to a stable state,
Figure 1.42.
The equilibrium of forces is valid in all kinds of solid–liquid systems. The geometrical force
balance is considered only in the X–Y plane. This assumes that the liquid does not affect the solid
surface (in any physical sense). This assumption is safe in most cases. However, only in very special
cases, if the solid surface is soft (such as contact lens), then one will expect that tangential forces
will also need to be included in Equation 1.119. There exist extensive data, which convincingly sup-
port the relation in Equation 1.119 for liquids and solids.
( γ S − γ LS )
cos(θ) = (1.121)
γL
which would then allow one to analyze the variation of γ with the change in the other terms. The
latter is important, because complete wetting occurs when there is no finite contact angle, and
thus, γL > γS − γLS. However, when γL > γS − γLS, then cos(θ) < 1, and a finite contact angle is pres-
ent. The latter is the case when water, for instance, is placed on hydrophobic solid, such as teflon,
polyethylene (PE), or paraffin. The addition of surfactants to water, of course, reduces γL; therefore,
θ will decrease on the introduction of such SASs (Adamson and Gast, 1997; Birdi, 1997, 2003a;
Chattoraj and Birdi, 1984; Eusthopoulos, 1999). The state of a fluid drop under dynamic conditions,
Introduction to Surface and Colloid Chemistry 89
such as evaporation, becomes more complicated (Birdi et al., 1988). However, we are interested
in the spreading behavior of a drop of one liquid when placed on another (immiscible) liquid. The
spreading phenomenon by introducing a quantity, spreading coefficient, Sa/b, is defined as follows
(Adamson and Gast, 1997; Birdi, 2003a; Harkins 1952):
Sa / b = γ a − ( γ b + γ ab ) (1.122)
where
Sa/b is the spreading coefficient for liquid b on liquid a
γa and γb are the respective surface tensions
γab is the IFT between the two liquids
If the value of Sb/a is positive, spreading will take place spontaneously, while if it is negative, liquid
b will rest as a lens on liquid a.
However, the value of γab needs to be considered as the equilibrium value, and therefore, if
one analyzes the system at a nonequilibrium state, then the spreading coefficients would be dif-
ferent. For example, the instantaneous spreading of benzene is observed to give a value of Sa/b as
8.9 dyn/cm, and therefore, benzene spreads on water. On the other hand, as the water phase gets
saturated with benzene with time, the value of the spreading coefficient decreases. This leads to the
formation of lenses. The short-chain hydrocarbons such as hexane and hexene also have positive
initial spreading coefficients and spread to give thicker films. Longer-chain alkanes, on the other
hand, do not spread on water, for example, the Sa/b for n-C16H34 (n-hexadecane)/water is −1.3 dyn/cm
at 25°C. It is also obvious that since impurities can have very drastic effects on the IFTs in Equation
1.122, the value of Sa/b would be expected to vary accordingly (see Table 1.7).
The spreading of a solid substance, for example, cetyl alcohol (C18H38OH), on the surface of
water has been investigated in some detail (Adamson and Gast, 1997; Birdi, 2003a; Gaines, 1966).
Generally, however, the detachment of molecules of the amphiphile into the surface film occurs
only at the periphery of the crystal in contact with the air–water surface. In this system, the diffu-
sion of amphiphile through the bulk water phase is expected to be negligible, because the energy
barrier now includes not only the formation of a hole in the solid, but also the immersion of the
hydrocarbon chain in the water. It is also obvious that the diffusion through the bulk liquid is a
rather slow process. Furthermore, the value of Sa/b would be very sensitive to such impurities as
regards spreading of one liquid upon another.
Another typical example is as follows: the addition of surfactants (detergents) to a fluid dramati-
cally affects its wetting and spreading properties. Thus, many technologies utilize surfactants for
the control of wetting properties (Birdi, 1997). The ability of surfactant molecules to control wetting
arises from their self-assembly at the liquid–vapor, liquid–liquid, solid–liquid, and solid–air interfaces
and the resulting changes in the interfacial energies (Birdi, 1997). These interfacial self-assemblies
exhibit rich structural detail and variation. In the case of oil spills on the seas, these considerations
TABLE 1.7
Calculation of Spreading Coefficients, Sa/b, for Air–Water Interfaces (20°C)
Oil γw/a − γo/a − γo/w = Sa/b Conclusion
n-C16H34 72.8 − 30.0 − 52.1 = −0.3 Will not spread
n-Octane 72.8 − 21.8 − 50.8 = +0.2 Will just spread
n-Octanol 72.8 − 27.5 − 8.5 = +36.8 Will spread
become very important. The treatment of such pollutant systems requires the knowledge of the state
of the oil. The thickness of the oil layer will be dependent on the spreading characteristics. The effect
on ecology (such as birds, plants) will depend on the spreading characteristics. Young’s equation at
liquid1–solid–liquid2 has been investigated for various systems. This is found in such systems where
the liquid1–solid–liquid2 surface tensions meet at a given contact angle. For example, the contact
angle of water drop on teflon is 50° in octane (Chattoraj and Birdi, 1984) (Figure 1.43):
water Teflon octane
In this system, the contact angle, θ, is related to the different surface tensions as follows:
or
( γ s − octane − γ water −s )
cos(θ) = (1.124)
γ octane − water
Experimental value of is found to be 50°. This agrees with the calculated value of θ = 50°, when
using the measured values of (γs–octane − γwater–s)/γoctane–water. This analysis showed that the assump-
tions made in derivation of Young’s equation are valid.
The most important property of a surface (solid or liquid) is its capability of interacting with
other materials (gases, liquids, or solids). All interactions in nature are governed by different kinds
of molecular forces (such as van der Waals, electrostatic, hydrogen bonds, dipole–dipole inter-
actions). Based on various molecular models, the surface tension, γ12, between two phases with
γ1 and γ2, was given as (Adamson and Gast, 1997; Chattoraj and Birdi, 1984):
where Φ12 is related to the interaction forces across the interface. The latter parameter will depend
on the molecular structures of the two phases. In the case of systems such as alkane (or paraf-
fin)–water, there Φ12 is found to be unity. Φ12 is unity since alkane molecule exhibits no hydrogen
bonding property, while water molecules are strongly hydrogen bonded. It is thus found convenient
that in all liquid–solid interfaces, there will be present different (apolar (dispersion) forces + polar
(hydrogen-bonding; electrostatic forces). Hence, all liquids and solids will exhibit γ composed of
different kinds of molecular forces:
Liquid surface tension:
γoctane–water
Octane Water
CA
γs–octane γwater–s
Solid
This means that γS for teflon arises only from dispersion (γSD) forces. On the other hand, a glass
surface shows γS that will be composed of both γS,D and γS,P. Hence, the main difference between
teflon and glass surface will arise from the γS,P component of glass. This criterion has been found
to be of importance in the case of application of adhesives. The adhesive used for glass will need
to bind to solid with both polar and apolar forces. This means that one has to design adhesives with
varying properties, which depend on the surfaces involved.
The values of γSD for different solids as determined from these analyses are given in the following.
In the case of polystyrene surfaces, it was found that the value of sp increased with the treatment of
sulfuric acid (due to the formation of sulfonic groups in the surface) (Birdi, 1997). This gave rise to
increased adhesion of bacteria cells to the surfaces (Birdi, 1981). The asymmetrical forces acting at
surfaces of liquids are much shorter than those expected on solid surface. This is due to the high ener-
gies, which stabilize the solid structures. Therefore, when one considers solid surface, the surface
roughness will need to be considered. Although this is a simple analysis of a complicated system,
one finds that the model as described here can describe many industrial and biological phenomena.
Water–teflon
Water–glass
For instance, let us analyze the state of a drop of water (10 μL) as placed on two different smooth
solid surfaces, for example, teflon and glass. Experiments show that the magnitudes of the contact
angles are different (Figure 1.44).
Since the surface tension of water is the same in the two systems, the difference in contact angle
can only arise due to the difference in surface tension of solids.
The surface tension of liquids can be measured directly. However, this is not possible in the case
of solid surfaces. Experiments show that when a liquid drop is placed on a solid surface, the contact
angle, θ, indicates that the molecules interact across the interface. This data of contact angle have
been used to estimate the surface tension of solids.
92 Handbook of Surface and Colloid Chemistry
Water drop
(a) (b)
FIGURE 1.44 Drop of water on smooth surface of (a) Teflon and (b) glass.
When θ is less than 90°, the surface is wetting (such as water on glass)
When θ is greater than 90°, the surface is nonwetting (such as water on teflon)
It is also known that by the treatment of glass surface by using suitable chemicals, the surface can
be rendered hydrophobic. This is the same technology as used in many utensils, which are treated
with teflon or similar items.
TABLE 1.8
Contact Angles, θ, of Water on Different Solid Surfaces (25°C)
Solid θ
Teflon (PTE) 108
Paraffin wax 110
Polyethylene 95
Graphite 86
AgI 70
Polystyrene 65
Glass 30
Mica 10
Next, one may consider the effect of surface charges on the contact angle of a solid–liquid inter-
face. Metal surfaces will exhibit varying degrees of charges at the surfaces. Biological cells will
exhibit charges that will affect adhesion properties to solid surfaces. However, in some applications,
one may change the surface properties of solids by chemical modifications of the surface. For instance,
polystyrene (PS) has some weak polar groups at the surface. For example, if one treats the PS surface
with H2SO4, one forms sulfonic groups. This leads to values of θ of water lower than 30° (depending
on the time of contact between sulfuric acid and PS surface) (Birdi, 2009). This treatment (or similar)
has been used in many applications where the solid surface is modified to achieve a specific property.
Since only the surface layer (a few molecules deep) is modified, the solid properties do not change.
This analysis shows the significant role of studying the contact angle of surfaces in relation to the
application characteristics. The magnitude of contact angle of water (e.g.) is found to vary depending
on the nature of the solid surface. The magnitude of θ is almost 100 on a waxed surface of a car paint.
The industry strives to create such surfaces to give θ > 150, the so-called superhydrophobic surfaces.
Large θ means that water drops do not wet the car polish and are easily blown off by wind.
In many industrial applications, one is both concerned with smooth as well as rough surfaces.
The analyses of θ on rough surfaces will be somewhat more complicated than on smooth surfaces.
The liquid drop on a rough surface, Figure 1.45, may show the real θ (solid line) or some lower value
(apparent) (dotted line), depending on the orientation of the drop.
However, no matter how rough the surface, the forces will be the same as that between a solid
and a liquid. In other words, at microscale (i.e., atomic scale), the balance of forces at the liquid–
solid and contact angle, surface roughness has no effect. The surface roughness may show contact
angle hysteresis if one makes the drop move, but this will arise from other parameters (e.g., wet-
ting and dewetting) (Birdi, 1993; Raj et al., 2012). A fractal approach has been used to achieve a
better understanding of this phenomenon (Barnsley, 1988; Birdi, 1993; Feder, 1988; Koch, 1993;
Mandelbrot, 1983).
Solid Liquid
FIGURE 1.45 Analysis of contact angle of a liquid drop on a rough solid surface.
94 Handbook of Surface and Colloid Chemistry
TABLE 1.9
Some Typical γcr Values for Solid Surfaces (Estimated from Zisman Plots)
Surface Group γcr
–CF2– 18
–CH2–CH3– 22
Phenyl– 30
Alkyl chlorine– 35
Alkyl hydroxyl 40
In spite of its basic assumptions, Young’s equation has been found to give useful analyses in a
variety of systems. For example, a typical data of cos(θ) versus various liquids on teflon gave an
almost straight line plot. The data can be analyzed by the following relation:
cos(θ) = k1 − k 2 γ L (1.128)
cos(θ) = 1 − k 3 ( γ L − γ cr ) (1.129)
where γcr is the critical value of γL at cos(θ) equal to 1 (i.e., =0). The values of γcr have been reported
for different solids using this procedure. The magnitude of γcr for teflon of 18 mN/m thus suggests
that –CF2– groups exhibit this low surface tension. The value of γcr for –CH2–CH3– alkyl chains
gave a higher value of 22 mN/m than for teflon. Indeed, from experience one also finds that teflon is
a better water-repellent surface than any other material (Starostina et al., 2014). The magnitudes of
γcr for different surfaces are seen to provide much useful information, Table 1.9.
These data show that the molecular groups of different molecules determine the surface
characteristics as related to γcr. In many cases, the surface of a solid may not behave as desired, and
therefore, the surface is treated accordingly, which results in a change of the contact angle of fluids.
The low surface energy polymer (PE) is found to change when treated with flame or corona:
Adsorption takes place spontaneously, which means that ΔGad is negative, which indicates that
ΔHad is negative (exothermic adsorption):
ΔG ad = ΔH ad − TΔSad (1.130)
The adsorption of gas can be of different types. The gas molecule may adsorb on the surface of
a solid (as a kind of condensation process). It may under other circumstances react with the solid
surface (chemical adsorption or chemisorption). In the case of chemoadsorption, one almost expects
a chemical bond formation. On carbon while oxygen adsorbs (or chemisorb), one can desorb CO
or CO2. The experimental data can provide information on the type of adsorption. On porous solid
surfaces, the adsorption may give rise to capillary condensation. This indicates that porous solid
surfaces will exhibit some specific properties. The most adsorption process in industry one finds in
the case of catalytic reactions (e.g., formation of NH3 from N2 and H2).
Further, it is thus apparent that in gas recovery from shale, the desorption of gas (mainly meth-
ane, CH4) will be determined by the surface forces.
The surface of a solid may differ in many ways from its bulk composition. Especially, solids
such as commercial carbon black may contain minor amounts of impurities (such as aromatics,
phenol, and carboxylic acid). This would render surface adsorption characteristics different from
pure carbon. It is therefore essential that in industrial production, one maintains quality control of
surface from different production batches. Otherwise, the surface properties will affect the applica-
tion. Another example, one may add, arises from the behavior of glass powder and its adsorption
character for proteins. It has been found that if glass powder is left exposed to air, then pollutants
from air may cover its surface. This leads to lower adsorption of proteins than on a clean surface.
Silica surface has been considered to exist as O–Si–O as well as hydroxyl groups formed with water
molecules. The orientation of the different groups may also be different at surface. Carbon black
has been reported to possess different kinds of chemical groups on their surfaces (Fiueiredo et al.,
1999). These different groups are as follows: aromatics, phenol, carboxylic, etc. These different
sites can be estimated by comparing the adsorption characteristics of different adsorbents (such as
hexane and toluene). When any clean solid surface is exposed to a gas, the latter may adsorb on
the solid surface to varying degrees (as found from experiments). It has been recognized for many
decades that gas adsorption on solid surfaces does not always stop at a monolayer state. Of course,
more than one layer (multilayer) adsorption will take place only if the pressure is reasonably high.
Experimental data show this when volume of gas adsorbed, vgas, is plotted against pgas.
From experimental analyses, it has been found that five different kinds of adsorption states exist.
These adsorption isotherms were classified based on extensive measurements of vgas versus pgas data.
The pores in a porous solid are found to vary from 2 to 50 nm (micropores).
Macropores are designated for larger than 50 nm. Mesopores are used for 2–50 nm range.
manometer, and the burette is used to control the levels of mercury in the apparatus. Calibration
involves measuring the volumes of the gas (vgas) lines and of the void space. The apparatus, including
the sample, is evacuated, and the sample is heated to remove any previously adsorbed gas. A gas such
as helium gas is usually used for the calibration, since it exhibits very low adsorption on the solid
surface. After helium gas is pushed into the apparatus, a change in volume is used to calibrate the
apparatus and corresponding change in pressure is measured. Different gas (such as nitrogen) is nor-
mally used as the adsorbate if one needs to estimate the surface area of a solid. The gas is cooled by
liquid nitrogen. The tap to the sample bulb is opened, and the drop in pressure is determined. In the
surface area calculations, one uses a value of 0.162 nm2 for the area of an adsorbed nitrogen molecule.
Ns ap
= (1.131)
N sm (1 + ap)
where
p is the pressure
a is dependent on the energy of adsorption
p ⎛ 1 p ⎞
=⎜ + ⎟ (1.132)
N s ⎝ (aN sm ) N sm ⎠
From the experimental data, one can plot p/Ns versus p. The plot will be linear, and the slope is equal
to 1/Nsm. The intersection gives the value of a. Charcoal is found to adsorb 15 mg of N2 as monolayer.
Another example is that of adsorption of N2 on mica surface (at 90 K). The following data were found:
The surface area of the solid can be estimated from the plot of p/Ns versus p. Most data fit this
equation under normal conditions and are therefore widely applied to analyze adsorption process.
Langmuir adsorption is found for the data of nitrogen on mica (at 90K). The data were found to be
as follows:
p = 1–2/Pa
Vs = 24–28 mm3
This shows that the amount of gas adsorbed increases by a factor 28/24 = 1.2 when the gas pressure
increases twofold.
ΠA = k BT (1.133)
In combination with Langmuir equation, one can derive the following relation between Ns and p:
N s = Kp (1.134)
where K is a constant. This is the well-known Henry’s law relation, and it is found to be valid for
most isotherms at low relative pressures. In those situations where the ideal equation (5.16) does not
fit the data, the van der Waals equation type of corrections has been suggested.
The adsorption–desorption process is of interest in many systems (such as cement). The water
vapor may condense in the pores after adsorption under certain conditions. This may be studied by
analyzing the adsorption–desorption data.
Multilayer gas adsorption: As mentioned elsewhere, the fracking process for gas recovery from
shale deposits is a process where adsorbed gas is released by the process. In some systems, adsorp-
tion of gas molecules proceeds to higher levels where multilayers are observed. From data analyses,
one finds that multilayer adsorption takes place, Figure 1.46.
The BET equation has originally been derived for multilayer adsorption only.
Solid
The enthalpy involved in multilayers is related to the differences and was defined by BET
theory as
⎡ (E − E v ) ⎤
E BET = exp ⎢ 1 ⎥ (1.135)
⎣ RT ⎦
where E1 and Ev are enthalpies of desorption. The BET equation thus after modification of the
Langmuir equation becomes
p 1 ⎡ (E BET − 1) p ⎤
= + (1.136)
(N s (po − p)) E BET N sm ⎢⎣ (E BET N sm ) po ⎥⎦
A plot of adsorption data of left-hand side of this equation versus relative pressure (p/po) allows one
to estimate Nsm and EBET.
fraction θ that is filled by one adsorbing solute. It will also be expected that there an adsorption–
desorption process will exist at the surface. The rate of adsorption will be given as
where
kads, kdes are the respective proportionality constants
Cbulk is the bulk solution concentration of solute
Cbulk 1
= Cbulk + (1.140)
θ K eq
and since θ = N/Nm, where N is the number of solute molecules adsorbed per gram of solid, one
can write
Cbulk Cbulk 1
= + (1.141)
N N m (K eq N m )
Thus, measurement of N for a range of concentrations (C) should give a linear plot of Cbulk /N against
Cbulk, where the slope gives the value of Nm and the intercept gives the value of the equilibrium
constant Keq. This model of adsorption was suggested by Irving Langmuir and is referred to as the
Langmuir adsorption isotherm.
A typical adsorption experiment is carried out as follows. The solid sample (e.g., activated char-
coal) is shaken in contact with a solution with known concentration of acetic acid. After equilibrium
is reached (approximately after 24 h), the amount of acetic adsorbed is determined. One can deter-
mine the concentration of acetic acid by titration with NaOH solution.
One may also use solutions of dyes (such as methylene blue), and after adsorption, the amount
of dye in solution is measured by any convenient spectroscopic method (VIS or UV or flourescence
spectroscopy).
In other words, if one needs 1000 m2 area for the adsorption when using charcoal, 1 g of solid will
be required. The estimation of surface area of finely divided solid particles from solution adsorp-
tion studies is subject to many of the same considerations as in the case of gas adsorption, but with
the added complication that larger molecules are involved, whose surface orientation and pore
penetrability may be uncertain. The first condition is that a definite adsorption model is obeyed,
which in practice means that area determination data are valid within the simple Langmuir equa-
tion relation.
In the literature, one has also used fatty acid adsorption for surface area estimation. In these
studies it was assumed that fatty acid molecules adsorbed with 20.5 Å2 area per molecule. This
seems to be true for adsorption on such diverse solids as carbon black and for TiO2. In all of these
cases, the adsorption is probably chemisorption in involving hydrogen bonding or actual salt for-
mation with surface oxygen. As another example, the adsorption of surfactants on polycarbonate
indicated that depending on the surfactant and concentration, the adsorbed molecules might be
lying flat on the surface perpendicular to it, or might form a bilayer. In some cases, one has also
used dyes as adsorbates. This method is appealing because of the ease with which analysis may
be made calorimetrically. The adsorption generally follows the Langmuir equation. An apparent
molecular area of 19.7 Å2 for methane blue on graphon was found (the actual molecular area is
17.5 Å2). The fatty acid adsorption method has been used by many investigators. Surface areas
of oxide powders were estimated from pyridine adsorption studies. The adsorption data followed
the Langmuir equation; the effective molecular area of pyridine is about 24 Å 2 per molecule. In
the literature many different approaches have been proposed to estimate the surface area of a
solid (Adamson and Gast, 1997; Pereira et al., 2003). Surface areas may be estimated from the
exclusion of like-charged ions from a charged interface. This method is intriguing in that no esti-
mation of either site or molecular area is needed. In general, however, surface area determination
by means of solution adsorption studies, while convenient experimentally, may not provide the
most correct information. Nonetheless, if a solution adsorption procedure has been standardized
for a given system, by means of independent checks, it can be very useful for determining rela-
tive areas of a series of similar materials. In all cases, it is also more real as it is what happens
in real life.
Area 55 × 10 4 cm 2 (1016 Å2 ) 55
= = = 127 Å2
Molecule of detergent 0.433 × 102 0 molecule 0.433 Å2
The adsorption isotherms obtained for various detergents showed a characteristic feature that an
equilibrium value was obtained when the concentration of detergent was over CMC. The adsorp-
tion was higher at 40°C than at 20°C. However, the shapes of the adsorption curves were the same
(Birdi, 2003a). Detergents adsorb on solids until CMC, after which no more adsorption is observed.
This shows that after the solid surface is covered by a monolayer of detergent, the adsorption stops
as the system reaches an equilibrium.
Introduction to Surface and Colloid Chemistry 101
Example
One can also calculate the amount of a small molecule, such as pyridine (mol. wt. 100), adsorbed
as a monolayer on charcoal with 1000 m2/g. In the following, these data are delineated:
q imm = E S − E SL (1.142)
where ES and ESL are the surface energies of solid surface and the solid surface in liquid, respec-
tively. The quantity qimm is measured from calorimetry where temperature change is measured after
a solid (in finely divided sate) is immersed in a given liquid. One will expect that when a polar
solid surface is immersed in a polar liquid, there will be larger qimm than if the liquid was an alkane
(nonpolar). Values of qimm for some typical systems are given in Table 1.10.
These data show further that such studies are sensitive to the surface purity of solids.
B Gas
Surface
Liquid
C
TABLE 1.10
Heats of Immersion (qimm) (erg/cm2 at 25°C) of Solids in Liquids
Liquid
Solid Polar (H2O–C2H5OH) Nonpolar (C6H14)
Polar 400–600 100
(TiO2; Al2O3; glass)
Nonpolar 6–30 50–100
(Graphon; Teflon)
Air injection
Electrolytic methods
Vacuum activation
Introduction to Surface and Colloid Chemistry 103
Froth apparatus
Air
Mineral
In a laboratory experiment (Adamson and Gast, 1997), one may use the following recipe. To a 1%
sodium bicarbonate solution, one can add a few grams of sand. Then if one adds some acetic acid
(or vinegar), the bubbles of CO2 produced cling to the sand particles and thus make these float on the
surface. It must be mentioned that in wastewater treatments, the flotation method is one of the most
important procedures. When rocks in crushed state are dispersed in water with suitable surfactants
(also called collectors in industry) to give stable bubbles on aeration, hydrophobic minerals will float
to the surface due to the attachment of bubbles, while the hydrophilic mineral particles will settle to
the bottom. The preferential adsorption of the collector molecules on a mineral makes it hydropho-
bic. Xanthates (alkyl–O–CS2) have been used for flotation of lead and copper.
TSA = At = Ao + Am
R ads = k a pA o (1.143)
⎛ −E ⎞
R des = k b A m exp ⎜ ads ⎟ (1.144)
⎝ RT ⎠
Ns
= A m = A t (1.146)
N sm
N sm
Ns = (1.147)
((ap) / (1 + ap))
where p is the pressure, and a is dependent on the energy of adsorption. Additionally, the heat of
adsorption has been investigated. For example, the amount of Kr adsorbed on AgI increases when
the temperature is decreased from 79 K (0.13 cc/g) to 77 K (0.16 cc/g). This data allow one to esti-
mate the isosteric heat of adsorption (Birdi, 2009; Jaycock and Parfitt, 1981):
d( ln P) q ads
= (1.148)
dt RT 2
Wetting
Adsorption
Desorption
Cleaning processes: garment industry, etc. (car or aeroplane washing)
Introduction to Surface and Colloid Chemistry 105
One finds that wetting characteristics of any solid surface play an important role in all kinds of
different systems. The next most important step is the process of adsorption of substances on solid
surfaces. These phenomena are the crucial steps for all kinds of cleaning processes.
These systems may be as follows: washing, coatings, adhesion, lubrication, oil recovery, etc.
The liquid–solid or liquid1–solid–liquid2 system is both a contact angle (Young’s equation) and
capillary phenomena (Laplace equation). These two parameters are
γ L cos(θ) = ( γ S − γ SL ) (1.149)
and
(2γ L cos(θ))
ΔP = (1.150)
Radius
1.7.1 Oil and Gas Recovery Technology (Oil/Gas Reservoirs and Shale Deposits)
(Fracking Technology and Methane Hydrates) and Surface Forces
Current energy demand (approximately 80 million barrels of oil per day + gas + coal + other forms
of energy sources) is known to have a very high priority as regards sources, such as oil (and gas).
Oil is normally found under high temperatures (80°C) and pressures (200 atm) depending on the
depth of the reservoir. The pressure needed for production depends primarily on the porosity of the
reservoir rock and the viscosity of the oil, among other factors (Figure 1.49). This creates flow of oil
through the rocks, which consists of pores of varying sizes and shapes. Roughly, one may compare
oil flow to the squeezing of water out of a sponge. The capillary pressure is lower in larger pores
than in pores of smaller diameter. Therefore, in the primary oil production, one recovers mostly
from the larger pores of the reservoir. Actually, one may consider it as an advantage that as technol-
ogy develops, one can anticipate old depleted reservoirs to become productive again in future.
The degree of oil recovery from reservoirs is never 100%, and therefore, some (more than 30%)
of the oil in the reservoir remains behind. In order to recover the residual oil, one needs to apply
EOR technology (Birdi, 2010a). Major factors arise from capillary forces, as well as adsorption and
flow hindrances. The same is valid for nonconventional oil reservoirs recovery from shale deposits.
On the other hand, one may consider this as an advantage, since in the long run, as the shortage of
5–10 km
Oil
FIGURE 1.49 Oil reservoir is generally found under large depths, ranging from few hundred meters to over
10 km (pressure increases approximately 100 atm/km depth).
106 Handbook of Surface and Colloid Chemistry
oil supplies comes nearer, one may be forced to develop technologies to recover the residual oil EOR
(Birdi, 2009; Green and Willhite, 1998; Holmberg, 2002).
EOR is a very important research area in energy supply industry.
Studies have shown that by using additives, one can recover the residual oil in the reservoirs.
Especially those additives that lower the IFT give rise to increase in oil production. The latter gives
increased production when water flooding (or similar method) is used.
There are various common EOR processes:
⎛ 1 1 ⎞
ΔPoilwater = 2IFT ⎜ ⎟ (1.151)
⎝ R1 R 2 ⎠
Introduction to Surface and Colloid Chemistry 107
Oil
Water bypass
FIGURE 1.50 Water bypass in an oil reservoir. Reservoir pressure is not high enough to push the oil inside
the narrow pore.
Thus by decreasing the value of IFT (with the help of surface-active agents) (from 50 to 1 mN/m),
the pressure needed for oil recovery would be decreased. In some reservoirs, the addition of CO2
has also increased the oil production. In water flooding process, one uses mixed emulsifiers.
Soluble oils are used in various oil well–treating processes, such as the treatment of water injec-
tion wells to improve water injectivity, to remove water blockage in producing wells. The same is
useful in different cleaning processes in the oil wells. This is known to be effective since water-
in-oil microemulsions are found in these mixtures, and with high viscosity. The micellar solution
is composed essentially of hydrocarbon, aqueous phase, and surfactant sufficient to impart micel-
lar solution characteristics to the emulsion. The hydrocarbon is crude oil or gasoline. Surfactants
are alkyl aryl sulfonates, more commonly known as petroleum sulfonates. These emulsions may
also contain ketones, esters, or alcohols as cosurfactants. In the case of square-shaped pores, one
will have to consider bypass in the corners, which are not found in circular-shaped pores (Birdi
et al., 1988).
Further, it has been known for decades that gas is found in adsorbed state in shale reserves
around the world. Until recently, the recovery of adsorbed gas from shale was very costly as com-
pared to other energy sources (EPA, 2013; Knaus et al., 2010). Shale consists of layered deposits
of river beds. The organic matter found in these layered structure over million years has evolved
into mostly methane gas. Natural gas is very tightly held within the shale structure and not easily
available. In the case of gas and oil recovery from shale, one bores down to the reservoir and pumps
water (under high pressure) with sand and additives (surfactants, salts) added. In this process, small
explosions are created to make fractures, through which shale releases gas/oil for recovery. The
process is specific for each shale reservoir type.
Shale gas fracking technology:
Layer in shale:
= = = = = = = = = = = = = = = = = = = =I = = = =I = = = = = = = = = = =
= = = = = = = = = = = = =I = = = = = =I = = = = = = = = = = = = = = = =
= = = = = = =I = = = = = I = = = = = = = = = = = = = = = = = = = = = = =
= = Gas/Oil = = = Gas/Oil = = = Gas/Oil = = = = = = = = = = = = = = = =
= = = = = = =I = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
= = = = = = = = = = = = = = = = = = =I = = = = = = = = = = = = = = = = =
The layered salt deposits (depicted as = = = =) contain tightly held gas or oil in shale reservoirs.
The recovery of gas is achieved by creating fractures (indicated as I: by inducing controlled
explosions) in the reservoir, thus releasing gas (mainly methane: as adsorbed on shale layers) for
recovery. High-pressure water has also been found to have an appreciable effect on gas release
from shale.
108 Handbook of Surface and Colloid Chemistry
1. Loss by evaporation
2. Loss by sinking to the bottom (as such or in conjunction with solids)
3. Emulsification
Oil spills on oceans are treated by various methods, depending on the region (in warmer seas or
around cold climate). It is also apparent that the oil spill in the Gulf of Mexico will be of completely
different nature than a similar accident near Greenland, for instance. The composition of oil differs
from place to place. The light fluids of oil will evaporate into the air. The oil that has adsorbed on
solid suspension will sink to the bottom. The remaining oil is skimmed off by suitable machines.
In some cases, one also uses surfactants to emulsify the oil and this emulsion sinks to the bottom
slowly. However, no two oil spills are the same because of the variation in oil types, locations, and
weather conditions involved. However, broadly speaking, there are four main methods of response.
1. Leave the oil alone so that it breaks down by natural means. If there is no possibility of the
oil polluting coastal regions or marine industries, the best method is to leave it to disperse
by natural means. A combination of wind, sun, current, and wave action will rapidly dis-
perse and evaporate most oils. Light oils will disperse more quickly than heavy oils. Of
course, the temperature of the sea water will also have an effect on the evaporation process.
2. Contain the spill with booms and collect it from the water surface using skimmer equip-
ment. Spilt oil floats on water and initially forms a slick that is a few millimeters thick.
3. Use dispersants to break up the oil and speed its natural biodegradation. Dispersants act
by reducing the surface tension that inhibits oil and water from mixing. Small droplets
of oil are then formed, which helps to promote a rapid dilution of the oil by water move-
ments. The formation of droplets also increases the oil surface area, thus increasing the
exposure to natural evaporation and bacterial action. Dispersants are most effective when
used within an hour or two of the initial spill. However, they are not appropriate for all oils
and all locations. Successful dispersion of oil through the water column can affect marine
organisms like deep water corals and sea grass. It can also cause oil to be temporarily
accumulated by subtidal seafood. Decisions on whether or not to use dispersants to combat
an oil spill must be made in each individual case. The decision will take into account the
time since the spill, weather conditions, the particular environment involved, and the type
of oil that has been spilt.
4. Introduce biological agents to the spill to hasten biodegradation. Most of the components
of oil washed up along a shoreline can be broken down by bacteria and other microorgan-
isms into harmless substances such as fatty acids and carbon dioxide. This action is called
biodegradation.
The oil spill control technology is very advanced and can operate under very divergent conditions
(from the cold seawaters to the tropic seas). The biggest difference arises from the oil that may con-
tain varying amounts of heavy components (such as tar.).
very highly sophisticated research and development for many decades; some of these are protected
by patents. Mankind is known to have used soaps for cleaning for many centuries.
The procedure for cleaning fabrics or metal surfaces, etc., is primarily to remove dirt, etc.,
from surfaces (fabrics: cotton, wool, synthetics, or mixtures). Second, one must make sure that
dirt does not redeposit after its removal. Dry cleaning is a different process, since here one uses
organic solvents. Dirt is adhering to the fabric through different forces (such as van der Waals
and electrostatic). Some components of dirt are water soluble, and some are water insoluble. The
detergents used are designed specifically for these particular processes by the industry and the
environment. The composition of soaps or detergents is mainly tuned to achieve the following
effects:
1. Water should be able to wet fibers as completely as possible; that is, θ should be less than
10. This is achieved by lowering the surface tension, γ, of the washing water, which thus
lowers the contact angle. The low value of surface tension also makes the washing liquid
to be able to penetrate the pores (if present), since from the Laplace equation, the pressure
needed would be much low.
For example, if the pore size of fabric (such as, any modern microcotton, Gortex) is 0.3 μm,
then it will require a certain pressure (= ΔP = 2γ/R) in order for water to penetrate the
fibers. In the case of water (γ = 72 mN/m) and using a contact angle of 105°, we obtain
2. The detergent then interacts with the dirty soil to start the process of removal from
the fibers and dispersion into the washing water. In order to be able to inhibit the soil
once removed to readsorb on the clean fiber, one uses polyphosphates or similar suit-
able inorganic salts. These salts also increase the pH (around 10) of the washing water.
In some cases, one also uses suitable polymeric anti-redeposition substances (such as
carboxymethyl cellulose).
3. After the fabric is cleaned, one uses special brighteners (fluorescent substances), which
give a bluish haze to the fabric. This enhances the whiteness (by depressing the yellow
tinge). Additionally, these also enhance the color perception. Brighteners as used for cot-
ton are different from those used for synthetic fabrics. Charges on the fabric determine
the adsorption characteristics of the brightener. Hence, the washing process is a series of
well-designed steps, which the industry has provided with much information and the state-
of-the-art technology. Further, all washing technology processes have changed all along
as the demands have changed. Washing machines are designed to operate in conjunction
with soap industry. The mechanical movement and agitation is coordinated with the soap/
detergent characteristics.
Typical compositions of different laundry detergents, shampoos, or dish-washing powders are given
as follows:
It is worth noticing that the aim of detergents in these different formulations is different in each
case. In other words, detergents are today tailor-made for each specific application (Ruiz, 2008).
The detergents in shampoo should give stable foam in order to increase the cleaning effect, at the
same time without any adverse effect on the structure of hair. On the other hand, in the case of laun-
dry detergents in the dish-washing should only give a lower surface tension but almost no foaming
(because foaming would reduce the cleaning effect). Hence in dish-washing machine formulations,
one uses nonionics that are very little soluble in water and thus produce very little (or none) foam.
These are sometimes of type EOEOEOPOPOPO (Ethylene oxide [EO]–propylene oxide [PO]). The
propylene group behaves as apolar, and the oxide group behaves (through hydrogen bonding) as
the polar part. These EOPO types can be tailor-made by combining various ratios of EO:PO in the
surfactant molecule. In some cases, even butylene oxide groups have been used.
Further, soil consists mainly of particulate, greasy matter, etc. Detergents are supposed to keep
the soil suspended in the solution and restrict the redeposition. Tests also show that detergents stabi-
lize the suspensions of carbon or other solids such as manganese oxide in water. This suggests that
detergents adsorb on the particles. Detergents are necessary also to remove the greasy part of soil.
The adsorption of detergents on soil particles is involved in the detergency process.
In these analyses, some assumptions have been made as regards the shape of the drops. The most
accurate data were obtained when weight method was used. Different analyses showed that the rate
of evaporation was linearly dependent on the radius of the drop. Further, the contact angle of water
drop on teflon (i.e., a nonwetting surface) remained constant under evaporation. On the other hand,
contact angle decreased as water drop evaporated on glass (i.e., a wetting surface).
It is almost fascinating to observe how dust or other fine particles remain in suspension in air. In fact,
once in a while, one observes that a particle gets a collision-like thrust. Already in the nineteenth
century (Brown), it was observed under microscope that a small microscopic particle suspended in
water made some erratic movements (as if hit by some other neighboring molecules) (Adamson and
Gast, 1997). This has since been called Brownian motion. The erratic motion arises from the kinetic
movement of the surrounding water molecules. Thus, colloidal particles would remain suspended
in solution through Brownian motion, only if the gravity forces did not drag these to the bottom (or
top). If one throws some sand into air, one finds that the particles fall to the earth rather quickly. On
the other hand, in the case of the talcum particles, one finds that these stay floating in the air for a
long time.
These characteristics will be described here. The size of particles may be considered from the
following data:
This shows that one has a range of size of particles. Actually, in colloids, it is the size of particles
that matters. The stability of such colloidal systems is something that may be compared to whether
the system stays as if energetically stable or it will take up a new state of more stable configuration.
One may very roughly compare this to a bottle, which is stable when standing up but if tilted beyond
a certain angle, it topples and comes to rest (stable state) on its side (Figure 1.51).
A colloidal suspension may be unstable and exhibit separation of particles within a very short
time. Or it may be stable for a very long time, such as over a year or more. And there will thus be
found a metastable state, which would be in between these two. This is an oversimplified example,
but it shows that one should proceed to analyze any colloidal system following these three criteria.
In fact, the most remarkable finding one can mention about colloidal suspensions is that these sys-
tems can exist at all! Especially, one finds that some solid suspensions can be stable for a very long
time. In pharmaceutical applications, one important example is the use of suspension of insulin
in pen injections. The insulin suspension is stable for long enough time for its application, which
provides very accurate dosage to the patient. In fact, there exists a wide variety of pharmaceutical
products that are based on suspended molecules. Nanoparticles have been applied in various phar-
maceutical treatments (Sotiriou et al., 2014).
Criteria of
colloidal
stability
As an example, one may consider the wastewater treatment process (Cherimisinoff, 2002). The
wastewater with colloidal particles is found to be a stable suspension. However, by treating it with
some definite methods (such as pH control, electrolyte concentration, etc.), one can change the sta-
bility of the system.
The wastewater treatment technology is one of the most important areas of surface chemistry
applications. Suspended materials are separated by coagulation and filtered away. Soluble pollutants
are removed by other procedures.
van der Waals forces: In colloidal systems, van der Waals forces play an important role
(Israelachvili, 2011; Adamson and Gast, 1997). When any two particles (neutral or with charges)
come very close to each other, van der Waals forces are strongly dependent on the surrounding
medium; in vacuum, two identical particles always exhibit an attractive force. On the other hand,
if two different particles are present in a medium (in water), then there may be repulsion forces.
This can arise from the particle–solvent adsorption characteristics. One example will be silica
particles in water medium and plastics (as in wastewater treatment). It is important to understand
under which experimental conditions it is possible that colloidal particles remain suspended for
a long time. For example, if paint aggregates in the container, then it is obviously useless. When
solid (inorganic) particles are dispersed in aqueous medium, ions are released in the medium. The
ions released from the surface of the solid are of opposite charge. For example, when glass powder
is mixed in water, one finds that conductivity increases with time.
The presence of same charge on particles in close proximity gives repulsion, which keeps the
particles apart, Figure 1.52. The charged positive–positive (or negative–negative) particles will
show repulsion. On the other hand, the positive–negative particles will attract each other. The ions
distribution will also depend on the concentration of any counterions or coions in the solution. Even
glass when dipped in water exchanges ions with its surroundings. Such phenomena can be easily
investigated by measuring the change in the conductivity of the water.
The force, F12, acting between these opposite charges is given by Coulomb’s law, with charges q1
and q2, separated at a distance R12, in a dielectric medium, De:
(q1q 2 )
F12 = (1.153)
(4πDeεo R12 )
Repulsion
Repulsion
Attraction
FIGURE 1.52 Solid particles with charges (positive–positive [repulsion] or negative–negative [repulsion] or
positive–negative [attraction]).
Introduction to Surface and Colloid Chemistry 113
The force would be attractive between opposite charges while repulsive in the case of similar
charges. Since De of water is comparatively very high (80 units) as compared to De of air (ca. 2), we
will expect very high dissociation in water, while there is hardly any dissociation in air or organic
liquids. Let us consider the F12 for Na+ and Cl− ions (with the charge of 1.6 × 10 −19 C = 4.8 × 10 −10 esu)
in water (De = 74.2 at 37°C), and at a separation (R12) of 1 nm:
In some colloidal systems, one may add large molecules (polymers) that when adsorbed on the solid
particles will impart special kind of stability criteria. It is well known that neutral molecules, such
as alkanes, attract each other mainly through van der Waals forces. van der Waals forces arise from
114 Handbook of Surface and Colloid Chemistry
the rapidly fluctuating dipoles moment (1015 s−1) of a neutral atom, which leads to polarization and
consequently to attraction. This is also called the London potential between two atoms in a vacuum
and is given as
⎛L ⎞
Vvdw = − ⎜ 116 ⎟ (1.155)
⎝R ⎠
where
L11 is a constant, which depends on the polarizability and the energy related to the dispersion frequency
R is the distance between the two atoms
Since the London interactions with other atoms may be neglected as an approximation, the total
interaction for any macroscopic bodies may be estimated by a simple integration.
When two similarly charged colloidal particles, under the influence of the EDL, come close to
each another, they will begin to interact. The potentials will feel each other, and this will lead to
consequences. The charged molecules or particles will be under both van der Waals and electro-
static interaction forces. van der Waals forces that operate at short distance between particles will
give rise to strong attraction forces. This kind of investigation is important in various industries:
Various theoretical models have been presented in the literature as regards the colloidal properties.
One of the most accepted theories was named after the scientists who developed the model, the so-
called DLVO (Derjaguin–Landau–Verwey–Overbeek) theory. The DLVO theory describes that the
stability of a colloidal suspension is mainly dependent on the distance between the particles (Adamson
and Gast, 1997; Birdi, 2009, 2010a; Grodzka and Pomianowski, 2005; Merk et al., 2014; Scheludko,
1966). DLVO theory has been modified in later years, and different versions are found in the current
literature. Electrostatic forces will give rise to repulsion at large distances. This arises from the fact that
electrical charge–charge interactions take place at a large distance of separation. The barrier height
determines the stability with respect to the quantity kT, the kinetic energy. DLVO theory predicts, in
most simple terms, that if the repulsion potential, W, exceeds the attraction potential by a value
W kT (1.156)
W ʺ kT (1.157)
then the suspension will be unstable and it will coagulate. It must be stressed that DLVO theory does
not provide comprehensive analyses. It is basically a very useful tool for such analyses of complicated
systems. Especially, it is a useful guidance theory in any new application or any industrial development.
in the case of neutral surfaces, one has only the van der Waals forces to be considered. This was clearly
seen in the case of micelles, where the addition of small amounts of NaCl to the solution showed
The addition of electrolytes produces a different kind of surface potential curve. This is easily veri-
fied in applications such as washing clothes, etc. Electrostatic and EDL forces are found to play a
very important role in a variety of systems as known in science and engineering (Baimpos et al.,
2014; Birdi, 2010b). It would be useful to consider a specific example, in order to understand these
phenomena. Let us take a surface with positive charge, which is suspended in a solution containing
both positive and negative ions. There will be a definite surface potential, ψo, which decreases to a
value zero as one moves away into solution. It is obvious that the concentration of positive ions will
decrease as one approaches the surface of the positively charged surface (charge–charge r epulsion).
On the other hand, the oppositely charged ions, negative, will be strongly attracted toward the
surface. This gives rise to the so-called Boltzmann distribution:
n − = n oe( z εψ / kT ) (1.158)
n + = n oe −( z εψ / kT ) (1.159)
This shows that positive ions are repelled, while negative ions are attracted to the positively charged
surface. At a reasonable distance from the particle, n+ = n− (as required by the electroneutrality).
In any aqueous solution when an electrolyte, such as NaCl, is present, it dissociates into positive
(Na+) and negative (Cl−) ions. Due to the requirement of electroneutrality (i.e., there must be same
positive and negative ions), each ion is surrounded by an appositively charged ion at some distance.
Obviously, this distance will decrease with an increasing concentration of the added electrolyte. The
expression 1/κ is called the Debye length. As expected, the D–H theory tells us that ions tend to
cluster around the central ion. A fundamental property of the counterion distribution is the thickness
of the ion atmosphere (Birdi, 2010b). This thickness is determined by the quantity Debye length or
Debye radius (1/κ). The magnitude of 1/κ has dimension in centimeters as follows:
1/ 2
⎛εε k T⎞
κ = ⎜ r o B2 ⎟ (1.160)
⎝ 2N A e I ⎠
The values of kB = 1.38 × 10 −23 J/molecule K, e = 4.8 × 10 −10 esu, εr is the dielectric constant, and εo
is the permittivity of free space. Thus, the quantity kBT/e = 25.7 mV at 25°C. As an example, with an
1:1 ion (such as NaCl, KBr) with concentration 0.001 M, one gets the value of 1/κ at 25°C (298 K):
which shows the change in ψ(r) with the distance between particles (r). At a distance 1/κ, the poten-
tial has dropped to ψo. This is accepted to correspond with the thickness of the double layer. This
116 Handbook of Surface and Colloid Chemistry
TABLE 1.11
Magnitude of the Debye Length ((1/k) nm) in Aqueous Salt Solutions
Salt Concentration (M) 1:1 1:2 2:2
0.0001 30.4 17.6 15.2
0.001 9.6 5.55 4.81
0.01 3.04 1.76 1.52
0.1 0.96 0.55 0.48
is the important analyses, since the particle–particle interaction is dependent on the change in ψ(r).
The decrease in ψ(r) at the Debye length is different for different ionic strengths (Table 1.11).
The data in Table 1.11 show values of D–H radius in various salt concentrations. The magnitude
of 1/κ decreases with λ and with the number of charges on the added salt. This means that the thick-
ness of the ion atmosphere around a reference ion will be much compressed with the increasing
value of λ and the magnitude of zion.
A trivalent ion such as Al3+ will compress the double layer to a greater extent in comparison with
a monovalent ion such as Na+. Further, inorganic ions can interact with charge surface in one of two
distinct ways:
1. Nonspecific ion adsorption where these ions have no effect on the isoelectric point IEP.
2. Specific ion adsorption, which gives rise to change in the value of the IEP.
Under those conditions where the magnitude of 1/κ is very small (e.g., in high electrolyte solution),
one can write
ψ = ψ oexp − ( κx ) (1.163)
where x is the distance from the charged colloid.
The value of ψo is found to be 100 mV (in the case of monovalent ions) (=4kBT/ze).
Both the experimental data and the theory show that surface potential, ψ, varies with electrolyte
concentration and the charge on the ions.
These data show
That the surface potential drops to zero at a faster rate if the ion concentration (C) increases
That the surface potential drops faster if the value of z goes from 1 to 2 or larger
In washing powders, for example, one uses multicharged phosphates, etc., to enable dirt particles
staying off the clean fabrics.
1.
Electrophoresis: This system refers to the movement of the colloidal particle under an
applied electric field. In biology, different proteins exhibit different charges and thus can
be separated using this property.
Negatively charged particle moves toward the positive electrode.
Positively charged particle moves toward the negative electrode.
Introduction to Surface and Colloid Chemistry 117
Number of charges
Size and shape of particle
2.
Electro-osmosis: This system is one where a fluid passes next to a charged material. This
is actually the complement of electrophoresis. The pressure needed to make the fluid flow
is called the electro-osmotic pressure.
Fluid movement through a charged material (like earth) gives rise to electroosmotic pressure. This
arises from asymmetrical charge distribution at the liquid–solid interface, which depends on the
magnitude of the surface potential.
3.
Streaming potential: If a fluid is made to flow past a charged surface, then an electric
field is created, which is called streaming potential. This system is thus opposite of the
electroosmosis.
4.
Sedimentation potential: A potential is created when charged particles settle out of a sus-
pension. This gives rise to sedimentation potential, which is the opposite of the streaming
potential. The reason for investigating electrokinetic properties of a system is to determine
the quantity known as the zeta potential.
Fe = qE (1.164)
The resulting electrophoretic migration is countered by forces of friction, Ff, such that the rate of
migration is constant in a constant and homogeneous electric field:
Ff = vfr (1.165)
where
v is the velocity
fr is the frictional coefficient
QE = vfr (1.166)
v q
μ= = (1.167)
E fr
The aforementioned expression can be applied only to ions at a concentration approaching 0 and in
a nonconductive solvent. Polyionic molecules are surrounded by a cloud of counterions that alter
the effective electric field applied on the ions to be separated. This renders the previous expression
a poor approximation of what really happens in an electrophoretic apparatus.
The mobility depends on both the particle properties (e.g., surface charge density and
size) and solution properties (e.g., ionic strength, electric permittivity, and pH). For high
118 Handbook of Surface and Colloid Chemistry
ionic strengths, an approximate expression for the electrophoretic mobility, μe, is given by the
Smoluchowski equation:
εεoη
μe = (1.168)
ζ
where
ε is the dielectric constant of the liquid
εo is the permittivity of free space
η is the viscosity of the liquid
ζ is the zeta potential (i.e., surface potential) of the particle
The most important system for mankind is of course wastewater treatment. However, natural
phenomena occurring around as in rivers, lakes, and oceans are of also of prime interest in ecology
and future life on the earth.
Particles in all kinds of suspensions or dispersions interact with two different kinds of forces (e.g.,
attractive forces and repulsive forces). One observes that lyophobic suspensions (sols) must exhibit a
maximum in repulsion energy in order to have a stable system. The total interaction energy, V(h), is
given as (Adamson and Gast, 1997; Birdi, 2003b, 2009, 2010a; Chattoraj and Birdi, 1984; Chiu and
Ducker, 2014; Gisler et al., 1994):
where Vel and Vvdw are electrostatic repulsion and van der Waals attraction components. Dependence
of the interaction energy V(h) on the distance h between particles has been ascribed to coagulation
rates as follows:
It is found that for large values of h, V(h) is negative (attraction), following the energy of attraction
Vvdw, which decreases more slowly with increasing distance (~1/h2). At short distances (small h),
the positive component Vel (repulsion), which increases exponentially with decreasing h, can
overcompensate Vvdw and reverse the sign of both dV(h)/dh and V(h) in the direction of repulsion.
In accordance with all that has been said before, coagulation will become fast starting from this
concentration. This is therefore the critical concentration, Ccc. In other words, the critical concen-
tration (Ccc) can be estimated from simultaneous solution of the following:
dV(h)
= 0 and V ( h ) = 0 (1.170)
dh
Introduction to Surface and Colloid Chemistry 119
Based on these assumptions, as related to h and C, this becomes (Schulze–Hardy Rule) (for solid
suspensions in water) (Adamson and Gast, 1997; Birdi, 2010a; Scheludko, 1966),
8.7 × 10 −39
Ccc =
Z6A2
where the constant includes (Hamaker constant is approximately 4.2 × 10 −19 J) all quantities
except Z. This shows that critical concentrations of ion to the sixth power of various valencies are
inversely proportion to valency
The flocculation concentrations of mono-, di-, and tri-valent gegen-ions should from this theory
expected as (experimental data shows good agreement with theory):
1 : (1/2)6 : (1/3)6
1 : 1/64 : 1/ 729
It thus becomes obvious that the colloidal stability of charged particles is dependent on
1. Concentration of electrolyte
2. Charge on the ions
3. Size and shape of colloids
4. Viscosity
The critical concentration (critical coagulation concentration) is thus found to depend on the
type of electrolyte used and on the valency of the counterion. It is seen that divalent ions are 60
times as effective as monovalent ions. Trivalent ions are several hundred times more effective
than monovalent ions. However, ions that specifically adsorb (such as surfactants) will exhibit
different behavior. Based on these observations, in the composition of washing powders, one has
used multivalent phosphates (or similar kinds of poly-ions), for instance, to keep the charged
dirt particles from attaching to the fabrics after having been removed off. Another example is the
wastewater treatment, where for coagulation purposes one uses multivalent ions (Cheremisinoff,
2002; Kim and Platt, 2007). Colloidal solutions are characterized by the degree of stability or
instability.
It is thus seen that from DLVO considerations the degree of colloidal stability will be dependent
on following factors:
The attraction force between two particles is proportional to the distance of separation and a
Hamaker constant (specific to the system). The magnitude of H is on the order of 10 −12 erg (Adamson
and Gast, 1997; Birdi, 2003b).
120 Handbook of Surface and Colloid Chemistry
The DLVO theory is thus found useful to predict and estimate the colloidal stability behavior.
Of course, in such systems with many variables, this simplified theory is to be expected to fit all
kinds of systems. In the past decade, much development has taken place as regards measuring the
forces involved in these colloidal systems. In one method, the procedure used is to measure the force
present between two solid surfaces at very low distances (less than a micrometer). The system can
operate under water, and thus, the effect of addictives has been investigated (Israelachvili, 2011).
These data have provided a verification of many aspects of DLVO theory. Recently, atomic force
microscope (AFM) has been used to measure directly these colloidal forces (Birdi, 2003b). Two
particle are brought closer and force (nano-Newton) is measured. In fact, commercially available
apparatus are designed to perform such analysis. The measurements can be carried out in fluids and
under various experimental conditions (such as added electrolytes, pH, etc.).
Coagulation can also be brought by adding suitable substances (coagulants) particularly for a
given system. The latter reduces the effective radius of the colloidal particle and leads to coagula-
tion. Flocculation is a secondary process after coagulation, and this leads to very large particle
(floccs) formation. Experiments show that coagulation takes place when the zeta potential is around
±0.5 mV. Coagulants such as iron and aluminum inorganic salts are effective in most cases. In
wastewater treatment plants, the zeta potential is used to determine the coagulation and flocculation
phenomena. The magnitude of zeta potential can be varied by changing the pH. Most of the solid
material in wastewater is negatively charged.
Methods needed to treat these pollution systems also depend on the quantitative amounts of sub-
stances present. The specific unit used depends on the amounts present. The unit used for trace
amounts, such as benzene, is given in ppm or ppb. The hardness of drinking water (mostly Na,
Ca, Mg) concentration is given as mg/L. The typical values as found are in the range of less than
10 mg/L (soft water) or hard water (over 20 mg/L).
The solids can be removed by filtration and precipitation methods. The precipitation (of charged
particles) is controlled by making the particles flocculate by controlling the pH and ionic strength.
The latter gives rise to decrease in charge–charge repulsion and thus can lead to precipitation and
removal of finely divided suspended solids. It is thus found that the most important factor that effects
zeta potential is pH. Imagine a particle in suspension with a negative zeta potential. If more alkali is
added to this suspension, then the particle will exhibit an increase in negative charge. On the other
hand, if acid is added to the colloidal suspension, then the particle will acquire increasing positive
charge. During this process, the particle will undergo a change from negative charge to zero charge
(where the number of positive charge is equal to negative charge [point-of-zero-charge: PZC]). In
other words, one can control the magnitude and sign of the surface charge by a potential determin-
ing ion. The stability is dependent on the magnitude of electrostatic potential at the surface of the
colloid, ψo. The magnitude of ψo is estimated by using the microelectrophoresis method. When
an electric field is applied across an electrolyte, charged particles suspended in the electrolyte are
attracted toward the electrode of opposite charge. Viscous forces acting on the particles tend to
oppose this movement. When equilibrium is reached between these two opposing forces, the par-
ticles move with constant velocity (Adamson and Gast, 1997; Israelachvili, 2011). In this technique,
the movement (or rather the speed) of a particle is observed under a microscope when subjected to
a given electric field. The field is related to the applied voltage, V, divided by the distance between
the electrodes (in cm). The velocity is dependent on the strength of electric field or voltage gradient,
the dielectric constant of the medium, the viscosity, and zeta potential.
Commercially available electrophoresis instruments are used where the quartz cells designed for any
specific system are available. The magnitude of zeta potential, ζ, is obtained from the following relation:
μη
ζ= (1.173)
εo D
where
η is the viscosity of the solution
εo is the permittivity of the free space
D is the dielectric constant
The velocity of a particle in a unit electric field is related to its electric mobility.
In another application, the magnitude of zeta potential is measured as a function of added counter-
ions. The variation in zeta potential is found to be related to the stability of the colloidal suspension.
The results of a gold colloidal suspension (gold sol) are reported as follows:
These data show that the charge on the colloidal particles changes from negative to zero (when
the particles do not show any movement) to positive, at high counter-ion concentration. This is a
very general picture. Therefore, in wastewater treatment plants, one adds counter-ions until the
movement of colloids is almost zero and thus one can achieve fast flocculation of pollutant particles.
The variation of ζ of silica particles has been investigated as a function of pH (Birdi, 2009). The
dissociation of the surface groups –Si(OH) is involved in this characteristic. Under these operations,
one constantly monitors the zeta potential by using a suitable instrument.
Microscopes have played a very important role in science, which reveals the structures of the
material. The degree of microscopic resolution determines the degree of information. Over many
decades, the ultimate aim has been to be able to see single atoms or molecules. During the end of
the twentieth century, a big surge in the development of very important techniques has become
available for science (nanoscience) and technology (self-assembly structures [micelles; monolayers;
vesicles]; biomolecules; biosensors; surface and colloid chemistry; nanotechnology). In fact, current
literature shows that these developments have no end to this trend as regards the vast expansion in
the sensitivity and level of information.
Typical of all humans, seeing is believing, so the microscope has attracted much interest for
many decades. All these inventions, of course, were basically initiated on the principles laid out
by the telescope (as invented by Galileo) and the light-optical microscope (as invented by Hooke).
Over the years, the magnification and the resolution of microscopes has improved. However, for
the man to understand the nature, the main aim of mankind has been to be able to see atoms or
molecules. This goal has been achieved and the subject as described here will explain the latest
developments that were invented only few decades ago.
The ultimate aim of scientists has always been to be able to see molecules while these are active.
In order to achieve this goal, the microscope should be able to operate under ambient conditions.
Further, all kinds of molecular interactions between a solid and its environment (gas or liquid or
solid), initially, can take place only via the surface molecules of the interface. It is obvious that
when a solid or a liquid interacts with another phase, knowledge of the molecular structures at these
interfaces is of interest. The term surface is generally used in the context of gas–liquid or gas–
solid phase boundaries, while the term interface is used for liquid–liquid or liquid–solid phases.
Furthermore, many fundamental properties of surfaces are characterized by morphology scales on
the order of 1–20 nm (1 nm = 10 −9 m = 10 Å [Angstrom = 10 −8 cm]).
Generally, the basic issues that should be addressed for these different interfaces are as
follows:
• What do the molecules of a solid surface look like, and how are the characteristics of these
different than the bulk molecules? In the case of crystals, one asks about the kinks and
dislocations.
• Adsorption on solid surfaces requires the same information about the structure of the
adsorbates and the adsorption site and configurations.
• Solid–adsorbate interaction energy is also required, as is known from the Hamaker theory.
• Molecular recognition in biological systems (active sites on the surfaces of macromolecule;
antibody–antigen) and biological sensors (enzyme activity; biosensors).
• Self-assembly structures at interfaces.
• Semiconductors and applications.
Most applications of microscopy are found in the case of surfaces and the study of molecules at the
surfaces. Generally, the study of surfaces is dependent on understanding not only the reactivity
Introduction to Surface and Colloid Chemistry 123
of the surface but also the underlying structures that determine that reactivity. Understanding the
effects of different morphologies may lead to a process for the enhancement of a given morphology
and hence to improved reaction selectivities and product yields. Atoms or molecules at the surface
of a solid have fewer neighbors as compared with atoms in the bulk phase, which is analogous to
the liquid surface. Hence, the surface atoms are characterized by an unsaturated, bond-forming
capability and accordingly are quite reactive. Until a decade ago, electron microscopy and some
other similarly sensitive methods provided some information about the interfaces. A few decades
ago, the best electron microscope images of globular proteins were virtually all little more than
shapeless blobs. However, these days, due to relentless technical advances, electron crystallogra-
phy is capable of producing images at resolutions close to those attained by x-ray crystallography
or multidimensional nuclear magnetic resonance (NMR). In order to improve upon some of the
limitations of the electron microscope, newer methods were needed. A few decades ago, a new
procedure for molecular microscopy was invented. The new scanning probe microscopes not
only provide new kind of information than hitherto as known from x-ray diffraction, for example,
but these also open up a new area of research (e.g., nanoscience and nanotechnology).
The basic method of these SPMs (Birdi, 2003b; Chen, 2008) was essentially to be able to move
a tip over the substrate surface with a sensor (probe) with molecular sensitivity (nm) in both longi-
tudinal and vertical directions (Figure 1.53). This may be compared with the act of sensing with a
finger over a surface or more akin to the old-fashioned vinyl record player with a metallic needle (a
probe for converting mechanical vibrations to music sound).
Scanning probe microscopy was invented by Binnig and Rohrer (Nobel prize-1986) (Birdi,
2003b, 2009; Rabe, 1989). Scanning tunneling microscope (STM) was based on scanning a probe
(metallic tip) since it is a sharp tip just above the substrate while monitoring some interactions
between the probe and the surface. The tip is controlled to within 0.1 Å (1 nm).
In SPM, various interactions between the tip and the substrate are as follows:
STM: The tunneling current between a metallic tip and a conducting substrate that are at very
close proximity but not actually touching in physical contact. This is controlled by piezo-
motors in a step-wise method.
AFM: The tip is brought closer to the substrate while van der Waals force is monitored.
At a given force, piezo-motor controls this setting while the surface is scanned in x–y
direction.
FFM: This is a modification of AFM, where force is measured (Birdi, 2003b).
A schematic description of the SPM with the tip (dimension: 0.2 mm) and sample is shown in
Figure 1.54.
The most significant difference between SPM and x-ray diffraction studies is that the former can
be carried out both in air and water (or any other fluid). Corrosion and similar systems have been
investigated by using STM. The tip is covered by a plastic material and this allows one to operate
STM under fluid environment. STM has been used to study the molecules adsorbed on solid sur-
faces. LB films have been extensively investigated by both STM and AFM.
Sensor (SPM)
Sample
FIGURE 1.53 SPMs. Probe can operate both in air and liquid media.
124 Handbook of Surface and Colloid Chemistry
Laser
Diode
Sample
FB
Scanner x and y
FIGURE 1.54 A schematic drawing of the sensor (tip/cantilever/optical/magnetic device) movement over a
substrate in x/y/z direction with nanometer sensitivity (controlled by piezomotor) (at solid–gas or solid–liquid
interface).
Silica
sphere
Colloid system studies by AFM: AFM has allowed scientists to be able to study molecular forces
between molecules at very small (almost molecular size) distances. Further, AFM is a very attrac-
tive and sensitive tool for such measurements. In a recent study, the colloidal force as a function
of pH of SiO2 immersed in aqueous phase was reported using AFM. The force between an SiO2
sphere (ca. 5 mm diameter) and a chromium oxide surface in aqueous phase of sodium phosphate
was measured (pH from 3 to 11). The SiO2 sphere was attached to the AFM sensor as shown in
Figure 1.55.
These data showed that the isoelectric point (IEP) of SiO2 was around pH 2, as expected. The
binding of phosphate ions to chrome surface were also estimated as a function of pH and ionic
strength (Birdi, 2003b). Further, both STM and AFM have been used to investigate corrosion mech-
anisms of metals exposed to aqueous phase. Since both STM and AFM can operate under water,
this gives rise to a variety of possibilities. It is found that one can use other setups where instead
of SiO2 one can use other molecules to investigate surface phenomena. For example, binding of
bacteria to surfaces can be investigated by SPM methods. In the coming decades, one expects much
advances in the industrial applications of SPM. As compared to ordinary microscopes, the SPM
provides 2D and 3D images. The 3D images allow one to see molecules with different diameters
(Birdi, 2003b). SPMs have contributed huge useful information about surface structures in the past
decades (Table 1.10).
solution (or in rivers, lakes, or oceans [mostly due to pollution!]). The formation and structure of
TLF such as in foams or bubbles is the most fascinating phenomena that mankind has studied over
many decades. TLF is thus the thinnest object one can see without the aid of any kind of micro-
scope. One of the most commonly known liquid thin film structure is the soap bubble or bubbles
formed on detergent solutions (as in dish-washing solutions). Everyone has enjoyed the formation of
soap bubbles and the view of the rainbow colors. It may look as if the bubble formation and stabil-
ity have not such a great consequence, but in fact, in everyday life bubbles play an important role
(e.g., from lung function to beer and champagne!). It is common observation that ordinary water
when shaken does not form any bubbles at the surface. On the other hand, all soap and detergent
solutions (shampoo; washing; beer; champagne; sea water) on shaking may form very extensive
bubbles. Further, even though one cannot see or observe the surface layer of a liquid directly, the
TLFs allow one to make some observations that provide much useful information (Adamson and
Gast, 1997; Birdi, 2003a, 2009, 2010a, 2010b; Cox, 2013; Hoher and Addad, 2006; Janaiaud et al.,
2014; MacDowell et al., 2014; Mager and Melosh, 2007).
Air bubble inside liquid phase—at surface bubble detaches and moves up under gravity.
The detergent molecule forms a bilayer in the bubble film. The water in between is the same
as the bulk solution. This may be depicted as follows:
Surface layer of detergent.
Bubble with air and a layer of detergent.
Bubble at the surface forms double layer of detergent with some water in between (TLF)
(varying from 10 to 100 μm).
Since even very minute amounts (around ppm) of SAS give rise to stable bubbles, it has been used to
detect the presence of the latter. A bubble is composed of a TLF with two surfaces, each with a polar
end pointed inward (towards water phase) and the hydrocarbon chains pointing outward. The water
inside the films will move away (due to gravity and due to evaporation), giving rise to the thinning
of the film. Since the thickness approaches the dimensions of the light wavelength, one observes
varying interference colors. The reflected ray will interfere with the incident wavelength. The con-
sequence of this will be that depending on the thickness of the film, one will observe colors. The
most amazing observation is found that the thickness of these films is comparable to the wavelength
of light. Especially, when the thickness of the film is approximately the same as the wavelength of
the light (i.e., between 400 and 1000 nm). The black film is observed when the thickness is the same
as the wavelength of the light (approx. 500–700 nm). Thus, this provides the closest visual observa-
tion of two molecule thick film by eyesight.
One of the most important roles of bubbles one finds is in the food industry (such as ice-cream;
champagne and beer industry). The stability and size of the bubbles determine the taste and the
looks of the product. Especially, in the case of champagne, both the size and the stability of bubbles
have been found to determine the impact of taste and flavor. It has been estimated that in a bottle of
volume 750 mL, there will be about 50 million bubbles (if the average radius of bubbles is 0.1 mm)
(Birdi, 2010a). This is a very rough estimate. However, more accurate estimates have been made
by using photography of bubbles, etc. In the wine industry, much research has been made on the
126 Handbook of Surface and Colloid Chemistry
determination of the factors that control the bubble formation and stability. Another example is ice
creams, where air bubbles are trapped in the frozen material.
The dimension of the detergent molecule is on the order of 15 Å, while the water phase can be many
thousands of molecules. It is worth mentioning that a merely one layer of detergent molecule can
contain a large volume of water phase. Of course, there is a physical limit to as to how much water
can be contained in the film. In thick films, one also observes the movement of water away from the
film, this process is called the thinning of the film. The orientation of detergent molecule in TLF
is such as that the polar group (OO) is pointing toward the water phase and the apolar alkyl part
(CCCCCCCCCC) is pointing toward the air. In the water phase one finds ions in the case of ionic
detergents. However, there will be no ions in the case of nonionic detergent films.
AirCCCCCCCCCCOOWATERWATEROOCCCCCCCCCCAir
AirCCCCCCCCCCOOWATERWATEROOCCCCCCCCCCAir
AirCCCCCCCCCCOOWATERWATEROOCCCCCCCCCCAir
AirCCCCCCCCCCOOWATERWATEROOCCCCCCCCCCAir
AirCCCCCCCCCCOOWATERWATEROOCCCCCCCCCCAir
AirCCCCCCCCCCOOWATERWATEROOCCCCCCCCCCAir
The thickness of water phase can vary from over 100 μm to less than 100 nm (Birdi, 2010a;
Scheludko, 1966). Foams are thermodynamically unstable, since there is a decrease in total free
energy when they collapse. As the thickness of the film decreases to around the wavelength of light
(nm), one starts to observe rainbow colors (arising from interference). The TLF at even smaller in
thickness (50 Å or 5 nm).
However, certain kinds of foams are known to persist for very long periods of time and many
attempts have been made to explain their metastability. The TLF may be regarded as a kind of
condenser. The repulsion between the two surfactant layers, Figure 1.17, will be determined by
the EDL. The effect of added ions to the solution is to make the EDL contract, and this leads
to thin films. It looks black-grey and the thickness is around 50 Å (5 nm), which is almost the
size of the bilayer structure of the detergent (i.e., twice the length [ca. 25 Å] of a typical deter-
gent molecule plus water). Actually, this is a remarkable fact that one can see two molecule
Introduction to Surface and Colloid Chemistry 127
thin structure with a naked eye. Rainbow colors are observed since the light is reflected by the
varying thickness of the TLF of the bubble.
It may not be obvious at first sight, but in beer industry, the foaming behavior is one of the most
important characteristics (Clark et al., 1994). Beer in a bottle is produced under high pressure of
CO2 gas. As soon as one opens a beer bottle, the pressure drops and the gas, CO2, is released, which
gives rise to foaming. In common behavior the foam stays inside the bottle. The foaming is caused
due to the presence of different amphiphilic molecules (fatty acids; lipids; proteins). This foam is
very rich as the liquid film is very thick and contains lots of aqueous phase (such foams are called
kugelschaum). The foam fills the empty space in the bottle, and under normal conditions, it barely
spills out. However, under some abnormal conditions the foam is highly stable and starts to pour out
of the bottle and is considered undesirable. In some reports, one has found that an addition of heavy
metal ions could change the foaming characteristics (Birdi, 1989).
As regards foam stability, it was recognized that the surface tension under film deformation
must always change in such a way as to resist the deforming forces. Thus, tension in the film where
expansion takes place will increase, while it will decrease in the part where contraction takes place.
There is, therefore, a force tending to restore the original condition. The film elasticity, Efilm, has
been defined as
⎛ dγ ⎞
E film = 2A ⎜ ⎟ (1.174)
⎝ dA ⎠
where
Efilm and A are the film elasticity and area of the film
γ is the surface tension of the surface deformed
One of the most important applications is the bubble formation and stability in champagne. The size
and the number of bubbles is found to be important for the impact of taste. The stability has also
much impact on the looks and taste as well. The taste of impact is related to the size and number of
bubbles, as well as how long the bubbles are stable.
Further, the stability of any foam film is related to the kinetics of the thinning of the TLF. As the
thickness reaches a critical value, the film becomes very unstable. It was recognized at a very early
stage by Gibbs that unstable state will conform when the film diverges from bulk system properties.
This thickness was mentioned to be in the range of 50–150 Å. This state is called the black film,
and the random motion of the molecules may easily give rise to a rupture of the TLF. The flow of
liquid is determined by gravity forces. It was found that assuming the fluid in film has the same
viscosity and density, the mean velocity will not exceed 1000 D2film, where latter is distance between
the lamella:
to stabilize the lamina, it should be able to maintain slight differences of tension in its different
regions. It is also therefore clear that a pure liquid, which has constant surface tension, cannot meet
this requirement. The stability of such foams or bubbles has been related to the monomolecular
films structures and stability. For instance, the foam stability has been shown to be related to the
surface elasticity or surface viscosity, ηs, besides other interfacial forces. Studies have indicated
that foam destabilization is related to the packing and orientation of mixed films, which can be
determined from monolayer studies. Since very small (µg = g) amounts are needed for such stud-
ies, the monolayer method has been useful in a very large variety of system studies. It is also worth
mentioning that foam formation from monolayers of amphiphiles constitutes the most fundamen-
tal process in everyday life. The other assemblies such as vesicles and BLM are somewhat more
complicated systems, which are also found to be in equilibrium with monolayers. It is important to
mention that foam does not form in organic liquids, such as methanol, ethanol, etc.
Although the surface potential, ψ, the electrical potential due to the charge on the monolayers,
will clearly affect the actual pressure required to thin the lamella to any given thickness, we shall
assume for the purpose of a simple illustration that 1/k, the mean Debye–Huckel thickness of the
ionic double layer, will influence the ultimate thickness when the liquid film is under a relatively
low pressure. Let us also assume that each ionic atmosphere extends only to a distance 3/k into the
liquid when the film is under a relatively low excess pressure from the gas in the bubbles; this value
corresponds to a repulsion potential of only a few millivolts. Thus at about 1 atm pressure,
6
h film = + 2(monolayer thickness) (1.175)
k
For charged monolayers adsorbed from 10 −3 n-sodium oleate, the final total thickness, hfilm, of the
aqueous layer should thus be on the order of 600 Å (i.e., 6/k or 18 Å). To this value, one needs to add
60 Å (60 × 10 −10 m) for the two films of oriented soap molecules, giving a total of 660 Å. The experi-
mental value is 700 Å (Birdi, 2010a; Scheludko, 1966). The thickness decreases on the addition of
electrolytes, as also suggested by the preceding equation. For instance, the value of hfilm is 120 Å
in the case of 0.1 M NaCl. Addition of a small amount of certain nonionic surface-active agents
(e.g., n-lauryl alcohol, n-decyl glycerol ether, laurylethanolamide, laurylsufanoylamide) to anionic
detergent solutions has been found to stabilize the foam. Measurements have been carried out on
the excess tensions, equilibrium thicknesses and compositions of aqueous foam films stabilized by
either n-decyl methyl sulfoxide or n-decyl trimethyl ammonium-decyl sulfate and containing inor-
ganic electrolytes. It was recognized at a very early stage (Birdi, 2003b, 2009) that the stability of a
liquid film must be greatest if the surface pressure strongly resists deforming forces.
It has been shown (Birdi, 2003b, 2009; Friberg et al., 2003) that there exists a correlation
between foam stability and the elasticity of the film, that is, monolayer elasticity. In order for
elasticity to be large, surface excess must be large. Maximum foam stability has been reported
in systems with fatty acid and alcohol concentrations well below the minimum in γ. Similar
conclusions have been observed with n-C12H25SO4Na (SDS) + n-C12H25OH systems, which give
minimum in γ versus concentration with maximum foam at the minimum point (Chattoraj and
Birdi, 1984), due to mixed monolayer formation. It has been found that SDS + C12H25OH (and
some other additives) make liquid-crystalline structures at the surface. This leads to a stable
foam (and liquid-crystalline structures). In fact, one deliberately uses in technical formulations
SDS with some (less than 1%) C12H25OH to enhance the foaming properties. The foam drain-
age, surface viscosity, and bubble size distributions have been reported for different systems
consisting of detergents and proteins. Foam drainage was investigated by using an incident light
interference microscope technique. The foaming of protein solution is of theoretical interest and
also has a wide application in the food industry (Friberg et al., 2003). Further, in the fermentation
industry where foaming is undesirable, the foam is generally caused by proteins. Since mechani-
cal defoaming is expensive due to the high power required, antifoam agents are generally used.
Introduction to Surface and Colloid Chemistry 129
On the other hand, antifoam agents are not desirable in some of these systems, as for instance in
food products. Further, antifoam agents deteriorate gas dispersion due to increased coalescence
of bubbles. It has been long time that foams are stabilized by proteins, and that these are depen-
dent on pH and electrolyte.
In stage II, the energy of the system is higher than in I, since the system has formed a contact area
(dAc). The energy difference between II and I is γdAc. When the final stage is reached, III, there
will be a decrease in total area by 41% (i.e., the sum of the area of two bubbles is larger than that of
one bubble). This means that system III is at a lower energy state than the initial state I (γdAII–III).
When three bubbles come into contact, the equilibrium angle is 120°. The angle of contact relates
to systems equilibrium state, which is 120° from simple geometrical considerations. If four bubbles
are attached to each other, then the angle will, at equilibrium, be 109°28′.
Many of the industrial wastewaters amenable to clarification by flotation are colloidal in nature,
for example, oil emulsions, pulp and paper wastes, and food processing. For the best results, such
wastes must be coagulated prior to flotation. In fact, flotation is always the last step in the treatment.
In order to aid the flotation effectivity, one uses surfactants. This leads to lower surface tension
and foaming. The latter helps in retaining the particles in the foam under flotation. Further, the
effectiveness of flotation is also dependent on what kind of gas is used to make the bubbles (e.g., air
and CO2). This is related to the dipole characteristics of the gas (CO2 has a stronger dipole than N2).
Bubbles
Air
Mineral
Glass
sinter
Bubbles are blown into the inverted funnel. Inside the funnel the bubble film is transported away
and collected. Since the bubble film consists of
SAS
Water (and salts, etc.)
it is seen that even very minute amounts (less than milligram per liter) of SAS will accumulate at the
bubble surface. As shown previously, it would require a large number of bubbles to remove a gram
of substance. However, since one can blow thousands of bubbles in a very short time, the method is
found to be very feasible.
In the shale fracking process, the wastewater is treated to remove pollutants (Knaus et al., 2010;
Yethiraj and Striolo, 2013). The bubble foam separation can be of importance in this process.
However, one finds that oil and water can be dispersed with the help of suitable emulsifiers (sur-
factants) to give emulsions (Becher, 2001; Birdi, 2008, 2010a; Friberg et al., 2003; James et al.,
2014; Sjoblom, 2008; Takahashi et al., 2014). This is well known in the emulsions as found in the
132 Handbook of Surface and Colloid Chemistry
household, such as milk, mayonnaise, etc. The basic reason is that the IFT between oil and water is
around 50 mN/m, which is high and leads to formation of large oil drops. However, addition of suit-
able emulsifier(s) reduces the value of IFT to very low values (even much less than 1 mN/m). Stable
emulsion formation means that oil drops remain dispersed for a given length of time (up to many
years). The stability and characteristics of these emulsions are related to the area of applications.
Emulsions are a mixture of two (or more) immiscible substances. Some everyday common examples
are milk, butter (fats, water, and salts), margarine, mayonnaise, and skin creams. In butter and mar-
garine, the continuous phase consists of lipids. These lipids surround water droplets (water in oil
emulsion). All emulsions are prepared by some convenient kind of mechanical agitation or mixing.
Remarkably, the natural product, milk (stable oil–water emulsion), is made by the organism without
any agitation but inside the glands (Birdi, 2010a; Bouchoux et al., 2014; Kristensen, 1997). Milk is
the most life-sustaining emulsion providing nutrition and other health-related support properties.
In general, one finds three different kinds of emulsion types:
Emulsions
Microemulsions
LC (and lyotropic LC)
The main criterion of interest in emulsions is that these systems consist of both water and oil. This
may be skin treatment or shoe shine or car polish, etc. In other words, one can apply both these com-
ponents (water and oil, which do not mix) simultaneously. This also allows one to perform functions
that are dependent on properties of water or oil. This, thus, needs information about the IFT and as
well as the solubility characteristics of SAS needed to stabilize the emulsions.
Microemulsions are microstructured mixtures of oil–water–emulsifiers–other substances. LCs
are substances that exhibit special melting characteristics. Further, some mixtures of surfactant–
water–co-surfactant may also exhibit LC (lyotropic crystals) properties. The emulsion technology is
basically thus concerned in preparing mixtures of two immiscible substances:
Oil
Water
by adding suitable surface-active agents (emulgators, co-surfactants, and polymers). The emulsion
technology is thus very varied, since one finds many simple systems (such as skin creams, etc.) and
there are also very complex systems (such as milk, etc.) (Kristensen, 1997).
Oil
Water + E
(a) (b)
Emulsion
night cream. One of the main differences in emulsions is whether oil droplets are dispersed in water
phase or water drops are dispersed in oil phase. One can determine this by measuring the conduc-
tivity, since it is higher for O/W than for W/O emulsion. Another useful property is that O/W will
dissolve water while W/O will not. This, thus, shows that one will choose W/O or O/W depending
on the application area. Especially in the case of skin emulsions the type is of much importance.
Oil-in-water emulsions: The main criterion for O/W emulsion is that if one adds water, then
it will be miscible with the emulsion. Further, after water evaporates, the oil phase will be
left behind. Thus, if one needs oil phase on the substrate (such as skin, metal, and wood),
then one shall use an O/W-type emulsion.
Water-in-oil emulsions: The criterion for W/O emulsion is that latter is miscible with oil.
That means that if one adds the emulsion to some oil, then one obtains a new but diluted
W/O emulsion. In some skin creams, one prefers to use W/O-type emulsion (especially if
one needs an oil-like feeling after its application). On the other hand, if one needs oil free
surface, then O/W emulsion is preferred.
TABLE 1.12
Magnitudes of Interfacial Tensions of Different Organic Liquids against Water (20°C)
Oil Phase IFT (mN/m)
Hexadecane 52
Tetradecane 52
Dodecane 51
Decane 51
Octane 51
Hexane 51
Benzene 35
Toluene 36
CCl4 45
CCl3 32
Oleic acid 16
Octanol 9
Hexanol 7
Butanol 2
W/O emulsions are formed by using HLB values between 3.6 and 6. This suggests that one gen-
erally uses emulgators that are soluble in oil phase.
O/W emulsions need HLB values around 8–18. This HLB criterion is only a very general obser-
vation. However, it must be noticed that HLB values alone do not determine the emulsion type.
Other parameters, such as temperature, properties of oil phase, and electrolytes in aqueous phase,
also affect the emulsion characteristics. HLB values have no relation to the degree of emulsion sta-
bility. HLB values of some surface-active agents are given in Table 1.13.
HLB values decrease as the solubility of surface-active agent decreases in water. The solubility of
cetyl alcohol in water (at 25°C) is less than 1 mg/L. It is thus obvious that in any emulsion cetyl alcohol
will be present mainly in the oil phase, while SDS will be mainly found in the water phase. Empirical
HLB values are found to have significant use in applications in emulsion technology. It was shown that
the HLB is related, in general, to the distribution coefficient, KD, of the emulsifier in oil and water phases:
C(water )
KD = (9.1)
C(oil)
where C(water) and C(oil) are the equilibrium molar concentrations of the emulsifier in water and
oil phases, respectively. Based on this definition of K D, one has found that the magnitude of HLB
for emulsifiers can be estimated as follows:
TABLE 1.13
HLB Values of Different Emulsifiers (Commonly Used in Emulsions)
Emulsifier Solubility in Water HLB Application
Very low solubility 0–2 W/O
Low solubility 4–8 W/O
Soluble 10–12 Wetting agent
High solubility 14–18 O/W
Introduction to Surface and Colloid Chemistry 135
TABLE 1.14
HLB Values of Some Typical Surface-Active Agents
SAA HLB
Na-lauryl sulfate 40
Na-oleate 18
Tween 80 (sorbitan monooleate EO20) 15
Tween 81 (sorbitan monooleate EO6) 10
Ca-dodecylbenzene sulfonate 9
Sorbitan monolaurate 9
Soya lecithin 8
Sorbitan monopalmiate 7
Glycerol monolaurate 5
Sorbitan monostearete 5
Span80 (sorbitan monooleate) 4
Glycerol monostearate 4
Glycerol monooleate 3
Sucrose distearate 3
Cetyl alcohol 1
Oleic acid 1
Based on these thermodynamic relations, one could then suggest the relation between HLB and
emulsion stability and structure. HLB values can also be estimated from the structural groups of
the emulsifier (Table 1.14). This table can be useful in those cases where one needs to estimate HLB
values.
There is found an extensive application of food emulsifiers. It is obvious that these emulsi-
fiers must satisfy special requirements (e.g., toxicity) in order to be useful in food industry.
One determines the toxicity from animal tests. The test determines the amount of a substance
which causes 50% (or more) of the test animals to die (lethal dosage; LD50). It is thus obvious
that food emulsions are subject to much strict controls and limitations (Friberg et al., 2003)
(Table 1.15).
TABLE 1.15
HLB Group Numbers
Group Group number
Hydrophilic
–SO4Na 39
–COOH 21
–COONa 19
Sulfonate 11
Ester 7
–OH 2
Lipophilic
–CH– 0.5
–CH2 0.5
–CH3 0.5
–CH2CH2O– 0.33
136 Handbook of Surface and Colloid Chemistry
Surfactant
rich
Micellar
Emulsions
Water Oil
FIGURE 1.58 Different phase equilibria in a water–surfactant (emulsifier)–oil mixture system (schematic).
Introduction to Surface and Colloid Chemistry 137
these conditions can be qualitatively described as follows. As two oil droplets approach each other,
the negative charge gives rise to repulsive effect. The repulsion will take place within the EDL
region. It can thus be seen that the magnitude of double-layer (EDL) distance will decrease if the
concentration of ions in the water phase increases (Birdi, 2010b). This is due to the fact that the EDL
region decreases. However, in all such cases where two bodies come closer, there exist two different
kinds of forces, which must be considered:
The attraction force arises from van der Waals forces. The kinetic movement will finally determine
whether the total force can maintain contact between two (or more) particles.
1.10.6 Microemulsions
As mentioned earlier, ordinary emulsions as prepared by mixing oil–water–emulsifier are thermo-
dynamically unstable (Birdi, 2009, 2010a; Guo et al., 1992; Meziani et al., 2003; Ohde et al., 2003).
In other words, such an emulsion may be stable over a long length of time, but finally it will separate
into two phases (oil phase and aqueous phase). All such emulsions can be separated into two phases,
that is oil phase and water phase, by centrifugation. These emulsions are opaque, which means that
the dispersed phase (oil or water) is present in the form of large droplets (over μm and thus visible
to the naked eye).
A microemulsion is defined as a thermodynamically stable and a clear isotropic mixture of
water–oil–surfactant–cosurfactant (in most systems it is short-chain alcohol). The co-surfactant is
the fourth component that gives rise to the formation of very small aggregates or drops that make the
microemulsion almost clear. Microemulsions are also therefore characterized as microstructured,
138 Handbook of Surface and Colloid Chemistry
Ratio of E:W
Microemulsion
region
Ratio of C:O
FIGURE 1.59 Four component system: oil (O)–water (W)–emulsifier (E)–co-surfactant (S) (ratio of O:S
versus S:W).
Emulsifier will be found in both these phases. On the other hand, in systems with four compo-
nents, Figure 1.59, consisting of oil–water–detergent–co-surfactant, there exists a region where
clear phase is found. This is the phase region where microemulsions are found. Microemulsions are
thermodynamically stable mixtures. The IFT is almost zero. The size of drops is very small, and
this makes microemulsions look clear. It has also been suggested that microemulsion may consist
of bicontinuous structures. This sounds more plausible in these four-component microemulsion
systems. It has been suggested that microemulsion may be compared to swollen micelles (i.e., if one
solubilizes oil in micelles). In such isotropic mixtures, there exists short range order between the
droplets. Microemulsions have been formed by one of the following procedures (Birdi, 2009, 2010a;
Friberg et al., 2003):
Oil–water mixture is added a surfactant. To this emulsion, one keeps adding a short-chain alco-
hol (with four to six carbon atoms) until a clear mixture (microemulsion) is obtained. It is thus obvi-
ous that microemulsion will exhibit very special properties, quite different from those exhibited by
the ordinary emulsions. The microdrops may be considered as large micelles.
As an example, a very typical microemulsion which has been extensively investigated consists
of a mixture of
The phase region is determined by mixing various mixtures (approximately 20 samples) and allow-
ing the system to reach equilibrium under controlled temperature. From the literature, one finds the
following microemulsion recipe (Birdi, 1982):
Mix 0.0032 mol (0.92 g) SDS (mol. wt. of SDS [C12H25SO4Na] = 288) with 0.08 mol (1.44 g)
water and add 40 ml of C6H6. This mixture is mixed under vigorous stirring, and one gets a creamy
Introduction to Surface and Colloid Chemistry 139
The HLB value of the SAS need be determined (for deciding the O/W or W/O type).
The phase diagram of the water–oil–surfactant (and co-surfactant) needs to be determined.
The effect of temperature is found to be very crucial.
The effect of added electrolytes is of additional importance.
In a recent report, the phase equilibria of a microemulsion were reported (Birdi, 2010a). Phase
behavior of a microemulsion formed with food grade surfactant sodium bis-(2-ethylhexyl) sulfosuc-
cinate (AOT) was studied. Critical microemulsion concentration was deduced from the dependence
of pressure on CP on the concentration of surfactant AOT at constant temperature and water con-
centration. The results show that there are transition points on the CP curve in a very narrow range
of concentration of surfactant AOT. The transition points were changed with the temperature and
water concentration. These phenomena show that lower temperature is suitable to forming micro-
emulsion droplet and the microemulsion with high water concentration is likely to absorb more
surfactants to structure the interface.
1. Increasing the viscosity of the water phase and thereby decreasing the rate of movement of
the oil drops
2. By decreasing IFT and thus the size of the oil drops
140 Handbook of Surface and Colloid Chemistry
The ionized surfactants will stabilize O/W emulsions by imparting surface EDL.
The curvature (i.e., size) of drops in emulsions is of most importance, as regards the stability and
application. The degree of stability of any emulsion is related to the rate of coagulation of two drops
(O/W: oil drops; W/O: water drops).
REFERENCES
Acree, W. E., J. Phys. Chem. Ref. Data, 5, 43, 2004.
Adam, N. K., The Physics and Chemistry of Surfaces, Clarendon Press, Oxford, U.K., 1930.
Adamson, A. W., Gast, A. P., Physical Chemistry of Surfaces, 6th edn., Wiley-Interscience, New York, 1997.
Attard, P., Adv. Chem. Phys., 1, 92, 1996.
Aveyard, R., Hayden, D. A., An Introduction to Principles of Surface Chemistry, Cambridge, London, U.K.,
1973.
Avnir, D., ed., The Fractal Approach to Heterogeneous Chemistry, Wiley, New York, 1989.
Bahadori, A., Energy Fuels, 5695, 25, 2011.
Bai, Z., Lodge, T. P., Langmuir, 8887, 26, 2010.
Baimpos, T., Shrestha, B. R., Raman, S., Valtiner, M., Langmuir, 4322, 30, 2014.
Bakker, G., Kapillaritat und Uberflachenspannung Handbuch der Eksperimentalphysik, 3d vi, Leipzig,
Germany, 1928.
Bancroft, W. D., Applied Colloid Chemistry, McGraw-Hill, New York, 1932.
Barnes, G., Interfacial Science (An Introduction), Oxford University Press, Oxford, U.K., 2011.
Barnsley, M., Fractals Everywhere, Academic Press, London, U.K., 1988.
Becher, P., Emulsions, Theory and Practice, 3rd edn., Oxford University Press, New York, 2001.
Ben-Naim, A., The Hydrophobic Interactions, Plenum Press, New York, 1980.
Birdi, K. S., J. Theor. Biol., 1, 93, 1981.
Birdi, K. S., J. Colloid Polym. Sci., 260, 8, 1982.
Birdi, K. S., Lipid and Biopolymer Monolayers at Liquid Interfaces, Plenum Press, New York, 1989.
Birdi, K. S., Self-Assembly Monolayer (SAM) Structures, Plenum Press, New York, 1999.
Birdi, K. S., Fractals (in Chemistry, Geochemistry and Biophysics), Plenum Press, New York, 1993.
Birdi, K. S., ed., Handbook of Surface & Colloid Chemistry, CRC Press, Boca Raton, FL, 1997.
Birdi, K. S., ed., Handbook of Surface & Colloid Chemistry—CD ROM, 2nd edn., CRC Press, Boca Raton,
FL, 2003a.
Birdi, K. S., Scanning Probe Microscopes (SPM), CRC Press, Boca Raton, FL, 2003b.
Introduction to Surface and Colloid Chemistry 141
Birdi, K. S., ed., Handbook of Surface & Colloid Chemistry, 3rd edn., CRC Press, Boca Raton, FL, 2009.
Birdi, K. S., Surface & Colloid Chemistry, CRC Press, Boca Raton, FL, 2010a.
Birdi, K. S., ed., Interfacial Electrical Phenomena, CRC Press, Boca Raton, FL, 2010b.
Birdi, K. S., Surface Chemistry Essentials, CRC Press, Boca Raton, FL, 2014.
Birdi, K. S., Ben-Naim, A., J. Chem. Soc., Faraday Trans., 741, 77, 1981.
Birdi, K. S., Dalsager, S. U., Backlund, S., J. Chem. Soc., Faraday Trans., 2035, 76(1), 1980.
Birdi, K. S., Fasman, G. D., J. Polym. Sci., 1099, 42, 1973.
Birdi, K. S., Nikolov, A., J. Phys. Chem., 365, 83, 1979.
Birdi, K. S., Vu, D. T., Winter, A., Naargard, A., J. Colloid Polym. Sci., 266, 5, 1988.
Biresaw, G., Mittal, K. L., Surfactants in Tribology, CRC Press, New York, 2008.
Bonzel, H. P., ed., Adsorption on Surfaces & Surface Diffusion of Adsorbates, Springer Publ., New York, 2014.
Bouchoux, A., Qu, P., Bacchin, P., Gesan-Guiziou, G., Langmuir, 22, 30, 2014.
Bouvrais, H., Duelund, L., Ipsen, J. H., Langmuir, 13, 30, 2014.
Boyd, D. A., Adleman, J. R., Goodwin, D. G., Psaltis, D., Anal. Chem., 2452, 80, 2008.
Boys, C. V., Soap Bubbles, Dover Publ., New York, 1959.
Butt, H. J., Surface and Interfacial Forces, Springer Publ., New York, 2010.
Butt, H. J., Graf, K., Kappi, M., Physics and Chemistry of Interfaces, Wiley-VCH, Berlin, Germany, 2013.
Castner, D. G., Ratner, B. D., Surf. Sci., 500, 28, 2002.
Chaibundit, C., Ricardo, N. M. P. S., Crothers, M., Booth, C., Langmuir, 4277, 18, 2002.
Chattoraj, D. K., Birdi, K. S., Adsorption and the Gibbs Surface Excess, Plenum Press, New York, 1984.
Chen, C. J., Introduction to Scanning Tunneling Microscopy, Oxford University Press, Oxford, U.K., 2008.
Chen, H., Ruckenstein, E., Langmuir, 3723, 30, 2014.
Cheremisinoff, N. P., Handbook of Water and Wastewater Treatment Technologies, Butterworth-Heinemann,
Oxford, U.K., 2002.
Chiu, C. W., Ducker, W. A., Langmuir, 140, 30, 2014.
Cini, R., Loglio, G., Ficalbi, A., J. Colloid Interf. Sci., 41, 287, 1972.
Clark, D. C., Wilde, P. J., Marion, D., J. Inst. Brewing, 23, 100, 1994.
Coertjens, S., Moldenaers, P., Vermant, J., Isa, L., Langmuir, 4289, 30, 2014.
Complete Guide to Hydraulic Fracturing for Shale Oil and Natural Gas, Environmental Protection Agency
(EPA), U.S. Geological Survey, Reston, VI, 2013.
Cox, S., ed., Foams: Structure and Dynamics, Oxford University Press, Oxford, U.K., 2013.
Danov, K. D., Kralchevsky, P., Ananthapadmanabhan, K. P., Adv. Colloid Interf. Sci., 17, 206, 2014.
David, R., Neumann, W., Adv. Colloid Interf. Sci., 46, 206, 2014.
Davies, J. T., Rideal, E. K., Interfacial Phenomena, Academic Press, New York, 1963.
De Gennes, P. G., Capillarity and Wetting Phenomena, Springer Publ., New York, 2003.
Defay, R., Prigogine, I., Bellemans, A., Everett, D. H., Surface Tension and Adsorption, Longmans, Green,
London, U.K., 1966.
Diebold, U., Surface Science Reports, Vol. 48, p. 53, Elsevier, Amsterdam, the Netherlands, 2003.
Do, D. D., Adsorption Analysis, Imperial College Press, London, U.K., 1998.
Dukhin, A. S., Goetz, P. J., Ultrasound Characterizing Colloids, Elsevier, New York, 2002
Eusthopoulos, N., Wettability at High Temperatures, Pergamon Press, Oxford, U.K., 1999.
Fainerman, V. B., Miller, R., Mohwald, H., J. Phys. Chem. B, 809, 106, 2002.
Fanum, M., Microemulsions, CRC Press, New York, 2008.
Feder, J., Fractals: Physics of Solids and Liquids, Plenum Press, New York, 1988.
Fendler, J. H., Fendler, E. J., Catalysis in Micellar and Macromolecular Systems, Academic Press, New York,
1975.
Fiueiredo, J. L., Pereira, M. F. R., Freitas, M. M. A., Orfao, J. J. M., Carbon, 37, 1379, 1999.
Franks, F., Water: A Comprehensive Treatise, Plenum Press, New York, 1975.
Fraxedas, J., Water at Interfaces, CRC Press, New York, 2014.
Friberg, S., Larsson, K., Sjoblom, J., Food Emulsions, CRC Press, Boca Raton, FL, 2003.
Frohn, A., Dynamics of Droplets, Springer Publ., New York, 2000.
Fuerstenau, M. C., Miller, J. D., Kuhn, M. C., Chemistry of Flotation, Society of Mining Engineers, New York, 1985.
Gagnon, B. P., Meli, M. V., Langmuir, 179, 30, 2014.
Gaines, G. L., Jr., Insoluble Monolayers at Liquid-Gas Interfaces, Wiley-Interscience, New York, 1966.
Ghatee, M. H., Moosavi, F., Zolghadr, A. R., Ind. Eng. Chem. Res., 12696, 49, 2010.
Ghosh, P., Chem. Eng. Res. Des., 849, 82, 2004.
Gisler, T., Schultz, S. F., Borkovec, M., Sticher, H., Schurtenberger, P., D’Aguanno, B., Klein, R., J. Chem.
Phys., 101, 1, 1994.
142 Handbook of Surface and Colloid Chemistry
Gitis, N., Sivamani, R., Tribology Transactions, Taylor & Francis, New York, 2004.
Global CCS (Carbon Capture and Storage), Institute News-letter, Docklands, VIC, Australia, April 2013.
Golub, T. P., de Keizer, A., Langmuir, 9506, 20, 2004.
Goodrich, F. C., Rusanov, A. I., Sonntag, H., The Modern Theory of Capillarity, Akademie Verlag, Berlin,
Germany, 1981.
Gotch, K., Ind. Eng. Chem. Fund., 287, 13, 1974.
Green, D. W., Willhite, G. P., eds., Enhanced Oil Recovery, Society of Petroleum Engineers, 1998.
Grodzka, J., Pomianowski, A., Physicochem. Probl. Min. Proc., 39, 11, 2005.
Guo, J. S., Sudol, E. D., Vanderhoff, J. W., El-Asser, M. S., Polymer latexes, Chapter 7: ACS Symposium Series,
Vol. 492, p. 99, American Chemical Soc., Washington, DC, 1992.
Hansen, M. C., Hansen Solubility Parameters, Taylor & Francis, New York, 2007.
Harkins, W. D., The Physical Chemistry of Surface Films, Reinhold, New York, 1952.
Harrison, P., Gardiner, C., Sargent, I. L., Extracellular Vesicles in Health & Disease, Pan Standard Publ.,
New York, 2014.
Hiemenz, P. C., Rajagopalan, R., Principles of Colloid and Surface Chemistry, 3rd edn., Marcel Dekker,
New York, 1997.
Hoher, S., Addad, S. C., Rheology of liquid foam, J. Phys. Condens. Matter, 1041, 17, 2006.
Holmberg, K., ed., Handbook of Applied Surface & Colloid Chemistry, John Wiley & Sons Ltd., New York,
2002.
Imae, T., Advanced Chemistry of Monolayers at Interfaces, Academic Press, New York, 2007.
Israelachvili, J., Intermolecular & Surface Forces, 2nd edn., Academic Press, London, U.K., 2011.
James, C., Hatzopoulos, M. H., Yan, C., Smith, G. N., Alexander, S., Rogers, S. E., Eastoe, J., Langmuir,
96, 30, 2014.
Janiaud, E., Bacri, J. C., Andreotti, B., Phys. Fluids, 37101, 26, 2014.
Jasper, J. J., J. Phys. Chem. Ref. Data, 841, 1, 1972.
Jaycock M. K., Parfitt, G. D., Chemistry of Interfaces, John Wiley & Sons, New York, 1981.
Jungwirth, P., Tobias, D. J., Chem. Rev., 1259, 106, 2006.
Kamperman, M., Synytska, A., J. Mater. Chem., 22, 19390, 2012.
Karapetsas, G., Sahu, K. C. Sefiane, K., Matar, O. K., Langmuir, 4310, 30, 2014.
Kelkar, A. D., Herr, D. J. C., Ryan, J. G., Nanoscience & Nanoengineering, CRC Press, New York, 2014.
Kendoush, A. A., Ind. Eng. Chem. Res., 9232, 46, 2007.
Kim, Y. J., Platt, U., Advanced Environmental Monitoring, Springer Publ., The Netherlands, 2007.
Kirmse, K., Morgner, B., Langmuir, 2193, 22, 2006.
Klimpel, R. R., in: Kawatra, S. K., ed., High Efficiency Coal Preparation, Society for Mining, Metallurgy, and
Exploration, Littleton, CO, 1995.
Knaus, E., Killen, J., Bigliarbigi, K., Crawford, P., Chapter 1: ACS Series, Vol. 1032, p. 3, American Chemical
Soc., Washington, DC, 2010.
Koch, J. P., Pharmazie, 48, 643, 1993.
Kolasinski, K. W., Surface Science, John Wiley & Sons, New York, 2008.
Kristensen, D., Milk—Fat globule structures, Ph.D. thesis, School of Pharmacy, Copenhagen, Denmark, 1997.
Krungleviclute, V., Migone, A. D., Yudasaka, M., Ijima, S., J. Phys. Chem. C, 116, 306, 2012.
Kuespert, D., Donohue, M. D., J. Phys. Chem., 99, 4805, 1995.
Kuo, A. T., Chang, C. H., Langmuir, 55, 30, 2014.
Kyte, J., Biophys. Chem., 193, 100, 2003.
Lara, J., Blunt, T., Kotvis, P., Riga, A., Tysoe, W. T., J. Phys. Chem. B, 102, 1703, 1998.
Li, J., Chen, H., Zhou, W., Wu, B., Stoyanov, S. D., Pelan, E. G., Surf. Chem. Colloid, 30, 4223, 2014.
Lin, Y., Boker, A., Skaff, H., Cookson, D., Dinsmore, A. D., Emrick, T., Russell, T. P., Langmuir, 191, 21,
2005.
Livingstone, D., Data Analysis for Chemists, Oxford University Press, New York, 1996.
Lovett, D., Science with Soap Films, Institute of Physics Publishing, Bristol, U.K., 1994.
Lu, J. R., Thomas, R. K., Penfold, J., Adv. Colloid Interf. Sci., 143, 84, 2000.
Lyklema, J., Fundamentals of Interfaces and Colloid Science, III, Liquid-Fluid Interfaces, Academic Press,
San Diego, CA, 2000.
MacDowell, L. G., Benet, J., Katcho, N. A., Palanco, M. G., Adv. Colloid Interf. Sci., 150, 206, 2014.
Mager, M. D., Melosh, N. A., Langmuir, 9369, 23, 2007.
Mandelbrot, B. B., The Fractal Geometry of Nature, W.H. Freeman, New York, 1983.
Masel, R. I., Principles of Adsorption and Reactions, John Wiley & Sons, New York, 1996.
Matijevic, E., ed., Surface and Colloid Science, Vols. 1–9, Wiley-Interscience, New York, 1969.
Introduction to Surface and Colloid Chemistry 143
Taylor, K. C., Automobile catalytic converters, in: Anderson, J. R., Boudart, M., eds., Catalysis (Science and
Technology), Vol. 5, Springer, Berlin, Germany, p. 120, 1984.
Taylor, R. P., Non-linear Dyn. Physiol. Life Sci., 15, 129, 2011.
Todorov, P. D., Kralchevsky, P. A., Denkov, N. D., Broze, G., Mehreteab, A., J. Colloid Interf. Sci., 371, 245,
2007.
Tung, W. S., Daoud, W. A., J. Mater. Chem., 7858, 21, 2011.
Woodson, M., Liu, C. J., Phys. Chem. Phys., 9, 207, 2007.
Wu, X. Z., Ocko, B. M., Sirota, C. B., Sinha, S. K., Deutsch, M., Cao, B. H., Kim, M. W., Science, 261, 1018, 1993.
Xu, W., Leetladhar, R., Kang, Y. T., Coi, C. H., Langmuir, 6032, 29, 2013.
Yan, H., Cui, P., Liu, C. B., Yuan, S. L., Langmuir, 4931, 28, 2012.
Yang, G., Rao, N., Yin, Z., Zhu, D. M., J. Colloid Interf. Sci., 104, 297, 2006.
Yates, J. T., Cambell, C. T., PNAS, 911, 108, 2011.
Yethiraj, A., Striolo, A., J. Phys. Chem. Lett., 687, 4, 2013.
Yoon, R. H., Luttrell, G. H., Adel, G. T., Final Report, Department of Energy, Washington, DC, August 1990.
Yoshida, N., Moroi, Y., Humphry-Baker, R., Gratzel, M., J. Phys. Chem., 3991, 106, 2002.
Yu, H. Z., Soolaman, D. M., Rowe, A. W., Banks, J. T., J. Chem. Phys. Chem., 1035, 5, 2004.
Zana, R., Giant Micelles, CRC Press, New York, 2008.
Zana, R., Langmuir, 2314, 11, 1995.
Zhou, N. F., J. Chem. Educ., 66, 137, 1989.
Zhuravlev, L. T., Colloid Surf., 1, 173, 2000.
Zoller, U., Sustainable Development of Detergents, CRC Press, New York, 2008.
2 Molecular Thermodynamics
of Hydrogen-Bonded Systems
Ioannis Tsivintzelis and Costas G. Panayiotou
CONTENTS
2.1 Introduction........................................................................................................................... 145
2.2 General Approach: The Basics of Thermodynamic Models................................................. 147
2.3 Combinatorial Formalism for Hydrogen Bonding................................................................ 148
2.3.1 Oligomers and Three-Dimensional Networks of Hydrogen-Bonded Molecules...... 150
2.3.2 Dimerization of Acids............................................................................................... 151
2.3.3 Intramolecular Hydrogen Bonding............................................................................ 152
2.4 Equation-of-State Approach: The NRHB Model.................................................................. 154
2.4.1 Applications of the NRHB Model............................................................................. 160
2.4.1.1 Pure Fluid Parameters................................................................................ 160
2.4.1.2 Mixture Parameters.................................................................................... 160
2.4.1.3 Applications................................................................................................ 162
2.4.2 Extension to Interfaces.............................................................................................. 172
2.4.2.1 Point-Thermodynamic Approach to Interfaces........................................... 173
2.4.2.2 Calculations of Interfacial Tension and Interfacial Profiles....................... 178
2.4.3 Partial Solubility Parameters..................................................................................... 186
2.5 Partial Solvation Parameter Approach.................................................................................. 189
2.6 Conclusions............................................................................................................................ 194
References....................................................................................................................................... 194
2.1 INTRODUCTION
The systems of fluids of practical interest to chemists and chemical engineers are as a rule c omplex,
departing significantly from ideal-solution behavior. The systems at interfaces are, in addition,
inhomogeneous exhibiting density gradients and, when multicomponent, they also exhibit com-
position gradients or peculiar composition profiles across the interfaces. Thus, the development of
thermodynamic models for complex fluids applicable over a wide range of conditions remains an
active and fascinating research area.
Recent advances in statistical thermodynamics and better understanding of intra- and intermo-
lecular interactions, thanks to accurate experimental measurements and molecular simulations using
realistic force fields, have contributed significantly to this end. Many of the recent thermodynamic
models based on statistical mechanics are rooted in the pioneering work of Guggenheim [1] and
Flory [2] on lattice models for complex fluids, including polymers. The lattice fluid (LF) theory of
Sanchez and Lacombe [3,4] is probably one of the most widely used and successful lattice models.
Significant improvement in the performance of these models is obtained by accounting explicitly
for the nonrandom distribution of molecular species and free volume and for highly specific forces
between neighboring molecules resulting in hydrogen bonding [5–8].
Nonrandomness is essentially omnipresent in fluids as the molecular species are, as a rule, dis-
tributed nonrandomly in their mixtures. In other words, the local composition in the immediate
145
146 Handbook of Surface and Colloid Chemistry
neighborhood of a molecule is, in general, different from the overall or bulk composition of the
mixture. Even in pure fluids, there is a degree of nonrandomness in the distribution of their func-
tional groups. In fact, modern experimental techniques, such as positron annihilation spectroscopy,
reveal significant nonrandomness in the distribution of free volume throughout the volume of the
pure fluid, even in nonpolar systems [9,10]. One of the principal causes of nonrandomness is of
course hydrogen bonding.
Hydrogen bonding is by itself a subject of remarkable diversity as it is present in and dictates
the behavior of an enormous type of systems including aqueous solutions, systems of biological/
biomedical interest, pharmaceuticals, colloids and surfactants, physical networks and gels, adhe-
sives and pastes, extractives and binders, and polymer alloys and blends. There are many reviews
of the subject in the open literature [11–26] each addressing, usually, one aspect or type of applica-
tion of hydrogen bonding. Because of its multifaceted character, unified approaches of treatment of
hydrogen bonding are particularly useful, especially in areas at the interface of various scientific
branches, such as the colloid and interface science.
For the treatment of hydrogen bonding in associated fluids and mixtures, a variety of different
approaches are popular. We could divide the overwhelming majority of these approaches into two
groups: the association models [5,18,27–31] and the combinatorial models [26,32–35]. Association
models invoke the existence of multimers or association complexes and seek expressions for their
population. Combinatorial models do not invoke the existence of association complexes, but, instead,
they focus on the donor–acceptor contacts and seek combinatorial expressions for the number of
ways of forming hydrogen bonds in systems of given proton–donor and proton–acceptor groups.
Both types of models imply that the molecules tend to be distributed in the system nonrandomly for
more efficient hydrogen-bonding interaction.
Earlier [26], we had compared the two approaches of hydrogen bonding and applied to the
description of phase equilibria and mixture properties of systems of fluids. The key conclusion was
that in the systems where both approaches apply, they prove to be, essentially, equivalent. However,
the combinatorial approach has a much broader field of applications as it can be applied even to
systems forming 3D hydrogen-bonding networks. In the second edition of this handbook [36], we
had presented an updated review on the thermodynamic aspects of hydrogen bonding in pure fluids
and their mixtures by focusing on the combinatorial hydrogen-bonding formalism. A variety of
examples were given in applications ranging from phase equilibria of simple aqueous mixtures to
(hydro)gel swelling, to intramolecular association, and to hydrogen-bonding cooperativity. In the
third edition of this handbook [37], besides hydrogen bonding for the study of bulk phases as well
as interfaces, emphasis was also given to the progress in accounting for nonrandomness in solution
thermodynamics.
The present review is, in a sense, a continuation of all three aforementioned reviews. The ther-
modynamic aspects of hydrogen bonding (combinatorial approach) in pure fluids and their mixtures
will be presented in a way that can be combined with any thermodynamic model. Here, it should
become clear at the outset that, in general, hydrogen bonding makes a contribution only and is not
sufficient for the complete evaluation of the various thermodynamic properties of fluids and their
mixtures. Thus, hydrogen-bonding formalisms are usually combined with thermodynamic models,
which account for all other contributions collectively called physical contributions. For the purposes
of this chapter, we will use two kinds of thermodynamic frameworks: equation-of-state theories
(EoS) and predictive (infinite dilution) activity coefficient models.
The former (EoS approach) is applicable to fluids over an extended range of external condi-
tions encompassing liquids, vapors, gases, supercritical fluids, amorphous and glassy polymers,
homogeneous as well as inhomogeneous systems, complex aqueous systems, associated polymer
mixtures, rubbers, and gels. Thus, the hydrogen-bonding formalism will be combined with two EoS
models, the quasichemical hydrogen-bonding (QCHB) theory [7] and its recent development the
nonrandom hydrogen-bonding (NRHB) model [8,38]. Both models take into account the nonran-
dom distribution of free volume in the system by using Guggenheim’s quasichemical approach [1].
Molecular Thermodynamics of Hydrogen-Bonded Systems 147
The latter (predictive approach) is based on the new concept of partial solvation parameters
(PSPs) [39–43] and focuses on the prediction of thermodynamic properties of complex systems with
minimal or almost no experimental data available.
In what follows, after an exposition of the essentials of the combinatorial hydrogen-bonding
formalism, we will present the EoS approach and characteristic applications to mixtures of
hydrogen-bonded fluids. Subsequently, we will present the extension of the model to interfaces
and to solubility parameters. Finally, we will present the essentials of the novel PSP approach.
Throughout the presentation, examples of calculations in systems of practical interest will be given.
Q( N1, N 2 ,…, N t , T , V ) = ∑Ω ei
−βEi
≅ QAthermalQThermal = QRQE (2.1)
i
The summation in the aforementioned equation spans all microstates i in the phase space of our
system characterized by a corresponding energy level Ei, while Ωi is the degeneration or multiplicity
factor for all microstates corresponding to the energy level Ei, β is the thermal energy factor, which
is equal to 1/kT, k being the Boltzmann’s constant. The random part QR represents the limiting value
of the partition function when all intermolecular interaction energies vanish and the molecules are
distributed randomly throughout the volume of the system. The energetic part, QE , is then a correc-
tion term for the nonrandom distribution of molecular species, which retains the original form of
the partition function and which should become equal to one in the limiting case of zero interaction
energies. In highly nonideal systems of molecules interacting with strong specific forces, such as
hydrogen-bonding forces, the energetic part may be factored, for convenience, into a nonrandom-
ness term, QNR, and a specific or hydrogen-bonding term, QH. Thus, the configurational partition
function can be written alternatively as
In this way, the partition function can be factored into a “physical,” QP, and a “chemical” or
hydrogen-bonding term, QH. This is a general approach.
The total partition function of the system in the P, T ensemble and in its maximum term
approximation is given by
⎛ − PV ⎞
Ψ(T , P,{N k}) = QP (T ,{N k})QH (T ,{N k},{N ij}) exp ⎜ ⎟ (2.3)
⎝ RT ⎠
where
Nij is the number of hydrogen bonds of i–j type
V is the total volume of the system, which is given by
m n
V = VP + VH = VP + ∑∑ N V
i j
0
ij ij (2.4)
Vij0 is the volume change accompanying a i–j hydrogen bond formation, while m and n are the num-
ber of types of proton donors and acceptors, respectively, which are able to hydrogen bond.
148 Handbook of Surface and Colloid Chemistry
G = −kT ln ψ (2.5)
A direct consequence of our approach for the factorization of the partition function is the division of
the Gibbs free energy, G, into a physical term, and a chemical or hydrogen-bonding term:
G = GP + GH (2.6)
For a system at equilibrium, the free energy is at a minimum. One may then write the following
minimization conditions:
⎛ ∂G ⎞
⎜ ∂υ ⎟ = 0 (2.7)
⎝ ⎠T , P ,{N k },{Nij }
⎛ ∂G ⎞
⎜ ⎟ = 0 (2.8)
⎝ ∂N ij ⎠T , P ,{N k },{Nrs }
The first one leads to a hydrogen-bonding contribution to the equation of state of the adopted model,
while the latter one leads to a system of equations that allow the calculation of the number of hydro-
gen bonds, Nij.
At equilibrium, the chemical potential of component k is given by
⎛ ∂G ⎞ ⎛ ∂G ⎞
μk = μk,P + μk,H = ⎜ P ⎟ +⎜ H ⎟ (2.9)
⎝ k ⎠T , P , N j ,υ ,{Nij } ⎝ ∂N k ⎠T , P , N j ,υ ,{Nij }
∂N
In the next section, we will present the combinatorial formalism for hydrogen bonding, which is
used for estimating the hydrogen-bonding contribution, QH, to the partition function. Such approach
is not bound to the models for QP and can be combined with any other appropriate thermodynamic
model able to describe the non-hydrogen-bonding contributions to the thermodynamic properties of
the studied systems. However, as it was already mentioned, the physical term, QP, of the partition
function will be estimated using the NRHB [38] and QCHB [7] formalism in the EoS approach,
presented in Section 2.4.1, and to the PSP [39–43] formalism, presented in Section 2.5, for the pre-
diction of thermodynamic properties of complex systems.
t
N di = ∑ d N (2.10)
k
i
k
k
Molecular Thermodynamics of Hydrogen-Bonded Systems 149
t
N aj = ∑ α N (2.11)
k
k
j k
The potential energy of the system due to hydrogen bonding is in excess of that due to physical
interactions. The total energy ΕΗ of the system due to hydrogen bonding is given by
m n
EH = ∑∑ N E
i j
ij
0
ij (2.12)
where
Nij is the number of hydrogen bonds between donors of type i and acceptors of type j
Eij0 the corresponding hydrogen-bonding energy of the i–j interaction
m n
NH = ∑∑ Ni j
ij (2.13)
What is now required is the number of ways Ω of distributing the Nij bonds among the functional
groups of the system. In order to find the different number of isoenergetic configurations of our
system (number of the different ways of forming or distributing the hydrogen bonds in the system),
we have to find different ways of
Having the number of ways, Ω, of distributing the Nij bonds among the functional groups of the
system, we may write the canonical partition function for hydrogen bonding as following:
m n
⎛ ρ ⎛ − N ij Fij0 ⎞⎞
QH (T ,{N k },{N ij }) = Ω ⎜ ∏∏ exp ⎜⎜ ⎟⎟ ⎟ (2.14)
⎜ i j
rN ⎝ RT ⎠ ⎟⎠
⎝
where
and Pij is the mean field probability that a specific acceptor j will be proximate to a given donor i,
since in order to form a hydrogen bond, the two interacting groups must be proximate. This term
is proportional to the volume of the acceptor group divided by the total system volume; that is,
Pij ~ 1/V. Even spatial proximity does not guarantee that a bond will form. Bond formation requires
that donor and acceptor adopt a unique spatial orientation with respect to one another. Formation of
the bond is also accompanied by a loss of rotational degrees of freedom. Steric considerations will
150 Handbook of Surface and Colloid Chemistry
also come into play in bond formation. In general and in this framework, for a donor i–acceptor j
pair, this probability is given by [26,35]
Sij0 / R
ρ
Pij = e (2.16)
rN
where Sij0 is the entropy loss (intrinsically negative) associated with hydrogen bond formation of an
/rN in Equation 2.16 comes from the estimation of the volume V by the
(i, j) pair. The last term ρ
model framework for the physical term. According to the NRHB model, which will presented next,
r is the number of molecular segments and ρ the reduced density of the system.
Next we will estimate the number of ways Ω of distributing the Nij bonds among the functional
groups of the system for three different cases.
N! N! N! N!
Ω= N11 ! = (2.17)
( N − N11 )! N11 ! ( N − N11 )! N11 ! N10 ! N10 ! N11 !
These arguments, when extended to the general case of multigroup molecules, lead to the following
equation [26,34–35]:
m n m n m n m n
N di ! N ai ! N di ! N aj ! 1
Ω= ∏ ∏ ∏∏ N ij ! = ∏ ∏ ∏∏ N !
i
N i 0 ! N i1 !… N in ! j
N 0 j ! N1 j !… N mj ! i j i
Ni 0 ! j
N0 j ! i j ij
(2.18)
where
Νi0 is the number of free (non-hydrogen-bonded) donor groups of type i
Ν0j is the respective number of free acceptor groups of type j
n
N i 0 = N di − ∑N j
ij (2.19)
m
N 0 j = N aj − ∑N i
ij (2.20)
Molecular Thermodynamics of Hydrogen-Bonded Systems 151
Equation 2.18, for a system with n proton donors and m proton acceptors and if we only account for
intramolecular association, leads to the following equations:
νij ⎛ G0 ⎞
exp ⎜ − ij ⎟ for all i, j (2.21)
=ρ
νi 0ν 0 j ⎝ RT ⎠
or
n m
⎡ ⎤⎡ ⎤ ⎛ G0 ⎞
νij = ⎢νid − ∑ νik ⎥ ⎢ν aj − ∑ν kj ⎥ρ exp ⎜ − ij ⎟ (2.22)
⎢⎣ k ⎥⎦ ⎢⎣ k ⎥⎦ ⎝ RT ⎠
where
N ij Ni 0 N di
νij ≡ νi 0 ≡ νid ≡ (2.23)
rN rN rN
and
This is a system of (m × n) quadratic equations for νij. This system must be solved in combination
with the equation of state, and thus, we are finally left with a system of (m × n + 1) coupled nonlinear
equations for the reduced density, ρ, and the ν .
ij
Furthermore, on the basis of the aforementioned data, we can estimate the hydrogen-bonding
contribution to the chemical potential from the following equation:
m n
μk,H νid ν aj
= rk ν H − ∑ dik ln − ∑ a kj ln (2.25)
RT i
νi 0 j
ν0 j
2.3.2 Dimerization of Acids
A second case that will be investigated here is the dimerization of organic acids [57]. In such sys-
tems, dimers are the overwhelming majority of the association species. Consequently, for simplicity,
we consider dimerization only in order to describe such hydrogen-bonding behavior.
Let Ndm be the number of dimers in the system. Then, the number of ways of selecting these
dimerized molecules out of the N acid molecules is
N!
(2.26)
(2 N dm )!( N − 2 N dm )!
N! N!
Ω= (2 N dm − 1)(2 N dm − 3)1 = (2.27)
(2 N dm )!( N − 2 N dm )! ( N − 2 N dm )! N dm ! 2 N dm
152 Handbook of Surface and Colloid Chemistry
N dm
N! ⎛ ρ ⎞ ⎛ N G ⎞
QH = ⎜ rN ⎟ exp ⎜ − dm dm ⎟ (2.29)
( N − 2 N dm )! N dm ! 2 N dm ⎝ ⎠ ⎝ RT ⎠
The equilibrium number of dimers per mol of segments of acid, νdm, is obtained from the aforemen-
tioned equation through the usual free-energy minimization condition, or
2 + (1 / K dm ) − (1 / K ) + (4 / K
2
dm dm )
ν dm = (2.30)
4r
where
ρ ⎛ −G ⎞
K dm = exp ⎜ dm ⎟ (2.31)
r ⎝ RT ⎠
μH 1
= rν dm − ln (2.32)
RT 1 − 2rν dm
{x!/[1! (x − 1)!]}B = xB ways. Having selected the B acceptor sites of type 2, we must now select,
out of the remaining (xN1 − B) acceptor type 2 population, the N12, which will participate in the
intermolecular bonds. This can be done in (xN1 − B)!/[(xN1 − B − N12)!N12!] ways. The N11 and N12
bonds can be done in N11!·N12! ways, while the B bonds in only one way after having selected,
both, the donor and the acceptor sites in each molecule. Thus, the number of configurations in the
hydrogen bonded system is
B
N1 ! N1 ! ⎛ x! ⎞ ( xN1 − B)!
Ω= ⎜ ⎟ N11 ! N12 !
B! N11 ! N12 ! N10 ! N11 !( N1 − N11 )! ⎝ ( x − 1)! ⎠ ( xN1 − B − N12 )! N12 !
( xN1 − B)!( N1 !)2 x B
= (2.34)
B! N11 ! N12 ! N10 !( N1 − N11 )!( xN1 − B − N12 )!
The number of the three types of hydrogen bonds can be obtained from the following coupled
equations:
B( xN1 − B) ⎛ G0 ⎞
= c exp ⎜ − B ⎟ = K B (2.35)
( xN1 − B − N12 ) N10 x ⎝ kT ⎠
N11
ρ ⎛ G0 ⎞ K
= exp ⎜ − 11 ⎟ = 11 (2.36)
( N1 − N11 ) N10 rN ⎝ kT ⎠ N
N12
ρ ⎛ G0 ⎞ K
= exp ⎜ − 12 ⎟ = 12 (2.37)
( xN1 − B − N12 ) N10 rN ⎝ kT ⎠ N
where
G110 and G120 are the free-energy changes upon formation of the intermolecular hydrogen bonds of
type 1–1 and 1–2, respectively
GB0 is the corresponding free-energy change for the formation of the intramolecular hydrogen
bonds
The free-energy change for the i–j bond can be resolved as follows:
K11
N11 = BN1 (2.40)
K B Nx + B( K11 − K12 )
( NxK B − K12 B)
B=
N
⎡ K12 xN1 − B K11B ⎤
⎢ N1 − B − B xK N
− N1
xNK B + B( K11 − K12 ) ⎥⎦
(2.41)
⎣ B
154 Handbook of Surface and Colloid Chemistry
The last equation contains only the unknown B, and it can be solved numerically by successive
substitutions. The solution for B can then be replaced in Equations 2.39 and 2.40 in order to obtain
N12 and N11, respectively.
Having obtained the numbers of hydrogen bonds in the system, the average number of hydrogen
bonds (intermolecular) per segment, vH, can be calculated by the following equation:
N H N11 + N12
νH = = (2.42)
rN rN
and the hydrogen-bonding contribution to the chemical potential of alkoxyalkanol is given by [58]
The corresponding contribution for the inert compound of the mixture is given by
μ2 H
= rvH (2.44)
RT
where N = N1 + N2 + ⋯ + Nt is the total number of molecules in the system and xi is the mole fraction
of component i. The average interaction energy per segment of molecule i is given by
⎛z⎞
ε*i = ⎜ ⎟ εii (2.46)
⎝2⎠
where εii is the interaction energy per i–i contact. If zqi is the number of external contacts per mol-
ecule i, a geometric characteristic of molecule, i, is its surface-to-volume ratio, si = qi/ri. This ratio
can be estimated for a very large number of different compounds by using the widely used group
contribution UNIFAC model [53,54]. In this way, si is not a fitted parameter but is calculated based
on well-established theories.
In a mixture, parameters r and q are calculated through the following simple mixing rules:
t
r= ∑ x r (2.47)
i
i i
Molecular Thermodynamics of Hydrogen-Bonded Systems 155
t
q= ∑xq
i
i i (2.48)
and so
q
s= (2.49)
r
ri N i xiri
φi = = i = 1, 2,…, t (2.50)
rN r
and
qi N i qi N i φi si φi si
θi = = = = i = 1, 2,…, t (2.51)
t
qN t
s
∑ k
qk N k ∑ k
φ k sk
zN q = zqN + zN 0 (2.52)
V = Nrv* + N 0 v* = N r v* = V * + N 0 v* (2.53)
where the same average segmental volume v* is assigned to an empty site as to an occupied site.
Furthermore, it is assumed that two neighboring empty sites on the quasilattice remain discrete and
do not coalesce. In earlier versions of the theory, v* was assumed to be a pure component parameter
adjusted to experimental data [7,8]. More often, however, v* is assumed to be constant for all fluids
[38,51] and is set equal to 9.75 cm3 mol−1.
We will ignore in this section the contributions to the partition function due to hydrogen bonding.
In the isothermal–isobaric statistical ensemble, the physical term, QP, of the partition function can
then be written as
⎛ E + PV ⎞
QP ( N , P, T ) = QRQNR = Ω R Ω NR exp ⎜ − ⎟ (2.54)
⎝ kT ⎠
where
QR is the combinatorial term of the partition function for a hypothetical system with a random
distribution of the empty and molecular segments
QNR is a correction term for the actual nonrandom distribution of the empty sites
For the random combinatorial term, earlier theories used the Flory expression [2], resulting in the LF
theory [3,4], or the Guggenheim expression [1], resulting in the PV model [51] and QCHB model [7].
156 Handbook of Surface and Colloid Chemistry
The PV model formed the basis for the development of many other lattice models in recent years. In
NRHB [8,38], the generalized Staverman expression [52] is adopted, according to which
t
t Nr !
∏ N ⎛⎜ N ! ⎞⎟ li Ni
r
z /2
ΩR = ∏i
ω Ni
i
i
N !∏ N ! ⎝ N ! ⎠
0
t
i
q
r
(2.55)
where ωi is a characteristic quantity for fluid i that accounts for the flexibility and symmetry of the
molecule. In phase equilibrium calculations, this quantity cancels out. Parameter li is calculated
from the expression
z
li = (ri − qi ) − (ri − 1) (2.56)
2
For the nonrandom correction, the Guggenheim’s quasichemical theory is used [1], as proposed in
the original model [51]:
2
N rr0 ! N 00
0 ⎡
! N r00 / 2 !⎤
( )
QNR = ⎣ ⎦ (2.57)
2
N rr ! N 00 ! ⎡⎣( N r 0 / 2 )!⎤⎦
In this equation, Nrr is the number of external contacts between the segments belonging to m
olecules,
N00 is the number of contacts between the empty sites, and Nr0 is the number of contacts between a
molecular segment and an empty site. Superscript 0 refers to the case of randomly distributed empty
sites.
In the random case, N rr0 takes the form
1 qN z
N rr0 = zqN = qNθr (2.58)
2 N 0 + qN 2
where
q /r
θr = 1 − θ0 = (2.59)
q /r + v − 1
V 1 1
v = = = (2.60)
V* ρ
∑ i
fi
where ρ is the reduced density. The site fractions, f0 and fi, for the empty sites and molecular
segments of component i, respectively, are related by the expression
f0 =
N0 Nr −
= i
ri N i∑= 1− ∑ f (2.61)
i
Nr Nr i
Molecular Thermodynamics of Hydrogen-Bonded Systems 157
In the random case, the number of contacts between empty sites is given by the equation
1 N z
0
N 00 = N 0 z 0 = N 0θ0 (2.62)
2 Nq 2
while the number of contacts between a segment and an empty site is given by
N0 qN
N r00 = zqN = zN 0 = zqNθ0 = zN 0θr (2.63)
Nq Nq
The total number of intersegmental contacts is calculated as the sum of contributions between like
molecules and between unlike molecules:
t t t
N rr0 = ∑i
N ii0 + ∑∑ N
i j >i
0
ij (2.64)
where
z qN z
N ii0 = qi N i i i = qi N i θi θr
2 Nq 2
(2.65)
qN
N ij0 = zqi N i j j = zqi N i θ j θr = zq j N j θi θr i≠ j
Nq
In order to calculate nonrandom distribution of molecular segments and empty sites, appropriate
nonrandom factors Γ are introduced, as explained in the following text. As a result, a total of 2t +
t(t − 1)/2 + 1 contact value expressions becomes
N ii = N ii0Γii i = 1,…, t
N i 0 = N i00Γi 0 i = 1,…, t
The nonrandom factors Γ should obey the following material balance expressions:
∑θ Γ i ij =1 j = 0,1,…, t (2.67)
i =0
Finally, in NRHB model, as well as in previous lattice models, it is assumed that only first-neighbor
segment–segment interactions contribute to the potential energy E of the system. Consequently, for
a mixture, it is
t t t
−E = ∑
i =1
N ii εii + ∑∑ N ε
i =1 j >i
ij ij (2.68)
158 Handbook of Surface and Colloid Chemistry
and
where kij (or ξij) is a binary interaction parameter between species i and j and is fitted to binary
experimental data.
By substituting Equation 2.54 to Equation 2.3, all thermodynamic properties can be derived. At
equilibrium, neglecting the hydrogen-bonding contribution, the reduced density of the system is
obtained from the following minimization condition:
⎛ ∂G ⎞
⎜ ⎟ = 0 (2.70)
⎝ ∂ρ ⎠T , P , N , N10 ,…, Nt 0
t
⎡ l z ⎡ q ⎤ z ⎤
P + T ⎢ ln(1 − ρ
) − ρ
∑ φi i − ln ⎢1 − ρ ⎥ + ln Γ 00 ⎥ = 0 (2.71)
+ ρ
⎢⎣ i =1
ri 2 ⎣ r ⎦ 2 ⎥⎦
where
T RT
T = =
T * ε*
(2.72)
P Pv*
P = =
P* RT *
(2.73)
t t
ε* = ∑ ∑ θ θ ε* (2.74)
i j ij
i =1 j =1
and
The 2t + t(t − 1)/2 + 1 different number of contacts Nij or, equivalently, the nonrandom factors Γij are
calculated from the following set of minimization conditions:
⎛ ∂G ⎞
⎜ ⎟ = 0 i = 0,1,…, t and j = i + 1,…, t (2.76)
⎝ ∂N ij ⎠T , P , N ,ρ
Γii Γ jj ⎛ Δε ⎞
2
= exp ⎜ ij ⎟ i = 0,1,…, t and j = i + 1,…, t (2.77)
Γij ⎝ RT ⎠
Molecular Thermodynamics of Hydrogen-Bonded Systems 159
where
and ε0 = 0. Equations 2.67 and 2.77 form a system of 2t + t(t − 1)/2 + 1 nonlinear algebraic equa-
tions, which is solved analytically for pure fluids and numerically for the case of multicomponent
mixtures. For this purpose, a robust algorithm was proposed by Abusleme and Vera [62] based on
the generalized Newton–Raphson method.
For phase equilibrium calculations, the chemical potential of each component i in the mix-
ture is needed. If we neglect the hydrogen-bonding contribution, it is obtained from the following
expression:
⎛ ∂G ⎞
μi = ⎜ ⎟ (2.79)
⎝ ∂N i ⎠T , P , N j , j ≠i , N10 ,…, Nt 0 ,ρ
By making the appropriate substitutions, Equation 2.79 results in the following expression for a non-
hydrogen-bonding component i (NRHB):
μi φ φ jl j
= ln i − ri ∑r + ri (v − 1) ln(1 − ρ
+ ln ρ )
RT ωiri j j
z ⎡ q ⎤ ⎡ q ⎤
− ri ⎢ v − 1 + i ⎥ ln ⎢1 − ρ
+ ρ
(2.80)
2 ⎣ ri ⎦ ⎣ r ⎥⎦
zqi ⎡ ri ⎤ Pv
qi
+
2 ⎢ ln Γii + q (v − 1) ln Γ 00 ⎥ + ri T − T
⎣ i ⎦ i
The expression for the chemical potential of a pure component, io, can be obtained from
Equation 2.80 by setting ϕi = θi = 1 and the number of components in the summations equal to 1.
It should be pointed out that the presence of hydrogen bonding influences the equation of state
through the minimization of Equation 2.7. As an example, Equation 2.71 becomes
⎡ t
⎛ li ⎞ z ⎡ q ⎤ z ⎤
P + T ⎢ ln(1 − ρ
) − ρ
⎜ ∑ φi − ν H ⎟ − ln ⎢1 − ρ ⎥ + ln Γ 00 ⎥ = 0 (NRHB) (2.81)
+ ρ
⎢⎣ ⎜ ri ⎟ 2 ⎣ r ⎦ 2 ⎥⎦
⎝ i =1 ⎠
In a similar manner, the QCHB model [7] is obtained by using the Guggenheim combinatorial term
[1] instead of the Staverman’s one. In this case, the equation of state becomes
⎡ ⎛ 1⎞ z ⎤
P + T ⎢ ln(1 − ρ) + ρ ⎜ 1 − ⎟ + ln Γ 00 ⎥ = 0 (QCHB) (2.82)
⎣ ⎝ r⎠ 2 ⎦
where
1 ⎛1 ⎞
= − vH ⎟ (2.83)
r ⎜⎝ r ⎠
160 Handbook of Surface and Colloid Chemistry
In the QCHB model, the segmental volume v*i was considered a characteristic property of compo-
nent i. In this case, then, the chemical potential in a binary mixture becomes (QCHB)
μ1 ⎛ r ⎞
= ln φ1 + ⎜ 1 − 1 X12θ22Γrr + r1 (v − 1) ln(1 − ρ
⎟ φ2 + r1ρ )
RT ⎝ r2 ⎠
r1ρ
ρ v∗ r s
Pv
+ ln − Γrr + r1 1 + 1 1 [ln Γrr + (v − 1) ln Γ 00 ]
ω1 T1 T v∗ 2
θ1 1− ρΓrr
− r1 (2.84)
φ1 T
Subscripts h and s in Equation 2.85 denote an enthalpic and an entropic contribution to the interac-
tion energy parameter.
The third scaling constant, v*sp,0, is used for the calculation of the close packed density, ρ* = 1/v*sp ,
as described by the following relation:
TABLE 2.1
NRHB Parameters for the Fluids That Were Used in This Study
Fluid ε*h (J mol−1) ε*s (J mol−1 K−1) * (cm3 g−1)
v sp,0 s References
n-Hexane 3957.1 1.6580 1.27753 0.857 [66]
n-Heptane 4042.0 1.7596 1.25328 0.850 [66]
n-Octane 4105.3 1.8889 1.23687 0.844 [66]
Benzene 5148.5 −0.2889 1.06697 0.753 [66]
Methanol 4202.3 1.52690 1.15899 0.941 [66]
Ethanol 4378.5 0.75100 1.15867 0.903 [66]
1-Propanol 4425.6 0.87240 1.13923 0.881 [66]
1-Butanol 4463.1 1.19110 1.13403 0.867 [66]
2-Butanol 4125.1 1.45711 1.11477 0.867 [68]
1-Hexanol 4522.7 1.64571 1.11545 0.850 [69]
1-Heptanol 4521.2 1.76223 1.12464 0.844 —
1-Octanol 4532.1 1.86863 1.12094 0.839 [63]
1-Nonanol 4576.7 1.87264 1.10261 0.836 —
Water 5336.5 −6.5057 0.97034 0.861 [66]
Carbon dioxide 3468.4 −4.5855 0.79641 0.909 [64]
Acetone 4909.0 −1.15000 1.14300 0.908 [66]
Acetic acid 6198.8 −2.0598 0.92434 0.941 [66]
Dichloromethane 5163.3 −1.3305 0.688 0.881 [66]
Benzoic acid 7228.5 −3.3473 0.8552 0.774 [64]
Acetanilide 7243.2 1.9411 0.8955 0.780 [63]
Phenacetin 6246.3 −0.5539 0.9333 0.827 [63]
[C4mim]Tf2N 5465.37 2.11139 0.66240 0.824 [69]
[C6mim+][Tf2N−] 5389.10 2.63354 0.69229 [69]
1,2 Propylene glycol (PG) 5088.9 1.05260 0.9084 0.903 [68]
Poly(ethylene glycol) 200 < 6273.0 0.71110 0.86620 − MW·7.98·10−08 0.829 [68]
MW < 100,000 g mol−1
Poly(vinyl acetate) 5970.4 2.59194 0.80924 0.825 [68]
Linear polyethylene 7434.9 −7.94296 1.20994 0.800 —
TABLE 2.2
Hydrogen-Bonding (HB) Sites in Each Functional Group
HB Sites
Group Positive (Proton Donors) Negative (Proton Acceptors)
–OH 1 1
Benzene 0 1
HOH 2 2
Benzoic acid 1 1
>NH 1 1
>C=O (in pharmaceuticals) 0 1
–O– 0 2
–O–C=O 0 1
ILs 5 5
162 Handbook of Surface and Colloid Chemistry
TABLE 2.3
Parameters for Hydrogen-Bonding (HB) Interactions
HB Groups Ehb (J mol−1) Shb (J mol−1 K−1) References
–OH⋯OH– (Methanol) −25,100 −26.5 [66]
–OH⋯OH– (Alkanols) −24,000 −27.50 [66]
–OH⋯OH– (Polymers) −22,500 −27.5 [68]
HOH⋯HOH −16,100 −14.70 [66]
Acetic Acid −14,682 −6.9 [66]
Benzoic Acid −16,380 −20.34 [64]
>NH⋯NH< −8,500 −10.25 [63]
>NH⋯O=C< −9,176 −6.94 [63]
>NH⋯–O– −10,900 −6.90 [63]
>NH⋯OH– −18,100 −16.00 [63]
–OH⋯O=C< −16,868 −13.25 [63]
–OH⋯–O– −10,943 −13.75 [63]
–OH⋯–O– (polymers) −14,900 −27.8 [68]
HOH⋯HOH −16,100 −14.70 [66]
HOH⋯benzene Combining rule (Equation 2.88) Combining rule (Equation 2.88) [67]
–OH⋯O=C–O– −22,100 −27.30 [68]
HOH⋯–O– −17,000 −27.0 [68]
HOH⋯OH– Combining rule (Equation 2.87) Combining rule (Equation 2.87) [64]
IL–IL −19,800 −26.5 [69]
IL–water −16,100 −20.0 [69]
estimating the cross-hydrogen-bonding parameters in cases where no experimental data exist. The
following combining rules were used for the cross-interaction between two self-associating groups:
3
Eaa
hb
+ Eββ
hb
⎛ S hb 1/ 3 + Sββ
hb1/ 3
⎞
Eαβ
hb
= , Sαβ
hb
= ⎜ αα ⎟ (2.87)
2 ⎝ 2 ⎠
while, for the cross-interaction between one self-associating and one non-self-associating group, the
combining rules were
Eaa
hb
S hb
Eαβ
hb
= , Sαβ
hb
= aa (2.88)
2 2
2.4.1.3 Applications
In this section, recent applications of the NRHB model to hydrogen-bonding systems will be
reviewed.
NRHB model can be used to predict the monomer fractions (fractions of non-hydrogen-bonded
molecules) of pure self-associating fluids and their mixtures with inert solvents. According to the
formalism of von Solms et al. [70], the fraction of nonbonded molecules for a pure self-associating
fluid is given by the following equation:
k
X1 = ∏X
i =1
K
(2.89)
where XK is the fraction of sites of type K that are not bonded.
Molecular Thermodynamics of Hydrogen-Bonded Systems 163
In the NRHB framework, alkanols are modeled assuming one proton donor and one proton
acceptor site on every molecule. Consequently, the aforementioned relation is transformed to X1 =
XdXa for alkanols, where Xa and Xd are the fraction of free acceptor and donor sites, respectively,
which can easily be calculated through vio and vjo fractions of Equations 2.21 through 2.23.
According to the aforementioned approach, the fraction of nonbonded molecules for alkanols is
obtained from the following equation [67]:
2
X1 = (2.90)
1 + 2ρK + 1 + 4ρK
⎛ −G hb ⎞
K = v * exp ⎜ ij ⎟ (2.91)
⎝ RT ⎠
Some characteristic calculations are presented in Figures 2.1 and 2.2 for pure alkanols and
1-hexanol–hexane mixture. From the first one, it is clear that the model captures the trend of mono-
mer fraction change with alkanol’s molecular weight and temperature. Relatively higher deviations
are observed for ethanol. However, the experimental data for ethanol seem to coincide with the
relevant data for methanol, indicating a phenomenon not accounted by the model, or an experimen-
tal error. Other EoS models result in similar predictions with the NRHB model [76]. According to
Tsivintzelis et al. [76], it is unclear whether it is the ethanol or the methanol data (or none of them)
that should be questioned, since these are the only experimental data that can be found in the
literature.
On the other hand, the NRHB predictions for the monomer fractions of binary mixtures with
alcohols and hydrocarbons seem to be in very good agreement with the experimental data as shown
in Figure 2.2 [67].
0.5
Methanol
Ethanol
0.4 1-Propanol
1-Octanol
NRHB
Monomer fraction
Octanol
0.2
0.1
Methanol
0.0
300 350 400 450 500 550
Temperature (K)
FIGURE 2.1 Experimental monomer fraction data for alcohols and NRHB predictions. Experimental data
are from Refs. [71–73] and were used as they were transformed by von Solms et al. [70]. (NRHB pure fluid
parameters were adapted from Tsivintzelis, I. et al., Ind. Eng. Chem. Res., 47, 5651, 2008.)
164 Handbook of Surface and Colloid Chemistry
1.0
0.8
Monomer fraction
0.6
0.4
0.2
298 K
0.0
0.0 0.1 0.2
x Hexanol
The NRHB model has been mainly applied to model the phase equilibrium of hydrogen-bonding
mixtures. Grenner et al. [66] used the model to describe the vapor–liquid equilibrium (VLE) of 104
binary mixtures of fluids with different polarity, including hydrogen-bonding systems. Furthermore,
Tsivintzelis et al. [67] used the model to describe the liquid–liquid equilibrium (LLE) of care-
fully selected binary systems, which include water–hydrocarbon, 1-alkanol–hydrocarbon, water–1-
alkanol, and glycol–hydrocarbon mixtures. Overall, the model resulted in satisfactory correlations
for the investigated systems. In both aforementioned studies [66,67], an extended parameterization
of the model was performed and tables with pure fluid and binary interaction NRHB parameters
are presented. Some characteristic calculations using such pure fluid and binary parameters are
shown in Figures 2.3 through 2.6. The phase behavior, which includes both VLE and LLE, of the
methanol–n-hexane system, is presented in Figure 2.3. The mutual solubility of water–n-hexane
and water–benzene is presented in Figure 2.4. Since n-hexane and benzene have similar molecu-
lar weight, the great difference in their solubility in water (benzene presents almost two orders of
magnitude higher solubility than n-hexane) originates from the different association behavior. Thus,
the higher solubility of benzene is attributed to the relatively strong specific interactions that occur
between water proton donors and the π-electrons of the benzene aromatic ring. Accounting for such
cross-association, the NRHB model is able to capture the different behavior of such molecules and,
as shown in Figure 2.4, results in satisfactory correlations. However, the model is not able to capture
the minimum in the hydrocarbon solubility, which is a very difficult task for all the EoS models [67].
Another characteristic calculation is presented in Figure 2.5 for the LLE of water–1,2-propylene
glycol. This system also presents a complicated hydrogen-bonding behavior, in which both self- and
cross-association interactions occur between the molecules of both fluids. In Figure 2.6, the VLE
of the acetic acid–dichloromethane system is presented. Here, the dimerization of acid molecules
is taken into account using the analysis of Section 2.3.2 and the hydrogen-bonding parameters of
Grenner et al. [66].
Recently, the NRHB model was applied to mixtures with pharmaceuticals [63,64]. Such
systems usually present a complicated hydrogen-bonding behavior, which strongly influences
drug’s solubility. A characteristic example is shown in Figure 2.7, where the solubility of two
Molecular Thermodynamics of Hydrogen-Bonded Systems 165
VLE
340
320
Temperature (K)
300
280
LLE
260
240
0.0 0.2 0.4 0.6 0.8 1.0
Methanol mole fraction
FIGURE 2.3 Methanol–n-hexane LLE and VLE. Experimental data [77,78] (points) and model correlations
(lines) using a temperature-independent binary interaction parameter, kij = 0.0304. (Pure fluid parameters
adapted from Tsivintzelis, I. et al., Ind. Eng. Chem. Res., 47, 5651, 2008.)
10–1
10–2
10–3
Mole fraction
10–4
10–5
Hexane in water
10–6 Water in hexane
Benzene in water
10–7 Water in benzene
NRHB correlations
10–8
280 320 360 400 440 480 520
Temperature (K)
FIGURE 2.4 Hexane–water and benzene–water LLE. Experimental data [79] (points) and model calcula-
tions (lines) using pure fluid and binary parameters. (From Tsivintzelis, I. et al., Ind. Eng. Chem. Res., 47,
5651, 2008.)
pharmaceutical molecules in 1-octanol is presented. Both molecules have more than one functional
groups that are able to hydrogen bond with the solvent molecules, and their solubility is defined by
the crystallization behavior and interplay of self- and cross-association interactions in the solution.
Accounting explicitly for all possible association interactions results in very satisfactory calculations
with the NRHB model, as shown in Figure 2.7.
However, the advantage of using an EoS model, against the popular activity coefficient models
that are usually applied in pharmaceutical systems, is the ability to model systems at high pressures.
166 Handbook of Surface and Colloid Chemistry
0.01
Mole fraction
1E–3
Glycol-rich phase
Heptane-rich phase
NRHB predictions
NRHB correlations
kij = 0.0166
1E–4
305 310 315 320 325 330 335 340 345
Temperatue (K)
FIGURE 2.5 1,2-Propylene glycol (PG)–heptane LLE. Experimental data [80] (points) and model calcula-
tions (lines) using pure fluid and binary parameters. (From Tsivintzelis, I. et al., Ind. Eng. Chem. Res., 47,
5651, 2008.)
400
P = 1 atm
380
Temperature (K)
360
340
320
FIGURE 2.6 Acetic acid–dichloromethane VLE. Experimental data [81] (points) and NRHB calculations
(lines) using pure fluid and binary parameters. (From Grenner, A. et al., Ind. Eng. Chem. Res., 47, 5636,
2008.)
Such pressures are not used in the great majority of industrial pharmaceutical applications. Recently,
however, many attractive alternative processes have been suggested, including the processing of
pharmaceuticals with supercritical fluids. The NRHB correlations for the solubility of benzoic acid
in supercritical CO2 and supercritical ethane are presented in Figure 2.8. All NRHB results for
systems with pharmaceuticals that are presented in Figures 2.7 and 2.8 were obtained using the pure
fluid and binary parameters of Refs. [63,64].
Molecular Thermodynamics of Hydrogen-Bonded Systems 167
0.28
0.24
0.20
Solute mole fraction
O
Solubility of acetanilide
0.16
HN CH3
O
0.12 HN CH3
Solubility of phenacetin
0.08
O
0.04
CH3
0.00
300 306 312
Temperature (K)
Exp. data
NRHB predictions (kij = 0)
NRHB correlations (kij = –0.0047
for acetanilide and kij = +0.0015
for phenacetine)
FIGURE 2.7 Solubility of acetanilide and phenacetin in 1-octanol. Experimental data [82] (points) and
NRHB calculations (lines) using pure fluid and binary parameters. (From Tsivintzelis, I. et al., AIChE J., 55,
756, 2009.)
0.01
Solute mole fraction
1E–3
5 10 15 20 25 30 35
Pressure (MPa)
FIGURE 2.8 Solubility of benzoic acid in supercritical CO2 and supercritical ethane at 318 K. Experimental
data [83] (points) and NRHB correlations (lines) with kij = −0.04417 for CO2 and −0.06974 for ethane. (Pure
fluid parameters adapted from Tsivintzelis, I. et al., AIChE J., 55, 756, 2009.)
168 Handbook of Surface and Colloid Chemistry
0.0375
0.0300 353.2 K
Pressure (MPa)
0.0225
333.2 K
0.0150
0.0075 313.2 K
0.0000
0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200
Solvent weight fraction
FIGURE 2.9 2-Butanol–poly(vinyl acetate) VLE. Experimental data [84] (points), NRHB predictions
(dotted lines, kij = 0), and correlations (solid lines, kij = 0.01372). (Pure fluid parameters were adopted from
Tsivintzelis and Kontogeorgis [68]).
0.0200
0.0175
333.1 K
0.0150
0.0125
Pressure (MPa)
0.0075
0.0050
313.1 K
0.0025
0.0000
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
Water weight fraction
FIGURE 2.10 Water–poly(ethylene glycol) VLE. Experimental data [85] (points), NRHB predictions
(dotted lines, kij = 0), and correlations (solid lines, kij = 0.01238). (Pure fluid parameters were adopted from
Tsivintzelis and Kontogeorgis [68]).
However, hydrogen bonding can influence the thermodynamic properties of systems with
macromolecules as well. Figures 2.9 and 2.10 present the VLE of polymer–solvent systems, in
which both self- and cross-association interactions occur between the solvent molecules and
between the solvent molecules and the polymer functional groups, respectively. All parameters
were adopted by Tsivintzelis and Kontogeorgis [68] who showed that the NRHB model is able
to satisfactorily predict (without the use of any binary adjustable parameter) the VLE of such
binary mixtures, while using one fitted binary interaction parameter, the model very accurately
Molecular Thermodynamics of Hydrogen-Bonded Systems 169
450
440
430 HDPE–Heptanol
Temperature (K)
420
410
HDPE–Nonanol
400
390
0.0 0.1 0.2 0.3 0.4
Polymer weight fraction
describes the solvent solubility in the liquid phase. The description of the LLE of hydrogen-
bonding polymer systems is shown in Figure 2.11.
Over the past decade, an increasing research interest concerning a new class of solvents, known
as ionic liquids (ILs), has arisen. Such solvents are considered to be “green,” since they present
extremely low vapor pressure, which allows for flexible separation processes and for their recycling.
Subsequently, the disposal of solvents in the environment can be minimized. Consequently, the
description of thermodynamic properties of mixtures with ILs is of great importance.
Recently, the NRHB model was applied to model systems with ILs and a new methodology was
suggested for the modeling of relevant systems with a nonelectrolyte model [69]. According to this,
the ILs’ pure fluid parameters are estimated using PVT data and the Hansen’s solubility parameters,
while all ionic, polar, and hydrogen-bonding interactions are treated as specific interactions. In this
way, the model was able to describe the LLE of aqueous IL systems and the VLE of IL systems with
organic solvents or supercritical CO2.
Some representative calculations, which were performed using the pure fluid and binary param-
eters of Tsioptsias et al. [69], are illustrated in Figures 2.12 and 2.13, for the LLE of water–1-
hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([C4mim+][Tf 2N−]) system, and the
VLE of acetone–1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide [C4mim+][Tf 2N−]
mixture.
Misopolinou et al. [60,61] used a slightly different version of the NRHB, which assumes that
the intersegmental energy, ε*, and the close packed volume, v*sp, are temperature independent,
while the segment volume v* is an adjustable fluid-specific parameter. They applied the model to
systems with alkoxyalkanols, compounds which are capable for both intra- and intermolecular asso-
ciation. Examples of these calculations for the separate contributions from non-hydrogen-bonding
(dispersion and polar) interactions, from intermolecular, as well as from intramolecular hydrogen
bonds are shown in Figures 2.14 and 2.15. The calculated heats of mixing are compared with the
experimental ones [61].
Interestingly, the behavior of the binary with benzene is different in two points: the maxi-
mum and the shape of the HE curve. The maximum value is about half the values of the systems
with the n-alkanes. The decreased maximum value for the benzene system could be explained
on the basis of the occurrence of σ–π hydroxyl–aromatic ring interactions (exothermic process).
170 Handbook of Surface and Colloid Chemistry
Water in IL
0.1
0.01
Mole fraction
1E–3
IL in water
1E–4
1E–5
285 290 295 300 305 310 315 320
Temperature (K)
FIGURE 2.12 [C6mim+][Tf2N−]–water LLE. Experimental data [87] (points) and NRHB calculations (lines).
0.24
353.15 K
0.20
0.16
Pressure (MPa)
0.12
0.08
0.04
0.00
0.0 0.2 0.4 0.6 0.8 1.0
Mole fraction of acetone in IL
FIGURE 2.13 Acetone–[C4mim]Tf2N VLE. Experimental data [87] (points) and NRHB calculations (lines).
The asymmetric shape of the HE curve could be explained by the difficulty of breaking the interac-
tions in the 2-ethoxyethanol-rich region: In this region, the low presence of benzene fails to break
the interactions of the pure 2-ethoxyethanol, while in the benzene-rich region, the breakage of the
intermolecular hydrogen bonding is rather extensive.
As observed in Figures 2.14 and 2.15, the contribution of the intramolecular hydrogen bonds to
the heats of mixing of the studied systems is by no means negligible. The important point is that its
contribution is negative (exothermic). An explanation of this negative contribution comes from our
previously published spectroscopic data [59], which indicated that at low concentration, the degree
of intramolecular hydrogen bonding is increasing with increasing mole fraction of 2-ethoxyethanol
much more than the degree of intermolecular hydrogen bonding. It is also known that, at the
Molecular Thermodynamics of Hydrogen-Bonded Systems 171
1800
1600
1400
FIGURE 2.14 Experimental (points) [61] and calculated (lines) heats of mixing for the system
x 2-ethoxyethanol + (1 − x) n-octane at 318.15 K. The contributions from intermolecular hydrogen bonds
(dash, ---), dispersive interactions (dot, ···), and intramolecular hydrogen bonds (dash dot, -·-·-) are shown by
separate lines.
1500
Excess enthalpy, HE (J mol–1)
1000
500
–500
0.0 0.2 0.4 0.6 0.8 1.0
Ethoxyethanol mole fraction, x
FIGURE 2.15 Experimental [61] (points) and calculated (lines) heats of mixing for the system
x 2-ethoxyethanol + (1 − x) benzene at 318.15 K. Symbols as in Figure 2.14.
low-concentration region, the inert solvent does not influence essentially the intramolecular hydro-
gen bonds, while it causes the destruction of the intermolecular bonds, as the interacting molecules
fail to come close together (proximity condition).
The prediction of the NRHB model for the contribution of intermolecular hydrogen bonding
is positive for all binaries, while the contribution of the van der Waals interactions is negative
(exothermic) for the system with benzene but positive (endothermic) for the other systems with
n-alkanes, which corroborates the previous established picture. In binary mixtures of benzene with
172 Handbook of Surface and Colloid Chemistry
1000
–500
0.0 0.2 0.4 0.6 0.8 1.0
Ethoxyethanol mole fraction, x
FIGURE 2.16 Comparison of the experimental [92] excess molar enthalpies, HE , at 298.15 K for the system
x 2-ethoxyethanol + (1 − x) benzene, ⚪, with the predictions of the NRHB model for the contributions from
intermolecular hydrogen bonds (---), dispersive interactions (···), intramolecular hydrogen bonds (-·-·-), and the
total HE (—). The contribution of dispersive interactions is compared with the experimental HE for the system
x 1,2-dimethoxyethane + (1 − x) benzene: ◻, data from Ref. [89]; ◊, data from Ref. [90].
plain ethers, there are no hydrogen-bonding interactions. It is worth pointing out that the enthalpies
of mixing for these binaries are negative. Typical examples are the mixtures of 1,4-dioxane with
benzene at different temperatures [88] and of 1,2-dimethoxyethane (an isomer to 2-ethoxyethanol)
with benzene [89,90]. The latter experimental data are compared with the predictions of the NRHB
model for the system x 2-ethoxyethanol + (1 − x) benzene in Figure 2.16 and, as is shown, the experi-
mental values for the binary with 1,2-dimethoxyethane coincide with the calculated contribution
of the dispersive (non-hydrogen-bonding) interactions. From literature data, we also see positive
enthalpies of mixing in binaries of n-alkanes or cycloalkanes with ethers. Typical examples are
the mixtures of n-hexane or cyclohexane with 1,2-dimethoxyethane [91]. This further supports the
validity of our model.
2.4.2 Extension to Interfaces
Methods for reliable estimations of interfacial properties in multicomponent systems are essen-
tial for a rational design of numerous processes, notably, emulsion and suspension processes in
food, pharmaceutical and material technology, in coating and adsorption processes, in industrial
separation processes, in microemulsions, and in tertiary oil recovery. In the era of nanotechnology,
these properties are also essential for controlling, among others, sizes and shapes of nanostructures.
There are many approaches for the study of fluid–fluid interfaces [92–105], but little progress has
been made for the calculations of interfacial tensions and interfacial profiles in highly nonideal
mixtures, such as the hydrogen-bonded and the azeotropic mixtures.
In a series of papers [7,106,107], we have combined our EoS model with the density gradient
approximation of inhomogeneous systems [99–105]. In Refs. [7,106,107], we have addressed in
three alternative ways the problem of consistency and equivalence of the various methods of calcu-
lating the interfacial tension. In the first case [106], we have simulated the number density profile
across the interface with the classical hyperbolic tangent expression [92] (Equation 2.138). In the
second case [7], this profile was obtained from the free-energy minimization condition [103,105].
Molecular Thermodynamics of Hydrogen-Bonded Systems 173
In the latter case, the internal consistency requirement resulted in a universal value of three for the
proportionality factor between the interfacial tension and the integral of the Helmholtz’s free-energy
density difference, Δψ0. In the third case [107], we have also incorporated density gradient contribu-
tions into, both, the equation for the chemical potential and the equation of state. The striking result
was that a universal value of four was then obtained for the aforementioned proportionality factor.
In the following text, we will present the basic formalism, starting from the second case [7], and we
will discuss some representative applications.
⎛ ∂Ψ ⎞
γ=⎜ ⎟ (2.93)
⎝ ∂A ⎠T ,V , N
Pe and ie in Equation 2.92 are the equilibrium pressure and chemical potentials, respectively.
Applying Euler’s theorem for homogeneous functions, we obtain from Equation 2.92
⎡ e t ⎤
⎛ μie N i − Ψ ⎞⎟ ⎥
⎢ P V − ⎜⎝
γ= ⎣
∑ i =1 ⎠⎦ Ψ − Ψ
=
e
(2.94)
A A
Let the z axis be normal to the plane interface, ρi(z) be the mean segment number density
of component i at the height z, and ψi(z) be the corresponding mean molecular contribution to Ψ of
component i at z. By integrating over the full height H of the system, we have
V=
∫ A dz
H
(2.95)
Ni =
∫ Aρ (z)dz
H
i (2.96)
t
Ψ= ∑ ∫ Aρ (z)ψ (z)dz
i =1 H
i i (2.97)
174 Handbook of Surface and Colloid Chemistry
γ = [ P e − Pʹ( z)] dz
∫
H
(2.98)
where
t t
Pʹ( z ) = ∑ ρi ( z ) ⎡⎣μie − ψ i ( z ) ⎤⎦ = ∑ ρ (z)μ − ψ (z) (2.99)
i
e
i 0
i =1 i =1
ψ0(z) is the local Helmholtz free-energy density of the fluid at T and ρ(z).
For the two bulk phases at equilibrium, we have from classical thermodynamics
∫ [P − Pʹ(z)]dz (2.101)
e
γ=
−∞
or
+∞
γ=
∫ Δψ(z)dz (2.102)
−∞
where
t
Ψe
Δψ = ψ 0 ( z ) − = ψ 0 (z) − ∑ ρ (z)μ + P
i
e
i
e
(2.103)
V i =1
In a one-component system, the curves ρ(z), ψ(z), P′(z) are completely determined by the t emperature,
and Pe, μe, γ are functions of temperature only. These properties are also functions of composition
in multicomponent systems.
Equation 2.101 can be used for the calculation of interfacial tension as long as we know the local
value of pressure P′ at the height z. In the frame of the local thermodynamic approach, we, now,
assume that the aforementioned EoS model(s) can be used to provide with the local quantities μi(z),
ψi(z), P′(z), and Vi ( z ) connected by the equation
Equilibrium, however, requires for the chemical potentials of each component to be equal through-
out the total volume of the system. This requirement gives the working form of Equation 2.100 for
the local quantities:
Equation 2.101 can be used now with the understanding that P′(z) is the pressure given by the EoS
model for given temperature and number densities (local density and composition). Alternatively,
Equation 2.102 can be used, but, now, the EoS model will be used to provide with the local Helmholtz
free-energy density ψ0(z) for given temperature, local density, and composition.
Let us now turn to the case where contributions of density gradients to Helmholtz free energy
are explicitly taken into account.
EoS models can also be used in the frame of the gradient approximation, such as the Cahn–
Hilliard theory [100] of inhomogeneous systems, for the description of surface properties. In the
frame of this theory, the Helmholtz’s free-energy density ψ in a one-component inhomogeneous
system can be expressed as an expansion of density ρ and its derivatives:
where the free-energy density, ψ0(ρ), of a uniform fluid of density ρ is again given by
t
ψ 0 ( z) = ∑ ρ (z)ψ (z) (2.107)
i i
i =1
For sufficiently slow variation of density, the free energy of the system may be approximated by
where
dc1
c=− + c2 (2.109)
dρ
With the aforementioned equations, the surface tension for a plane interface can be approximated by
+∞ 2
⎡ ⎛ dρ ⎞ ⎤
⎢ Δψ + c ⎜ ⎟ ⎥ dz (2.110)
γ=
∫ ⎢
−∞ ⎣
⎝ dz ⎠ ⎥⎦
Here again, Δψ(z) is the free-energy density difference given by Equation 2.103. The interaction
or influence parameter c can be estimated if the intermolecular interaction potential of the fluid is
known [103]. Most often, however, it is treated as an adjustable parameter for each fluid.
Minimization of Equation 2.110 yields [100,103,105]
∂Δψ d 2ρ
− 2c 2 = 0 (2.111)
∂ρ dz
2
⎛ dρ ⎞
Δψ = c ⎜ ⎟ (2.112)
⎝ dz ⎠
176 Handbook of Surface and Colloid Chemistry
or
c
dz = dρ (2.113)
Δψ
This is a useful equation, since it may be used to obtain the density profile ρ(z) across the interface.
Substituting it into Equation 2.110, we obtain
ρβ
ρα and ρβ are the densities of the homogeneous phases α and β, respectively, at equilibrium.
Substituting Equation 2.112 into Equation 2.110, we also obtain
+∞
γ = 2 Δψ( z )dz
∫ (2.115)
−∞
Ψe
Δψ = ψ 0 ( z ) − = ρ( z )ψ( z ) − ρ( z )μ e + P e = ρ( z )(μʹ(ρ( z ), P e ) − μ e (ρe (∞), P e )) (2.116)
V
d 2ρ
μʹ(ρ( z ), P e ) − 2c = μ e (ρe (∞), P e ) (2.117)
dz 2
This equation indicates how the chemical potential, calculated at the local density ρ and the equilib-
rium (normal) pressure of the system, must be corrected in order to make it equal to the equilibrium
chemical potential.
In the case of systems of t components with plane interfaces, Equation 2.108 can be generalized
as follows [45,48,50]:
⎡ dρi dρ j ⎤
Ψ=A ⎢ψ 0 (ρ) + ∑∑c ⎥ dz (2.118)
∫ H⎢
⎣ i j
ij
dz dz ⎥
⎦
Substituting into Equation 2.94 and recalling that dρi/dz → 0 as z → ∞, we obtain for the interfacial
tension
+∞
⎡ d ρi d ρ j ⎤
⎢ Δψ( z ) + ∑∑ cij ⎥ dz (2.119)
γ=
∫ ⎢
−∞ ⎣ i j
dz dz ⎥
⎦
Molecular Thermodynamics of Hydrogen-Bonded Systems 177
where Δψ(z) is again given by Equation 2.103. The equilibrium interfacial tension can be
obtained now by minimizing Equation 2.119. Such minimization results to t coupled differential
equations:
∂Δψ d 2ρ j
− ∑ cij = 0 i, j = 1, 2,…, t (2.120)
∂ρi j
dz 2
which are the conditions subject to which the integral in Equation 2.119 must be evaluated.
Multiplying Equations 2.120 by dρi /dz, summing over all components i, and integrating, we
obtain
dρi dρ j
Δψ( z) = ∑∑c ij (2.121)
i j
dz dz
Substituting Equation 2.121 into Equation 2.119 results in an equation identical in form to
Equation 2.115, or
+∞ +∞
dρi dρ j
γ = 2 Δψ( z)dz = 2
∫ ∫ ∑∑c i j
ij
dz dz
dz (2.122)
−∞ −∞
An appropriate combining rule for cij is needed when i ≠ j. With cijs available and an EoS model
available, the earlier formalism is complete for the calculation of interfacial tensions at fluid–fluid
plane interfaces.
Equation 2.120 is not very useful in this form. However, they can turn to a set of most useful
equations if we replace the first terms by the derivatives obtained from Equation 2.103, which for
convenience is rewritten here in an alternative form:
t t
Δψ = ψ 0 ( z) − ∑ ρi ( z)μie + P e = ∑ ρ (z) ⎡⎣μʹ(ρ(z), P ) − μ ⎤⎦ (2.123)
i i
e e
i
i =1 i =1
d 2ρ j
⎡μʹi (ρ( z), P e ) − μie ⎤ − ∑ cij = 0 i, j = 1, 2,…, t (2.124)
⎣ ⎦
j
dz 2
Since we have accepted that the equations of the aforementioned EoS model(s) can be used in the
inhomogeneous region and in terms of the local variables, we may further simplify Equation 2.124
by using one of the equations for the chemical potential, such as Equation 2.80 or 2.84 (plus
Equation 2.25 for the hydrogen-bonding contribution) and the requirement for equilibrium—
Equation 2.105:
∂Δψ
= μʹi (ρ( z ), P e ) − μi (ρ( z ), Pʹ( z )) = ( P e − Pʹ( z ))vv
i∗ i = 1, 2,…, t (2.125)
∂ρi
178 Handbook of Surface and Colloid Chemistry
or
1 ∂Δψ
= ( P e − Pʹ( z ))v i = 1, 2,…, t (2.126)
vi∗ ∂ρi
or
d 2ρ j
( P e − Pʹ( z ))vv
i∗ = ∑c ij i, j = 1, 2,…, t (2.127)
j
dz 2
Equation 2.126 is the most useful equation for the evaluation of the composition at the height z. For
this purpose, it may be used in the alternative form
Equation 2.123 can also be written in a useful form by transforming it from the z space to the ρ
space. From Equation 2.121, we may write
⎡ ⎤
dρ1 Δψ( z )
= ±⎢ ⎥ (2.129)
dz ⎢c + 2 c1 j (dρ j /dρ1 ) + c jj (dρ j /dρ1 )
⎢⎣ 11 ∑ j >1 ∑ j >1
2⎥
⎥⎦
This equation can be integrated to give the interfacial profile z(ρ) or ρ(z):
ρ1 ( z )
c11 + 2 ∑ c1 j (dρ j /dρ1 ) + ∑ c jj (dρ j /dρ1 )2
j >1 j >1
z − z0 = dρ1 (2.130)
∫
ρ1 ( z0 )
Δψ
5/3
cii = ε*i vi∗ ( ) κii (2.132)
Molecular Thermodynamics of Hydrogen-Bonded Systems 179
TABLE 2.4
QCHB Scaling Constants for Pure Fluids
Fluid ε* = RT* (J mol−1) v* = ε*P*−1 (cm3 mol−1) * = ρ* −1 (cm3 g−1)
v sp s = q/r
Methane 1809 7.100 2.0630 0.980
Ethane 2967 7.824 1.5600 0.965
Propane 3168 9.131 1.4201 0.955
n-Pentane 4043 12.24 1.2614 0.950
n-Hexane 4281 12.80 1.2162 0.947
n-Heptane 4479 13.04 1.1879 0.933
n-Decane 5180 16.92 1.2576 0.900
Carbon monoxide 1378 6.683 1.1250 0.914
Carbon dioxide 2569 4.546 0.6172 1.020
Acetone 4982 9.138 1.1573 0.871
Benzene 5372 12.40 1.0610 0.892
Toluene 5392 12.87 1.0661 0.898
Cyclohexane 5102 13.99 1.1722 0.902
Methanol 5508 15.84 1.3459 0.795
Ethanol 5222 14.69 1.2791 0.795
1-Propanol 5405 15.36 1.2422 0.795
1-Butanol 5700 16.70 1.2347 0.795
1-Pentanol 5878 17.40 1.2249 0.795
1-Hexanol 5911 18.67 1.2131 0.824
1-Octanol 6068 19.25 1.2037 0.835
1-Decanol 6068 19.63 1.1980 0.848
Water 4350 13.92 1.0310 0.830
This equation satisfies the required dimensionality for cii and lets us treat κii as a dimensionless
adjustable parameter. Values of this parameter are given in Table 2.5 for some representative fluids.
For simplicity, the following geometric mean rule can be adopted for κij:
With this combining rule, there are no adjustable parameters for the mixture and, thus, the model
can be used for the prediction of interfacial tensions in multicomponent systems. For these systems,
TABLE 2.5
Influence Parameter of Representative Fluids
Fluid κ
n-Hexane 0.792
Cyclohexane 0.835
Toluene 0.720
Methanol 1.230
Ethanol 0.980
1-Propanol 0.875
Poly(dimethyl siloxane) 0.470
Polystyrene 0.530
180 Handbook of Surface and Colloid Chemistry
however, apart from the density profile ρ(z), we need the profiles ρi(z) for each component i.
Equivalently, we need the profiles of the compositions xi(z) or, in the case of multimer molecules,
the segment fractions φi(z). For these profiles, the following equations hold:
t t
i =1 i =1
10
PO
–10
PT (MPa)
–20 PEoS
–30
X1,liq = 0.0525
–40
–25 –20 –15 –10 –5 0 5 10 15
z, 10–10 m
FIGURE 2.17 The pressure profile at the interfacial layer as calculated by the equation of state (PEoS) and
by the chemical potential (PCP) for the system 1-propanol (1)–n-hexane (2) at 298.15 K near the azeotropic
composition.
Molecular Thermodynamics of Hydrogen-Bonded Systems 181
TABLE 2.6
Influence Parameter and the Shift Factor of Common Fluids
Fluid κ β
Carbon dioxide 0.550 1.000
Propane 0.370 1.000
n-Pentane 0.380 1.000
n-Hexane 0.355 1.000
Cyclohexane 0.375 1.000
Benzene 0.342 1.000
Toluene 0.322 1.000
Acetone 0.255 1.000
Poly(Dimethyl Siloxane) 0.215 1.000
Polyethylene–(linear) 0.615 1.000
Polyisobutylene 0.484 1.000
Polystyrene 1.005 1.000
Methanol 0.570 0.965
Ethanol 0.523 0.980
1-Propanol 0.400 0.987
In Ref. [106], we had considered that this discrepancy is primarily due to the earlier
truncation of higher-order contributions to the Helmholtz free energy. Of course, if we had
the complete form of Ψ, we could also get the complete forms of the equation of state and
the chemical potential in the local thermodynamic approach. In this case, we would have, quite
generally [106],
+∞ +∞ +∞
[ P e − PEoS
ʹ ]dz = (2 + β) Δψ dz = (2 + β) [ P e − PCP
ʹ ]dz (2.135)
γ=
∫ ∫ ∫
−∞ −∞ −∞
where β in this equation is a shift factor required to make all different estimations of γ equal. This
is not an external parameter. It is an internal parameter of the model and, at each set of external
conditions, it takes a value that fulfils the consistency requirement, namely, that Equation 2.135
holds true. Neither its value nor its dependence on external conditions is a priori known. We have
applied Equation 2.135 to a number of fluids of varying polarity. The striking result is that β not
only is independent of external conditions, but it appears that it has a universal value equal to
unity—at least for the nonstrongly polar fluids. The parameters κii, or simply κ, are different for
different fluids, however. These parameters are reported in Table 2.6 for a number of common
fluids and should be contrasted with the κ values reported in Table 2.5, where the consistency
requirement was ignored. They have been determined at one temperature by fitting Equation 2.135
to the known surface tension at this temperature. Calculations at other temperatures are just pre-
dictions. The integrals in Equation 2.135 were evaluated numerically by summation over a finely
subdivided interval.
Τhe following equation has been adopted for the influence parameter of alkanols [7,106]:
5/3
where λΗ and λV were set equal to one. νH is again the number of hydrogen bonds per molecular seg-
ment. If needed (as in the case of mixtures with alkanols), the following combining rule is adopted
for simplicity:
The values for the consistency parameter β follow the “universality” shown in Table 2.6 if Equations
2.113 and 2.130 are used for the density profile across the interface. The universality is lost and
lower values are obtained [106] if a different ad hoc profile equation is adopted, such as the classical
equation [92]:
ρα + ρβ ρα − ρβ ⎛ 2( z − z0 ) ⎞
ρ( z ) = − tanh ⎜ (2.138)
2 2 ⎝ D ⎟⎠
where
z0 is the arbitrary origin
D is a measure of the interfacial layer thickness
Quantities ρα and ρβ are the densities of the two bulk phases at equilibrium
In Figure 2.18, the experimental [108] surface tensions of three representative fluids with the
calculated ones are compared. Similar comparison is made in Figure 2.19 for representative
high polymers. As observed, the agreement is rather satisfactory over an extended range of
temperatures.
In Figure 2.20, the experimental [110,111] surface tensions of the 1-propanol + n-hexane mixture
at 298.15 K with the predicted ones by the present combined model are compared. In view of the
strong nonideality of this system, the agreement is again rather satisfactory. For the same system,
we have calculated the interfacial layer thickness as a function of composition of the liquid phase.
The calculations are shown in Figure 2.21. This is an azeotropic system, and it is worth observing
that the calculated maximum in Figure 2.21 is close to the azeotropic composition. In Ref. [106],
we have shown that the interfacial layer thickness increases with the vapor pressure of the system.
25
20
2
Y (mN m–1)
15
10 1
3
5
0
250 300 350 400 450 500 550
T (K)
FIGURE 2.18 Experimental [108] (symbols) and calculated (lines) surface tensions for acetone (1), ethanol
(2), and n-hexane (3) as a function of temperature.
Molecular Thermodynamics of Hydrogen-Bonded Systems 183
34
32 Polystyrene
Polyiso-butylene
30
28
Y (mN m–1) 26
Polyethylene(lin.)
24
22
20
18 Poly(dimethyl siloxane)
16
14
FIGURE 2.19 Experimental [109] (symbols) and calculated (solid lines) surface tensions of pure polymers.
24
23
22
Y (mN m–1)
21
20
19
18
17
0.0 0.2 0.4 0.6 0.8 1.0
Xpropanol
FIGURE 2.20 Experimental [110,111] (symbols) and predicted (lines) surface tensions of 1-propanol(1)–n-
hexane (2) mixture at 298.15 K.
Since this azeotrope exhibits a maximum in vapor pressure, one may easily explain the occurrence
of a maximum in Figure 2.21.
Of interest is the composition profile across the interface for such an azeotropic system. These
profiles are shown in Figure 2.22. As shown, near the azeotropic composition, the interfacial com-
position profile exhibits a deep minimum, indicating that 1-propanol tends to avoid the interface
and stay, instead, in the bulk liquid phase where it may form easier its hydrogen bonds or in the
bulk vapor phase where it has much freedom of motion. The composition profiles in Figure 2.22
are qualitatively similar to the corresponding profiles calculated in Ref. [106] by simulating the
density profiles with Equation 2.138. The density profiles obtained with Equation 2.130 are indeed
of the form suggested by Equation 2.138 [7]. In general, the profiles obtained with Equation 2.130
are somewhat broader compared to those obtained with Equation 2.138. Qualitatively, however, we
expect similar calculations of composition profiles by the two equations.
184 Handbook of Surface and Colloid Chemistry
36
34
32
30
Thickness, 10–10 m
28
26
24
22
20
18
FIGURE 2.21 The calculated interfacial layer thickness as a function of composition for 1-propanol
(1)–n-hexane (2) mixture at 298.15 K.
1-Propanol(1)-n-Hexane(2) at 298.15
0.060
X1,liq = 0.0245 X1,liq = 0.0525
0.04 Near Azeotrope
0.045
0.03
X1
X1
0.030
0.02
0.01 0.015
–15 –10 –5 0 5 10 15 –15 –10 –5 0 5 10
–10
z, 10 m z, 10–10 m
0.12 0.8
X1,liq = 0.105 X1,liq = 0.709
0.10
0.6
0.08
X1
0.4
X1
0.06
0.04 0.2
0.02
0.0
–15 –10 –5 0 5 10 15 –15 –10 –5 0 5 10
z, 10–10 m
z, 10–10 m
FIGURE 2.22 The evolution of the composition profiles across the interface for the 1-propanol(1)–n-hexane
(2) mixture at 298.15 K. The liquid phase composition, x1, increases from top left to bottom right.
Molecular Thermodynamics of Hydrogen-Bonded Systems 185
0.8 0.80
X1,liq = 0.4188
X1,liq = 0.2988 0.75
0.7
0.70
0.6 0.65
0.60
X1
X1
0.5
0.55
0.50
0.4
0.45
0.3 0.40
0.90 0.98
X1,liq = 0.8045
(Near Azeotrope) 0.96 X1,liq = 0.9485
0.85
0.94
0.80
0.92
X1
X1
0.75 0.90
0.88
0.70
0.86
0.65
0.84
–15 –10 –5 0 5 10 15 20 –10 –5 0 5 10 15 20
z, 10–10 m z, 10–10 m
FIGURE 2.23 Methanol (1)–toluene (2) at 308.15 K. The evolution of composition profiles across the inter-
face, as predicted by the equation-of-state model [106].
Very interesting is Figure 2.23 where the evolution of composition profiles is shown for another
azeotropic mixture. The calculations were done [106] by adopting Equation 2.138.
In summarizing the results of this section, it is worth focusing on the important aspect that
emerged from the extension of the EoS model(s) to interfaces, namely, the unification of three,
essentially, different approaches for calculating interfacial properties through the equation
+∞
Δψ( z)dz. By disregarding density gradient contributions to Helmholtz free energy (and to
γ=α
∫ −∞
derived properties), one may use this equation with α = 1. By adopting the density gradient approxi-
mation and the truncated form of Equation 2.108, this equation should be used with α = 2. However,
it appears that internal consistency (in the sense that the various ways of calculating γ give the same
result) is fulfilled only when α = 3, or near 3. It should be pointed out that, in the latter case, the pre-
cise value of α is an internal parameter of the model and not an external adjustable parameter. Less
than 3 values have been obtained only for the hydrogen-bonded 1-alkanols, but this is not a proof of
lack of universality: It should be stressed that this is an approximate model and Equation 2.136 for
the evaluation of the influence parameter of hydrogen-bonded fluids is just a simple approximation
for the OH–OH interaction.
Work must be done with fluids exhibiting hydrogen bonds between other pairs of proton donors
and acceptors. General rules and conclusions for α could then be possible.
As already mentioned, in the earlier approach, we have disregarded density gradient contributions
in the chemical potential and in the equation of state. But the alternative approach, of taking into
consideration the full density gradient contributions, is by no means an easy task. However, a first
186 Handbook of Surface and Colloid Chemistry
approximation has been attempted in Ref. [107] by recognizing that incorporation of these contribu-
tions would imply that the total Helmholtz free energy may be written, in general, as
⎧⎪ 1 ⎫⎪ ⎧⎪ ⎡ ⎤ ⎫⎪
Ψ = V * ⎨ψ 0 (ρ) + ∑ ρi λ nm,i (ρ)(∇ nρ)m ⎬ = V * ⎨ ∑ ρi ⎢ ψ i,0 (ρ) + ∑ δnm,i (ρ)(∇ nρ)m ⎥ ⎬ (2.139)
ρ n , m ,i ⎢⎣ ⎥⎦ ⎭⎪
⎩⎪ ⎭⎪ ⎩⎪ i n, m
This equation, in turn, implies that there may be density gradient contributions to, both, the chemi-
cal potential and to the equation of state. As an example, for the QCHB model, we would have
⎡ 1 z ⎤
P + T ⎢ ln(1 − ρ ⎛⎜ 1 − ⎞⎟ + ln Γ 00 ⎥ +
) + ρ ∑ω nm ,i (ρ)(∇ nρ)m = 0 (2.141)
⎣ ⎝ r ⎠ 2 ⎦ n, m
where μi,0 in Equation 2.140 is the chemical potential of component i without the density gradient
contribution and is assumed to be obtained from the EoS model as before. By using the last three
equations and by following a procedure entirely analogous to the one described in this section, one
ends up [107] in the striking result of α = 4. In this sense, the truncation used in the density gradient
approximation approach [100,103,105] is, rather, severe. It should be repeated, however, that all the
aforementioned approaches may give similar estimations of interfacial tension. The key differences
appear in the calculated profiles of the various properties across the interface. Since there are no
experimental data available for these profiles, we cannot judge safely the various approaches. We
hope, however, that such data, though challenging, will be available soon, since they may create new
perspectives in the theoretical description of fluid–fluid interfaces.
The approach presented in this work can be extended to mixtures of macromolecules, including
systems with strong specific interactions, and to aqueous systems. Of course, entropic corrections
must be introduced in the case of polymer systems due to the effects of density and concentration
gradients on the spatial conformations of macromolecules in the interfacial regions. The aforemen-
tioned interesting results indicate the potential of the EoS approach in studying interfaces. It should
be stressed, however, once again that this density-gradient approach is an approximate approach
with ad hoc assumptions and simplifications.
Recently [57], our previous approach to solubility parameter estimation [120,121] was extended,
in an effort to account for all three components of the solubility parameter. This was done by
adopting the NRHB EoS framework, modified in order to explicitly account for dipole–dipole
interactions and, thus, explicitly calculate the polar component, δp. In what follows, we summarize
these developments and show some representative applications.
For the purposes of this section, it is convenient to rewrite Equation 2.2 in the following manner:
− Ed −E p −E − PV
Q( N , T , P ) = QRQNRQhb = Ω R Ω NR Ωhb exp exp exp hb exp (2.142)
kT kT kT kT
Ed, Ep, and, Ehb are the dispersion, polar, and hydrogen-bonding components, respectively, of the
potential energy of the system. From this equation, we may extract in the usual manner the follow-
ing equation for the chemical potential:
μ μ μ
= dp + H (2.143)
RT RT RT
The first term is the chemical potential for the dispersion and polar interactions, and the s econd
chemical potential for the hydrogen-bonding interactions. The key problem now is the separation of
the contributions from dispersion and polar forces. Following the rationale of Twu and Gubbins [122]
and Nezbeda and Pavlicek [123–125], it was proposed [57] to replace ε* by ε *[1 + (4 /π)(m /r )2 (s 2 /T)],
where m is the dipole moment of the fluid. In this way, the potential energy of the fluid may be
written as
⎡ 4 ⎛ m ⎞2 s 2 ⎤
− E = Γrr qNθr ε * ⎢1 + ⎜ ⎟ ⎥ − N H E H (2.144)
⎢⎣ π ⎝ r ⎠ T ⎥⎦
On the basis of the aforementioned equation, the partial solubility parameters are given by
Γrr qNθr ε∗
δd =
V (2.145)
− N H EH
δhb = (2.147)
V
Thus, by knowing the scaling constants of a fluid along with its dipole moment and hydrogen-
bonding interaction energies, one may estimate (essentially, predict) the three partial solubility
parameters over an extended range of external conditions. On the other hand, if the partial solubility
parameters are known, they may be used in order to extract information about the scaling constants
of the fluid and its hydrogen-bonding interaction, which in turn could be used for predicting all
basic thermodynamic properties of the fluid. Such calculations have been done for a large number
of fluids including polymers and extensive tabulations are reported in Ref. [57]. One representative
sample from these tabulations is shown in Table 2.7.
188 Handbook of Surface and Colloid Chemistry
TABLE 2.7
Total and Partial Solubility Parameters (MPa1/2) of Pure Fluids
Polar/Hydrogen-Bonded Fluids
δTotal δhb δp
Fluid Exp Calc
a Exp Exp Calca
Acetone 19.95 21.29(20.04) 6.95 6.95 10.43 10.44(10.14)
Ethyl acetate 18.48 18.66(18.31) 9.20 9.20 5.32 5.70(6.07)
1,4-Dioxane 20.47 19.67(20.08) 7.36 7.36 1.84 1.84(1.91)
CHCl3 18.94 19.46(19.18) 5.73 5.73 3.07 3.34(3.79)
Methanol 29.61 30.86(29.89) 22.30 24.15(24.08) 12.27 12.11(11.34)
Ethanol 26.50 26.26(26.08) 19.43 20.08(19.98) 8.80 8.49(8.24)
1-Butanol 23.35 22.92(22.90) 15.80 15.84(15.80) 5.70 5.70(5.72)
1-Octanol 20.87 20.30(20.27) 11.86 11.97(11.94) 3.27 3.27(3.75)
Phenol 24.63 24.59(24.69) 14.90 14.90 5.90 6.31(5.16)
Ethylene glycol 33.70 33.97(33.64) 25.77 25.93 11.05 16.01(12.2)
Glycerol 34.12 34.14(34.34) 29.25 30.91 12.07 12.10(14.31)
Diethylamine 16.61 16.97(16.80) 6.10 6.10 2.30 2.49(2.89)
n-Butylamine 18.31 18.32(18.48) 8.00 8.00 4.50 4.86(4.75)
Tetrahydrofuran (THF) 19.46 20.05(18.93) 8.00 8.00 5.70 6.04(6.98)
Ammonia 27.40 28.73(26.52) 17.80 17.80 15.70 16.10(15.70)
Water 47.82 47.80(48.68) 42.82 43.15 16.00 18.70(16.00)
2
⎡ ⎛m⎞ ⎤
a Values in parenthesis obtained when ε* is replaced by ε* ⎢1 + π ⎜ ⎟ s 2 ⎥ .
⎢⎣ ⎝ r ⎠ ⎥⎦
In Figure 2.24, the calculated components of the solubility parameter of water over an extended
range of saturation temperatures are shown. As was expected, the main contribution to δ of water,
especially at low temperatures, is the hydrogen bonding. This type of diagrams is most useful for
designing applications involving subcritical or supercritical water.
A most useful concept, which quantifies the “similarity” of two substances 1 and 2, especially
the similarity of a polymer, 2, and a potential solvent, 1, for it, is the solubility parameter distance,
Ra, defined by [116]
The idea is: the smaller the Ra, the better is the solvent for the polymer. A sphere with radius Ra
encompasses the good solvents for this polymer. A refined discussion on Ra and the related quanti-
ties Ro and RED = Ra/Ro is provided by Hansen [117].
In Figure 2.25, the partial solubility parameters for (bisphenol A) polycarbonate as functions
of temperature are plotted. As shown, all three components are nonnegligible, and there is a
crossover in the polar and hydrogen-bonding components for this polymer. In Figure 2.26, the
distances Ra of this polymer with three common solvents are compared. According to the calcu-
lations, chloroform is the best of the three solvents for this polymer, followed by tetrahydrofuran
(THF). Heptane is the worst and, essentially, a nonsolvent for the polymer, and all these findings
agree with experiment.
In a similar manner, by comparing the distances Ra for polypropylene with three solvents, THF,
chloroform, and tetralin, it is shown that the best solvent for polypropylene is tetralin, which is again
Molecular Thermodynamics of Hydrogen-Bonded Systems 189
50
δ
40
δhb
δ (MPa1/2) 30
20 δd
δp
10
0
300 400 500 600
T (K)
FIGURE 2.24 Fractional solubility parameters for water as calculated when ε* is replaced by ε* [1 + π(m/r)2 s2].
21
δ
18
δd
15
δ (MPa1/2)
12
δp
6
δhb
3
280 320 360 400 440 480
T (K)
corroborated by the experiment. This type of figures is most useful not only for the mere selection
of the solvent, but also for the selection of the external conditions (especially, temperature) for the
dissolution of the polymer or any other solute.
10
n-Heptane
8
Ra (MPa1/2) 6
THF
4
2 Chloroform
0
300 310 320 330 340
T (K)
FIGURE 2.26 The estimated solubility parameter distance, Ra, of (bisphenol A) polycarbonate (MW =
100,000) with three common solvents, as a function of temperature.
approaches. In what follows, we will briefly present the PSP essentials. Details may be found in the
recent relevant literature [39–43].
There are two schemes of PSPs: the s-scheme and the σ-scheme. Due to space limitations, we
will confine ourselves here to the σ-scheme. According to this scheme, each compound/molecule is
characterized by three major σ PSPs, which have the same dimensions as the solubility parameters
(MPa1/2). The first descriptor, called van der Waals PSP, is defined as follows:
EvdW
σW = (2.149)
Vmol
EvdW and Vmol are the van der Waals interaction energy and the molar volume of the compound,
respectively. The van der Waals energy is already available for thousands of compounds [136], but it
may be obtained also from alternative straightforward calculation schemes. The hydrogen-bonding
PSP, σhb, is subdivided into a donor or acidic component, σa, and an acceptor or basic component,
σb, through the relation
m-SUM
σ2hb = σ2a + σ2b = 4 (2.150)
Vcos m
where Vcosm is the COSMO volume [134–136] of the molecule and m-SUM is obtained from the third
hydrogen-bonding σ-COSMOments [134–136] through the equation:
and
The third PSP, σpz, reflects interactions due to polarity/polarizability/refractivity of the molecule, and
it may also be subdivided into two components for the polarity, and the refractivity although, for all
Molecular Thermodynamics of Hydrogen-Bonded Systems 191
practical purposes and at this stage, the combined σpz PSP is sufficient. Its mathematical definition
will be better understood after we divert for the introduction of a key concept underlying PSPs.
The cohesive interactions in a pure compound are, in general, different from its solvation inter-
actions with solvents. Some molecules may self-associate in their pure state via hydrogen-bonding
interactions, which others can only cross-associate with different molecules. The cohesive energy,
then, or the heat of vaporization cannot be used to determine the hydrogen-bonding capacity
of these latter molecules. This is why PSPs are focusing on solvation rather than on cohesion.
The compounds are divided into two major classes, the homosolvated and the heterosolvated ones.
The homosolvated molecules are the molecules in which all types of intermolecular interactions are
operational in pure state as well as in solution. In contrast, in heterosolvated molecules, the types
of operational intermolecular interactions in solution may be different from those in pure state.
Typical representatives of homosolvated and heterosolvated compounds are aspirin and acetone,
respectively, as shown graphically in the upper part of Figure 2.27. The surface charge (σ) profile of
aspirin contains, both, acidic as well as basic surface areas and may self-associate, while acetone
does not possess acidic areas and can only cross-associate with another acidic compound. This dis-
tinction leads to the distinction between the well-known cohesive energy density and the solvation
energy density.
In view of the aforementioned text, the σpz PSP is obtained from the following balance equations:
(a) (b)
(c) (d)
FIGURE 2.27 Surface charge distribution in representative molecules. (a) Acetone, (b) aspirin, (c) vinyl
phenol trimer, and (d) nylon-6 pentamer.
192 Handbook of Surface and Colloid Chemistry
Strictly speaking, the σa and σb hydrogen-bonding PSPs cannot be mapped one to one in the corre-
sponding A and B acidic and basic LSER descriptors since the latter reflect interaction free energies
rather than interaction energies. A direct mapping would introduce the free-energy hydrogen-
bonding PSP, σGhb, and its components, σGa and σGb, satisfying the relations
2 2 2
σGa + σGb = σGhb (2.155)
2
σGb σ2 B
= 1 − 2Ga = (2.156)
2
σGhb σGhb m-SUMG
and
m-SUMG
σGhb = 5.5 (2.157)
Vcosm
where
Having, however, hydrogen-bonding energy PSPs (from COSMOments) and free-energy PSPs
(from LSER), one may calculate energies, entropies, and free energies of hydrogen bond formation
of type αβ between molecules of type i and j, respectively, using the following equation:
Eαβ
H
,ij = −hσ a ,i σ b, j Vmol ,iVmol , j (2.159)
α β
Gαβ
H
,ij = − fσGa ,i σGb, j Vmol ,iVmol , j + g
α β
(2.160)
and
Sαβ
H (E H
αβ,ij − Gαβ
H
,ij ) (2.161)
,ij =
T
The proposed [40] values for the universal coefficients h, f, and g are h = 1.86; f = 1.70; g = 2.50 kJ/mol.
Having, however, these energies and free energies, we may apply directly the aforementioned pre-
sented hydrogen-bonding and combinatorial formalism and have a predictive thermodynamic
model for hydrogen-bonded systems. In a free-energy expression for a binary mixture, the σW and
σpz PSPs contribute via the term [41–43]
where
1
Δσ = (σW 1 − σW 2 )2 + (σ pz1 − σ pz 2 )2 (2.163)
3
Molecular Thermodynamics of Hydrogen-Bonded Systems 193
This PSP formalism can be used for both concentrated as well as infinitely dilute systems, for small
molecules as well as for high polymers and copolymers. It may, in particular, be used for the surface
characterization of polymer and other solid surfaces. In fact, it may be used in order to obtain the
corresponding surface energy components, γw, γpz, γa, and γb, of the solid surfaces [42]. The widely
used van der Waals–Lifshitz surface tension component is nearly the sum of γw and γpz, while the
contact angle θ of a liquid L on a polymer surface P is given by the quadratic equation [42]:
γ L (1 + cos θ) = 2 ( )
γ WP γ WL + γ Ppz γ Lpz + γ aP γ bL + γ bP γ aL (2.164)
The range of applications of the PSP approach is vast. It may be used for the prediction of phase equilib-
ria in highly nonideal systems, for the solubility of solids, such as drugs, in various solvent systems over
an extended range of temperatures, or for the sorption of solvent vapors on polymer surfaces. Various
examples are reported in Ref. [41]. It may, in particular, be used for the complete thermodynamic char-
acterization of solid surfaces via a novel method exposed in Ref. [42]. With this method, the PSPs of
polymers may be obtained graphically, as shown in Figure 2.28. Inverse gas chromatography may now
be used for the thermodynamic characterization of polymers or solid powders (drugs, dyes, clays, etc.).
Having the PSPs of a polymer, one may predict its Flory–Huggins interaction parameter, χ12,
not only with a solvent but also with another polymer. This then may be used for the prediction of
polymer–polymer miscibility—a subject of high technological importance. It may, as an example,
predict the effect of copolymer composition on its miscibility with another polymer or copolymer as
shown in Figure 2.29. It may also use experimental or theoretical information on oligomers (cf. lower
part of Figure 2.27) or low molecular weight analogs in order to characterize high p olymers. The
detailed formalism for these calculations may be found in Ref. [43].
Due to its simplicity, the PSP approach can be turned into an EoS approach along the lines of the
previous paragraphs. This may lead to a very useful predictive model over a broad range of external
conditions of temperature, pressure, and composition. Work is underway in our laboratory on these
developments.
Alkanes
Aromatics
9
CCl4
Cyclohexane
6
σpz
0
12 15 18
σW
FIGURE 2.28 Determination of the poly(dimethyl siloxane) (PDMS) PSPs. The PSP lines for various
non-hydrogen-bonded solute probes intersect at σW = 14.5 and σpz = 3.7 MPa1/2. Symbols are experimental [137]
χ12 parameters.
194 Handbook of Surface and Colloid Chemistry
0.6
Miscible
0.4 (1-phase)
Immiscible
0.2 (2-phase)
2-phase
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Vinyl phenol fraction in copolymer
FIGURE 2.29 The miscibility map for the system styrene-co-vinyl phenol (S-co-VPH) (1)–poly(methyl
methacrylate) (PMMA) (2) at 40°C as given by the spinodal lines (dash-dotted lines) and the line of zero χ12
interaction parameter (solid line) as functions of the vinyl phenol content (FVPH) of the copolymer.
2.6 CONCLUSIONS
In the previous sections, we have exposed the rationale and the formalism of a unified treatment of
hydrogen bonding in fluid systems ranging from small gas molecules up to high molecular weight
polymeric glasses, in homogeneous as well as heterogeneous systems and interfaces. There is no
limit, in principle, in the number and types of different hydrogen bonds that can be treated simul-
taneously. They can be intermolecular, intramolecular, or cooperative. They are incorporated into
an EoS framework—the NRHB, and this facilitates the systematic study of the effect of density or
pressure and temperature on the degree of hydrogen bonding or the fraction of hydrogen-bonded
donor or acceptor groups in the system. The presented applications are only indicative and by no
means exhaustive. Our more recent work on PSPs is a new type of predictive thermodynamic mod-
els, which combines elements from quantum mechanics, statistical thermodynamics, and QSPR
approaches and has the potential to remove some of the key limitations of the successful COSMO-
type thermodynamic models. Its capacity to handle surface characterization of polymers/solids is
one of its promising features of particular interest to the colloids and surface science community.
Much of its strength resides on the way hydrogen bonding is implemented in both concentrated and
infinitely dilute systems.
REFERENCES
1. Guggenheim, E.A., Mixtures, Oxford University Press: Oxford, U.K., 1952.
2. Flory, P.J., Principles of Polymer Chemistry, Cornell University Press: Ithaca, NY, 1953.
3. Sanchez, I.C. and Lacombe, R.H., J. Phys. Chem., 80, 2352, 1976.
4. Sanchez, I.C. and Lacombe, R.H., Macromolecules, 11, 1145, 1978.
5. Panayiotou, C.G., J. Phys. Chem., 92, 2960, 1988.
6. Taimoori, M. and Panayiotou, C., Fluid Phase Equilib., 205, 249, 2003.
7. Panayiotou, C., J. Chem. Thermodyn., 35, 349, 2003.
8. Panayiotou, C., Pantoula, M., Stefanis, E., Tsivintzelis, I., and Economou, I.G., Ind. Eng. Chem. Res.,
43, 6592, 2004.
9. Abdel-Hady, E.E., Polym. Degrad. Stabil., 80, 363, 2003.
Molecular Thermodynamics of Hydrogen-Bonded Systems 195
10. Kilburn, D., Dlubek, G., Pionteck, J., Bamford, D., and Alam, M.A., Polymer, 46, 869, 2005.
11. Frank, H.S. and Wen, W.-Y., Discuss. Faraday Soc., 24, 133, 1957.
12. Pimentel, G.C. and McClellan, A.L., The Hydrogen Bond, W.H. Freeman: San Francisco, CA, 1960.
13. LaPlanche, L.A., Thompson, H.B., and Rogers, M.T., J. Phys. Chem., 69(5), 1482, 1965.
14. Vinogradov, S. and Linnell, R., Hydrogen Bonding, Van Nostrand Reinhold: New York, 1971.
15. Joesten, M.D. and Saad, L.J., Hydrogen Bonding, Marcel Dekker: New York, 1974.
16. Huyskens, P.L., J. Am. Chem. Soc., 99, 2578, 1977.
17. Walter, H., Brooks, D., and Fisher, D. (Eds.), Partitioning in Aqueous Two-Phase Systems, Academic
Press: New York, 1985.
18. Marsh, K. and Kohler, F., J. Mol. Liquids, 30, 13, 1985.
19. Kleeberg, H., Klein, D., and Luck, W.A.P., J. Phys. Chem., 91, 3200, 1987.
20. Bourrel, M. and Schechter, R.S. (Eds.), Microemulsions and Related Systems, Marcel Dekker: New York,
1988.
21. Hobza, P. and Zahradnik, R., Intermolecular Complexes, Academia: Prague, Czech Republic, 1988.
22. Reichardt, C., Solvent and Solvent Effects in Organic Chemistry, VCH Verlag GmbH: Weinheim,
Germany, 1988.
23. Burchard, W. and Ross-Murphy, S.B. (Eds.), Physical Networks, Polymers, and Gels, Elsevier Applied
Science: London, U.K., 1990.
24. Coleman, M.M., Graf, J.F., and Painter, P.C., Specific Interactions and the Miscibility of Polymer Blends,
Technomic: Lancaster, PA, 1991.
25. Maes, G. and Smets, J., J. Phys. Chem., 97, 1818, 1993.
26. Sanchez, I.C. and Panayiotou, C. Equation of State Thermodynamics of Polymer and Related Systems,
In Models for Thermodynamic and Phase Equilibria Calculations, S. Sandler (Ed.), Marcel Dekker:
New York, 1994.
27. Acree, W.E., Thermodynamic Properties of Nonelectrolyte Solutions, Academic Press: New York, 1984.
28. Prausnitz, J.M., Lichtenthaler, R.N., and de Azevedo, E.G., Molecular Thermodynamics of Fluid Phase
Equilibria, 2nd edn., Prentice-Hall: New York, 1986.
29. Heintz, A., Ber. Bunsenges. Phys. Chem., 89, 172, 1985.
30. Panayiotou, C., J. Solut. Chem., 20, 97, 1991.
31. Panayiotou, C. and Sanchez, I.C., Macromolecules, 24, 6231, 1991.
32. Levine, S. and Perram, J.W. In Hydrogen Bonded Solvent Systems, A.K. Covington and P. Jones (Eds.),
Taylor & Francis: London, U.K., 1968.
33. Luck, W.A.P., Angew. Chem., 92, 29, 1980.
34. Veytsman, B.A., J. Phys. Chem., 94, 8499, 1990.
35. Panayiotou, C. and Sanchez, I.C., J. Phys. Chem., 95, 10090, 1991.
36. Panayiotou, C. Hydrogen Bonding in Solutions: The Equation of State Approach, In Handbook of Surface
and Colloid Chemistry, Birdi, K.S. (Ed.), CRC Press: Boca Raton, FL, 2003.
37. Panayiotou, C. Hydrogen Bonding and Non-randomness in Solution Thermodynamics, In Handbook of
Surface and Colloid Chemistry, Birdi, K.S. (Ed.), CRC Press: Boca Raton, FL, 2009.
38. Panayiotou, C., Tsivintzelis, I., and Economou, I.G., Ind. Eng. Chem. Res., 46, 2628, 2007.
39. Panayiotou, C., Phys. Chem. Chem. Phys., 14, 3882, 2012.
40. Panayiotou, C., J. Chem. Thermodynamics, 51, 172, 2012.
41. Panayiotou, C., J. Phys. Chem. B, 116, 7302, 2012.
42. Panayiotou, C.G., J. Chromatogr. A, 1251, 194, 2012.
43. Panayiotou, C., Polymer, 54, 1621, 2013.
44. Yan, Q., Liu, H., and Hu, Y., Fluid Phase Equilib., 218, 157, 2004.
45. Wilson, G., J. Am. Chem. Soc., 86, 127, 1964.
46. Panayiotou, C. and Vera, J.H., Can. J. Chem. Eng., 59, 501, 1981.
47. Panayiotou, C., Fluid Phase Equilib., 237, 130, 2005.
48. Wang, W. and Vera, J.H., Fluid Phase Equilib., 85, 1, 1993.
49. Vera, J.H., Fluid Phase Equilib., 145, 217, 1998.
50. Hill, T.L., An Introduction to Statistical Thermodynamics, Dover Publications: New York, 1986.
51. Panayiotou, C. and Vera, J.H., Polymer J., 14, 681, 1982.
52. Staverman, A.J., Rec. Trav. Chim. Pays-Bas., 69, 163, 1950.
53. Fredenslund, A., Jones, R.L., and Prausnitz, J.M., AIChE J., 21, 1086, 1975.
54. Fredenslund, A. and Sorensen, M.J. Group Contribution Estimation Methods, In Models for Thermo
dynamic and Phase Equilibria Calculations, Sandler, S. (Ed.), Marcel Dekker: New York, 1994.
55. Kemeny, S., Balog, G., Radnai, G., Savinsky, J., and Rezessy, G., Fluid Phase Equilib., 54, 247, 1990.
196 Handbook of Surface and Colloid Chemistry
56. Li, X. and Zhao, D., J. Chem. Phys., 117, 6803, 2002.
57. Stefanis, E., Tsivintzelis, I., and Panayiotou, C., Fluid Phase Equilib., 240, 144, 2006.
58. Missopolinou, D. and Panayiotou, C., J. Phys. Chem. A, 102, 3574, 1998.
59. Missopolinou, D., Ioannou, K., Prinos, I., and Panayiotou, C., Z. Phys. Chem., 216, 905, 2002.
60. Missopolinou, D., Tsivintzelis, I., and Panayiotou, C., Fluid Phase Equilib., 238, 204, 2005.
61. Missopolinou, D., Tsivintzelis, I., and Panayiotou, C., Fluid Phase Equilib., 245, 89, 2006.
62. Abusleme, J.A. and Vera, J.H., Can. J. Chem. Eng., 63, 845, 1985.
63. Tsivintzelis, I., Economou, I.G., and Kontogeorgis, G.M., AIChE J., 55, 756, 2009.
64. Tsivintzelis, I., Economou, I.G., and Kontogeorgis, G.M., J. Phys. Chem. B, 113, 6446, 2009.
65. Tsivintzelis, I., Spyriouni, T., and Economou, I.G., Fluid Phase Equilib., 253, 19, 2007.
66. Grenner, A., Tsivintzelis, I., Kontogeorgis, G.M., Economou, I.G., and Panayiotou, C., Ind. Eng. Chem.
Res., 47, 5636, 2008.
67. Tsivintzelis, I., Grenner, A., Economou, I.G., and Kontogeorgis, G.M., Ind. Eng. Chem. Res., 47, 5651,
2008.
68. Tsivintzelis, I. and Kontogeorgis, G.M., Fluid Phase Equilib., 280, 100, 2009.
69. Tsioptsias, C., Tsivintzelis, I., and Panayiotou, C., Phys. Chem. Chem. Phys., 12, 4843, 2010.
70. von Solms, N., Michelsen, M.L., Passos, C.P., Derawi, S.O., and Kontogeorgis, G.M., Ind. Eng. Chem.
Res., 45, 5368, 2006.
71. Luck, W.A.P., Angew. Chem. (Int. Ed. Engl.), 19, 28,1980.
72. Lien, T.R., A study of the thermodynamic excess functions of alcohol solutions by IR spectroscopy.
Applications to chemical solution theory. PhD thesis, University of Toronto, Toronto, Ontario, Canada,
1972.
73. Fletcher, A.N. and Heller, C.A., J. Phys. Chem., 71, 3742, 1967.
74. Aspiron, N., Hasse, H., and Maurer, G., Fluid Phase Equilib., 186, 1, 2001.
75. Gupta, R.B. and Brinkley, R.L., AIChE J., 44, 207, 1998.
76. Tsivintzelis, I., Bøgh, D., Karakatsani, E., and Kontogeorgis, G.M., Fluid Phase Equilibria, 365, 112,
2014.
77. Sørensen, J.M. and Arlt, W., Liquid–Liquid Equilibrium Data Collection (Binary Systems), Vol. V, Part 1,
DECHEMA Chemistry Data Series: Frankfurt/Main, Germany, 1995.
78. Raal, J.D., Best, D.A., Code, R.K., J. Chem. Eng. Data, 17, 211, 1972.
79. Tsonopoulos, C. and Wilson, G.M., AIChE J., 29, 990, 1983.
80. Derawi, S.O., Kontogeorgis, G.M., Stenby, E.H., Haugum, T., and Fredheim, A.O., J. Chem. Eng. Data,
47, 169, 2002.
81. Dobroserdov, L.L. and Shakhanov, V.D., Zh. Prikl. Khim. (Leningrad), 44, 45, 1971.
82. Baena, Y., Pinzon, J.A., Barbosa, H.J., and Martinez, F., Phys. Chem. Liquids, 42, 603, 2004.
83. Schmitt, W.J. and Reid, R.C., J. Chem. Eng. Data, 31, 204, 1986.
84. Wibawa, G., Hatano, R., Sato, Y., Takishima, S., and Masuoka, H., J. Chem. Eng. Data, 47, 1022, 2002.
85. Herskowltz, M. and Gottlleb, M., J. Chem. Eng. Data, 30, 233, 1985.
86. Nakajima, A., Fujiwara, H., and Hamada, F., J. Polym. Sci. A-2, 4, 507, 1966.
87. Freire, M.G., Carvalho, P.J., Gardas, R.L., Marrucho, I.M., Santos, L.M.N.B.F., Coutinho, J.A.P., J. Phys.
Chem. B, 112, 1604, 2008.
88. Andrews, A.W. and Morcon, K.W., J. Chem. Thermodyn., 3, 519, 1971.
89. Kehiaian, H.V., Sosnkowska-Kehiaian, K., and Hryniewicz, R., J. Chim. Phys. Phys. Chim. Biol., 68,
929, 1971.
90. Ohji, H. and Tamura, K., J. Chem. Thermodyn., 35, 1591, 2003.
91. De Torre, A., Velasco, I., Otin, S., and Gutierrez Losa, C., J. Chem. Thermodyn., 12, 87, 1989.
92. Rowlinson, J.S. and Widom, B., Molecular Theory of Capillarity, Clarendon Press: Oxford, U.K., 1982.
93. Gibbs, J.W., Collected Works, Vol. I, Longmans, Green: New York, 1906.
94. Sprow, F.B. and Prausnitz, J.M., Can. J. Chem. Eng., 45, 25, 1967.
95. Rusanov, A.I., Pure Appl. Chem., 64, 111, 1992.
96. Hill, T.L., J. Chem. Phys., 56, 526, 1952.
97. Toxvaerd, S., J. Chem. Phys., 55, 3116, 1971.
98. Haile, J.M., Gray, C.G., and Gubbins, K.E., J. Chem. Phys., 64, 2569, 1976.
99. van der Waals, J.D., Z. Phys. Chem., 657, 1894.
100. Cahn, J.W. and Hilliard, J.E., J. Chem. Phys., 30, 1121, 1958.
101. Widom, B., J. Chem. Phys., 43, 3892, 1965.
102. Yang, A.J.M., Fleming, P.D., III, and Gibbs, J.H., J. Chem. Phys., 64, 3732, 1976.
103. Bongiorno, V., Scriven, L.E., and Davis, H.T., J. Colloid Interface Sci., 57, 462, 1976.
Molecular Thermodynamics of Hydrogen-Bonded Systems 197
CONTENTS
3.1 Introduction........................................................................................................................ 199
3.2 Choice of Solvents..............................................................................................................200
3.2.1 Rules of Thumb Based on Solubility Parameters................................................... 201
3.2.2 Rule of Thumb Based on the Infinite Dilution Activity Coefficient......................202
3.2.3 Rule of Thumb Based on the Flory–Huggins Model............................................. 203
3.3 Free-Volume Activity Coefficient Models..........................................................................206
3.3.1 The Free-Volume Concept......................................................................................206
3.3.2 Entropic-FV Model................................................................................................207
3.3.3 Results and Discussion...........................................................................................208
3.3.4 Correlative Versions of the Entropic-FV model..................................................... 213
3.3.4.1 Entropic-FV/UNIQUAC Model............................................................... 213
3.3.4.2 Extension of Free-Volume Models to Gas Solubilities in Elastomers..... 214
3.4 SAFT Family of Equations of State................................................................................... 215
3.4.1 Introduction............................................................................................................ 215
3.4.2 SAFT Equation of State—History and Introduction.............................................. 217
3.4.3 Original SAFT........................................................................................................ 217
3.4.4 “Original” SAFT.................................................................................................... 220
3.4.5 Simplified SAFT..................................................................................................... 223
3.4.6 LJ-SAFT................................................................................................................. 223
3.4.7 SAFT-VR................................................................................................................ 225
3.4.8 PC-SAFT................................................................................................................ 226
3.4.9 Simplified PC-SAFT.............................................................................................. 227
3.4.10 Extensions to SAFT-Type Equations...................................................................... 229
3.4.11 Applications of SAFT............................................................................................. 230
3.5 Concluding Remarks and Future Challenges..................................................................... 239
3.A Appendix............................................................................................................................240
3.A.1 An Expression of the FH Model for Multicomponent Mixtures........................................240
Abbreviations.................................................................................................................................. 241
References....................................................................................................................................... 242
3.1 INTRODUCTION
Knowledge of phase behavior (thermodynamics) of polymer solutions is important for the design of
many processes and products, including many specific applications in colloid and surface chemistry.
Among the many applications, we can mention the following:
1. The design of separations for the removal of unreacted monomers, colorants, by-products,
and additives after solution or emulsion polymerizations1
199
200 Handbook of Surface and Colloid Chemistry
2. Selection of appropriate mixed solvents for paints and coatings, which can meet strict pro-
duction and environmental criteria (fewer VOCs, water-based products)2,3
3. Numerous applications in the coatings industry, for example, in the control of emissions
during production and application, the effect of temperature, and swelling of film or sorp-
tion of gases and chemicals from the atmosphere3–6
4. Novel recycling methods for polymer waste based on physicochemical methods, the so-
called selective dissolution7,8
5. Design of advanced materials based on compatible polymer blends9 or novel structures, for
example, star or hyberbranched polymers10,11
6. Permeabilities of gases in the flexible polymeric pipelines used, for example, in the North
Sea for transporting hydrocarbons from the seabed to the surface12,13
7. Use of CO2 as user-friendly solvent for many polymer-related applications including those
involving paints and coatings14–16
8. Simultaneous representation of bulk and surface thermodynamic properties17
9. Migration of plasticizers from PVC18–21
10. Separation of proteins based on the so-called aqueous two-phase systems using polymers
like PEG or dextran1,22,23
11. Choice of suitable solvents for polymers and especially copolymers used as stabilizers in
colloidal dispersions
The aforementioned list shows some of the many applications where polymer thermodynamics plays
a key role. Polymer solutions and blends are complex systems: frequent existence of liquid–liquid
equilibria (LLE) (UCST, LCST, closed loop, etc.), the significant effect of temperature and polymer
molecular weight including polydispersity in phase equilibria, free-volume (FV) effects, and other
factors may cause difficulties. For this reason, many different models have been developed for poly-
mer systems and often the situation may seem rather confusing to the practicing engineer.
The choice of a suitable model will depend on the actual problem and depends, specifically, on
This chapter will present key tools in polymer thermodynamics at three different levels from the
“simpler” but also “more easy-to-use” methods up to the “more advanced” but also “more complex”
and potentially more accurate approaches:
1. Rules of thumb for choosing solvents including a discussion of the Flory–Huggins (FH)
approach (Section 3.2)
2. FV activity coefficient models based on UNIFAC (Section 3.3), which are often predictive,
and have found widespread applications at low pressures
3. The advanced statistical-associating fluid theory (SAFT) (Section 3.4), which is a
theoretically based equation of state, which can be rigorously applied, unlike the activity
coefficient models, at both low and high pressures
(the FH parameter). It can be roughly stated that a chemical (1) will be a good solvent for a specific
polymer (2), or in other words, the two compounds will be miscible if one (or more) of the following
“rules of thumb” are valid:
1. If the polymer and the solvent have similar hydrogen bonding degrees:
A negative value implies that a solvent–polymer system forms a homogeneous solution, that is, the
two components are miscible. Since the contribution of the entropic term (−TΔS) is always negative,
it is the heat of mixing term that determines the sign of the Gibbs energy. The heat of mixing can be
estimated from various theories, for example, the Hildebrand regular solution theory for nonpolar
systems, which is based on the concept of the solubility parameter. For a binary solvent (1)–polymer
(2) system, according to the regular solution theory,
where φi is the so-called volume fraction of component i. This is defined via the mole fractions xi
and the molar volumes Vi as (for binary systems):
xiVi
ϕi = (3.5)
xiVi + x jV j
According to Equation 3.4, the heat of mixing is always positive. For some systems with specific
interactions (hydrogen bonding), the heat of mixing can be negative and Equation 3.4 does not hold.
202 Handbook of Surface and Colloid Chemistry
Thus, the regular solution theory is strictly valid for nonpolar–slightly polar systems, without any
specific interactions.
According to Equations 3.3 and 3.4, if solvent and polymer have the same solubility parameters,
the heat of mixing is zero and they are therefore miscible at all proportions. The lower the solubility
parameter difference, the larger the tendency for miscibility. Many empirical rules of thumb have
been proposed based on this observation. Seymour24 suggests that if the difference of solubility
parameters is below 1.8 (cal/cm3)1/2, then polymer and solvent are miscible (Equation 3.1).
Similar rules can be applied for mixed solvent–polymer systems, which are very important in
many practical applications, for example, in the paints and coatings industry and for the separation
of biomolecules using aqueous two-phase systems.
The solubility parameter of a mixed solvent is given by the equation:
δ= ∑ ϕ δ (3.6)
i
i i
Barton25,26 provides empirical methods based on solubility parameters for ternary solvent systems.
Charles Hansen introduced the concept of 3D solubility parameters, which offers an exten-
sion of the regular solution theory to polar and hydrogen bonding systems. Hansen observed
that when the solubility parameter increments of the solvents and polymers are plotted in 3D
plots, then the “good” solvents lie approximately within a sphere of radius R (with the poly-
mer being in the center). This can be mathematically expressed as shown in Equation 3.2. The
quantity under the square root is the distance between the solvent and the polymer. Hansen
found empirically that a universal value 4 should be added as a factor in the dispersion term to
approximately attain the shape of a sphere. This universal factor has been confirmed by many
experiments. Hansen 27 in his book provides a review of the method together with extensive
tables of parameters.
The Hansen method is very valuable. It has found widespread use particularly in the paints
and coatings industry, where the choice of solvents to meet economical, ecological, and safety
constraints is of critical importance. It can explain some cases in which polymer and solvent solu-
bility parameters are almost perfectly matched and yet the polymer will not dissolve. The Hansen
method can also predict cases where two nonsolvents can be mixed to form a solvent. Still, the
method is approximate; it lacks the generality of a full thermodynamic model for assessing mis-
cibility and requires some experimental measurements. The determination of R is typically based
on visual observation of solubility (or not) of 0.5 g polymer in 5 cm3 solvent at room temperature.
Given the concentration and the temperature dependence of phase boundaries, such determination
may seem a bit arbitrary. Still the method works out pretty well in practice, probably because the
liquid–liquid boundaries for most polymer–solvent systems are fairly “flat.”
γ i∞ = lim γ i
xi →0
(3.7)
⎛xγ ⎞ M
Ω = lim ⎜ i i ⎟ = γ1∞ pol
∞
i
wi→0 ⎝ wi ⎠ M solv
Thermodynamics of Polymer Solutions 203
∞
The latter part of the equation is valid for a binary solvent–polymer solution and γ1 is the infinite
dilution activity coefficient of the solvent.
The weight-based infinite dilution activity coefficient, Ω1∞, which can be determined experimen-
tally from chromatography, is a very useful quantity for determining good solvents. As mentioned
in the previous section, low values (typically below 6) indicate good solvents, while high values
(typically above 10) indicate poor solvents according to rules of thumb discussed by several investi-
gators.28,29 The derivation of this rule of thumb is based on the FH model.
This method for solvent selection is particularly useful, because it avoids the need for direct
liquid–liquid measurements and it makes use of the existing databases of solvent infinite dilution
activity coefficients, which is quite large (e.g., the DECHEMA and DIPPR databases30,31).
Moreover, in the absence of experimental data, existing thermodynamic models (such as the FH,
the Entropic-FV, and the UNIFAC-FV discussed later) can be used to predict the infinite dilution
activity coefficient. Since, in the typical case today, existing models perform much better for VLE
and activity coefficient calculations than directly for LLE calculations, this method is quite valuable
and successful, as shown by sample results in Table 3.1.
This rule of thumb makes use of either experimental or predicted, by a model, infinite dilution
activity coefficients. However, the results depend not only on the accuracy of the model, but also
on the rule of thumb, which in turns depends on the assumptions of the FH approach. A thermo-
dynamically more correct method is to employ the activity–concentration diagram. The maximum
indicates phase split, while a monotonic increase of activity with concentration indicates a single
liquid phase (homogeneous solutions).
ϕ1 ϕ ϕ ⎛ 1⎞
ln γ1 = ln + 1 − 1 + χ12ϕ22 = ln 1 + ⎜ 1 − ⎟ ϕ2 + χ12ϕ22 (3.8)
x1 x1 x1 ⎝ r ⎠
where
φi can be volume or segment fractions
r is the ratio of the polymer volume to the solvent volume V2/V1 (approximately equal to the
degree of polymerization)
Appendix 3.A presents the general expression for the FH model suitable for multicomponent
systems.
Using standard thermodynamics and Equation 3.8, it can be shown that for high molecular
weight polymer–solvent systems, the polymer critical concentration is close to zero and the interac-
tion parameter has a value equal to 0.5. Thus, a good solvent (polymer soluble in the solvent at all
proportions) is obtained if χ12 ≤ 0.5, while values greater than 0.5 indicate poor miscibility. Since
the FH model is only an approximate representation of the physical picture and particularly the FH
parameter is often not a constant at all, this empirical rule is certainly subject to some uncertainty.
Nevertheless, it has found widespread use and its conclusions are often in good agreement with
experiment.
204 Handbook of Surface and Colloid Chemistry
TABLE 3.1
Choice of Suitable Solvents Using the Ω1∞-Rule of Thumb for PBMA Systems
(See Section 3.2)
Solvent S/NS EFV UFV GCFL
Hexane NS 7.1 7.0 10.7
n-Octane NS 6.7 6.3 10.4
n-Decane NS 6.5 6.0 10.7
n-Dodecane NS 6.6 6.0 11.3
n-Hexadecane NS 6.8 6.1 13.2
Toluene S 3.2 4.4 4.7
Xylene S 2.3 3.6 5.7
Methylene dichloride S 3.3 2.5 3.0
Chloroform S 1.9 2.1 1.7
Carbon tetrachloride S 2.2 2.2 2.9
Ethylene dichloride S 3.5 3.0 —
Trichloroethylene S 2.5 2.9 33.9
Chlorobenzene S 2.5 3.0 3.0
o-Dichlorobenzene S 1.3 2.5 2.7
Acetone S 10.9 14.2 11.1
MEK S 8.4 10.5 8.2
MIBK S 6.3 7.7 5.7
Acetophenone S 8.1 9.3 8.6
Ethyl acetate S 6.7 6.7 60.3
Butyl acetate S 5.3 5.1 31.4
Diethyl ether S 5.2 5.8 11.6
THF S 3.8 4.0 —
1,4-Dioxane S 4.1 4.4 159.4
Methanol NS 43.7 57.7 35.7
Ethanol NS 29.2 31.3 17.3
1-Butanol NS 18.1 17.1 8.1
Cyclohexanol NS 24.3 20.1 3.0
Ethylene glycol NS 277.8 — 15,947.0
Propylene glycol NS 212.6 — 1,879.2
1,3-Butanediol NS 158.5 — 525.5
Glycerol NS 294.6 — 2,282.4
Isopropanol NS 23.4 21.6 10.6
Isobutanol S 19.0 17.9 7.9
Diethylene glycol NS 240.1 — 2,470.4
Dipropylene glycol NS 127.9 945.7 287.8
Nitromethane NS 16.7 17.2 —
1-Nitropropane S 4.7 5.2 —
N,N-dimethylformamide S 3.8 — —
Source: Modified from Lindvig, Th. et al., AIChE J., 47(11), 2573, 2001.
Notes: EFV, Entropic-FV; UFV, UNIFAC-FV; GCFL, GC-Flory; S, solvent; NS, nonsolvent (experimental observation).
Thermodynamics of Polymer Solutions 205
There are several, still rather obscure issues about the FH model, which we summarize here
together with some recent developments:
1. The FH parameter is typically not a constant and should be estimated from experimental
data. Usually, it varies with both temperature and concentration, which renders the FH
model useful only for describing experimental data. It cannot be used for predicting phase
equilibria for systems for which no data is available. Moreover, when fitted to the critical
solution temperature, the FH model cannot yield a good representation of the whole shape
of the miscibility curve with a single parameter.
2. Accurate representation of miscibility curves is possible with the FH model using suitable
(rather complex) equations for the temperature and the concentration dependence of the
FH parameter.34
3. In some cases, a reasonable value of the FH parameter can be estimated using solubility
parameters via the equation:
V1
χ12 = χ s + χh = 0.35 + (δ1 − δ2 )2 (3.9)
RT
Equation 3.9, without the empirical 0.35 factor, is derived from the regular solution the-
ory. The constant 0.35 is added for correcting for the deficiencies of the FH combina-
torial term. These deficiencies become evident when comparing experimental data for
athermal polymer and other asymmetric solutions to the results obtained with the FH
model. A systematic underestimation of the data is observed, as discussed extensively in
the literature,28,35 which is often attributed to the inability of the FH model in accounting
for the FV differences between polymers and solvents or between compounds differing
significantly in size such as n-alkanes with very different chain lengths. The term, which
contains the “0.35 factor,” corrects in an empirical way for these FV effects. However,
and although satisfactory results are obtained in some cases, we cannot generally rec-
ommend Equation 3.9 for estimating the FH parameter. Moreover, for many nonpolar
systems with compounds having similar solubility parameters, the empirical factor 0.35
should be dropped.
4. Lindvig et al.36 proposed an extension of the FH equation using the Hansen solubility
parameters for estimating activity coefficients of complex polymer solutions.
ϕ1 ϕ
ln γ1 = ln + 1 − 1 + χ12ϕ22
x1 x1
V
χ12 = 0.6 1 ⎡⎣(δd1 − δd 2 )2 + 0.25(δ p1 − δ p22 )2 + 0.25(δh1 − δh 2 )2 ⎤⎦ (3.10)
RT
In order to achieve that, Lindvig et al.36 have, as shown in Equation 3.10, employed
a universal correction parameter, which has been estimated from a large number of
polymer–solvent VLE data. Very good results are obtained, especially when the vol-
ume-based combinatorial term of FH is employed, also for ternary polymer–solvent
systems.37
5. Based on the FH model, several techniques have been proposed for interpreting and for cor-
relating experimental data for polymer systems, for example, the so-called Schultz–Flory
206 Handbook of Surface and Colloid Chemistry
(SF) plot. Schultz and Flory38 have developed, starting from the FH model, the following
expression, which relates the critical solution temperature (CST), with the theta tempera-
ture and the polymer molecular weight:
1 1⎡ 1⎛ 1 1 ⎞⎤
= ⎢1 + ⎜ + ⎟ ⎥ (3.11)
CST Θ ⎣ ψ ⎝ r 2r ⎠ ⎦
where
ψ = ((1/2)−χs) is the entropic parameter of the FH model (Equation 3.9)
r is the ratio of molar volumes of the polymer to the solvent
This parameter is evidently dependent on the polymer’s molecular weight. The SF plot can
be used for correlating data of critical solution temperatures for the same polymer–solvent
system, but at different polymer molecular weights. This can be done, as anticipated from
Equation 3.11, because the plot of 1/CST against the quantity in parentheses in Equation
3.11 is linear. The SF plot can also be used for predicting CST for the same system but at
different molecular weights than those used for correlation as well as for calculating the
theta temperature and the entropic part of the FH parameter. It can be used for correlating
CST/molecular weight data for both the UCST and LCST areas. Apparently different coef-
ficients are needed.
V f = V − V * = V − Vw (3.12)
originally proposed by Bondi39 and later adopted by Elbro et al.35 and Kontogeorgis et al.40
in the so-called Entropic-FV model (described later). According to this equation, FV is just
Thermodynamics of Polymer Solutions 207
the “empty” volume available to the molecule when the molecules’ own (hard-core or closed-
packed V*) volume is subtracted.
The FV is not the only concept, which is loosely defined in this discussion. The hard-core volume
is also a quantity difficult to define and various approximations are available. Elbro et al.35 suggested
using V* = Vw, that is, equal to the van der Waals volume (Vw), which is obtained from the group
increments of Bondi and is tabulated for almost all existing groups in the UNIFAC tables. Other
investigators41 interpreted somewhat differently the physical meaning of the hard-core volume in
the development of improved FV expressions for polymer solutions, which employ Equation 3.12 as
basis, but with V* values higher than Vw (about 1.2–1.3Vw).
3.3.2 Entropic-FV Model
The original UNIFAC model does not account for the FV differences between solvents and p olymers
and, as a consequence of that, it highly underestimates the solvent activities in polymer solutions.
One of the most successful and earliest such models for polymers is the UNIFAC-FV by Oishi and
Prausnitz.42 The UNIFAC-FV model was developed for solvent activities in polymers, but it cannot
be successfully applied to LLE.
A similar to UNIFAC-FV but somewhat simpler approach, which can be readily extended to
multicomponent systems and LLE, is the so-called Entropic-FV model proposed by Elbro et al.35
and Kontogeorgis et al.40:
ln γ i = ln γ icomb − FV + ln γ res
i
φiFV φFV
ln γ icomb − FV = ln +1− i
xi xi
xiVi, FV xi (Vi − Vwi )
φiFV = =
∑ xV i
j j , FV ∑ x (V − V
j
j j wj )
ln γ res
i → UNIFAC (3.13)
As can been seen from Equation 3.13, the FV definition given by Equation 3.12 is employed. The
combinatorial term of Equation 3.13 is very similar to that of FH. However, instead of volume or seg-
ment fractions, FV fractions are used. In this way, both combinatorial and FV effects are combined
into a single expression. The combinatorial–FV expression of the Entropic-FV model is derived
from statistical mechanics, using a suitable form of the generalized van der Waals partition function.
The residual term of Entropic-FV is taken by the so-called new or linear UNIFAC model, which
uses a linear-dependent parameter table43:
This parameter table has been developed using the combinatorial term of the original UNIFAC
model. As with UNIFAC-FV, no parameter re-estimation has been performed. The same group
parameters are used in the “linear UNIFAC” and in the Entropic-FV models.
A common feature for both UNIFAC-FV and Entropic-FV is that they require values for the
volumes of solvents and polymers (at the different temperatures where application is required). This
can be a problem in those cases where the densities are not available experimentally and have to be
estimated using a predictive group-contribution or other method, for example, GCVOL44,45 or van
Krevelen methods. These two estimation methods perform quite well and often similarly even for
low molecular weight compounds or oligomers such as plasticizers.
208 Handbook of Surface and Colloid Chemistry
Both UNIFAC-FV and Entropic-FV, especially the former, are rather sensitive to the density
values used for the calculations of solvent activities.
As already mentioned, the Entropic-FV model has been derived from the van der Waals partition
function. The similarity of the model with the van der Waals equation of state P = RT/(V − b) − a/V 2
becomes apparent if the van der Waals equation of state is written (when the classical van der Waals
one fluid mixing and classical combining rules are used) as an activity coefficient model:
⎛ ϕiFV ϕFV ⎞ ⎛ V ⎞
ln γ i = ln γ icomb − FV + ln γ res
i = ⎜ ln + 1 − i ⎟ + ⎜ i (δi − δi )2 ϕ2j ⎟
⎝ xi xi ⎠ ⎝ RT ⎠
xi (Vi − bi )
ϕiFV = (3.15)
∑ x j (V j − b j )
j
ai
δi =
Vi
where φi is the volume fraction as defined in Equation 3.5. The first term in Equation 3.15 is the same
as in Entropic-FV with Vw = b, while the latter term is a regular solution theory or van Laar-type term.
1. Satisfactory predictions are obtained for solvent activities, even at infinite dilution, and for
nonpolar, as well as for complex polar and hydrogen bonding systems including solutions
of interest to paints and coatings. Rather satisfactory predictions are also achieved when
mixed solvents and copolymers are present.
TABLE 3.2
Applications of the Entropic-FV Model
Application References
VLE binary solutions [40,41]
VLE complex polymers–solvents [48]
VLE ternary polymer–solvents [37]
Paints [61]
Dendrimers
VLE copolymers [47]
VLE athermals systems [28,35,50,62]
SLE hydrocarbons [51,52]
SLE polymer–solvents [49]
Comparison with other models [40,46,53,60]
LLE polymer–solvents [58]
LLE ternary polymer–mixed solvents [8]
Polymer blends [59]
EFV + UNIQUAC
Thermodynamics of Polymer Solutions 209
TABLE 3.3
Prediction of the Solubility for Characteristic Polymer–Solvent Systems
Using the Ω ∞1 Rule of Thumb and Two FV Models for Solvent Selection
System Experiment Entropic-FV UNIFAC-FV
PBMA/nC10 NS 6.5 (–) 6.1 (–)
PBMA/xylene S 2.3 (S) 3.6 (S)
PBMA/CHCl3 S 1.9 (S) 9.1 (NS)
PBMA/acetone S 0.2 (NS) 14.1 (NS)
PBMA/ethyl acetate S 6.7 (–) 6.7 (–)
PBMA/ethanol NS 29.2 (NS) 31.3 (NS)
PMMA/acetone S 10.0 (NS) 16.5 (NS)
PMMA/ethyl acetate S 6.6 (–) 8.4 (NS)
PMMA/butanol NS 26.8 (NS) 14.4 (NS)
PEMA/MEK S 8.1 (NS) 11.7 (NS)
PEMA/diethyl ether S 5.8 (S) 7.6 (–)
PEMA/nitropropane NS 4.5 (S) 1.4 (S)
PVAc/hexane NS 38.7 (NS) 38.6 (NS)
PVAc/methanol S 18.9 (NS) 19.4 (NS)
PVAc/ethanol NS 15.2 (NS) 38.9 (NS)
PVAc/nitromethane S 3.9 (S) 3.8 (S)
PVAc/THF S 8.4 (NS) 5.6 (S)
Notes: S, good solvent; NS, nonsolvent; –, no answer according to the rule of thumb.
TABLE 3.4
Prediction of Infinite Dilution Activity Coefficients for Polyisoprene (PIP)
Systems with Two Predictive Group Contribution Models
PIP Systems Exper. Value Entropic-FV UNIFAC-FV
+Acetonitrile 68.6 47.7 (31%) 52.3 (24%)
+Acetic acid 37.9 33.5 (12%) 17.7 (53%)
+Cyclohexanone 7.32 5.4 (27%) 4.6 (38%)
+Acetone 17.3 15.9 (8%) 13.4 (23%)
+MEK 11.4 12.1 (6%) 10.1 (12%)
+Benzene 4.37 4.5 (2.5%) 4.4 (0%)
+1,2-Dichloroethane 4.25 5.5 (29%) 6.5 (54%)
+CCl4 1.77 2.1 (20%) 1.8 (0%)
+1,4-Dioxane 6.08 6.3 (4%) 5.9 (2%)
+Tetrahydrofurane 4.38 4.9 (14%) 3.9 (10%)
+Ethylacetate 7.47 7.3 (2%) 6.6 (11%)
+n-Hexane 6.36 5.1 (20%) 4.6 (27%)
+Chloroform 2.13 3.00 (41%) 2.6 (20%)
TABLE 3.5
Average Absolute Deviations between Experimental and Calculated
Activity Coefficients of Paint-Related Polymer Solutions Using the FH/
Hansen Method and Two FV Models
% AAD (Systems in % AAD
Model Database) Araldite 488 % AAD Eponol-55
FH/Hansen, Volume (Equation 3.10) 22 31 28
Entropic-FV 35 34 30
UNIFAC-FV 39 119 62
Source: Adapted from Lindvig, Th. et al., Fluid Phase Equilib., 203, 247, 2002.
The second column presents the systems used for optimization of the universal parameter (solutions
containing acrylates and acetates). The last two columns show predictions for two epoxy resins.
TABLE 3.6
Mean Percentage Deviations between Experimental and Calculated
Activity Coefficients of Solvents in Various Nearly Athermal Solutions
% AAD Infinite Dilution Conditions
γ 1∞ (Ω
Ω1∞ for Polymers) Entropic-FV UNIFAC-FV Flory-FV
Short n-alkanes/long alkanes 8 15 20
Short branched, cyclic alkanes/long alkanes 10 17 20
Alkanes/polyethylene 9 23 19
Alkanes/polyisobutylene 16 12 38
Organic solvent/PDMS, PS, PVAc 20 29 26
Overall 13 19 25
Source: Kouskoumvekaki, I. et al., Fluid Phase Equilib., 202(2), 325, 2002. With permission.
TABLE 3.7
Mean Percentage Deviations between Experimental and Calculated
Activity Coefficients of Heavy Alkanes Solutes in Alkane Solvents
Entropic-FV UNIFAC-FV Flory-FV
∞
% AAD infinite dilution conditions, γ
2
Source: Kouskoumvekaki, I. et al., Fluid Phase Equilib., 202(2), 325, 2002. With permission.
Thermodynamics of Polymer Solutions 211
TABLE 3.8
Average Absolute Logarithmic Percentage Deviations between Experimental and
Predicted Equilibrium Pressures and Average Absolute Deviation (×100) between
Calculated and Experimental Vapor Phase Compositions (Mole Fractions) for Various
Ternary Polymer–Mixed Solvent Systems
Sys. No. Variable SAFT EFV/UQ FH Pa-Ve EFV/UN UFV GCFL GCLF FHHa
1 P 11 6 6 6 2 1 21 8 2
y 4 3 3 3 3 3 3 3 3
2 P 4 2 14 8 2 2 — 2 11
y 5 2 5 1 2 2 — 4 5
3 P — — — — 3 3 12 5 18
y — — — — 3 3 2 4 4
4 P — — — — 4 4 13 9 5
y — — — — 18 18 19 19 15
5 P 14 16 11 4 17 19 93 5 52
y 17 17 16 11 17 13 2 18 14
Sources: Katayama, T. et al., Kagaku Kogaku, 35, 1012, 1971; Matsumara, K. and Katayama, T., Kagaku Kogaku, 38, 388,
1974; Tanbonliong, J.O. and Prausnitz, J.M., Polymer, 38, 5775, 1997.
Based on the results shown by Lindvig et al.37
Notes: 1, PS–toluene–ethylbenzene at 303 K; 2, PS–toluene–cyclohexane at 303 K; 3, PVAc–acetone–ethyl acetate at
303 K; 4, PVAc–acetone–methanol at 303 K; 5, PS–chloroform–carbon tetrachloride at 323.15 K.
TABLE 3.9
Prediction of Liquid–Liquid Equilibria for Polymer–Solvent Systems with Various
Thermodynamic Models
Polymer System Molecular Weight Entropic-FV N0065w UNIFAC GC-Flory
PS/acetone 4,800 84 21 75
PS/acetone 10,300 98 8 42
PS/cyclohexane 20,400 38 62 —
PS/cyclohexane 37,000 26 59 —
PS/cyclohexane 43,600 24 63 —
PS/cyclohexane 89,000 11 60 —
PS/cyclohexane 100,000 15 62 —
PS/cyclopentane 97,200 27 105 —
PS/cyclopentane 200,000 12 103 —
HDPE/n-butyl acetate 13,600 10 82 72
HDPE/n-butyl acetate 20,000 22 97 70
HDPE/n-butyl acetate 61,100 29 107 71
PMMA/1-chloro butane 34,760 53 — —
PBMA/n-pentane 11,600 Hourglass — —
PBMA/n-octane 11,600 155 — —
Absolute difference between experimental and predicted UCST for several polymer solutions using various models (based
on results from Ref. [58]). All results are predictions. The three last models are based on group contributions. The new
UNIFAC model is a combination of FH with the UNIFAC residual term.
Notes: PS, polystyrene; HDPE, high-density polyethylene; PMMA, poly(methyl methacrylate); PBMA, poly(butyl
methacrylate).
212 Handbook of Surface and Colloid Chemistry
600.00
500.00
19,800
400.00
Temperature (K)
10,300
300.00
4,800
200.00
FIGURE 3.1 Experimental and predicted LLE diagram for the system polystyrene/acetone at three polymer
molecular weights (4,800, 10,300, 19,800). The points are the experimental data and the lines are the predic-
tions with the Entropic-FV model. (From Kontogeorgis, G.M. et al., Ind. Eng. Chem. Res., 34, 1823, 1995.
With permission.)
Methyl cyclohexane
Experimental data
Entropic-FV model
HA model
Acetone PS
FIGURE 3.2 Ternary LLE for PS(300,000)/methyl cyclohexane/acetone, T = 298.15 K. (Reprinted from
Pappa, G.D. et al., Ind. Eng. Chem. Res., 36, 5461, 1997. With permission.)
1. Athermal solutions, for example, polyolefins with alkanes offer a way of testing FV terms
and numerous such investigations have been presented. FV models perform generally bet-
ter than those that do not contain volume-dependent terms.60 Better FV terms than that of
Thermodynamics of Polymer Solutions 213
Entropic-FV (Equation 3.13) have been proposed41,51,52,57,60 and they may serve as the basis
for future developments resulting in even better activity coefficient models for polymer
solutions. A rigorous test for newly developed FV expressions is provided by athermal
alkane systems, especially the activity coefficients of heavy alkanes in short-chain ones.
2. A difference of 10°–30° should be expected in UCST predictions with Entropic-FV, see
Table 3.9.
3. An alternative to Entropic-FV successful approach is the FH/Hansen model presented
previously (Equation 3.10). In this case, best results are obtained when the original FH
combinatorial rather than a FV term is used together with HSP. It seems that HSP incorpo-
rate some FV effects.
440
420
400
380
T (K)
360
340
320
300
0.00 0.04 0.08 0.12 0.16 0.20
Weight fraction of polymer
FIGURE 3.3 Correlation of liquid–liquid equilibria for the PVAL/water system with the Entropic-FV/
UNIQUAC model. (•) Exp. data (Mn = 140,000 g/mol); (—) correlation. (From Bogdanic, G. and Vidal, J.,
Fluid Phase Equilib., 173, 241, 2000. With permission.)
214 Handbook of Surface and Colloid Chemistry
475
450
425
400
375
T (K)
350
325
300
275
250
225
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Weight fraction of polymer
FIGURE 3.4 Correlation and prediction of liquid–liquid equilibria for the PBD/1-octane system with the
Entropic-FV/UNIQUAC model. (◾) Exp. data (Mv = 65,000 g/mol), (—) correlation; (▴) Exp. data (Mv =
135,000 g/mol), (⎯ ⎯) prediction; and (⦁) Exp. data (Mw = 44,500 g/mol), (- - - -) prediction. (From Bogdanic, G.
and Vidal, J., Fluid Phase Equilib., 173, 241, 2000. With permission.)
correlate both UCST/LCST and closed-loop behavior and even show the pressure dependency of
critical solution temperatures (UCST and LCST).
ln γ i = ln γ iR + ln γ Ci + FV (3.16)
where
δ1 is the solvent solubility parameter
δ2 is the gas solubility parameter
x2 is the gas mole fraction in liquid/polymer
Φ2 is the “apparent” volume fraction of solvent, given by
x2 ⋅ V2L
Φ2 =
x1 ⋅ V1L + x2 ⋅ V2L
Thermodynamics of Polymer Solutions 215
Φ 2FV =
(
x2 ⋅ V2L − V2VW )
(
x2 ⋅ V − V
2
L
2
VW
) + x ⋅ (V
1 1
L
− V1VW )
V2L is a hypothetical liquid volume of the (gaseous) solute.
1 fl V L ⋅ (δ1 − δ2 )2 ⋅ Φ 21
= 2g ⋅ exp 2 (3.18)
x2 fˆ2 R ⋅T
f̂2g is the fugacity of the gas
f2l is the fugacity of the hypothetical liquid, which can be estimated from the equation:
f2l 4.74547
ln = 3.54811 − + 1.60151 ⋅ Tr − 0.87466 ⋅ Tr2 + 0.10971 ⋅ Tr3 (3.19)
Pc Tr
Finally, the gas solubility in the polymer is estimated from the equation:
1 fl ⎛ V2L ⋅ ( δ1 − δ2 )2 ⋅ Φ12 Φ FV Φ FV ⎞
= 2g ⋅ exp ⎜ + ln 2 + 1 − 2 ⎟ (3.20)
ˆ
x2 f2 ⎜ R ⋅T x2 x2 ⎟
⎝ ⎠
Calculations showed that the hypothetical gas “liquid” volumes are largely independent to the polymer
used, and moreover, for many gases (H2O, O2, N2, CO2 and C2H2), these are related to the critical
volume of the gas by the equation:
TABLE 3.10
Summary of the Performance of the Models Tested at T = 298 K; P = 101.3 kPa
Hildebrand Hildebrand
Polymer Gas Michaels/Bixler Tseng/Lloyd Hildebrand/Scott Entropic-FV-1 Entropic-FV-2
PIP N2 14.7 73 3.9 −7.9 −4.6
O2 −16.1 −4 14 10.8 11.8
Ar −32.5 −23 — 29.4 −22.2
CO2 −3.2 13 4.5 8.7 4.6
PIB N2 −2.5 — 6.8 3.1 5.0
O2 −6.1 — −1.7 −8.3 1.7
Ar — — — — —
CO2 32.8 — −1.9 41.1 35.2
PBD N2 22.3 — 8.1 8.1 12.6
O2 14.9 — −6 8.7 10.8
Ar 12.1 — — 111.1 24.0
CO2 −9.7 — −4.6 0.4 −4.0
PDMB N2 — — −23 −7.5 −3.1
O2 — — −32 −16.8 −15.9
Ar — — — — —
CO2 — — −24 2.3 −2.2
PCP N2 58.1 — 49 −7.0 −4.2
O2 43.7 — 60 −1.4 −1.4
Ar — — — — —
CO2 8.8 — 27 −13.3 −17.1
AAD 19.8 28 18 16.8 10.6
where
aseg is the Helmholtz energy of the segment, including both hard-sphere reference and dispersion
terms
achain is the contribution from chain formation
aassoc is the contribution from association
218 Handbook of Surface and Colloid Chemistry
Chain Association
side side
(a)
(b) (c)
(e) (d)
FIGURE 3.5 Procedure to form a molecule in the SAFT model. (a) The proposed molecule. (b) Initially
the fluid is a hard-sphere fluid. (c) Attractive forces are added. (d) Chain sites are added and chain molecules
appear. (e) Association sites are added and molecules form association complexes through association sites.
(From Fu, Y.-H. and Sandler, S.I., Ind. Eng. Chem. Res., 34, 1897, 1995. With permission.)
It is this separation of the Helmholtz energy into additive components that is partly respon-
sible for the fact that SAFT has so many variants—each of the contributions can be considered
(and modified) separately, or new terms may be added (such as polar, electrostatic, or other con-
tributions). The individual terms may also be used outside of the context of SAFT. Thus, the
term that accounts for association has been combined (with minor modification) with the SRK
equation to create CPA, an equation of state, which has had remarkable success in a number of
applications.81
It is worth noting that both the chain formation and the association term derive from Wertheim’s
TPT. However, chains (i.e., covalent bonds) are formed in the limit of complete association. It is
these two terms that make up the innovative development in SAFT—the first because now we
have available a rational method for considering polymer molecules (linear chains with very many
bonded segments) and the second because we can now consider associating molecules in a way that
more closely resembles the actual physical picture. The calculation of useful thermodynamic prop-
erties from aassoc was initially a complex procedure, although Michelsen and Hendriks82 simplified
the computations considerably by recasting the equations in a simpler, although mathematically
equivalent form.
We now consider each of the terms individually. We follow the original notation of Chapman
et al.74 Each pure component is characterized by a chain length m, a size-parameter σ, and
an energy parameter, ε. If the molecule is self-associating, there are two further parameters,
Thermodynamics of Polymer Solutions 219
which characterize the volume (κ AiBi) and energy (εAiBi) of association. The association term is
given for mixtures by
a assoc ⎡ ⎡ X Ai ⎤ 1 ⎤
∑ ∑ Xi ⎢ ⎢ ln X − ⎥ + Mi ⎥ (3.23)
Ai
=
RT i ⎢⎣ Ai ⎣ 2 ⎦ 2 ⎥
⎦
where
X Ai is the mole fraction of molecules i not bonded at site A
Mi is the number of association sites on molecule i
We have
−1
⎡ ⎤
X Ai
= ⎢1 + N AV ∑∑ρ X j
Bj
Δ Ai B j
⎥ (3.24)
⎢ j Bj
⎥
⎣ ⎦
where
NAV is the Avogadro’s number
ρj is the molar density of j
Δ AiBj is the association strength given by
⎡ ⎛ ε Ai B j ⎞ ⎤
Δ Ai B j = dij3 gij (dij )seg κ Ai B j ⎢exp ⎜⎜ ⎟⎟ − 1⎥ (3.25)
⎢⎣ ⎝ kT ⎠ ⎥⎦
⎛ kT ⎞
d = σf ⎜ , m ⎟ (3.26)
⎝ ε ⎠
where
⎛ kT ⎞ 1 + 0.2977kT /ε
f⎜ ,m⎟ = (3.27)
⎝ ε ⎠ 1 + 0. 33163 kT /ε + f (m)(kT /ε)2
and
m −1
f (m) = 0.0010477 + 0.025337 (3.28)
m
This temperature dependence is incorporated to account for the fact that real molecules are not
hard spheres, but rather there is some degree of interpenetration between molecules, particularly
at high temperatures. Thus, the “effective” hard-sphere diameter of a segment is smaller at higher
temperatures. The radial distribution function in Equation 3.25 is given by the mixture version of
the Carnahan–Starling equation of state for hard-sphere mixtures:
2
1 ⎛ dd ⎞ 3ζ 2 ⎛ d d ⎞ 2ζ 22
gijseg (dij ) ≈ gijHS dij+ =
( ) +⎜ i j ⎟ +⎜ i j ⎟ (3.29)
1 − ζ 3 ⎝ di + d j ⎠ (1 − ζ 3 ) ⎝ di + d j ⎠ (1 − ζ 3 )
2 3
220 Handbook of Surface and Colloid Chemistry
where
πN AV
ζk =
6
ρ ∑X md
i
i
k
i ii (3.30)
a chain
RT
= ∑ X (1 − m ) ln(g (d )
i
i i ii ii
HS
) (3.31)
where the 0 subscript indicates a nonassociated segment. The segment energy consists of a hard-
sphere reference and a dispersion contribution:
The Carnahan–Starling equation66 is used for both pure components and mixtures to give
a0HS 4η − 3η2
= (3.34)
RT (1 − η)2
where for mixtures η = ζ3 as defined by Equation 3.30. The dispersion term is given by
εR ⎛ disp a02
disp
⎞
a0disp = ⎜ a01 + ⎟ (3.35)
k ⎝ TR ⎠
where
a01
disp
= ρ R ⎣⎡ −0.85959 − 4.5424ρ R − 2.1268ρ2R + 10.285ρ3R ⎤⎦ (3.36)
a02
disp
= ρ R ⎡⎣ −1 − 9075 − 9.9724ρ R − 22.216ρ2R + 15.904ρ3R ⎤⎦ (3.37)
differences. Thus, instead of a size parameter σ, they use a volume parameter v 00, which is
related through the equation
πN AV 3
v 00 = σ (3.38)
6τ
Here τ = 0.74048 is the highest possible packing fraction for a system of pure hard spheres. They also
use the notation u 0 instead of ε for the energy parameter, although these terms are completely
equivalent.
The rather complex temperature dependence of the hard-sphere diameter given by Equations
3.26 through 3.28 was simplified by Huang and Radosz, following Chen and Kreglewski83 to
⎡ ⎡ −3u0 ⎤ ⎤
d = σ ⎢1 − 0.12 exp ⎢ ⎥ ⎥ (3.39)
⎢⎣ ⎣ kT ⎦ ⎥⎦
The dispersion term is also different from that of Chapman et al.74 and is given by
i j
a0disp ⎡ u ⎤ ⎡η⎤
RT
= ∑∑
i j
Dij ⎢ ⎥ ⎢ ⎥ (3.40)
⎣ kT ⎦ ⎣ τ ⎦
where
Dij are universal constants
u is the temperature-dependent energy parameter given by u = u 0(1 + e/kT) and e/k is a constant
set to −10, with a few exceptions
Another important contribution of Huang and Radosz is the presentation of detailed tables discuss-
ing bonding schemes for different associating fluids. These schemes are presented as Tables 3.11
and 3.12 and have been widely adopted in the literature of SAFT and other equations of state for
associating fluids.
In their paper on mixture properties,79 Huang and Radosz also use the full mixture version of the
Carnahan–Starling equation for the hard-sphere mixtures reference system:
a HS 1 ⎡ 3ζ ζ ζ 32 ⎛ ζ 32 ⎞ ⎤
= ⎢ 1 2+ + ⎜ 2 − ζ 0 ⎟ ln(1 − ζ 3 ) ⎥ (3.41)
RT ζ 0 ⎢⎣ 1 − ζ 3 ζ 3 (1 − ζ 3 ) ⎝ ζ 3
2
⎠ ⎥⎦
One of the reasons that the Huang and Radosz version of SAFT has been adopted (and is widely
referred to as “original” SAFT) is that they undertook an extensive pure-component parameter-
ization for over 100 pure fluids. This meant that their equation of state could be used imme-
diately for real fluids of industrial interest without any intermediate steps. Both Huang and
Radosz were employed by Exxon during the development of SAFT. The fact that the model
had the backing of a major oil company may also help explain its rapid adoption and use as an
engineering tool.
Besides it is interesting to note that the first paper to appear describing PC-SAFT also had pure-
component parameters for about 100 species.84 This fact, coupled with the success of the model, has
certainly been partly responsible for the rapid adoption of PC-SAFT in the few years it has been in
existence, both in industry and in academia.
222 Handbook of Surface and Colloid Chemistry
TABLE 3.11
Unbonded Site Fractions XA for Different Bonding Types
Type Δ Approximations XA Approximations XA
−1 + (1 + 4ρΔ)1/ 2
1 ΔAA ≠ 0
2ρΔ
−1 + (1 + 8ρΔ)1/ 2
2A ΔAA = ΔAB = ΔBB ≠ 0 XA = XB
4ρΔ
2B ΔAA = ΔBB = 0 XA = XB −1 + (1 + 4ρΔ)1/ 2
ΔAB ≠ 0 2ρΔ
−1 + (1 + 12ρΔ)1/ 2
3A ΔAA = ΔAB = ΔBB = ΔAC = ΔBC = ΔCC ≠ 0 XA = XB = XC
6ρΔ
3B ΔAA = ΔAB = ΔBB = ΔCC = 0 XA = XB −(1 − ρΔ) + ((1 + ρΔ)2 + 4ρΔ)1/ 2
ΔAC = ΔBC ≠ 0 XC = 2XA − 1 4ρΔ
4A ΔAA = ΔAB = ΔBB = ΔAC = ΔBC = ΔCC = ΔAD = XA = XB = XC = XD −1 + (1 + 16ρΔ)1/ 2
ΔBD = ΔCD = ΔDD ≠ 0 8ρΔ
4B ΔAA = ΔAB = ΔBB = ΔAC = ΔBC = ΔCC = ΔDD = 0 XA = XB = XC −(1 − 2ρΔ) + ((1 + 2ρΔ)2 + 4ρΔ)1/ 2
ΔAD = ΔBD = ΔCD ≠ 0 XD = 3XA − 2 6ρΔ
4C ΔAA = ΔAB = ΔBB = ΔCC = ΔCD = ΔDD = 0 XA = XB = XC = XD −1 + (1 + 8ρΔ)1/ 2
ΔAC = ΔAD = ΔBC = ΔBD ≠ 0 4ρΔ
Source: Huang, S.H. and Radosz, M., Ind. Eng. Chem. Res., 29, 2284, 1990. With permission.
TABLE 3.12
Types of Bonding in Real Associating Fluids
Species Formula Rigorous Type Assigned Type
Acid O HO 1 1
C C
OH O
Alkanol A 3B 2B
O
B
C H
Water A 4C 3B
B O H c
H D
Amines
Tertiary A 1 Non-self-associating
N
Secondary A 2B 2B
N
HB
Primary C 3B 3B
N HB
HA
Ammonia D 4B 3B
AH N HB
HC
Source: Huang, S.H. and Radosz, M., Ind. Eng. Chem. Res., 29, 2284, 1990. With permission.
Thermodynamics of Polymer Solutions 223
a disp ⎛ vs ⎞
= mZ M ln ⎜ ⎟ (3.42)
RT ⎜ vs + v*Y ⎟
⎝ ⎠
where
“average” chain length m = ∑ i
xi mi
Z M = 36 is the maximum coordination number
vs = 1/ρm is the total molar volume of a segment
v*Y =
N AV ∑ ∑ x x m m ( d 2 ) ⎡⎣exp(u /kT ) − 1⎤⎦ (3.43)
i j
i j i j
3
ij ij
∑ ∑ xx mm i j
i j i j
The remaining terms have the same meaning as in Huang and Radosz.78 Generally, simplified SAFT
performs as well as Huang and Radosz SAFT, although it requires refitting all the pure-component
parameters. Fu and Sandler provide parameters for 10 nonassociating and 8 associating fluids.
Table 3.13 is reproduced from Fu and Sandler. It is interesting because it presents different types of
cross-association in some detail. This scheme is completely general and applicable to any equation
of state incorporating association. Combining rules for cross-association also need to be introduced,
however, and are far from self-evident.
3.4.6 LJ-SAFT85,86
The main change in the SAFT version of Kraska and Gubbins85,86 is that they use Lennard–Jones
(LJ) spheres for the reference term, rather than hard spheres. The remaining terms are unchanged,
except that the radial distribution function used in the calculation of the chain and association
contributions in Equations 3.25 and 3.31 is the radial distribution function for LJ spheres rather
than hard spheres. Thus, an equation of state for LJ spheres is required. The equation used is that
of Kolafa and Nezbeda.87 The Helmholtz energy for the reference (LJ) system is (for a pure fluid)
⎛ ⎞
Aseg = m ⎜ AHS + exp(− γρ*2 )ρTΔB2,hBH + ∑ CijT i / 2ρ* j ⎟ (3.44)
⎜ i, j
⎟
⎝ ⎠
where
0
ΔB2,hBH = ∑C T
i =−7
i
i /2
(3.46)
224 Handbook of Surface and Colloid Chemistry
TABLE 3.13
Type of Association for Cross-Associating Mixtures
Mixture Component 1 Component 2 Association Type
Acid–acid O H O ∈A1A1 ≠ 0
C Site A1 ∈A2A2 ≠ 0
Site A2 C
O H ∈A1A2 ≠ 0
O
Acid–alcohol
O Site B2 ∈A1B1 ≠ 0,∈A2B2 ≠ 0
C Site A1 C Ö ∈A1A2 =∈A1B2 ≠ 0
O H ∈A2A2 =∈B2B2 = 0
H
Site A2
Water–acid
Site C1 H O ∈A1C1 =∈B1C1 ≠ 0,∈A2A2 ≠ 0
H Ö Site A2 C ∈A1A2 =∈B1A2 =∈C1A2 ≠ 0
Site B1 O ∈A1A1 =∈B1B1 =∈C1C1 =∈A1B1 = 0
H
Site A1
Source: Fu, Y.-H. and Sandler, S.I., Ind. Eng. Chem. Res., 34, 1897, 1995. With permission.
mb
ρ* = (3.47)
Vm
π * 3
η= ρ σ BH (3.48)
6
1
σ BH = ∑DT i
i /2
+ Dln ln T (3.49)
i =−2
Apart from using a different reference system, the notation in Kraska and Gubbins does not follow
the customary SAFT notation, nor are the pure-component parameters defined in the same way
(e.g., the energy parameter with units of 1/K is defined inversely as a temperature parameter with
units K) and care should be taken in using it. They also incorporate a term to account for dipole–
dipole interactions.
Thermodynamics of Polymer Solutions 225
3.4.7 SAFT-VR88
SAFT-VR is the version of SAFT developed by George Jackson and coworkers first at the University
of Sheffield and currently at Imperial College (Gil-Vilegas et al.88; McCabe et al.89). SAFT-VR is
identical to the Huang and Radosz version except in the dispersion contribution. This term incor-
porated attraction in the form of a square-well potential. Thus, in addition to a segment being
characterized by a size and an energy parameter, the square-well width (λ) is also included as a
pure-component parameter. Thus, changing the parameter λ changes the range of attraction of the
segment (hence the name VR for “variable range”). It is the introduction of this extra term that
gives SAFT-VR greater flexibility, since we now have an extra pure component to “play with.”
Although it is generally desirable to describe pure-component liquid densities and vapor pressure
with the minimum number of parameters, the extra variable-range parameter may be necessary for
the description of certain anomalous behaviors in systems containing water. The Helmholtz energy
for the dispersion energy is given by
a1 a
a disp = + 2 (3.50)
kT (kT )2
where
a1 = −ρs ∑∑ x x αVDW
s ,i s , j
ij g HS ⎡⎣σ x ; ζ eff
x ⎤ ⎦ (3.51)
i j
VDW
2πεij σ3ij λ 3ij − 1
( ) (3.52)
α ij =
3
and gHS is the radial distribution function for hard spheres as before except that the arguments are
different:
2 3
x = c1ζ x + c2 ζ x + c3ζ x (3.53)
ζ eff
σ3x = ∑∑ x
i j
x σij3 (3.54)
s ,i s , j
π
ζx = ρs σ3x (3.55)
6
n n
1 ∂a
a2 = ∑∑ x
i =1 j =1
x
s ,i s , j
2
K HS εijρs 1 (3.57)
∂ρs
where
ζ 0 (1 − ζ 3 )4
K HS = (3.58)
ζ 0 (1 − ζ 3 ) + 6ζ1ζ 2 (1 − ζ 3 ) + 9ζ 32
2
3.4.8 PC-SAFT84
A recent version of SAFT that has appeared is that due to Gross and Sadowski84 developed at the
Technical University of Berlin. Once again, most of the terms in PC-SAFT are the same as those
in the Huang and Radosz version. The term that is different is the dispersion term. However, it is
not simply a different way of expressing the dispersion attraction between segments, but rather
it tries to account for dispersion attraction between whole chains. Referring to Figure 3.5 should
make this clear. Instead of adding the dispersion to hard spheres and then forming chains, we first
form hard-sphere chains and then add a chain dispersion term, so the route in Figure 3.5 would be
(b)–(d)–(c)–(e). To do this, we require interchain rather than intersegment radial distribution func-
tions. These are given by O’Lenick et al.90 The Helmholtz energy for the dispersion term is given as
the sum of a first- and second-order term:
Adisp A A
= 1 + 2 (3.59)
kTN kTN kTN
where
∞
A1 ⎛ ε ⎞ 3 hc ⎛ xσ ⎞ 2
= −2πρm 2 ⎜ ⎟ σ u( x )g ⎜ m; d ⎟ x dx (3.60)
kTN ⎝ kT ⎠ 1 ∫ ⎝ ⎠
−1 2
A2 ⎛ ∂Z hc ⎞ ⎡ ∞ xσ ⎞ 2 ⎥⎤
2⎛ ε ⎞ 3 ∂ 2 hc ⎛
= −πρm ⎜ 1 + Z + ρ
hc
⎟ m ⎜ u ( x ) g m; x dx (3.61)
kTN ⎝ ∂ρ ⎠ ⎟
⎝ kT ⎠
σ ⎢
∂ρ ⎢
ρ
⎣ 1
∫
⎜
⎝ d ⎟⎠ ⎥
⎦
where
x = r/σ
u( x ) = u( x )/ε is the reduced intermolecular potential
The radial distribution function ghc is now an interchain function rather than a segment function as
before. This is a key point in PC-SAFT. The term involving compressibilities is given by
We still need to solve the integrals in Equations 3.60 and 3.61. Setting
∞
⎛ xσ ⎞ 2
I1 = u( x )g hc ⎜ m; x dx (3.63)
∫
1
⎝ d ⎟⎠
∞
∂ ⎡⎢ ⎛ xσ ⎞ 2 ⎥⎤
I2 = ρ u( x )2 g hc ⎜ m; x dx (3.64)
∂ρ ⎢
⎣ 1
∫ ⎝ d ⎟⎠ ⎥
⎦
we can substitute the Lennard–Jones potential and the radial distribution function of O’Lenick
et al.90 This was done for the series of n-alkanes and the integrals were fit as a power series:
6
I1 = ∑ a η (3.65)
i =0
i
i
6
I2 = ∑ b η (3.66)
i =0
i
i
with
m −1 m −1 m − 2
ai = a0i + a1i + a2i (3.67)
m m m
m −1 m −1 m − 2
bi = b0i + b1i + b2i (3.68)
m m m
Equations 3.67 and 3.68 require a total of 42 constants, which are adjusted to fit experimen-
tal pure-component data of n-alkanes. This direct fitting to experimental data to some extent
accounts for errors in the reference equation of state, the perturbing potential, and the radial
distribution function, which appear in the integrals of Equations 3.63 and 3.64. The dispersion
potential given by Equations 3.60 and 3.61 is readily extended to mixtures using the van der
Waals one-fluid theory.
Since this first PC-SAFT paper appeared, the authors rapidly published a series of further papers
applying PC-SAFT to polymers,91,92 associating fluids93 and copolymers.94
1 − η/ 2
g HS (d + ) = (3.69)
(1 − η)3
and
4η − 3η2
aHS = (3.70)
(1 − η)2
respectively. In the initial paper,95 use of Equation 3.69 only is called modification 1, while use
of both Equations 3.69 and 3.70 is called modification 2. This and subsequent work has shown
that there is no loss of accuracy using the most simplified version of PC-SAFT, so modification
2 is used throughout and is called “simplified PC-SAFT.” Simplified PC-SAFT has since been
applied in our group to several polymer systems including: polymer VLE,96 polymer–solvent
binary LLE,97 ternary polymer–solvent and blend systems,98 and high-pressure gas solubility in
polymers.99,100
One of the interesting points about SAFT in general is the ability to extrapolate the properties of
higher molecular weight substances from knowledge of similar shorter chain compounds. This is
most evident for the n-alkane series. Pure-component parameters for polyethylene can in principle
be predicted by extrapolating from the properties of the n-alkanes. In practice, this is problematic,
since for very long chains, effects such as entanglement of the individual polymer chains start to
influence the behavior. Figure 3.6 shows PC-SAFT parameters for the alkanes up to eicosane (C20).
The parameters m, mσ3, and mε/k are all linear with molecular weight. Tables 3.14 and 3.15 show
computing times for simplified PC-SAFT compared with original PC-SAFT, as well as with other
equations of state.
2000 9
1800 mε/k
8
mσ3
1600 m 7
1400
6
1200
5
1000
4
800
3
600
2
400
200 1
0 0
0 100 200 300
Molecular weight
FIGURE 3.6 The groups m, mσ3, and mε/k vs. molecular weight for linear alkanes up to eicosane. Points are
PC-SAFT parameters, lines are linear fits to these points, excluding methane. (From von Solms, N. et al., Ind.
Eng. Chem. Res., 42, 1098, 2003. With permission.)
Thermodynamics of Polymer Solutions 229
TABLE 3.14
Comparison of Computing Times for 36-Component Phase-Envelope Calculation with
PC-SAFT and with Simplified PC-SAFT
Calculation Computing Time (ms)
Phase-envelope calculation, 36-component mixture, full PC-SAFT 48
Phase-envelope calculation, 36-component mixture, PC-SAFT modification 2 32
Source: von Solms, N. et al., Ind. Eng. Chem. Res., 42, 1098, 2003. With permission.
Computations were performed on a 2.0 GHz Pentium IV machine with DVF compiler.
TABLE 3.15
Comparison of Computing Times for Various Models and Mixtures
Computing Time (μs)
All Derivatives (T, P, Composition
Mixture Model Fugacity Coefficients Only Residual Heat Capacity)
SRK, 6 components 1.9 3.4
SRK, 6 components 3.1 5.6
CPA, 15 components, 0 sites 15 20.1
CPA, 15 components, 2 sites 18.5 24.8
CPA, 15 components, 4 sites 24.4 32.2
CPA, 15 components, 6 sites 32.3 40.9
CPA, 15 components, 8 sites 40.7 50.9
Modification 2, 15 components, 0 sites 15.6 23.1
Modification 2, 15 components, 6 sites 39.1 54.4
Source: von Solms, N. et al., Ind. Eng. Chem. Res., 42, 1098, 2003. With permission.
Number of sites refers to number of association sites on a molecule when employing an equation of state with
association.
3.4.11 Applications of SAFT
The remainder of this discussion looks at some polymer applications of SAFT. Figure 3.7 shows
the results of using parameters extrapolated based on Figure 3.6. The system is methane–
tetratetracontane (C44), where the C44 parameters are obtained by extrapolation of the lines in
Figure 3.6. While this is not a polymer system, the method of obtaining parameters by extrapola-
tion is applicable to polymer systems. A more sophisticated method for finding polymer parameters
based on extrapolation of the monomer properties and polymer density data has been published
recently.108
Figures 3.8 and 3.9 show a comparison of the various modifications and original PC-SAFT
for VLE in the systems polystyrene–propyl acetate (Figure 3.8) and polypropylene–diisopropyl
ketone (Figure 3.9). In general, PC-SAFT and simplified PC-SAFT performed similarly, as can be
seen from Table 3.16, which gives the errors in prediction for a large number of polymer–solvent
VLE systems. Figure 3.10 is a pressure–weight fraction diagram (VLE) for the polymer poly(vinyl
acetate) in the associating solvent 2-methyl-1-propanol. A small value of the binary interaction
parameter correlates the data well.
Figure 3.11 is an illustration of a novel method developed by von Solms et al.97 for finding LLE in
polymer systems, known as the method of alternating tangents. This figure shows the Gibbs energy
of mixing for two binary systems as a function of the mole fraction of component 1. The method
will be illustrated with reference to the system methanol(1)–cyclohexane(2), since this curve clearly
shows the existence of two phases. The composition of methanol in each phase is found by locating
a single line, which is a tangent to the curve in two places (the common tangent). In Figure 3.11,
these compositions are given by x1eq1 and x1eq 2. In fact the curve for the system PS(1)–acetone(2) also
shows the existence of two phases, although this is not visible. The first step in the procedure is to
determine whether a spinodal point exists (this is a necessary condition for phase separation). In
the figure, the two spinodal points are given by the compositions x1sp1 and x1sp 2. The spinodal condi-
tion is given by ((∂2gmix/RT)/∂x2) = 0, that is, an inflection point on the curve. Once a spinodal point
has been found (using a Newton–Raphson method), the next step is to find the point of tangent of a
5
kij = 0.04
P (MPa)
2 kij = 0.0
1
Darwish et al.
0
0 0.1 0.2 0.3 0.4
X1
FIGURE 3.7 Vapor pressure curve for the system methane(1)–tetratetracontane(2) at T = 423.2 K. The
dashed line is PC-SAFT, the solid line is PC-SAFT with a binary interaction parameter kij = 0.04. The points
are experimental data. (From von Solms, N. et al., Ind. Eng. Chem. Res., 42, 1098, 2003. With permission.)
Thermodynamics of Polymer Solutions 231
40
35
30
25
Pressure (kPa)
20
15
Modification 2
10
Modification 1
5 PC-SAFT
Bawn and Wajid
0
0 0.2 0.4 0.6 0.8 1
Propyl acetate weight fraction
FIGURE 3.8 Vapor pressure curve for the system propyl acetate(1)–polystyrene(2) at T = 343.15 K. The
dashed line is PC-SAFT (the lowest line on the plot), the solid line is modification 1, and the dotted line is
modification 2. Points are experimental data. All lines are pure predictions (kij = 0). (From von Solms, N. et al.,
Ind. Eng. Chem. Res., 42, 1098, 2003. With permission.)
4
Pressure (kPa)
2
Brown et al. (1964)
1 Simplified PC-SAFT
PC-SAFT
0
0 0.2 0.4 0.6 0.8 1
Solvent weight fraction
TABLE 3.16
Comparison of the Performance of the Simplified against the Original PC-SAFT in
Predicting Vapor–Liquid Equilibria of Polymer Solutions (kij = 0 in All Cases)
PC-SAFT
% AAD Simplified Version Original Version
Cyclic hydrocarbons
PS–cyclohexane 13 16
PS–benzene 28 13
PS–ethyl benzene 3 6
PS–m-xylene 25 16
PS–toluene 18 7
PVAc–benzene 6 10
Chlorinated hydrocarbons
PS–carbon tetrachloride 18 12
PS–chloroform 30 11
PP–dichloromethane 59 74
PP–carbon tetrachloride 55 47
Esters
PS–propyl acetate 5 21
PS–butyl acetate 3 25
PVAc–methyl acetate 3 2
PVAc–propyl acetate 19 18
Ketones
PS–acetone 6 26
PS–diethyl ketone 7 28
PS–methyl ethyl ketone 14 12
PVAc–acetone 4 7
PP–diethyl ketone 16 27
PP–diisopropyl ketone 4 11
PVAc–methyl ethyl ketone 7 6
Amines
PVAc–propylamine 4 3
PVAc–isopropyl amine 17 16
Alcohols
PVAc–1-propanol 56 54
PVAc–2-propanol 84 73
PVAc–1-butanol 59 59
PVAc–2-butanol 39 36
PVAc–2-methyl-1-propanol 29 29
Overall 23 24
Source: Kouskoumvekaki, I.A. et al., Fluid Phase Equilib., 215, 71, 2004. With permission.
Average percentage deviation between experimental and predicted equilibrium pressure curves.
line originating at the spinodal point. In the figure, this is the line connecting x1sp1 and x1. This point
is just to the left of x1eq 2 (i.e., we are not yet quite at the equilibrium concentration after one step).
Once the first tangent has been found, the point of tangent opposite is then found in a similar way.
This process is repeated until the change in the composition at the tangent point is within a specified
tolerance. At this point, the equilibrium values have been calculated.
Thermodynamics of Polymer Solutions 233
12
10
Pressure (kPa)
6
0
0 0.1 0.2 0.3 0.4 0.5
Solvent weight fraction
0 0
Methanol–cyclohexane
–0.02 PS–acetone
–0.5
–0.04
sp1
1
–0.06
–1
gmix/RT
sp2
–0.08 1
eq1
1 –1.5
–0.1
eq2
1
–0.12
–2
–0.14
1
–0.16 –2.5
0 0.2 0.4 0.6 0.8 1
Mole fraction (x1)
FIGURE 3.11 Illustration of the method of alternating tangents. The solid line is the system methanol(1)–
cyclohexane(2). The dotted line is the system PS(1)–acetone. The two spinodal points are indicated by x1sp1
and x1sp 2. The equilibrium (binodal) points are indicated by x1eq1 and x1eq 2. Starting from a spinodal point, the
equilibrium values can be calculated by solving for only one point at a time. (From Fluid Phase Equilib.,
222–223, von Solms, N., Kouskoumvekaki, I.A., Lindvig, T., Michelsen, M.L., and Kontogeorgis, G.M.,
A novel approach to liquid-liquid equilibrium in polymer systems with application to simplified PC-SAFT,
87–93, Copyright 2004, with permission from Elsevier.)
234 Handbook of Surface and Colloid Chemistry
Figures 3.12 through 3.15 are examples of binary LLE for polymer systems. Figure 3.12 shows
results for the system polystyrene–methylcyclohexane for different molecular weights of polysty-
rene. The experimental data are from the classic work of Dobashi et al.109 The lines are simpli-
fied PC-SAFT correlations with kij = 0.0065 for polystyrene molecular weights 10,200, 46,400
and 719,000 in order of increasing critical solution temperature. The data are reasonably well cor-
related over a very large range of molecular weight with a single value of the binary interaction
parameter, kij. The binary interaction parameter was adjusted to give the correct upper critical
solution temperature. However, the correct critical solution concentration is not obtained, although
the experimental trends are correctly predicted by the model: The critical solution temperature
increases, and the polymer weight fraction at the critical solution temperature decreases with
increasing molecular weight.
Figure 3.13 shows results for the system polyisobutylene–diisobutyl ketone at different polymer
molecular weights. The experimental data are from Shultz and Flory.110 The lines are simplified
PC-SAFT correlations. A single binary interaction parameter (kij = 0.0053) was used for all three
systems, although it seems that there is a weak dependence of molecular weight on kij. Incorporating
a functional dependence of kij on molecular weight (e.g., a linear fit) would improve the correlation.
It should also be noted that these three systems represent a very large range of molecular weights.
Figure 3.14 shows the results for a single molecular weight of HDPE in five different n-alkanol
solvents from n-pentanol up to n-nonanol. The results are well correlated using simplified PC-SAFT
with a small value of the binary interaction parameter kij. A kij value of around 0.003 gives a good cor-
relation for all the systems, except HDPE–n-pentanol. In the figure, a small value (kij = 0.0006) was
used to correlate the data, although the data is also well predicted by simplified PC-SAFT (kij = 0),
giving an error in the upper critical solution temperature of 3 K in the case of HDPE–n-pentanol.
Figure 3.15 shows the results for the system HDPE–butyl acetate. This system displays both
UCST and LCST behaviors. A single binary interaction parameter (kij = 0.0156) was used to cor-
relate the data for both molecular weights shown. The binary interaction parameter was adjusted to
340
330
320
310
Temperature (K)
300
719,000
290 181,000
109,000
280 46,400
34,900
270 20,200
17,300
Dobashi et al., (1980) 17,200
260
Dobashi et al., (1984) (MW 17,200) 16,100
10,200
250
0 0.1 0.2 0.3 0.4 0.5 0.6
Weight fraction polymer
FIGURE 3.12 Liquid–liquid equilibrium in the system polystyrene–methyl cyclohexane for different
molecular weights of polystyrene. The experimental data are from Dobashi et al.109 The lines are simpli-
fied PC-SAFT correlations with kij = 0.0065 for polystyrene molecular weights 10,200, 46,400 and 719,000
in order of increasing critical solution temperature. (From Fluid Phase Equilib., 222–223, von Solms, N.,
Kouskoumvekaki, I.A., Lindvig, T., Michelsen, M.L., and Kontogeorgis, G.M., A novel approach to liquid-
liquid equilibrium in polymer systems with application to simplified PC-SAFT, 87–93, Copyright 2004, with
permission from Elsevier.)
Thermodynamics of Polymer Solutions 235
340
330
320
310
290
280
270
6,000,000
260 285,000
22,700
250
0 0.1 0.2 0.3 0.4
Weight fraction polymer
FIGURE 3.13 Liquid–liquid equilibrium in the system polyisobutylene–diisobutyl ketone. PC-SAFT param-
eters for diisobutyl ketone were obtained by fitting to experimental liquid density and vapor pressure data in
the temperature range 260–600 K. This data was taken from the DIPPR database. The parameters were:
m = 4.6179, ε/k = 243.72 K, and σ = 3.7032 Å. Average percent deviations were 1.03% for vapor pressure and
0.64% for liquid density. Lines are simplified PC-SAFT correlations with kij = 0.0053, the same at all three
molecular weights. (From Fluid Phase Equilib., 222–223, von Solms, N., Kouskoumvekaki, I.A., Lindvig, T.,
Michelsen, M.L., and Kontogeorgis, G.M., A novel approach to liquid–liquid equilibrium in polymer systems
with application to simplified PC-SAFT, 87–93, Copyright 2004, with permission from Elsevier.)
450
440
430
Temperature (K)
420
410
400
Pentanol
Hexanol
390 Heptanol
Octanol
Nonanol
380
0 0.1 0.2 0.3 0.4
Weight fraction polymer
FIGURE 3.14 Liquid–liquid equilibrium for HDPE with n-alkanols. Lines are simplified PC-SAFT corre-
lations for each of the five solvents (pentanol highest, nonanol lowest). Polymer molecular weight is 20,000.
Binary interaction parameters are as follows: pentanol: 0.0006; hexanol: 0.003; heptanol: 0.0025; octanol:
0.0033; and nonanol: 0.0029. (From Fluid Phase Equilib., 222–223, von Solms, N., Kouskoumvekaki, I.A.,
Lindvig, T., Michelsen, M.L., and Kontogeorgis, G.M., A novel approach to liquid–liquid equilibrium in
polymer systems with application to simplified PC-SAFT, 87–93, Copyright 2004, with permission from
Elsevier.)
236 Handbook of Surface and Colloid Chemistry
540
520
500
Temperature (K)
480
460
440
420
400 13,600
64,000
380
0 0.05 0.1 0.15
Weight fraction HDPE
FIGURE 3.15 Liquid–liquid equilibrium in the system HDPE–butyl acetate. The system displays both
upper and lower critical solution temperature behaviors. The experimental data for molecular weights 13,600
and 64,000. Lines are simplified PC-SAFT correlations with kij = 0.0156 for both molecular weights. (From
Fluid Phase Equilib., 222–223, von Solms, N., Kouskoumvekaki, I.A., Lindvig, T., Michelsen, M.L., and
Kontogeorgis, G.M., A novel approach to liquid–liquid equilibrium in polymer systems with application to
simplified PC-SAFT, 87–93, Copyright 2004, with permission from Elsevier.)
give a good correlation for the UCST curve at the higher molecular weight (64,000). As mentioned
earlier, the LCST curve is rather insensitive to kij. Nevertheless, the LCST curve is reasonably well
correlated using this value. The prediction (kij = 0) is almost as good for the LCST curve, although
the UCST will then be substantially underpredicted.
Figure 3.16 shows a ternary phase diagram for the system polystyrene–acetone–methylcyclohexane.
The binary interaction parameters were obtained by fitting to the individual binary systems.
Polystyrene(1)
100
90
80
70
60
k12 = –0.005
50
k13 = 0.006
40
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
Methylcyclohexane(3) Acetone(2)
FIGURE 3.16 Ternary phase diagram for the system polystyrene–acetone–methylcyclohexane. The binary
interaction parameters were obtained by fitting to the individual binary systems. The ternary coexistence
curves are predictions.
Thermodynamics of Polymer Solutions 237
1500
1000
P (bar) 68% MA
58% MA
0% MA
500 25% MA
0
75 95 115 135 155 175
T (C)
The ternary coexistence curves are predictions. The algorithm for finding ternary LLE in systems
containing polymers is an extension of the binary algorithm discussed earlier and was developed
by Lindvig et al.98
Figure 3.17 from Gross et al.94 shows high pressure equilibrium for mixtures of poly(ethylene-
co-methyl acrylate) (EMA) and propylene for different repeat unit compositions of EMA. As repeat
units of methyl acrylate (MA) are added to the polyethylene chain, the demixing pressure at first
declines, but then increases as the composition of MA increases. This effect is correctly predicted
by PC-SAFT.
Finally, Figure 3.18 shows a comparison of SAFT-VR and simplified PC-SAFT in a recent study100
where the two models were compared in their ability to model multicomponent phase equilibrium
in systems typical of real polyethylene reactors. The system shown here is polyethylene/nitrogen/1-
butene. The results of the simplified PC-SAFT and SAFT-VR calculations for this ternary are
consistent. From the figure, one can see that as butene in the vapor is replaced by nitrogen, the
calculated absorption of butene decreases (not surprisingly—there is less of it to absorb). However,
it is also clear that as nitrogen in the vapor is replaced by butene, absorption of nitrogen increases,
even though there is less nitrogen to absorb. This suggests that there may be some enhancement/
inhibition of absorption effect.
In the time since the previous edition of this work (2008), around 150 articles have appeared
applying SAFT to various polymer systems. Certain newer themes become apparent here:
The use of density functional theory combined with equations of state to examine surface and
structural properties is a relatively new phenomenon, resulting in equations such as iSAFT.111,112 For
example, Jain et al.111 used SAFT and density functional theory to probe the structure of tethered
polymer chains.
Perhaps the most frequently occurring, relatively new idea is that of group-contribution SAFT,
where there has been a great deal of activity.107,113–122 Part of the motivation for a group-contribution
method is the need for an equation that can be readily extended to complex molecules, where pure-
component properties may be unavailable, and for which SAFT types of equation are well suited.
Examples of complex systems where SAFT has been applied are biopolymers,123 hyperbranched
polymers,124 and asphaltenes.125
238 Handbook of Surface and Colloid Chemistry
150 150
100% nC4= 75% nC4=
nC4 =
120 VR 25% N2
PC 120
Solubility (g/100 g PE)
30 30 N2
0 0
0 1 2 3 4 0 1 2 3 4
(a) p (MPa) (b) p (MPa)
150 6
50% nC4= 50% nC4 =
50% N2 50% N2
120
Solubility (g/100 g PE)
90
nC4 = nC4 = N2
3
60
N2 1.5
30
0 0
0 1 2 3 4 0 1 2 3 4
(c) p (MPa) (d) p (MPa)
6 6
nC4 = 20% nC4 = 100% N2
80% N2 VR
Solubility (g/100 g PE)
4.5
Solubility (g/100 g PE)
4.5 PC
3 3
N2 N2
1.5 1.5
0 0
0 1 2 3 4 0 1 2 3 4
(e) p (MPa) (f) p (MPa)
FIGURE 3.18 Gas absorptions in amorphous PE calculated with SAFT-VR and simplified-PC-SAFT for a
range of vapor compositions of the ternary mixture of (but-1-ene + nitrogen + the reference PE [MW = 12,000
g mol−1]) at T = 80°C: (a) 100 vapor mol% butene (binary mixture); (b) 75% but-1-ene, 25% nitrogen; (c) 50%
but-1-ene, 50% nitrogen (with a vertical scale chosen to highlight butene absorption); (d) 50% but-1-ene, 50%
nitrogen (with a vertical scale chosen to highlight nitrogen absorption); (e) 20% butene, 80% nitrogen; and
(f) 100% nitrogen (binary mixture). In each case, solid curves represent SAFT-VR calculations and dashed
curves represent simplified-PC-SAFT calculations. (From Haslam, A.J. et al., Fluid Phase Equilib., 243, 74,
2006. With permission).
Thermodynamics of Polymer Solutions 239
Another encouraging feature is the use of SAFT-type models in industrial applications. In addi-
tion to the asphaltene example noted earlier,125 work has appeared relevant to refrigeration126,127 and
catalytic polymerization processes.128
Finally, it can be mentioned that since the review article by Müller and Gubbins,77 several mono-
graphs reviewing various aspects of SAFT have appeared.129–131
1. Many databases (some available in computer form) and reliable group-contribution meth-
ods are available for estimating many pure polymer properties and phase equilibria of
polymer solutions such as densities, solubility parameters, glass and melting temperatures,
and solvent activity coefficients.
2. Simple group-contribution methods based on UNIFAC, containing corrections for the FV
effects, satisfactorily predict the solvent activities and vapor–liquid equilibria for binary
and ternary polymer solutions. They are less successful for the prediction of liquid–liquid
equilibria if the parameters are based on VLE. They are much more successful if the
parameters are based on LLE data. The combination of a simple FV expression such as that
employed in the Entropic-FV model and a local-composition energetic term such as that of
UNIQUAC seems to be a very promising tool for both VLE and LLE in polymer solutions.
We expect that such tools may find widespread use in the future for practical applications.
3. The SAFT model will continue to be a very successful tool for polymer systems. The
low-pressure and especially the high-pressure results for systems including solvents and
nonpolar polymers (with emphasis to those of interest to the polyolefin industry) are very
satisfactory. However, the extension to polar systems represents so far a limitation of the
model. In many ways, PC-SAFT has fulfilled the early promise of SAFT—it is reasonably
simple to implement (compared with many versions of SAFT), it is relatively undemanding
computationally, and it has proved successful in predicting and correlating phase equilibria
in many systems containing polymers and/or associating compounds. There is still work
to be done, however. One area is in finding parameters for pure polymers. Since polymers
are nonvolatile, one cannot use vapor-pressure regression and the energy parameter is rela-
tively insensitive within the range of experimental P–V–T data. Regressing pure-component
parameters from binary systems is problematic as this leads to nonunique pure-component
parameters—a situation best avoided. Another area where research is needed is in the
modeling of water. There was been a great deal of effort in this area, but as yet no satisfac-
tory model, fitting within the existing SAFT framework, has been developed for modeling
water-containing systems. Examples of such deficiencies are predicting the density maxi-
mum for pure water, as well as modeling water–hydrocarbon mutual solubilities.
4. Most theoretical/modeling studies in polymer thermodynamics are limited to
a. Organic polymers
b. Binary systems often involving monodisperse polymers and single solvents
c. Rather “simple” polymers (polyolefins, polystyrene, PVC, etc.)
d. Vapor–liquid equilibria and activity coefficients
e. “Rules of thumb” estimates of miscibility (solubility parameters, theta parameters, etc.)
Some of the future challenges in the area of polymer thermodynamics will involve
3.A APPENDIX
3.A.1 AN EXPRESSION OF THE FH MODEL FOR
MULTICOMPONENT MIXTURES
The FH model was originally developed as a model for the entropy of mixing for mixtures contain-
ing molecules of different size, but it was soon modified also to account for energetic interactions.
The model can be formulated in terms of the excess Gibbs energy as follows (Lindvig et al.37):
G E = G E ,comb + G E ,res
N
G E ,comb ϕi
RT
= ∑ n ln x
i =1
i
i
N N
G E ,res
RT
= ∑∑ϕ ϕ a
i =1 j =1
i j ij
χij = 2aij vi
Using basic thermodynamics, the following expression for the activity coefficient is obtained:
ln γ i = ln γ icomb + ln γ res
i
ϕi ϕ
ln γ icomb = ln +1− i
xi xi
NC NC NC
ln γ res
i = 2 vi ∑ ϕ j aij − vi ∑∑ϕ ϕ a j k jk
j =1 j =1 k =1
Thermodynamics of Polymer Solutions 241
The aforementioned formulation of the FH model is slightly different from the conventionally used
formulation using the FH interaction parameter (χ12), although there is an interrelationship based on
the simple equation shown above.
For a binary mixture, the multicomponent equation reduces to the traditional FH residual term:
ln γ1res = χ12ϕ22
ABBREVIATIONS
BR butadiene rubber
CPA cubic plus association
CST critical solution temperature
EAC ethyl acetate
EoS equation of state
EFV entropic-free volume
EMA poly(ethylene-co-methyl acrylate)
FH Flory–Huggins (model/equation/interaction parameter)
FV free-volume
GC group contribution (method/principle)
GC Fl(ory) group contribution Flory equation of state
GCVOL group contribution volume (method for estimating the density)
HDPE high density polyethylene
LCST lower critical solution temperature
LJ-SAFT Lennard–Jones SAFT
LLE liquid–liquid equilibria
MA methyl acrylate
MCSL Mansoori–Carnahan–Starling–Leland
PBMA polybutyl methacrylate
PC-SAFT perturbed chain-SAFT
PDMS polydimethylsiloxane
PE polyethylene
PEMA polyethyl-methacrylate
PEO polyethylene oxide
PHSC perturbed hard-sphere chain
PIB polyisobutylene
PS polystyrene
PVAC polyvinyl acetate
PVC polyvinyl chloride
SAFT statistical associating fluid theory
SAFT-VR SAFT-variable range
SLE solid–liquid equilibria
SLLE solid–liquid–liquid equilibria
SRK Soave–Redlich–Kwong
TPT thermodynamic perturbation theory
UCST upper critical solution temperature
UNIFAC UNIQUAC functional activity coefficient (a method for estimating activity coefficients)
UFV UNIFAC-FV
vdW1f van der Waals one fluid (mixing rules)
VLE vapor–liquid equilibria
VOC volatile organic content
242 Handbook of Surface and Colloid Chemistry
REFERENCES
1. Prausnitz, J.M., Lichtenthaler, R.N., and Azevedo, E.G.D., 1999. Molecular Thermodynamics of Fluid
Phase Equilibria, 3rd edn. Prentice-Hall International, Upper Saddle River, NJ.
2. Holten-Andersen, J., 1987. Physical models and coatings technology. Report T12-87, Scandinavian Paint
and Printing Ink Research Institute, Hørsholm, Denmark. Also Presented at the International Conference
in Organic Coatings Science and Technology, Athens, Greece, pp. 13–30, 1986.
3. Holten-Andersen, J., 1986. Heat balance and solvent balance in the drying of coatings. JOCCA, 69(12): 324–331.
4. Holten-Andersen, J. and Hansen, C.M., 1983. Solvent and Water Evaporation from coatings, Prog. Org.
Coat., 11(3): 219.
5. Holten-Andersen, J. and Eng, K., 1988. Activity coefficients in polymer solutions. Prog. Org. Coat., 16: 77.
6. Doong, S.J. and Ho, W.S., 1991. Sorption of organic vapors in polyethylene. Ind. Eng. Chem. Res., 30:
1351–1361.
7. Kampouris, E.M., Diakoulaki, D.C., and Papaspyrides, C.D., 1986. Solvent recycling of rigid poly(vinyl
chloride) bottles. J. Vinyl Technol., 8(2): 79.
8. Pappa, G.D., Kontogeorgis, G.M., and Tassios, D.P., 1997. Prediction of ternary liquid-liquid equilibria
in polymer-solvent-solvent systems. Ind. Eng. Chem. Res., 36: 5461.
9. Coleman, M.M., Graf, J.F., and Painter, P.C., 1995. Specific Interactions and the Miscibility of Polymer
Blends. Technomic Publishing Co., Inc., Lancaster, PY.
10. Mio, C., Kiritsov, S., Thio, Y., Brafman, R., Prausnitz, J.M., Hawker, C., and Malmstrom, E.E., 1998.
Vapor-liquid equilibria for solutions of dendritic polymers. J. Chem. Eng. Data, 43: 541–550.
11. Lieu, J.G., Liu, M., Frechet, J.M.J., and Prausnitz, J.M., 1999. Vapor-liquid equilibria for dendritic-
polymer solutions. J. Chem. Eng. Data, 44: 613–620.
12. von Solms, N., Nielsen, J.K., Hassager, O., Rubin, A., Dandekar, A.Y. Andersen, S.I., and Stenby, E.H.,
2004. Direct measurement of gas solubilities in polymers with a high-pressure microbalance. J. Appl.
Polym. Sci., 91: 1476–1488.
13. Patterson, R., Yampolskii, Y.P., Fogg, P.G.T., Bokarev, A., Bondar, V., Illinich, O., and Shishatskii, S.,
1999. IUPAC-NIST solubility data series 70. The solubility of gases in glassy polymers. J. Phys. Chem.
Ref. Data, 28(5): 1255–1452.
14. Leitner, W., May 2000. Designed to dissolve. Nature, 405: 129–130.
15. Teja, A.S. and Eckert, C.A., 2000. Commentary on supercritical fluids: Research and applications. Ind.
Eng. Chem. Res., 39: 4442–4444.
16. Perrut, M., 2000. Supercritical fluid applications: Industrial developments and economic issues. Ind. Eng.
Chem. Res., 39: 4531–4535.
17. Hansen, C.M., 1967. The three-dimensional solubility parameter and solvent diffusion coefficient, PhD
Thesis. Technical University of Denmark, Copenhagen, MI.
18. Shashoua, Y., 2000. Permanence of plasticizers in polyvinylchloride objects in the museum environment.
Polym. Preprints, 41(2): 1796–1797.
19. Su, C.S., Patterson, D., and Schreiber, H.P., 1976. Thermodynamic interactions and the properties of the
PVC-plasticiser systems. J. Appl. Polym. Sci., 20: 1025–1034.
20. Demertzis, P.G., Riganakos, K.A., and Akrida-Demertzi, K., 1990. Study of compatibility of PVC and
polyester-type plasticizer blends by inverse gas chromatography. Eur. Polym. J., 26(2): 137–140.
21. Demertzis, P.G., Riganakos, K.A., and Akrida-Demertzi, K., 1991. An inverse gas chromatographic study
of the compatibility of food grade PVdC copolymer and low volatility plasticisers. Polym. Int., 25: 229–236.
22. Prausnitz, J.M., 1995. Some new frontiers in chemical engineering thermodynamics. Fluid Phase
Equilib., 104: 1–20.
23. Prausnitz, J.M., 1989. Biotechnology: A new frontier for molecular thermodynamics. Fluid Phase
Equilib., 53: 439–451.
24. Seymour, R.B., 1982. Plastics vs. Corrosives. SPE Monograph Series. Wiley, New York.
25. Barton, A.F.M., 1983. Handbook of Solubility Parameters and Other Cohesion Parameters. CRC Press,
Boca Barton, FL.
26. Barton, A.F.M., 1990. CRC Handbook of Polymer–Liquid Interaction Parameters and Solubility
Parameters. CRC Press, Boca Raton, FL.
27. Hansen, C.M., 2000. Hansen Solubility Parameters. A User’s Handbook. CRC Press, Boca Raton, FL.
28. Elbro, H.S., 1992. Phase equilibria of polymer solutions—With special emphasis on free volumes, Ph.D.
Thesis. Department of Chemical Engineering, Technical University of Denmark, Copenhagen, MI.
29. Klein, J. and Jeberien, H.E., 1980. Chainlength dependence of thermodynamic properties of polyethyl-
ene (glycol). Makromol. Chem., 181: 1237.
Thermodynamics of Polymer Solutions 243
30. Wen, H., Elbro, H.S., and Alessi, P., 1991. Polymer Solution Data Collection. Dechema Chemistry Data
Series, Frankfurt, Germany.
31. High, M.S. and Danner, R.P., 1992. Polymer Solution Handbook; DIPPR 881 Project. Design Institute for
Physical Property Data, PY.
32. Flory, P.J., 1941. Thermodynamics of high polymer solutions. J. Chem. Phys., 9: 660.
33. Huggins, M.L., 1941. Solutions of long chain compounds. J. Chem. Phys., 9: 440.
34. Qian, C., Mumby, S.J., and Eichinger, B.E., 1990. Application of the theory of phase diagrams to binary
polymer solutions and blends. Polym. Preprints, 31: 621.
35. Elbro, H.S., Fredenslund, Aa., and Rasmussen, P., 1990. A new simple equation for the prediction of
solvent activities in polymer solutions. Macromolecules, 23: 4707.
36. Lindvig, Th., Michelsen, M.L., and Kontogeorgis, G.M., 2002. A Flory–Huggins model based on the
Hansen solubility parameters. Fluid Phase Equilib., 203: 247.
37. Lindvig, Th., Economou, I.G., Danner, R.P., Michelsen, M.L, and Kontogeorgis, G.M., 2004. Modeling of
multicomponent vapour-liquid equilibria for polymer-solvent systems. Fluid Phase Equilib., 220: 11–20.
38. Schultz, A.R. and Flory, P.J., 1953. Phase equilibria in polymer-solvent systems. II. Thermodynamic
interaction parameters from critical miscibility data. J. Am. Chem. Soc., 75: 496.
39. Bondi, A., 1968. Physical Properties of Molecular Crystals, Liquids and Glasses. John Wiley & Sons,
New York.
40. Kontogeorgis, G.M., Fredenslund, Aa., and Tassios, D.P., 1993. Simple activity coefficient model for the
prediction of solvent activities in polymer solutions. Ind. Eng. Chem. Res., 32: 362.
41. Kouskoumvekaki, I., Michelsen, M.L., and Kontogeorgis, G.M., 2002. An improved entropic expression
for polymer solutions. Fluid Phase Equilib., 202(2): 325.
42. Oishi, T. and Prausnitz, M., 1978. Estimation of solvent activities in polymer solutions using a group-
contribution method. Ind. Eng. Chem. Proc. Des. Dev., 17(3): 333.
43. Hansen, H.K., Coto, B., and Kuhlmann, B., 1992. UNIFAC with lineary temperature-dependent group-
interaction parameters, IVC-SEP Internal Report 9212, Technical University of Denmark.
44. Elbro, H.S., Fredenslund, Aa., and Rasmussen, P., 1991. Group contribution method for the prediction of
liquid densities as a function of temperature for solvents, oligomers, and polymers. Ind. Eng. Chem. Res.,
30: 2576.
45. Tsibanogiannis, I.N., Kalospiros, N.S., and Tassios, D.P., 1994. Extension of the GCVOL method and
application to some complex compounds. Ind. Eng. Chem. Res., 33: 1641.
46. Bogdanic, G. and Fredenslund, Aa., 1994. Revision of the GC-Flory EoS for phase equilibria in mixtures
with polymers. 1. Prediction of VLE for polymer solutions. Ind. Eng. Chem. Res., 33: 1331.
47. Bogdanic, G. and Fredenslund, Aa., 1995. Prediction of VLE for mixtures with co-polymers. Ind. Eng.
Chem. Res., 34: 324.
48. Lindvig, Th., Hestkjær, L.L., Hansen, A.F., Michelsen, M.L., and Kontogeorgis, G.M., 2002. Phase equi-
libria for complex polymer solutions. Fluid Phase Equilib., 663: 194.
49. Harismiadis, V.I. and Tassios, D.P., 1996. Solid-liquid-liquid equilibria in polymer solutions. Ind. Eng.
Chem. Res., 35: 4667.
50. Kontogeorgis, G.M., Coutsikos, Ph., Tassios, D.P., and Fredenslund, Aa., 1994. Improved models for the
prediction of activity coefficients in nearly athermal mixtures. Part I. Empirical modifications of free-
volume models. Fluid Phase Equilib., 92: 35.
51. Coutinho, J.A.P., Andersen, S.I., and Stenby, E.H., 1995. Evaluation of activity coefficient models in
prediction of alkane SLE. Fluid Phase Equilib., 103: 23.
52. Polyzou, E.N., Vlamos, P.M., Dimakos, G.M., Yakoumis, I.V., and Kontogeorgis, G.M., 1999. Assessment
of activity coefficient models for predicting solid-liquid equilibria of asymmetric binary alkane systems.
Ind. Eng. Chem. Res., 38: 316–323.
53. Kontogeorgis, G.M., Fredenslund, Aa., Economou, I.G., and Tassios, D.P., 1994. Equations of state and
activity coefficient models for vapor-liquid equilibria of polymer solutions. AIChE J., 40: 1711.
54. Bogdanic, G. and Vidal, J., 2000. A segmental interaction model for liquid-liquid equilibrium calcula-
tions for polymer solutions. Fluid Phase Equilib., 173: 241–252.
55. Bogdanic, G., 2001. The FV-UNIQUAC segmental interaction model for liquid-liquid equilibrium calculations
for polymer solutions. Part 2. Extension to solutions containing polystyrene. Fluid Phase Equilib., 4791: 1–9.
56. Panagou, E., Vidal, J., and Bogdanic, G., 1998. A segmental interaction model for LLE correlation and
prediction—Application to the poly(vinyl alcohol)/water system. Polym. Bull., 40: 117.
57. Pappa, G.D., Voutsas, E.C., and Tassios, D.P., 2001. Liquid-liquid phase equilibrium in polymer-solvent
systems: Correlation and prediction of the polymer molecular weight and the pressure effect. Ind. Eng.
Chem. Res., 40(21): 4654.
244 Handbook of Surface and Colloid Chemistry
58. Kontogeorgis, G.M., Saraiva, A., Fredenslund, Aa., and Tassios, D.P., 1995. Prediction of liquid-liquid
equilibrium for binary polymer solutions with simple activity coefficient models. Ind. Eng. Chem. Res.,
34: 1823.
59. Harismiadis, V.I., van Bergen, A.R.D., Saraiva, A., Kontogeorgis, G.M., Fredenslund, Aa., and Tassios,
D.P., 1996. Miscibility of polymer blends with engineering models. AIChE J., 42: 3170.
60. Pappa, G.D., Voutsas, E.C., and Tassios, D.P., 1999. Prediction of solvent activities in polymer solutions
with simple group-contribution models. Ind. Eng. Chem. Res., 38: 4975.
61. Lindvig, Th., Michelsen, M.L., and Kontogeorgis, G.M., 2001. Thermodynamics of paint related systems
with engineering models. AIChE J., 47(11): 2573–2584.
62. Kontogeorgis, G.M., Nikolopoulos, G.I., Tassios, D.P., and Fredenslund, Aa., 1997. Improved models for
the prediction of activity coefficients in nearly athermal mixtures. Part II. A theoretically based GE-model
based on the van der Waals partition function. Fluid Phase Equilib., 127: 103.
63. Thorlaksen, P., Abildskov, J., and Kontogeorgis, G.M., 2003. Prediction of gas solubilities in elastomeric
polymers for the design of thermopane windows. Fluid Phase Equilib., 211: 17.
64. McQuarrie, D.A., 1976. Statistical Mechanics. Harper Collins, New York.
65. Wertheim, M.S., 1963. Exact solution of Percus-Yevick integral equation for hard spheres. Phys. Rev.
Lett., 10: 321.
66. Carnahan, N.F. and Starling, K.E., 1970. Thermodynamic properties of a rigid-sphere fluid. J. Chem.
Phys., 53: 600.
67. Economou, I.G., 2003. Molecular simulation for industrial applications. In: Kontogeorgis, G. and Gani,
R., eds., Computer Aided Property Estimation for Process and Product Design, Elsevier, Amsterdam,
the Netherlands. pp. 279–308.
68. Wertheim, M.S., 1984. Fluids with highly directional attractive forces. I. Statistical thermodynamics.
J. Stat. Phys., 35: 19.
69. Wertheim, M.S., 1984. Fluids with highly directional attractive forces. II. Thermodynamic perturbation
theory and integral equations. J. Stat. Phys., 35: 35.
70. Wertheim, M.S., 1986. Fluids with highly directional attractive forces. III. Multiple attraction site. J. Stat.
Phys., 42: 459.
71. Wertheim, M.S., 1986. Fluids with highly directional attractive forces. IV. Equilibrium polymerization.
J. Stat. Phys., 42: 477.
72. Chapman, W.G., Jackson, G., and Gubbins, K.E., 1988. Phase equilibria of associating fluids: Chain
molecules with multiple bonding sites. Mol. Phys., 65: 1057.
73. Chapman, W.G., Gubbins, K.E., Jackson, G., and Radosz, M., 1989. SAFT: Equation-of-state solution
model for associating liquids. Fluid Phase Equilib., 52: 31.
74. Chapman, W.G., Gubbins, K.E., Jackson, G., and Radosz, M., 1990. New reference equation of state for
associating liquids. Ind. Eng. Chem. Res., 29, 1709.
75. Jackson, G., Chapman, W.G., and Gubbins, K.E., 1988. Phase equilibria of associating fluids: Spherical
molecules with multiple bonding sites. Mol. Phys., 65: 1.
76. Donohue, M.D. and Prausnitz, J.M., 1978. Perturbed hard chain theory for fluid mixtures—Thermodynamic
properties for mixtures in natural-gas and petroleum technology. AIChE J., 24: 849–860.
77. Müller, E.A. and Gubbins, K.E., 2001. Molecular-based equations of state for associating fluids: A review
of SAFT and related approaches. Ind. Eng. Chem. Res., 40: 2193.
78. Huang, S.H. and Radosz, M., 1990. Equation of state for small, large, polydisperse and associating
molecules. Ind. Eng. Chem. Res., 29: 2284.
79. Huang, S.H. and Radosz, M., 1991. Equation of state for small, large, polydisperse and associating
molecules: Extension to fluid mixtures. Ind. Eng. Chem. Res., 30: 1994.
80. Fu, Y.-H. and Sandler, S.I., 1995. A simplified SAFT equation of state for associating compounds and
mixtures. Ind. Eng. Chem. Res., 34: 1897.
81. Kontogeorgis, G.M., Voutsas, E.C., Yakoumis, I.V., and Tassios, D.P., 1996. An equation of state for
associating fluids. Ind. Eng. Chem. Res., 35: 4310.
82. Michelsen, M.L. and Hendriks, E.H., 2001. Physical properties of association models. Fluid Phase
Equilib., 180: 165.
83. Chen, S.S. and Kreglewski, A., 1977. Applications of augmented van der Waals theory of fluids. 1. Pure
fluids. Ber. Bunsen Phys. Chem., 81: 1048.
84. Gross, J. and Sadowski, G., 2001. Perturbed-chain SAFT: An equation of state based on a perturbation
theory for chain molecules. Ind. Eng. Chem. Res., 40: 1244.
85. Kraska, T. and Gubbins, K.E., 1996. Phase equilibria calculations with a modified SAFT equation of
state. 1. Pure alkanes, alkanols and water. Ind. Eng. Chem. Res., 35: 4727.
Thermodynamics of Polymer Solutions 245
86. Kraska, T. and Gubbins, K.E., 1996. Phase equilibria calculations with a modified SAFT equation of
state. 2. Binary mixtures of n-alkanes, 1-alkanols and water. Ind. Eng. Chem. Res., 35: 4738.
87. Kolafa, J. and Nezbeda, I., 1994. The Lennard-Jones fluid: An accurate analytic and theoretically-based
equation of state. Fluid Phase Equilib., 100: 1.
88. Gil-Vilegas, A., Galindo, A., Whitehead, P.J., Mills, S.J., Jackson, G., and Burgess, A.N., 1997. Statistical
associating fluid theory for chain molecules with attractive potentials of variable range. J. Chem. Phys.,
106: 4168.
89. McCabe, C., Gil-Vilegas, A., and Jackson, G., 1999. Gibbs ensemble computer simulation and SAFT-VR
theory of non-conformal square-well monomer-dimer mixtures. Chem. Phys. Lett., 303: 27.
90. O’Lenick, R., Li, X.J., and Chiew, Y.C., 1995. Correlation functions of hard-sphere chain mixtures:
Integral equation theory and simulation results. Mol. Phys., 86: 1123.
91. Gross, J. and Sadowski, G., 2002. Modeling polymer systems using the perturbed-chain statistical asso-
ciating fluid theory equation of state. Ind. Eng. Chem. Res., 41: 1084.
92. Tumakaka, F., Gross, J., and Sadowski, G., 2002. Modeling of polymer phase equilibria using the
perturbed-chain SAFT. Fluid Phase Equilib., 194–197: 541.
93. Gross, J. and Sadowski, G., 2002. Application of the perturbed-chain SAFT equation of state to associat-
ing systems. Ind. Eng. Chem. Res., 41: 5510.
94. Gross, J., Spuhl, O., Tumakaka, F., and Sadowski, G., 2003. Modeling copolymer systems using the
perturbed-chain SAFT equation of state. Ind. Eng. Chem. Res., 42: 1266.
95. von Solms, N., Michelsen, M.L., and Kontogeorgis, G.M., 2003. Computational and physical perfor-
mance of a modified PC-SAFT equation of state for highly asymmetric and associating mixtures. Ind.
Eng. Chem. Res., 42: 1098.
96. Kouskoumvekaki, I.A., von Solms, N., Michelsen, M.L., and Kontogeorgis, G.M., 2004. Application of a sim-
plified perturbed chain SAFT equation of state to complex polymer systems. Fluid Phase Equilib., 215: 71–78.
97. von Solms, N., Kouskoumvekaki, I.A., Lindvig, T., Michelsen, M.L., and Kontogeorgis, G.M., 2004. A
novel approach to liquid-liquid equilibrium in polymer systems with application to simplified PC-SAFT.
Fluid Phase Equilib., 222–223: 87–93.
98. Lindvig, T., Michelsen, M.L., and Kontogeorgis, G.M., 2004. Liquid-liquid equilibria for binary and
ternary polymer solutions with PC-SAFT. Ind. Eng. Chem. Res., 43: 1125–1132.
99. von Solms, N., Michelsen, M.L., and Kontogeorgis, G.M., 2005. Prediction and correlation of high-
pressure gas solubility in polymers with simplified PC-SAFT. Ind. Eng. Chem. Res., 44: 3330.
100. Haslam, A.J, von Solms, N., Adjiman, C.S., Galindo, A., Jackson, G., Paricaud, P., Michelsen, M.L., and
Kontogeorgis, G.M., 2006. Predicting enhanced absorption of light gases in polyethylene using simpli-
fied PC-SAFT and SAFT-VR. Fluid Phase Equilib., 243: 74.
101. Jog, P.K., Sauer, S.G., Blaesing, J., and Chapman, W.G., 2001. Application of dipolar chain theory to the
phase behavior of polar fluids and mixtures. Ind. Eng. Chem. Res., 40: 4641.
102. Gross, J., 2005. An equation-of-state contribution for polar components: Quadrupolar molecules. AIChE
J., 51: 2556–2568.
103. Gross, J. and Vrabec, J., 2006. An equation-of-state contribution for polar components: Dipolar
molecules. AIChE J., 52: 1194–1204.
104. Kleiner, M. and Gross, J., 2006. An equation of state contribution for polar components: Polarizable
dipoles. AIChE J., 52: 1951–1961.
105. Behzadi, B., Patel, B.H., Galindo, A., and Ghotbi, C., 2005. Modeling electrolyte solutions with the
SAFT-VR equation using Yukawa potentials and the mean-spherical approximation. Fluid Phase
Equilib., 236: 241–255.
106. Cameretti, L.F., Sadowski, G., and Mollerup, J.M., 2005. Modeling of aqueous electrolyte solutions with
perturbed-chain statistical associated fluid theory. Ind. Eng. Chem. Res., 44, 3355–3362.
107. Le Thi, C., Tamouza, S., Passarello, J.P., Tobaly, P., de Hemptinne, J.-C., 2006. Modeling phase
equilibrium of H-2+n-alkane and CO2+n-alkane binary mixtures using a group contribution statisti-
cal association fluid theory equation of state (GC-SAFT-EOS) with a k(ij) group contribution method.
Ind. Eng. Chem. Res., 45: 6803–6810.
108. Kouskoumvekaki, I.A., von Solms, N., Lindvig, T., Michelsen, M.L., and Kontogeorgis, G.M., 2004.
Novel method for estimating pure-component parameters for polymers: Application to the PC-SAFT
equation of state. Ind. Eng. Chem. Res., 43: 2830–2838.
109. Dobashi, T., Nakata, M., and Kaneko, M., 1984. Coexistence curve of polystyrene in methylcyclohexane.
3. Asymptotic-behavior of ternary-system near the Plait point. J. Chem. Phys., 80: 948–953.
110. Shultz, A.R. and Flory, P.J., 1952. Phase equilibria in polymer-solvent systems. J. Am. Chem. Soc.,
74: 4760–4767.
246 Handbook of Surface and Colloid Chemistry
111. Jain, S., Jog, P., Weinhold, J., Srivastava, R., and Chapman, W.G., 2008. Modified interfacial statistical
associating fluid theory: Application to tethered polymer chains. J. Chem. Phys., 128: 154910.
112. Bymaster, A. and Chapman, W.G., 2010. An iSAFT density functional theory for associating polyatomic
molecules. J. Phys. Chem. B, 114: 12298.
113. Tamouza, S., Passarello, J.P., Tobaly, P., de Hemptinne, J.C., 2004. Group contribution method with SAFT
EOS applied to vapor liquid equilibria of various hydrocarbon series. Fluid Phase Equilib., 222–223: 67.
114. Grandjean, L., de Hemptinne, J.-C., and Lugo, R., 2014. Application of GC-PPC-SAFT EoS to ammonia
and its mixtures. Fluid Phase Equilib., 367: 159.
115. Tihic, A., von Solms, N., Michelsen, M.L., Kontogeorgis, G.M., and Constantinou, L., 2009. Application
of sPC-SAFT and group contribution sPC-SAFT to polymer systems—Capabilities and limitations.
Fluid Phase Equilib., 281: 70.
116. Tihic, A., von Solms, N., Michelsen, M.L., Kontogeorgis, G.M., and Constantinou, L., 2009. Analysis
and applications of a group contribution sPC-SAFT equation of state. Fluid Phase Equilib., 281: 60.
117. Tihic, A., Kontogeorgis, G.M., von Solms, N., and Michelsen, M.L., 2008. A predictive group-contribution
simplified PC-SAFT equation of state: Application to polymer systems. Ind. Eng. Chem. Res., 47: 5092.
118. Peng, Y., Goff, K.D., dos Ramos, M.C., McCabe, C., 2010. Predicting the phase behavior of polymer
systems with the GC-SAFT-VR approach. Ind. Eng. Chem. Res., 49: 1378.
119. Clark, G.N.I., Galindo, A., Jackson, G., Rogers, S., and Burgess, A.N., 2008. Modeling and under-
standing closed-loop liquid-liquid immiscibility in aqueous solutions of poly(ethylene glycol) using the
SAFT-VR approach with transferable parameters. Macromolecules, 41: 6582.
120. Sanchez, F.A., Pereda, S., and Brignole, E.A., 2011. GCA-EoS: A SAFT group contribution model-extension
to mixtures containing aromatic hydrocarbons and associating compounds. Fluid Phase Equilib., 306: 112.
121. Peters, F.T., Laube, F.S., Sadowski, G., 2012. Development of a group contribution method for polymers
within the PC-SAFT model. Fluid Phase Equilib., 324: 70.
122. Papaioannou, V., Lafitte, T., Avendano, C., Adjiman, C.S., Jackson, G., Muller, E.A., and Galindo, A.,
2014. Group contribution methodology based on the statistical associating fluid theory for heteronuclear
molecules formed from Mie segments. J. Chem. Phys., 140: 054107.
123. Cameretti, L.F. and Sadowski, G., 2008. Modeling of aqueous amino acid and polypeptide solutions with
PC-SAFT. Chem. Eng. Proc., 47: 1018.
124. Kozlowska, M.K., Jurgens, B.F., Schacht, C.S., Gross, J., and de Loos, T.W., 2009. Phase behavior of
hyperbranched polymer systems: Experiments and application of the perturbed-chain polar SAFT equa-
tion of state. J. Phys. Chem. B, 113: 1022.
125. Vargas, F.M., Gonzalez, D.L., Hirasaki, G.J., and Chapman, W.G., 2009. Modeling asphaltene phase
behavior in crude oil systems using the perturbed chain form of the statistical associating fluid theory
(PC-SAFT) equation of state. Energy Fuels, 23: 1140.
126. von Solms, N. and Kristensen, J., 2010. Refrigeration plants using carbon dioxide as refrigerant:
Measuring and modelling the solubility, diffusivity and permeability of carbon dioxide in polymers used
as packing and sealing materials. Int. J. Refrig., 33: 19.
127. Neela, V. and von Solms, N., 2014. Permeability, diffusivity and solubility of carbon dioxide in
fluoropolymers: An experimental and modeling study. J. Polym. Res., 21: 401.
128. Krallis, A. and Kanellopoulos, V., 2013. Application of Sanchez-Lacombe and perturbed-chain statisti-
cal associating fluid theory equation of state models in catalytic olefins (co)polymerization industrial
applications. Ind. Eng. Chem. Res., 52: 9060.
129. Tan, S.P., Adidharma, H., and Radosz, M., 2008. Recent advances and applications of statistical associat-
ing fluid theory. Ind. Eng. Chem. Res., 47: 8063.
130. Sadowski, G., 2011. Modeling of polymer phase equilibria using equations of state. In: Enders, S. and
Wolf, B.A., eds. Polymer Thermodynamics: Liquid Polymer-Containing Mixtures. Advances in Polymer
Science, Vol. 238, p. 389, Springer-Verlag, Berlin, Heidelberg.
131. Kleiner, M., Turnakaka, F., and Sadowski, G., 2009. Thermodynamic modeling of complex systems. In: Lu,
X. and Hu, Y., eds. Molecular Thermodynamics of Complex Systems. Structure and Bonding, Vol. 131, p. 75,
Springer-Verlag, Berlin, Heidelberg.
132. Tanbonliong, J.O. and Prausnitz, J.M., 1997. Vapour-liquid equilibria for some binary and ternary poly-
mer solutions. Polymer, 38: 5775.
133. Matsumara, K. and Katayama, T., 1974. Katayama, T. Mateumura, K. and Urahama, Y.1971. Measurement
of Vapor-Liquid Equilibria of Binary and, ternary Solutions containing polystyrene as a Component.
Kagaku Kogaku, 38: 388.
134. Tanbonliong, J.O. and Prausnitz, J.M., 1997. Vapor-Liquid Equilibria of Binary and Ternary Solutions
Containing Polyvinylacetate as a Component. Polymer, 38: 5775.
4 Chemical Physics of Colloid
Systems and Interfaces
Peter A. Kralchevsky and Krassimir D. Danov
CONTENTS
4.1 Introduction...........................................................................................................................248
4.2 Surface Tension of Surfactant Solutions................................................................................ 249
4.2.1 Static Surface Tension............................................................................................... 249
4.2.1.1 Nonionic Surfactants.................................................................................. 250
4.2.1.2 Ionic Surfactants......................................................................................... 257
4.2.2 Dynamic Surface Tension..........................................................................................266
4.2.2.1 Adsorption under Diffusion Control........................................................... 267
4.2.2.2 Small Initial Perturbation........................................................................... 268
4.2.2.3 Large Initial Perturbation........................................................................... 269
4.2.2.4 Generalization for Ionic Surfactants........................................................... 271
4.2.2.5 Adsorption under Barrier Control............................................................... 274
4.2.2.6 Dynamics of Adsorption from Micellar Surfactant Solutions.................... 276
4.3 Capillary Hydrostatics and Thermodynamics....................................................................... 283
4.3.1 Shapes of Fluid Interfaces......................................................................................... 283
4.3.1.1 Laplace and Young Equations.................................................................... 283
4.3.1.2 Solutions of Laplace Equations for Menisci of Different Geometry.......... 285
4.3.1.3 Gibbs–Thomson Equation.......................................................................... 289
4.3.1.4 Kinetics of Ostwald Ripening in Emulsions............................................... 291
4.3.2 Thin Liquid Films and Plateau Borders.................................................................... 293
4.3.2.1 Membrane and Detailed Models of a Thin Liquid Film............................ 293
4.3.2.2 Thermodynamics of Thin Liquid Films..................................................... 294
4.3.2.3 The Transition Zone between Thin Film and Plateau Border.................... 297
4.3.2.4 Methods for Measuring Thin Film Contact Angles................................... 301
4.3.3 Lateral Capillary Forces between Particles Attached to Interfaces..........................302
4.3.3.1 Particle–Particle Interactions......................................................................302
4.3.3.2 Particle–Wall Interactions...........................................................................307
4.3.3.3 Electrically Charged Particles at Liquid Interfaces....................................308
4.4 Surface Forces....................................................................................................................... 314
4.4.1 Derjaguin Approximation.......................................................................................... 314
4.4.2 van der Waals Surface Forces.................................................................................... 315
4.4.3 Electrostatic Surface Forces...................................................................................... 318
4.4.3.1 Two Identically Charged Planes................................................................. 318
4.4.3.2 Two Nonidentically Charged Planes........................................................... 321
4.4.3.3 Two Charged Spheres................................................................................. 322
4.4.4 DLVO Theory............................................................................................................ 323
4.4.5 Non-DLVO Surface Forces........................................................................................ 324
4.4.5.1 Ion Correlation Forces................................................................................ 324
4.4.5.2 Steric Interaction......................................................................................... 325
247
248 Handbook of Surface and Colloid Chemistry
4.1 INTRODUCTION
A colloidal system represents a multiphase (heterogeneous) system in which at least one of the
phases exists in the form of very small particles: typically smaller than 1 μm but still much larger
than the molecules. Such particles are related to phenomena like Brownian motion, diffusion, and
osmosis. The terms “microheterogeneous system” and “disperse system” (dispersion) are more
general because they also include bicontinuous systems (in which none of the phases is split into
separate particles) and systems containing larger, non-Brownian particles. The term dispersion is
often used as a synonym of colloidal system.
A classification of the colloids with respect to the state of aggregation of the disperse and
continuous phases is shown in Table 4.1. Some examples are the following.
1. Examples for gas-in-liquid dispersions are the foams or the boiling liquids. Gas-in-solid
dispersions are the various porous media like filtration membranes, sorbents, catalysts,
isolation materials, etc.
2. Examples for liquid-in-gas dispersions are the mist, the clouds, and other aerosols. Liquid-
in-liquid dispersions are the emulsions. At room temperature there are only four types of
mutually immiscible liquids: water, hydrocarbon oils, fluorocarbon oils, and liquid metals
(Hg and Ga). Many raw materials and products in food and petroleum industries exist in
the form of oil-in-water or water-in-oil emulsions. The soil and some biological tissues can
be considered as examples of liquid-in-solid dispersions.
Chemical Physics of Colloid Systems and Interfaces 249
TABLE 4.1
Types of Disperse Systems
Continuous Phase
Disperse Phase Gas Liquid Solid
Gas — G in L G in S
Liquid L in G L1 in L2 L in S
Solid S in G S in L S1 in S2
3. Smoke, dust, and some other aerosols are examples for solid-in-gas dispersions. The solid-
in-liquid dispersions are termed suspensions or sols. The pastes and some glues are highly
concentrated suspensions. The gels represent bicontinuous structures of solid and liquid.
The pastes and some glues are highly concentrated suspensions. Examples for solid-in-
solid dispersions are some metal alloys, many kinds of rocks, some colored glasses, etc.
In the following section, we will consider mostly liquid dispersions, that is, dispersions with liquid con-
tinuous phase, like foams, emulsions and suspensions. Sometimes these are called “complex fluids.”
In general, the area of the interface between the disperse and continuous phases is rather large.
For instance, 1 cm3 of dispersion phase with particles of radius 100 nm and volume fraction 30%
contains interface of area about 10 m2. This is the reason why the interfacial properties are of cru-
cial importance for the properties and stability of colloids.
The stabilizing factors for dispersions are the repulsive surface forces, the particle thermal
motion, the hydrodynamic resistance of the medium, and the high surface elasticity of fluid par-
ticles and films.
On the contrary, the factors destabilizing dispersions are the attractive surface forces, the factors
suppressing the repulsive surface forces, the low surface elasticity, gravity, and other external forces
tending to separate the phases.
In Sections 4.2 and 4.3 we consider effects related to the surface tension of surfactant solution
and capillarity. In Section 4.4 we present a review on the surface forces due to intermolecular
interactions. In Section 4.5 we describe the hydrodynamic interparticle forces originating from
the effects of bulk and surface viscosity and related to surfactant diffusion. Finally, Section 4.6 is
devoted to the kinetics of coagulation in dispersions.
dσ = − ∑ Γ dμ (4.1)
i
i i
where
Γi is the surface concentration (adsorption) of the ith component
μi is its chemical potential
250 Handbook of Surface and Colloid Chemistry
The summation in Equation 4.1 is carried out over all components. Usually an equimolecular divid-
ing surface with respect to the solvent is introduced for which the adsorption of the solvent is set zero
by definition [4,5]. Then the summation is carried out over all other components. Note that Γi is an
excess surface concentration with respect to the bulk; Γi is positive for surfactants, which decreases
σ in accordance with Equation 4.1. On the contrary, Γi is negative for aqueous solutions of electro-
lytes, whose ions are repelled from the surface by the electrostatic image forces [5]; c onsequently,
the addition of electrolytes increases the surface tension of water [6]. For surfactant concentrations
above the critical micellization concentration (CMC) μi = constant and, consequently, σ = constant
(see Equation 4.1).
dσ = −kTΓ1 d ln c1 (4.2)
The surfactant adsorption isotherms, expressing the connection between Γ1 and c1, are usually
obtained by means of some molecular model of adsorption. Table 4.2 contains the six most pop-
ular surfactant adsorption isotherms, those of Henry, Freundlich [7], Langmuir [8], Volmer [9],
Frumkin [10] and van der Waals [11]. For c1 → 0 all isotherms (except that of Freundlich) reduce
to the Henry isotherm: Γ1/Γ∞ = Kc1. The physical difference between the Langmuir and Volmer
isotherms is that the former corresponds to a physical model of localized adsorption, whereas the
latterto nonlocalized adsorption. The Frumkin and van der Walls isotherms generalize, respec-
tively, the Langmuir and Volmer isotherms for case, in which the interaction between neighboring
adsorbed molecules is not negligible. (If the interaction parameter β is set zero, the Frumkin and
van der Walls isotherms reduce to the Langmuir and Volmer isotherms, correspondingly.) The
comparison between theory and experiment shows that for air–water interfaces β > 0, whereas for
oil–water interfaces we can set β = 0 [12,13]. The latter facts lead to the conclusion that for air–water
interfaces β takes into account the van der Waals attraction between the hydrocarbon tails of the
adsorbed surfactant molecules across air; such attraction is missing when the hydrophobic phase
is oil. (Note that in the case of ionic surfactants it is possible to have β < 0, see the next s ection.)
The adsorption parameter K in Table 4.2 characterizes the surface activity of the surfactant: the
greater K, the higher the surface activity. K is related to the standard free energy of adsorption,
ΔG o = μ1( 0s ) − μ1( 0 ), which is the energy gain for bringing a molecule from the bulk of the aqueous
phase to a diluted adsorption layer [14,15]:
δ1 ⎛ μ( 0 ) − μ1( 0s ) ⎞
K= exp ⎜ 1 ⎟ (4.3)
Γ∞ ⎝ kT ⎠
The parameter δ1 characterizes the thickness of the adsorption layer; δ1 can be set (approximately)
equal to the length of the amphiphilic molecule. Γ∞ represents the maximum possible value of the
adsorption. In the case of localized adsorption (Langmuir and Frumkin isotherms), 1/Γ∞ is the area
Chemical Physics of Colloid Systems and Interfaces 251
TABLE 4.2
Types of Adsorption and Surface Tension Isotherms
Surfactant Adsorption Isotherms
Type of Isotherm (for Nonionic Surfactants: a1s ≡ c1)
Γ1
Henry Ka1s =
Γ∞
1/ m
⎛Γ ⎞
Freundlich Ka1s = ⎜ 1 ⎟
⎝ Γ∞ ⎠
Γ1
Langmuir Ka1s =
Γ ∞ − Γ1
Γ1 ⎛ Γ1 ⎞
Volmer Ka1s = exp ⎜ ⎟
Γ ∞ − Γ1 ⎝ Γ ∞ − Γ1 ⎠
Γ1 ⎛ 2βΓ1 ⎞
Frumkin Ka1s = exp ⎜ − ⎟
Γ ∞ − Γ1 ⎝ kT ⎠
Γ1 ⎛ Γ1 2βΓ1 ⎞
van der Waals Ka1s = exp ⎜ − ⎟
Γ ∞ − Γ1 ⎝ Γ ∞ − Γ1 kT ⎠
Note: The surfactant adsorption isotherm and the surface tension isotherm, which are combined to fit experimental data,
obligatorily must be of the same type.
per adsorption site. In the case of nonlocalized adsorption (Volmer and van der Waals isotherms),
1/Γ∞ is the excluded area per molecule.
The standard free energy of surfactant adsorption, ΔGo, can be determined by nonlinear fits
of surface tension isotherms with the help of a theoretical model of adsorption. The models of
Frumkin, van der Waals, and Helfant–Frisch–Lebowitz have been applied, and the results have
been compared [16]. Irrespective of the differences between these models, they give close values
for the standard free energy because all of them reduce to the Henry isotherm for diluted adsorp-
tion layers. The results from the theoretical approach have been compared with those of the most
popular empirical approach [17]. The latter gives values of the standard free energy, which are
considerably different from the respective true values, with ca. 10 kJ/mol for nonionic surfactants,
and 20 kJ/mol for ionic surfactants. These differences are due to contributions from interactions
between the molecules in dense adsorption layers. The true values of the standard free energy can
be determined with the help of an appropriate theoretical model. The van der Waals model was
252 Handbook of Surface and Colloid Chemistry
found to give the best results, especially for the determination of the standard adsorption enthalpy
ΔHo and entropy ΔSo from the temperature dependence of surface tension [16].
As already mentioned, the Freundlich adsorption isotherm, unlike the other ones in Table 4.2,
does not become linear at low concentrations, but remains convex to the concentration axis.
Moreover, it does not show saturation or limiting value. Hence, for the Freundlich adsorption
isotherm in Table 4.2 Γ∞ is a parameter scaling the adsorption (rather than saturation adsorption).
This isotherm can be derived assuming that the surface (as a rule solid) is heterogeneous [18,19].
Consequently, if the data fit the Freundlich equation, this is an indication, but not a proof, that the
surface is heterogeneous [6].
The adsorption isotherms in Table 4.2 can be applied to both fluid and solid interfaces. The
surface tension isotherms in Table 4.2, which relate σ and Γ1, are usually applied to fluid interfaces,
although they could also be used also for solid–liquid interfaces if σ is identified with the Gibbs [4]
superficial tension. (The latter is defined as the force per unit length which opposes every increase
of the wet area without any deformation of the solid.)
The surface tension isotherms in Table 4.2 are deduced from the respective adsorption isotherms
in the following way. The integration of Equation 4.2 yields
σ = σ0 − kTJ (4.4)
c1 Γ1
dc d ln c1
J ≡ Γ1 1 = d Γ1 (4.5)
∫
0
c1 ∫Γ
0
1
dΓ1
The derivative d ln c1/dΓ1 is calculated for each adsorption isotherm, and then the integration in
Equation 4.5 is carried out analytically. The obtained expressions for J are listed in Table 4.2. Each
surface tension isotherm, σ(Γ1), has the meaning of a 2D equation of state of the adsorption mono-
layer, which can be applied to both soluble and insoluble surfactants [6,20].
An important thermodynamic property of a surfactant adsorption monolayer is its Gibbs (surface)
elasticity
⎛ ∂σ ⎞
EG ≡ −Γ1 ⎜ ⎟ (4.6)
⎝ ∂Γ1 ⎠T
Expressions for EG, corresponding to various adsorption isotherms, are shown in Table 4.3. The
Gibbs elasticity characterizes the lateral fluidity of the surfactant adsorption monolayer. At high
values of the Gibbs elasticity the adsorption monolayer behaves as tangentially immobile. In such
case, if two emulsion droplets approach each other, the hydrodynamic flow pattern, and the hydro-
dynamic interaction as well, is almost the same as if the droplets were solid. For lower values of the
surfactant adsorption the so-called “Marangoni effect” appears, which is equivalent to appearance
of gradients of surface tension due to gradients of surfactant adsorption: ∇sσ = −(EG /Γ1)∇sΓ1 (here ∇s
denotes surface gradient operator). The Marangoni effect can considerably affect the hydrodynamic
interactions of fluid particles (drops, bubbles), (see Section 4.5).
TABLE 4.3
Elasticity of Adsorption Monolayers at a Fluid Interface
Type of Isotherm
(cf. Table 4.2) Gibbs Elasticity EG
Henry EG = kTΓ1
Γ
Freundlich EG = kT 1
m
Γ∞
Langmuir EG = kT Γ1
Γ ∞ − Γ1
Γ 2∞
Volmer EG = kT Γ1
(Γ ∞ − Γ1 )2
⎛ Γ∞ 2βΓ1 ⎞
Frumkin EG = kT Γ1 ⎜ − ⎟
⎝ Γ ∞ − Γ1 kT ⎠
⎡ Γ 2∞ 2βΓ1 ⎤
van der Waals EG = kT Γ1 ⎢ − ⎥
(
⎣ ∞ Γ − Γ1 ) 2
kT ⎦
Note: The above expressions are valid for both nonionic and ionic surfactants.
1. With the help of statistical mechanics an expression is obtained, say, for the canonical
ensemble partition function, Q, from which the surface free energy Fs is determined [11]:
where
A is the interfacial area
N1 is the number of adsorbed surfactant molecules; see Table 4.4
2. Differentiating the expression for Fs, we derive expressions for the surface pressure, πs,
and the surface chemical potential of the adsorbed surfactant molecules, μ1s [11]:
⎛ ∂F ⎞ ⎛ ∂F ⎞
πs ≡ σ0 − σ = − ⎜ s ⎟ , μ1s = ⎜ s ⎟ (4.8)
⎝ ∂A ⎠T , N1 ⎝ ∂N1 ⎠T , A
Combining the obtained expressions for πs and μ1s, we can deduce the respective form of
the Butler equation [21], see Equation 4.16.
3. The surfactant adsorption isotherm (Table 4.2) can be derived by setting the obtained
expression for the surface chemical potential μ1s equal to the bulk chemical potential of the
surfactant molecules in the subsurface layer (i.e., equilibrium between surface and subsur-
face is assumed) [11]:
⎛a δ ⎞
μ1s = μ1( 0 ) + kT ln ⎜ 1s 1 ⎟ (4.9)
⎝ Γ∞ ⎠
254 Handbook of Surface and Colloid Chemistry
TABLE 4.4
Free Energy and Chemical Potential for Surfactant Adsorption Layers
Surface Free Energy Fs(T, A, N1)
Type of Isotherm (M = Γ∞ A)
⎡ ⎛N ⎞ ⎤
Henry Fs = N1μ1( 0s ) + kT ⎢ N1 ln ⎜ 1 ⎟ − N1 ⎥
⎣ ⎝ M ⎠ ⎦
kT ⎡ ⎛ N1 ⎞ ⎤
Freundlich Fs = N1μ1( 0s ) + ⎢ N1 ln ⎜ ⎟ − N1 ⎥
m ⎣ ⎝M⎠ ⎦
Langmuir Fs = N1μ1( 0s ) + kT [ N1 ln N1 + ( M − N1 ) ln( M − N1 ) − M ln M ]
Volmer Fs = N1μ1( 0s ) + kT [ N1 ln N1 − N1 − N1 ln( M − N1 )]
β Γ ∞ N12
Frumkin Fs = N1μ1( 0s ) + kT [ N1 ln N1 + ( M − N1 ) ln( M − N1 ) − M ln M ] +
2M
β Γ ∞ N12
van der Waals Fs = N1μ1( 0s ) + kT [ N1 ln N1 − N1 − N1 ln( M − N1 )] +
2M
Here a1s is the activity of the surfactant molecule in the subsurface layer; a1s is scaled with the
volume per molecule in a dense (saturated) adsorption layer, v1 = δ1/Γ∞, where δ1 is interpreted
as the thickness of the adsorption layer, or the length of an adsorbed molecule. In terms of the
subsurface activity, a1s, Equation 4.9 can be applied to ionic surfactants and to dynamic processes.
In the simplest case of nonionic surfactants and equilibrium processes we have a1s ≈ c1, where c1 is
the bulk surfactant concentration.
First, let us apply the general scheme mentioned earlier to derive the Frumkin isotherm, which
corresponds to localized adsorption of interacting molecules. (Expressions corresponding to the
Langmuir isotherm can be obtained by setting β = 0 in the respective expressions for the Frumkin
isotherm.) Let us consider the interface as a 2D lattice having M adsorption sites. The corresponding
partition function is [11]
M! ⎛ n wN 2 ⎞
Q(T , M , N1 ) = [q(T )]N1 exp ⎜ − c 1 ⎟ (4.10)
N1 ! ( M − N1 )! ⎝ 2kTM ⎠
The first multiplier in the right-hand side of Equation 4.10 expresses the number of ways N1 indis-
tinguishable molecules can be distributed among M labeled sites; the partition function for a single
Chemical Physics of Colloid Systems and Interfaces 255
adsorbed molecule is q = qxqyqz, where qx, qy, and qz are 1D harmonic-oscillator partition functions.
The exponent in Equation 4.10 accounts for the interaction between adsorbed molecules in the
framework of the Bragg–Williams approximation [11]. w is the nearest-neighbor interaction energy
of two molecules and nc is the number of nearest-neighbor sites to a given site (for example nc = 4 for
a square lattice). Next, we substitute Equation 4.10 into Equation 4.7 and using the known Stirling
approximation, ln M! = M ln M − M, we get the expression for the surface free energy correspond-
ing to the Frumkin model:
nc wN12
Fs = kT [ N1 ln N1 + ( M − N1 ) ln( M − N1 ) − M ln M − N1 ln q(T )] + (4.11)
2M
Note that
M = Γ ∞ A, N1 = Γ1 A (4.12)
where Γ ∞−1 is the area per one adsorption site in the lattice. Differentiating Equation 4.11 in accor-
dance with Equation 4.8, we deduce expressions for the surface pressure and chemical potential [11]:
θ
μ1s = μ1( 0s ) + kT ln †−2βΓ1 (4.14)
1− θ
Γ1 nw
θ= , β = − c , μ1( 0s ) = −kT ln q(T ) (4.15)
Γ∞ 2Γ ∞
We can check that Equation 4.13 is equivalent to the Frumkin’s surface tension isotherm in Table 4.2
for a nonionic surfactant. Furthermore, eliminating ln(1 − θ) between Equations 4.13 and 4.14, we
obtain the Butler [21] equation in the form
⎡ βΓ θ(2 − θ) ⎤
γ1s = exp ⎢ − ∞ ⎥ †(for Frumkin isotherm ) (4.17)
⎣ kT ⎦
(In the special case of Langmuir isotherm we have β = 0, and then γ1s = 1.) The Butler equation
is used by many authors [12,22–24] as a starting point for the development of thermodynamic
adsorption models. It should be kept in mind that the specific form of the expressions for πs and γ1s,
which are to be substituted in Equation 4.16, is not arbitrary, but must correspond to the same ther-
modynamic model (to the same expression for Fs —in our case Equation 4.11). At last, substituting
Equation 4.16 into Equation 4.9, we derive the Frumkin adsorption isotherm in Table 4.2, where K
is defined by Equation 4.3.
Now, let us apply the same general scheme, but this time to the derivation of the van der Waals
isotherm, which corresponds to nonlocalized adsorption of interacting molecules. (Expressions
corresponding to the Volmer isotherm can be obtained by setting β = 0 in the respective expressions
256 Handbook of Surface and Colloid Chemistry
for the van der Waals isotherm.) Now the adsorbed N1 molecules are considered as a 2D gas.
The corresponding expression for the canonical ensemble partition function is
1 N1 ⎛ n wN 2 ⎞
Q(T , M , N1 ) = q exp ⎜ − c 1 ⎟ (4.18)
N1 ! ⎝ 2kTM ⎠
where the exponent accounts for the interaction between adsorbed molecules, again in the frame-
work of the Bragg–Williams approximation. The partition function for a single adsorbed molecule
is q = qxyqz, where qz is 1D (normal to the interface) harmonic-oscillator partition function. On the
other hand, the adsorbed molecules have free translational motion in the xy-plane (the interface);
therefore we have [11]
2πmkT
Aˆ (4.19)
q xy =
hp2
where
m
is the molecular mass
hp is the Planck constant
 = A − N1Γ ∞−1 is the area accessible to the moving molecules; the parameter Γ ∞−1 is the excluded
area per molecule, which accounts for the molecular size
Having in mind that M ≡ Γ∞ A, we can bring Equation 4.18 into the form
1 N1 ⎛ n wN 2 ⎞
Q(T , M , N1 ) = q0 ( M − N1 ) N1 exp ⎜ − c 1 ⎟ (4.20)
N1 ! ⎝ 2kTM ⎠
where
2πmkT
q0 (T ) ≡ 2
qz (T ) (4.21)
hpΓ ∞
Further, we substitute Equation 4.20 into Equation 4.7 and, using the Stirling approximation, we
determine the surface free energy corresponding to the van der Waals model [11,20,25]:
nc wN12
Fs = kT [ N1 ln N1 − N1 − N1 ln q0 (T ) − N1 ln( M − N1 )] + (4.22)
2M
Again, having in mind that M ≡ Γ∞ A, we differentiate Equation 4.22 in accordance with Equation
4.8 to deduce expressions for the surface pressure and chemical potential:
θ
πs = Γ ∞ kT − βΓ12 (4.23)
1− θ
⎛ θ θ ⎞
μ1s = μ1( 0s ) + kT ⎜ + ln †−2βΓ1 (4.24)
⎝ 1− θ 1 − θ ⎟⎠
where
μ1( 0s ) = −kT ln q0 (T )
β is defined by Equation 4.14
Chemical Physics of Colloid Systems and Interfaces 257
We can check that Equation 4.23 is equivalent to the van der Waals surface tension isotherm in
Table 4.2 for a nonionic surfactant. Furthermore, combining Equations 4.23 and 4.24, we obtain
the Butler Equation 4.16, but this time with another expression for the surface activity coefficient
1 ⎡ βΓ θ(2 − θ) ⎤
γ1s = exp ⎢ − ∞ ⎥ (for van der Waals isotherm ) (4.25)
1− θ ⎣ kT ⎦
[In the special case of Volmer isotherm we have β = 0, and then γ1s = 1/(1 − θ).] Finally, substituting
Equation 4.24 into Equation 4.9, we derive the van der Waals adsorption isotherm in Table 4.2, with
K defined by Equation 4.3.
In Table 4.4 we summarize the expressions for the surface free energy, Fs, and chemical potential
μ1s, for several thermodynamic models of adsorption. We recall that the parameter Γ∞ is defined in
different ways for the different models. On the other hand, the parameter K is defined in the same
way for all models, viz. by Equation 4.3. The expressions in Tables 4.2 through 4.4 can be general-
ized for multicomponent adsorption layers [20,26].
At the end of this section, let us consider a general expression, which allows us to obtain the
surface activity coefficient γ1s directly from the surface pressure isotherm πs(θ). From the Gibbs
adsorption isotherm, dπs = Γ1dμ1s, it follows that
⎛ ∂μ1s ⎞ 1 ⎛ ∂πs ⎞
⎜ ∂Γ ⎟ = Γ ⎜ ∂Γ ⎟ (4.26)
⎝ 1 ⎠T 1⎝ 1 ⎠T
By substituting μ1s from the Butler’s Equation 4.16 into Equation 4.26 and integrating, we can derive
the sought for expression:
θ
⎛ (1 − θ) ∂πs ⎞ dθ
ln γ1s = ⎜ − 1⎟
∫
0
⎝ Γ ∞ kT ∂θ ⎠ θ
(4.27)
We can check that a substitution of πs from Equations 4.13 and 4.23 into Equation 4.27 yields,
respectively, the Frumkin and van der Waals expressions for γ1s, viz. Equations 4.17 and 4.24.
μi = μ(i 0 ) + kT ln ai + Z i eψ (4.28)
where
e is the elementary electric charge
ψ is the electric potential
Zi is the valence of the ionic component “i”
ai is its activity
In the EDL (Figure 4.1), the electric potential and the activities of the ions are dependent on the
distance z from the phase boundary: ψ = ψ(z), ai = ai(z). At equilibrium the electrochemical
258 Handbook of Surface and Colloid Chemistry
Aqueous phase
– + –
+
–
Nonaqueous phase
Coions
+ –
–
+
– +
+
–
– Counterions
+
–
+ –
–
Surfactant Diffuse layer
adsorption layer Stern layer
(a) of adsorbed
counterions
Nonionic concentration
Counterions
c∞
Coions
0 z
(b)
FIGURE 4.1 Electric double layer in the vicinity of an adsorption layer of ionic surfactant. (a) The diffuse
layer contains free ions involved in Brownian motion, while the Stern layer consists of adsorbed (bound)
counterions. (b) Near the charged surface there is an accumulation of counterions and a depletion of coions.
potential, μi, is uniform throughout the whole solution, including the EDL (otherwise diffusion
fluxes would appear) [31]. In the bulk of solution (z → ∞) the electric potential tends to a constant
value, which is usually set equal to zero, that is ψ → 0 and ∂ψ/∂z → 0 for z → ∞. If the expression for
μi at z → ∞ and that for μi at some finite z are set equal, from Equation 4.28 we obtain a Boltzmann-
type distribution for the activity across the EDL [31]:
⎡ Z eψ( z ) ⎤
ai ( z ) = ai∞ exp ⎢ − i (4.29)
⎣ kT ⎥⎦
where ai∞ denotes the value of the activity of ion “i” in the bulk of solution. If the activity in the
bulk, ai∞, is known, then Equation 4.29 determines the activity ai(z) in each point of the EDL.
A good agreement between theory and experiment can be achieved [12,13,26] using the following
expression for ai∞:
where
ci∞ is the bulk concentration of the respective ion
Chemical Physics of Colloid Systems and Interfaces 259
A Z+Z− I
log γ ± = − + bI (4.31)
1 + Bdi I
which originates from the Debye–Hückel theory; I denotes the ionic strength of the solution:
1
I≡
2 ∑Z c
i
2
i i∞ (4.32)
where the summation is carried out over all ionic species in the solution. When the solution contains
a mixture of several electrolytes, then Equation 4.31 defines γ± for each separate electrolyte, with Z +
and Z − being the valences of the cations and anions of this electrolyte, but with I being the total ionic
strength of the solution, accounting for all dissolved electrolytes [32]. The log in Equation 4.31 is
decimal, di is the ionic diameter, A, B, and b are parameters, whose values can be found in the book
by Robinson and Stokes [32]. For example, if I is given in moles per liter (M), the parameters values
are A = 0.5115 M−1/2, Bdi = 1.316 M−1/2, and b = 0.055 M−1 for solutions of NaCl at 25°C.
The theory of EDL provides a connection between surface charge and surface potential (known
as the Gouy equation [33,34] of Graham equation [35,36]), which can be presented in the form
[26,37]
1/ 2
N N
2 ⎧⎪ ⎫⎪
∑ zi Γi = ⎨ ∑ ai∞ [exp(− zi Φ s ) − 1]⎬ (Gouy equatiion) (4.33)
i =1
κc ⎩⎪ i =1 ⎭⎪
where
Γi (i = 1, …, N) are the adsorptions of the ionic species
zi = Zi/Z1
2 Z12e2 Z eψ
κ2c ≡ , Φ s ≡ 1 s (4.34)
ε0εkT kT
The adsorption of the coions of the nonamphiphilic salt is expected to be equal to zero, Γ3 = 0,
because they are repelled by the similarly charged interface [26,38–40]. However, the adsorption
of surfactant at the interface, Γ1, and the binding of counterions in the Stern layer, Γ2, are different
from zero (Figure 4.1). For this system the Gouy Equation 4.33 acquires the form
4 ⎛Φ ⎞
Γ1 − Γ 2 = a2 ∞ sinh ⎜ s ⎟ ( Z1:Z1 electrolyte ) (4.36)
κc ⎝ 2 ⎠
∞
i = Γi + Λ i , where Λ i ≡ [ ai ( z ) − ai∞ ]dz (4.37)
Γ
∫
0
i represents a surface excess of component “i” with respect to the uniform bulk solution.
and Γ
N N
Because the solution is electroneutral, we have i = 0. Note, however, that
zi Γ∑ zi Γi ≠ 0, ∑
i =1 i =1
see the Gouy Equation 4.33. Expressions for Λi can be obtained by using the theory of EDL. For
example, because of the electroneutrality of the solution, the right-hand side of Equation 4.36 is
equal to Λ2 − Λ1 − Λ3, where
⎡ ⎛Φ ⎞ ⎤ ⎡ ⎛ −Φ s ⎞ ⎤
Λ 2 = 2a2∞ κ −1 ⎢exp ⎜ s ⎟ − 1⎥ ; Λ j = 2a j∞ κ −1 ⎢exp ⎜ ⎟ − 1⎥ , j = 1, 3 (4.38)
⎣ ⎝ 2 ⎠ ⎦ ⎣ ⎝ 2 ⎠ ⎦
(κ2 = κ2c I ; Z1:Z1 electrolyte). In analogy with Equation 4.37, the interfacial tension of the solution, σ,
can be expressed as a sum of contributions from the adsorption and diffuse layers [26,27,34]:
σ = σa + σd (4.39)
where
∞ 2
⎛ dψ ⎞
σa = σ0 − kTJ and σd = −ε0ε ⎜ ⎟ dz (4.40)
0
⎝ dz ⎠∫
Expressions for J are given in Table 4.2 for various types of isotherms. Note that Equations 4.39 and
4.40 are valid under both equilibrium and dynamic conditions. In the special case of SDS + NaCl
solution (see the text explained earlier), at equilibrium, we can use the theory of EDL to express
dψ/dz; then from Equation 4.40 we derive [26,27,34]
8kT ⎡ ⎛Φ ⎞ ⎤
σd = − a2 ∞ ⎢cosh ⎜ s ⎟ − 1⎥ ( Z1:Z1 electrolyte, at equiliibrium) (4.41)
κc ⎣ ⎝ 2 ⎠ ⎦
Analytical expressions for σd for the cases of 2:1, 1:2, and 2:2 electrolytes can be found in
Refs. [26,36].
Chemical Physics of Colloid Systems and Interfaces 261
In the case of ionic surfactants, Equation 4.1 can be presented in two alternative, but equivalent
forms [26,37]:
N
dσ = −kT ∑ Γ d ln a
i =1
i i∞ (T = const.) (4.42)
N
dσa = −kT ∑ Γ d ln a
i =1
i is (T = const.) (4.43)
where ais = ai (z = 0) is the “subsurface” value of activity ai. From Equations 4.29 and 4.34, we
obtain
The comparison between Equations 4.42 and 4.43 shows that the Gibbs adsorption equation can be
i, and ai∞, or in terms of σa, Γi, and ais. Note that Equations 4.42 and
expressed either in terms of σ, Γ
4.44 are valid under equilibrium conditions, whereas Equation 4.43 can also be used for the descrip-
tion of dynamic surface tension (Section 4.2.2) in the case of surfactant adsorption under diffusion
control, assuming local equilibrium between adsorptions Γi and subsurface concentrations of the
respective species.
The expression σa = σ0 − kTJ, with J given in Table 4.2, can be used for description of both static
and dynamic surface tension of ionic and nonionic surfactant solutions. The surfactant adsorption
isotherms in this table can be used for both ionic and nonionic surfactants, with the only difference
that in the case of ionic surfactant the adsorption constant K depends on the subsurface concentra-
tion of the inorganic counterions [26], see Equation 4.48.
where as before, the indices “1” and “2” refer to the DS− and Na+ ions, respectively. The differentials
in the right-hand side of Equation 4.45 are independent (we can vary independently the concentra-
tions of surfactant and salt), and moreover, dσa is an exact (total) differential. Then, according to the
Euler condition, the cross derivatives must be equal [26]:
∂Γ1 ∂Γ 2
= (4.46)
∂ ln a2 s ∂ ln a1s
Γ2 K 2 a2 s
= (Stern isotherm ) (4.47)
Γ1 1 + K 2 a2 s
where K2 is a constant parameter. The latter equation, termed the Stern isotherm [41], describes
Langmuirian adsorption (binding) of counterions in the Stern layer. It can be proven that a sufficient
262 Handbook of Surface and Colloid Chemistry
condition Γ2 form Equation 4.47 to satisfy the Euler’s condition 4.46, together with one of the
surfactant adsorption isotherms for Γ1 in Table 4.2, is [26]
K = K1 (1 + K 2 a2 s ) (4.48)
where K1 is another constant parameter. In other words, if K is expressed by Equation 4.48, the Stern
isotherm 4.47 is thermodynamically compatible with each of the surfactant adsorption isotherms
in Table 4.2. In analogy with Equation 4.3, the parameters K1 and K2 are related to the respective
standard free energies of adsorption of surfactant ions and counterions Δμ(i 0 ):
δi ⎛ Δμ( 0 ) ⎞
Ki = exp ⎜ i ⎟ (i = 1, 2) (4.49)
Γ∞ ⎝ kT ⎠
Γ1 Γ2
= θ1 + θ2 ; = θ2 (4.50)
Γ∞ Γ∞
Following Kalinin and Radke [28], we consider the “reaction” of adsorption of S− ions:
A 0 + S− = A 0S− (4.51)
where A0 symbolizes an empty adsorption site. In accordance with the rules of the chemical k inetics,
we can express the rates of adsorption and desorption in the form:
where
c1s is the subsurface concentration of surfactant
K1,ads and K1,des are constants
In view of Equation 4.50, we can write θ 0 = (Γ∞ − Γ1)/Γ∞ and θ1 = (Γ1 − Γ2)/Γ∞. Thus, with the help
of Equation 4.52 we obtain the net adsorption flux of surfactant:
(Γ ∞ − Γ1 ) (Γ − Γ 2 )
Q1 ≡ r1,ads − r1,des = K1,adsc1s − K1,des 1 (4.53)
Γ∞ Γ∞
where
K2,ads and K2,des are the respective rate constants
c2s is the subsurface concentration of counterions
Having in mind that θ1 = (Γ1 − Γ2)/Γ∞ and θ2 = Γ2/Γ∞, with the help of Equation 4.55 we deduce an
expression for the adsorption flux of counterions:
(Γ1 − Γ 2 ) Γ
Q2 ≡ r2,ads − r2,des = K 2,adsc2s − K 2,des 2 (4.56)
Γ∞ Γ∞
If we can assume that the reaction of counterion binding is much faster than the surfactant adsorp-
tion, then we can set Q2 ≡ 0, and Equation 4.56 reduces to the Stern isotherm, Equation 4.47, with
K2 ≡ K2,ads/K2,des. Next, a substitution of Γ2 from Equation 4.47 into Equation 4.53 yields [37]
(Γ ∞ − Γ1 ) Γ
Q1 ≡ r1,ads − r1,des = K1,adsc1s − K1,des (1 + K 2c2s )−1 1 (4.57)
Γ∞ Γ∞
Equation 4.57 shows that the adsorption flux of surfactant is influenced by the subsurface concen-
tration of counterions, c2s. At last, if there is equilibrium between surface and subsurface, we have
to set Q1 ≡ 0 in Equation 4.57, and thus we obtain the Langmuir isotherm for an ionic surfactant:
Γ1 K
Kc1s = , with K ≡ 1,ads (1 + K 2c2s ) (4.58)
Γ ∞ − Γ1 K1,des
Note that K1 ≡ K1,ads /K1,des. This result demonstrates that the linear dependence of K on c2s (Equation
4.48) can be deduced from the reactions of surfactant adsorption and counterion binding, Equations
4.51 and 4.54. (For I < 0.1 M we have γ± ≈ 1 and then activities and concentrations of the ionic spe-
cies coincide.)
80
NaCl
concentration
70
0 mM
FIGURE 4.2 Plot of the surface tension σ vs. the concentration of SDS, c1∞, for 11 fixed NaCl concentra-
tions. The symbols are experimental data by Tajima et al. [40,42,43]. The lines represent the best fit [42] with
the full set of equations specified in the text, involving the van der Waals isotherms of adsorption and surface
tension (Table 4.2).
simultaneously [44]. In other words, the parameters β, Γ∞, K1, and K2 are the same for all curves.
The value of Γ∞, obtained from the best fit of the data in Figure 4.2, corresponds to 1/Γ∞ = 29.8 Å2.
The respective value of K1 is 99.2 m3/mol, which in view of Equation 4.49 gives a standard free
energy of surfactant adsorption Δμ1( 0 ) = 12.53kT per DS− ion, that is 30.6 kJ/mol. The determined
value of K2 is 6.5 × 10−4 m3/mol, which after substitution in Equation 4.49 yields a standard free
energy of counterion binding Δμ(20 ) = 1.64kT per Na+ ion, that is, 4.1 kJ/mol. The value of the param-
eter β is positive, 2βΓ∞/kT = +2.73, which indicates attraction between the hydrocarbon tails of the
adsorbed surfactant molecules. However, this attraction is too weak to cause 2D phase transition.
The van der Waals isotherm predicts such transition for 2βΓ∞/kT > 6.74.
Figure 4.3 shows calculated curves for the adsorptions of surfactant, Γ1 (the full lines), and coun-
terions, Γ2 (the dotted lines), versus the SDS concentration, c1∞. These lines represent the variations
1.0
DS– adsorption
0.8 Na+ adsorption
Dimensionless adsorption
115 mM NaCl
0.6
No salt
0.4
0.2
0.0
0.01 0.1 1 10
SDS concentration (mM)
FIGURE 4.3 Plots of the dimensionless adsorptions of surfactant ions Γ1/Γ∞ (DS–, the full lines), and coun-
terions Γ2/Γ∞ (Na+, the dotted lines), vs. the surfactant (SDS) concentration, c1∞. The lines are calculated [44]
for NaCl concentrations 0 and 115 mM using parameter values determined from the best fit of experimental
data (Figure 4.2).
Chemical Physics of Colloid Systems and Interfaces 265
of Γ1 and Γ2 along the experimental curves, which correspond to the lowest and highest NaCl con-
centrations in Figure 4.2, viz. c3∞ = 0 and 115 mM. We see that both Γ1 and Γ2 are markedly greater
when NaCl is present in the solution. The highest values of Γ1 for the curves in Figure 4.3 are
4.2 × 10−6 and 4.0 × 10−6 mol/m2 for the solutions with and without NaCl, respectively. The latter
two values compare well with the saturation adsorptions measured by Tajima et al. [42,43] for the
same system by means of the radiotracer method, viz. Γ1 = 4.3 × 10−6 and 3.2 × 10−6 mol/m2 for the
solutions with and without NaCl, respectively.
For the solution without NaCl the occupancy of the Stern layer, Γ2/Γ1, rises from 0.15 to
0.73 and then exhibits a tendency to level off. The latter value is consonant with data of other
authors [45–47], who have obtained values of Γ2/Γ1 up to 0.70−0.90 for various ionic surfactants;
pronounced evidences for counterion binding have also been obtained also in experiments with
solutions c ontaining surfactant micelles [48–52]. As it could be expected, both Γ1 and Γ2 are higher
for the solution with NaCl. These results imply that the counterion adsorption (binding) should be
always be taken into account.
The fit of the data in Figure 4.2 also gives the values of the surface electric potential, ψs. For
the solutions with 115 mM NaCl the model predicts surface potentials varying in the range |ψs| =
55−95 mV within the experimental interval of surfactant concentrations, whereas for the solution
without salt the calculated surface potential is higher: |ψs| = 150–180 mV (for SDS ψs has a negative
sign). Thus it turns out that measurements of surface tension, interpreted by means of an appropri-
ate theoretical model, provide a method for determining the surface potential ψs in a broad range of
surfactant and salt concentrations. The described approach could be also applied to solve the inverse
problem, viz. to process data for the surface potential. In this way, the adsorption of surfactant on
solid particles can be determined from the measured zeta-potential [53].
It is remarkable that the minimal (excluded) area per adsorbed surfactant molecule, α ≡ 1/Γ∞,
obtained from the best fit of surface tension data by the van der Waals isotherm practically c oincides
with the value of α estimated by molecular size considerations (i.e., from the maximal cross-sectional
area of an amphiphilic molecule in a dense adsorption layer); see for example Figure 7.1 in Ref. [36].
This is illustrated in Table 4.5, which contains data for alkanols, alkanoic acids, (SDS), (DDBS),
cocamidopropyl betaine (CAPB), and Cn-trimethyl ammonium bromides (n = 12, 14, and 16).
The second column of Table 4.5 gives the group whose cross-sectional area is used to calculate α.
For molecules of circular cross section, we can calculate the cross-sectional area from the expres-
sion α = πr 2, where r is the respective radius. For example [54], the radius of the SO42− ion is r =
3.09 Å, which yields α = πr 2 = 30.0 Å2. In the fits of surface tension data by the van der Waals
isotherm, α was treated as an adjustable parameter, and the value α = 30 Å2 was obtained from the
best fit. As seen in Table 4.5, excellent agreement between the values of α obtained from molecular
size and from surface tension fits is obtained also for many other amphiphilic molecules [54–61].
TABLE 4.5
Excluded Area per Molecule, α, Determined in Two Different Ways
α from Molecular α from Surface
Amphiphile Group Size (Å2) Tension Fitsa (Å2) References
Alkanols Paraffin chain 21.0 20.9 [54]
Alkanoic acids COO− 22–24 22.6 [55,56]
SDS SO42− 30.0 30 [44,57]
DDBS Benzene ring 35.3 35.6 [58]
CAPB CH3–N+–CH3 27.8 27.8 [59]
CnTAB (n = 12, 14, 16) N(CH3)4+ 37.8 36.5–39.5 [57,60]
It should be noted the result mentioned earlier holds only for the van der Waals (or Volmer)
isotherm. Instead, if the Frumkin (or Langmuir) isotherm is used, the value of α obtained from
the surface tension fits is about 33% greater than that obtained from molecular size [44]. A pos-
sible explanation of this difference could be the fact that the Frumkin (and Langmuir) isotherm
is s tatistically derived for localized adsorption and is more appropriate to describe adsorption at
solid interfaces. In contrast, the van der Waals (and Volmer) isotherm is derived for nonlocalized
adsorption, and they provide a more adequate theoretical description of the surfactant adsorption at
liquid–fluid interfaces. This conclusion refers also to the calculation of the surface (Gibbs) elasticity
by means of the two types of isotherms [44].
The fact that α determined from molecular size coincides with that obtained from surface tension
fits (Table 4.5) is very useful for applications. Thus, when fitting experimental data, we can use the
value of α from molecular size, and thus to decrease the number of adjustable parameters. This
fact is especially helpful when interpreting theoretically data for the surface tension of surfactant
mixtures, such as SDS + dodecanol [54]; SDS + CAPB [59], and fluorinated + nonionic surfac-
tant [61]. An additional way to decrease the number of adjustable parameters is to employ the
Traube rule, which states that Δμ1( 0 ) increases with 1.025kT when a CH2 group is added to the paraf-
fin chain; for details see Refs. [54,55,60].
1. Diffusion of surfactant molecules from the bulk solution to the subsurface layer;
2. Transfer of surfactant molecules from the subsurface to the adsorption layer; the rate of
transfer is determined by the height of the kinetic barrier to adsorption.
(In the case of desorption the processes have the opposite direction.) Such interfacial expansions
are typical for foam generation and emulsification. The rate of adsorption relaxation determines
whether the formed bubbles/drops will coalesce upon collision, and in final reckoning—how large
will be the foam volume and the emulsion drop size [62,63]. In the following section, we focus our
attention on the relaxation time of surface tension, τσ, which characterizes the interfacial dynamics.
The overall rate of surfactant adsorption is controlled by the slowest stage. If it is stage (1), we
deal with diffusion control, while if stage (2) is slower, the adsorption occurs under barrier
(kinetic) control. The next four sections are dedicated to processes under diffusion control (which
are the most frequently observed), whereas in Section 4.2.2.5 we consider adsorption under barrier
control. Finally, Section 4.2.2.6 is devoted to the dynamics of adsorption from micellar surfactant
solutions.
Various experimental methods for dynamic surface tension measurements are available. Their
operational time scales cover different time intervals [64,65]. Methods with a shorter characteristic
operational time are the oscillating jet method [66–68], the oscillating bubble method [69–72], the
fast-formed drop technique [73,74], the surface wave techniques [75–78], and the maximum bubble
pressure method (MBPM) [57,79–84]. Methods of longer characteristic operational time are the
inclined plate method [85], the drop-weight/volume techniques [86–90], the funnel [91] and over-
flowing cylinder [60,92] methods, and the axisymmetric drop shape analysis (ADSA) [93,94]; see
Refs. [64,65,95] for a more detailed review.
In this section, devoted to dynamic surface tension, we consider mostly nonionic surfactant
solutions. In Section 4.2.2.4 we address the more complicated case of ionic surfactants. We will
Chemical Physics of Colloid Systems and Interfaces 267
restrict our considerations to the simplest case of relaxation of an initial uniform interfacial dila-
tation. The more complex case of simultaneous adsorption and dilatation is considered elsewhere
[57,64,80,84,92,95].
∂c1 ∂ 2c
= D1 21 ( z > 0, t > 0) (4.59)
∂t ∂z
where D1 is the diffusion coefficient of the surfactant molecules. The exchange of surfactant between
the solution and its interface is described by the boundary conditions
dΓ1 ∂c
c1 (0, t ) = c1s (t ), = D1 1 ( z = 0, t > 0) (4.60)
dt ∂z
The latter equation states that the rate of increase of the adsorption Γ1 is equal to the diffusion influx
of surfactant per unit area of the interface. Integrating Equation 4.59, along with 4.60, we can derive
the equation of Ward and Tordai [96]:
t
D1 ⎢⎡ c1s (τ) ⎤⎥
Γ1 (t ) = Γ1 (0) + 2c1∞ t − dτ (4.61)
π ⎢
⎣
∫
0
t−τ ⎥
⎦
Solving Equation 4.61 together with some of the adsorption isotherms Γ1 = Γ1(c1s) in Table 4.2, we
can in principle determine the two unknown functions Γ1(t) and c1s(t). Because the relation Γ1(c1s)
is nonlinear (except for the Henry isotherm), this problem, or its equivalent formulations, can be
solved either numerically [97], or by employing appropriate approximations [80,98].
In many cases, it is convenient to use asymptotic expressions for the functions Γ1(t), c1s(t), and σ(t)
for short times (t → 0) and long times (t → ∞). A general asymptotic expression for the short times
can be derived from Equation 4.61 substituting c1s ≈ c1s(0) = constant:
D1
Γ1 (t ) = Γ1 (0) + 2 [c1∞ − c1s (0)] t (t → 0) (4.62)
π
Analogous asymptotic expression can also be obtained for long times, although the derivation is not
so simple. Hansen [99] derived a useful asymptotics for the subsurface concentration:
Γ1e − Γ(0)
c1s (t ) = c1∞ − (t → ∞) (4.63)
πD1t
where Γ1e is the equilibrium value of the surfactant adsorption. The validity of Hansen’s Equation 4.63
was confirmed in subsequent studies by other authors [100,101].
In the following section, we continue our review of the asymptotic expressions considering
separately the cases of small and large initial perturbations.
268 Handbook of Surface and Colloid Chemistry
⎛ ∂Γ ⎞
Γ1 (t ) − Γ1,e ≈ ⎜ 1 ⎟ [c1s (t ) − ce ] (4.64)
⎝ ∂c1 ⎠e
Here and hereafter, the subscript “e” means that the respective quantity refers to the equilibrium
state. The set of linear Equations 4.59, 4.60, and 4.64 has been solved by Sutherland [102]. The
result, which describes the relaxation of a small initial interfacial dilatation, reads:
σ(t ) − σe Γ1 (t ) − Γ1,e ⎛ t ⎞ ⎛ t ⎞
= = exp ⎜ ⎟ erfc ⎜ (4.65)
σ(0) − σe Γ1 (0) − Γ1,e ⎝ τσ ⎠ ⎜ τσ ⎟⎟
⎝ ⎠
where
2
1 ⎛ ∂Γ1 ⎞
τσ ≡ (4.66)
D1 ⎜⎝ ∂c1 ⎟⎠e
∞
2
erfc( x ) ≡
∫ exp(− x )dx (4.67)
2
π
x
is the so-called complementary error function [103,104]. The asymptotics of the latter function for
small and large values of the argument are [103,104]:
2
2 e− x ⎡ ⎛ 1 ⎞⎤
erfc( x ) = 1 − x + O( x 3 ) for x 1; erfc( x ) = ⎢1 + O ⎜ x 2 ⎟ ⎥ for x 1 (4.68)
π πx ⎣ ⎝ ⎠⎦
Combining Equations 4.65 and 4.68, we obtain the short-time and long-time asymptotics of the
surface tension relaxation:
σ(t ) − σe Γ1 (t ) − Γ1,e 2 t ⎡ ⎛ t ⎞3 / 2 ⎤
= = 1− + O ⎢⎜ ⎟ ⎥ (t τσ ) (4.69)
σ(0) − σe Γ1 (0) − Γ1,e π τσ τ
⎣⎢⎝ σ ⎠ ⎥⎦
σ(t ) − σe Γ1 (t ) − Γ1,e τσ ⎡⎛ τ ⎞3 / 2 ⎤
= = + O ⎢⎜ σ ⎟ ⎥ t τσ (4.70)
σ(0) − σe Γ1 (0) − Γ1,e πt ⎢⎣⎝ t ⎠ ⎥⎦
Equation 4.70 is often used as a test to verify whether the adsorption process is under diffusion
control: data for σ(t) are plotted versus 1/ t and it is checked if the plot complies with a straight line;
moreover, the intercept of the line gives σe. We recall that Equations 4.69 and 4.70 are valid in the
case of a small initial perturbation; alternative asymptotic expressions for the case of large initial
perturbation are considered in the next section.
Chemical Physics of Colloid Systems and Interfaces 269
With the help of the thermodynamic Equations 4.2 and 4.6, we derive
2
1 ⎛ Γ12 kT ⎞
τσ = ⎜ ⎟ (4.72)
D1 ⎝ c1EG ⎠e
Substituting EG from Table 4.3 into Equation 4.72, we can obtain expressions for τσ corresponding to
various adsorption isotherms. In the special case of Langmuir adsorption isotherm, we can present
Equation 4.72 in the form [37]
1 ( KΓ ∞ ) 2 1 ( KΓ ∞ ) 2
τσ = = (for Langmuir isootherm) (4.73)
D1 (1 + Kc1 ) 4
D1 (1 + EG / (Γ ∞ kT ))4
Equation 4.73 visualizes the very strong dependence of the relaxation time τσ on the surfac-
tant concentration c1; in general, τσ can vary with many orders of magnitude as a function of c1.
Equation 4.73 shows also that high Gibbs elasticity corresponds to short relaxation time, and vice
versa.
As a quantitative example let us take typical parameter values: K1 = 15 m3/mol, 1/Γ∞ = 40 Å2,
D1 = 4.5 × 10−6 cm2/s, and T = 298 K. Then with c1 = 6.5 × 10−6 M, from Table 4.3 (Langmuir
isotherm) and Equation 4.73, we calculate EG ≈ 1.0 mN/m and τσ ≈ 5 s. In the same way, for c1 =
6.5 × 10−4 M we calculate EG ≈ 100 mN/m and τσ ≈ 5 × 10−4 s.
To directly measure the Gibbs elasticity EG, or to precisely investigate the dynamics of surface
tension, we need an experimental method, whose characteristic time is smaller compared to τσ.
Equation 4.73 and the latter numerical example show that when the surfactant concentration is
higher, the experimental method should be faster.
Γ1,e
c1s (t ) = c1∞ − (t → ∞) (4.75)
πD1t
By substituting c1s(t) for c1 in the Gibbs adsorption Equation 4.2, and integrating, we obtain the
long-time asymptotics of the surface tension of a nonionic surfactant solution after a large initial
perturbation:
1/ 2
⎛ Γ 2 kT ⎞ ⎛ 1 ⎞
σ(t ) − σe = ⎜ 1 ⎟ ⎜ ⎟ (large initial perturbatiion) (4.76)
⎝ c1 ⎠e ⎝ πD1t ⎠
270 Handbook of Surface and Colloid Chemistry
with the help of Equation 4.72, we can bring Equation 4.76 into another form:
1/ 2
⎛τ ⎞
σ(t ) − σe = EG ⎜ σ ⎟ (large initial perturbation) (4.77)
⎝ πt ⎠
where E G is given in Table 4.3. It is interesting to note that Equation 4.77 is applicable to
both nonionic and ionic surfactants with the only difference that for nonionics τσ is given
by Equation 4.66, whereas for ionic surfactants the expression for τσ is somewhat longer, see
Refs. [37,105].
The equations mentioned earlier show that in the case of adsorption under diffusion control the
long-lime asymptotics can be expressed in the form
σ = σe + St −1/ 2 (4.78)
In view of Equations 4.70 and 4.77, the slope S of the dependence σ versus t−1/2 is given by the
expressions [105]
1/ 2
⎛τ ⎞
Ss = [σ(0) − σe ] ⎜ σ ⎟ (small perturbation) (4.79)
⎝ π⎠
1/ 2
⎛τ ⎞
Sl = EG ⎜ σ ⎟ (large perturbation) (4.80)
⎝ π⎠
As known, the surfactant adsorption Γ1 monotonically increases with the rise of the surfactant con-
centration, c1. In contrast, the slope Sl is a nonmonotonic function of c1: Sl exhibits a maximum at a
certain concentration. To demonstrate that we will use the expression
Γ12,e kT
Sl = (4.81)
c1 πD1
which follows from Equations 4.76 and 4.78. In Equation 4.81 we substitute the expressions for c1
stemming from the Langmuir and Volmer adsorption isotherms (Table 4.2 with c1 = a1s); the result
reads
⎛ θ ⎞
Sl = θ(1 − θ) exp ⎜ − (for Volmer isotherm ) (4.83)
⎝ 1 − θ ⎟⎠
where θ and Sl are the dimensionless adsorption and slope coefficient, respectively:
Γ1,e S πD1
θ= , Sl = l (4.84)
Γ∞ kT KΓ 2∞
Chemical Physics of Colloid Systems and Interfaces 271
0.25
Langmuir
0.20
Dimensionless slope
0.15
0.10 Volmer
0.05
0.29 0.5
0.00
0.0 0.2 0.4 0.6 0.8 1.0
Dimensionless adsorption, Γ1e/Γ∞
FIGURE 4.4 Plot of the dimensionless slope, Sl, vs. the dimensionless equilibrium surfactant adsorption, θ =
Γ1e /Γ∞, in accordance with Equations 4.82 and 4.83, corresponding to the cases of localized and nonlocalized
adsorption.
Figure 4.4 compares the dependencies Sl (θ) given by Equations 4.82 and 4.83: we see that the
former is symmetric and has a maximum at θ = 0.5, whereas the latter is asymmetric with a maxi-
mum at θ ≈ 0.29. We recall that the Langmuir and Volmer isotherms correspond to localized and
nonlocalized adsorption, respectively (see Section 4.2.1.1.2). Then Figure 4.4 shows that the symme-
try/asymmetry of the plot Sl versus θ provides a test for verifying whether the adsorption is localized
or nonlocalized. (The practice shows that the fits of equilibrium surface tension isotherms do not
provide such a test: theoretical isotherms corresponding to localized and nonlocalized adsorption
are found to fit surface tension data equally well!)
From another viewpoint, the nonmonotonic behavior of Sl(θ) can be interpreted as follows.
Equation 4.80 shows that Sl ∝ EG τσ ; then the nonmonotonic behavior stems from the fact that EG
is an increasing function of c1, whereas τσ is a decreasing function of c1. This qualitative conclusion
is also valid for the case of ionic surfactant, as demonstrated in the next section.
The first special case refers to adsorption at an immobile interface that has been initially per-
turbed, and to the maximum bubble pressure method (MBPM). The generalization of Equations
4.78 and 4.81 for this case reads [114]:
Sl kT Γ1,2 eλ ⎛ 1 1 ⎞
σ = σe + , Sl ≡ ⎜ + ⎟ (4.85a)
(tage )1/ 2
(πDeff ) γ ± ⎝ c1∞ c2∞
1/ 2
⎠
As usual, the subscript ‘e’ denotes equilibrium values; tage is the age of the interface, which is
defined as the period of time between the minimum pressure (bubble formation) and the maximum
pressure (bubble detachment) in the case of MBPM; λ is a dimensionless parameter; λ = 1 for immo-
bile interfaces; in the case of MBPM, λ is an apparatus constant that can be determined by calibra-
tion experiments [57]; as mentioned earlier, c1∞ and c2∞ are the bulk concentrations of surfactant
ions and counterions, respectively; γ± is the activity coefficient; Deff is an effective diffusivity that
depends on the diffusivities and bulk concentrations of surfactant ions, counterions, and inorganic
coions: Deff = Deff(D1, D2, D3, c1∞, c2∞, c3∞). The latter dependence is described by explicit formulas
derived in Ref. [114]; see Equations 6.19 through 6.26 therein.
In the case of the cationic surfactant dodecyl trimethyl ammonium bromide (C12TAB), the cal-
culated dependence of Deff on the surfactant and salt concentrations, c1∞ and c3∞, is illustrated in
Figure 4.4. Because the range c1∞ ≤ CMC is considered, the calculated curves end at the CMC.
At very low surfactant concentrations, c1∞ → 0, in the presence of salt (c3∞ > 0), the effective dif-
fusivity approaches its limiting value for diluted solutions, Deff → D1. We see that Deff increases
with the rise of c1∞, except the case without added salt (c3∞ = 0), for which Deff is a constant:
1/Deff = (1/D1 + 1/D 2)/2. The curves in Figure 4.5 show that Deff decreases with the rise of salt
concentration, c3∞, and becomes ≈D1 for c3∞ = 100 mM. Note that the salt concentration affects
the dynamic surface tension, σ, also through Γ1,e and through the factor (1/c1∞ + 1/c2∞) in Equation
4.85a; see Ref. [114] for details.
The accuracy of Equation 4.85a can be verified in the following way. Each of the dynamic
surface tension isotherms for C12TAB in Figure 4.6a are fitted by means of the equation σ = σe +
Sl/[aσ + (tage)1/2], and the parameters σe, aσ, and Sl are determined from the best fit. Next, for each
6.2
No NaBr
Diffusion coefficient, Deff × 106 (cm2/s)
6.0 1 mM NaBr
5.8
5 mM NaBr
5.6
5.4
10 mM NaBr Dependence of the
5.2 effective diffusivity of
100 mM NaBr
C12TAB on the surfactant
5.0 and salt concentrations
0 2 4 6 8 10 12
Surfactant concentration (mM)
FIGURE 4.5 Dependence of the effective diffusivity, Deff, on the surfactant concentration, c1∞, for vari-
ous salt concentrations, c3∞, denoted in the figure. The curves are calculated by using the values of D1, D 2,
and D 3 given in Ref. [114] for the cationic surfactant C12TAB. The end points of some curves correspond to
the CMC.
Chemical Physics of Colloid Systems and Interfaces 273
value of Sl, we calculate the equilibrium surfactant adsorption, Γ1,e, using Equation 4.85a and the
theoretical value of Deff from Figure 4.5; see the points in Figure 4.6b. For the used MBPM set up
[57], the apparatus constant is λ = 6.07. The solid line in the latter figure represents the equilibrium
surfactant adsorption independently calculated from the fit of equilibrium surface tension data by
means of the van der Waals isotherm [114]. The excellent agreement between the points with the
line in Figure 4.6b (no adjustable parameters) confirms the accuracy of Equation 4.85a.
The case of adsorption at an interface that is subjected to stationary expansion needs a special
theoretical description. This case is experimentally realized with the strip method [95,116], and the
overflowing cylinder method [60,92]. It could be realized also by a Langmuir trough. The interfa-
cial expansion is characterized by the quantity α = dA /(Adt ), which represents the relative rate of
1 mM
65
2 mM
60
3 mM
55
5 mM
50 7 mM
45 10 mM
12 mM
40
10 100 1000 10000
Surface age, tage (ms)
(a)
4
Equilibrium adsorption, Г1,e (µmol/m2)
C12TAB + 5 mM NaBr
1
0
0 2 4 6 8 10 12
(b) C12TAB concentration (mM)
FIGURE 4.6 (a) Data for the dynamic surface tension, σ, vs. the surface age, tage, measured by MBPM [57]
at concentrations of C12TAB denoted in the figure; the solid lines are fits (see the text). (b) Dependence of the
equilibrium surfactant adsorption, Γ1,e, on the C12TAB concentration The points are calculated by means of
Equation 4.85a for Sl determined from the fits in Figure 4.6a. The solid lines are calculated independently
from fits of surface tension data, σe vs. c1∞, by means of the van der Waals adsorption model. (From Danov,
K.D. et al., J. Colloid Interface Sci., 303, 56, 2006.)
274 Handbook of Surface and Colloid Chemistry
1/ 2
⎛ πα ⎞ 1 ⎛ 1 1 ⎞
σ = σe + kTΓ12,e ⎜ ⎟ ⎜c +c ⎟ (4.85b)
⎝ 2 Deff ⎠ γ± ⎝ 1∞ 2∞ ⎠
where Deff is given by Equations 6.19 through 6.26 in Ref. [114]. Equation 4.85b does not contain the
time, t, as it should be for a stationary process. For nonionic surfactants and for ionic surfactants at
high salt concentrations the term 1/c2∞ in Equation 4.85b disappears and Deff = D1.
dΓ1
= Q ≡ rads (c1, Γ1 ) − rdes (Γ1 ) (4.86)
dt
where rads and rdes are the rates of surfactant adsorption and desorption, respectively. The concept
of barrier-limited adsorption originates from the works of Bond and Puls [118], and Doss [119], and
has been further developed by other authors [120–127]. Table 4.6 summarizes some expressions for
the total rate of adsorption under barrier control, Q. The quantities Kads and Kdes in Table 4.6 are
TABLE 4.6
Rate of Surfactant Adsorption for Different Kinetic Models
Rate of Reversible Adsorption
Type of Isotherm Q = rads(c1, Γ1)−rdes(Γ1)
Γ1
Henry Q = K ads c1 − K des
Γ∞
Γ1
Freundlich Q = K ads K m−1c1m − K des
Γ∞
⎛ Γ ⎞ Γ
Langmuir Q = K ads c1 ⎜ 1 − 1 ⎟ − K des 1
⎝ Γ∞ ⎠ Γ∞
⎛ Γ ⎞ Γ ⎛ 2β Γ1 ⎞
Frumkin Q = K ads c1 ⎜ 1 − 1 ⎟ − K des 1 exp ⎜ − ⎟
⎝ Γ∞ ⎠ Γ∞ ⎝ kT ⎠
Γ1 ⎛ Γ1 ⎞
Volmer Q = K ads c1 − K des exp ⎜ ⎟
Γ ∞ − Γ1 ⎝ Γ ∞ − Γ1 ⎠
Γ1 ⎛ Γ1 2β Γ1 ⎞
van der Waals Q = K ads c1 − K des exp ⎜ − ⎟
Γ ∞ − Γ1 ⎝ Γ ∞ − Γ1 kT ⎠
Chemical Physics of Colloid Systems and Interfaces 275
the rate constants of adsorption and desorption, respectively. Their ratio is equal to the equilibrium
constant of adsorption
K ads
= K (4.87)
K des
The parameters Γ∞ and K are the same as in Tables 4.2 through 4.4. Setting Q = 0 (assuming equi-
librium at surface–subsurface), from each expression in Table 4.6 we deduce the respective equi-
librium adsorption isotherm in Table 4.2. In addition, for β = 0 the expressions for Q related to the
Frumkin and van der Waals model reduce, respectively, to the expressions for Q in the Langmuir
and Volmer models. For Γ1 ≪ Γ∞ both the Frumkin and Langmuir expressions in Table 4.6 reduce
to the Henry expression.
Substituting Q from Table 4.6 into Equation 4.86, and integrating, we can derive explicit expres-
sions for the relaxation of surfactant adsorption:
σ(t ) − σe Γ1 (t ) − Γ1,e ⎛ t ⎞
≈ = exp ⎜ − ⎟ (4.88)
σ(0) − σe Γ1 (0) − Γ1,e ⎝ τσ ⎠
Equation 4.88 holds for σ(t) only in the case of small deviations from equilibrium, whereas there is
not such a restriction concerning Γ1(t); the relaxation time in Equation 4.88 is given by the expressions
−1
⎛K ⎞
τσ = ⎜ des ⎟ (Henry and Freundlich) (4.89)
⎝ Γ∞ ⎠
−1
⎛K K c ⎞
τσ = ⎜ des + ads 1 ⎟ (Langmuir ) (4.90)
Γ
⎝ ∞ Γ∞ ⎠
Equation 4.88 predicts that the perturbation of surface tension, Δσ(t) = σ(t) − σe, relaxes exponentially.
This is an important difference with the cases of adsorption under diffusion and electro-diffusion
control, for which Δσ(t ) ∝ 1/ t , cf. Equations 4.70, 4.76, and 4.78. Thus, a test to check whether
or not the adsorption occurs under purely barrier control is to plot data for ln[Δσ(t)] versus t and to
check if the plot complies with a straight line.
In the case of ionic surfactants, the adsorption of surfactant ions is accompanied by binding of
counterions. In addition, the concentrations of the ionic species vary across the EDL (even at equi-
librium). These effects are taken into account in Equation 4.57, which can be used as an expression
for Q in the case of Langmuir barrier adsorption of an ionic surfactant.
In fact, a pure barrier regime of adsorption is not frequently observed. It is expected that the bar-
rier becomes more important for substances of low surface activity and high concentration in the
solution. Such adsorption regime was observed with propanol, pentanol, and 1,6 hexanoic acid [95],
as well as with proteins [128].
It may happen that the characteristic times of diffusion and transfer across the barrier are
comparable. In such case we deal with mixed kinetic regime of adsorption [129]. Insofar as the
stages of diffusion and transfer are consecutive, the boundary conditions at the interface are
dΓ1 ⎛ ∂c ⎞
= rads (c1, Γ1 ) − rdes (Γ1 ) = D1 ⎜ 1 ⎟ (4.91)
dt ⎝ ∂z ⎠ z =0
276 Handbook of Surface and Colloid Chemistry
The formal transition in Equation 4.91 from mixed to diffusion control of adsorption is not trivial
and demands application of scaling and asymptotic expansions. The criterion for occurrence of
adsorption under diffusion control (presence of equilibrium between subsurface and surface) is
aK des ⎛ ∂Γ1 ⎞
1 (4.92)
D1 ⎜⎝ ∂c1 ⎟⎠e
Equation 4.93 reflects the fact that in diffusion regime the surface is always assumed to be equili-
brated with the subsurface. In particular, if Γ1 = 0, then we must have c1s = 0. In contrast, Equation
4.94 stems from the presence of barrier: for time intervals shorter than the characteristic time of
transfer, the removal of the surfactant from the interface (Γ1 = 0) cannot affect the subsurface layer
(because of the barrier) and then c1s(0) = c1∞. This purely theoretical consideration implies that
the effect of barrier could show up at the short times of adsorption, whereas at the long times the
adsorption will occur under diffusion control [129,130]. The existence of barrier-affected adsorption
regime at the short adsorption times could be confirmed or rejected by means of the fastest methods
for measurement of dynamic surface tension.
Release of Adsorption
monomers Subsurface
layer
Diffusion Diffusion
Micelles convection convection
Release of
Bulk
monomers
Assembly
z Monomers
FIGURE 4.7 In the neighborhood of an expanded (nonequilibrium) adsorption monolayer, the micelles (the
aggregates) release monomers to restore the equilibrium concentrations of surfactant monomers at the surface
and in the bulk. The concentration gradients give rise to diffusion of both monomers and micelles. As a rule,
the adsorbing component are the surfactant monomers, whereas the micelles are repelled by the interface and
do not adsorb.
When surfactant molecules adsorb at an interface, the concentration of monomers in the subsur-
face layer decreases, which leads to release of monomers from the neighboring micelles, or to their
complete decomposition. The decrease in the concentrations of monomers and micelles gives rise to
corresponding diffusion fluxes from the bulk of solution toward the subsurface layer (Figure 4.7). In
general, the role of the micelles as sources and carriers of monomers leads to a marked acceleration
of surfactant adsorption.
The first models of micellar kinetics in spatially uniform solutions have been developed by
Kresheck et al. [140] and Aniansson and Wall [141]. The existence of “fast” and “slow” processes
of the micellar dynamics has been established. The fast process represents exchange of separate
monomers between micelles and the surrounding solution. If the micelle releases monomers, its
aggregation number could decrease to a critical value, after which a complete decomposition of
the micelle to monomers takes place. This decomposition is known as the slow demicellization
process [141].
The first theoretical model of surfactant adsorption from micellar solutions, proposed by
Lucassen [142], uses the simplifying assumptions that the micelles are monodisperse and that
the micellization happens as a single step, which is described as a reversible reaction of order n (the
micelle aggregation number). Later, more realistic models, which account for the multi-step charac-
ter of the micellar process, were developed [143–145]. The assumption for a complete local dynamic
equilibrium between monomers and micelles makes possible to use the equilibrium mass action
law for the micellization reaction [142,146,147]. In such a case, the surfactant transfer corresponds
to a conventional diffusion-limited adsorption characterized by an effective diffusion coefficient,
Deff, which depends on the micelle diffusivity, concentration, and aggregation number. Deff is inde-
pendent of the rate constants of the fast and slow demicellization processes: km and kS. Joos et al.
[146,147] confirmed experimentally that in some cases the adsorption from micellar solutions could
be actually described as a diffusion-limited process characterized by an apparent diffusivity, Deff.
In other experiments, Joos et al. [95,148] established that sometimes the dynamics of adsorption
from micellar solutions exhibits a completely different kinetic pattern: the interfacial relaxation
is exponential, rather than inverse square root, as it should be for diffusion-limited kinetics.
278 Handbook of Surface and Colloid Chemistry
The theoretical developments [95,129,148] revealed that the exponential relaxation is influenced
by the kinetics of micellization, and from the data analysis we could determine the rate constant
of the fast process, km. The observation of different kinetic regimes for different surfactants and/or
experimental methods makes the physical picture rather complicated.
A realistic model of the micellar kinetics was proposed [149] and applied to investigate the
dynamics of adsorption at quiescent [150] and expanding [57,151] interfaces. The theoretical analy-
sis reveals the existence of four different consecutive relaxation regimes (stages) for a given micellar
solution: two exponential regimes and two inverse-square-root regimes, following one after another
in alternating order. The results of these studies are briefly described in the following section, and
the agreement between theory and experiment is illustrated.
ha h hc
ξ1 ≡ c1, p ; ξc ≡ a Cm, p ; ξm ≡ 2a 1,eq m p (4.95)
Γ p,0 β Γ p,0 seq Γ p,0
Here
c1,p, Cm,p, and mp are, respectively, the perturbations in the monomer concentration, c1, micelle
concentration, Cm, the micelle mean aggregation number, m, the respective dimensionless
perturbations are ξ1, ξc, and ξm;
Γp,0 is the perturbation in the surfactant adsorption at the initial moment (t = 0);
seq is the halfwidth of the equilibrium micelle size distribution modeled by a Gaussian bell-like
curve;
β and ha are, respectively, the dimensionless bulk micelle concentration and the characteristic
adsorption length, defined as follows:
Ctot − CMC ⎛ dΓ ⎞
β≡ ; ha = ⎜ ⎟ (4.96)
CMC ⎝ dc1 ⎠eq
where
Ctot is the total surfactant concentration
Γ is the surfactant adsorption
The dimensionless fluxes of the fast and slow demicellization processes, denoted by ϕm and ϕs,
respectively, can be expressed as follows [150]:
ϕm = ξ1 − ξm (4.97)
(Some small terms are neglected in Equations 4.97 and 4.98.) Here meq is the equilibrium micelle
aggregation number w = (meq – nr)/seq, where nr is an aggregation number at the boundary between
the regions of the rare aggregates and the abundant micelles [150]
Chemical Physics of Colloid Systems and Interfaces 279
A 1,0
100 exp(–τ/τF)
FIGURE 4.8 Time dependence of the perturbations in the subsurface monomer concentration, ξ1,0, micelle
concentration, ξc,0, and mean aggregation number, ξm,0, for β = 100. The curves are obtained by numerical
solution of the general system of equations. (From Danov, K.D. et al., Adv. Colloid Interface Sci., 119, 17,
2006.)
Figure 4.8 shows results obtained by solving numerically the general system of equations in Ref.
[150] for a relatively high micelle concentration, β = 100. The calculated curves ξ1,0(τ), ξc,0(τ), and
ξm,0(τ) represent the subsurface values (at z = 0, Figure 4.7) of the perturbations ξ1, ξc, and ξm, plotted
versus the dimensionless time, τ = ( D1 /ha2 )t, where D1 is the diffusion coefficient of the surfactant
monomers. Note that ξ1,0 expresses not only the perturbation in the subsurface monomer concentra-
tion, but also the perturbations in the surface tension and adsorption [150]:
σ(t ) − σe Γ(t ) − Γ e
= = ξ1,0 (τ) (4.99)
σ(0) − σe Γ(0) − Γ e
where
σ(t) and Γ(t) are the dynamic surface tension and adsorption, respectively
σ(0) and Γ(0) are their initial values
σe and Γe are their final equilibrium values
A typical value, km /kS = 107, of the ratio of the rate constants of the fast and slow demicellization
processes is used to calculate the curves in Figure 4.8.
The most important feature of the relaxation curves in Figure 4.8, which represents a kinetic dia-
gram, is that ξm,0 merges with ξ1,0 at a given point, denoted by B, while ξc,0 merges with ξ1,0 (and ξm,0)
at another point, denoted by D. The time moments, corresponding to the points B and D, are denoted
by τB and τD, respectively. As seen in Figure 4.8, for τ > τB, we have ξ1,0 = ξm,0. In view of Equation
4.97, this means that for τ > τB the flux of the fast micelle relaxation process, ϕm is equal to zero. In
other words, for τ > τB the monomers and micelles are equilibrated with respect to the fast micellar
process. For a regular relaxation process, the theoretical analysis [150] yields the expression τB =
seqha(2km /D1)1/2. In addition, for τ > τD we have ξc,0 = ξ1,0 = ξm,0, and then Equation 4.98 indicates that
ϕs = 0, that is, the monomers and micelles are equilibrated with respect to the slow micellar process.
The computer modeling [150] shows that ξ1,0(τ) exhibits two exponential (kinetic) regimes, AB
and CD, and two inverse-square-root (diffusion) regimes, BC and DE, see Figure 4.8. In particular,
the point C corresponds to the moment τC = ( D1 /ha2 )tc ≈ (βD1σ2eq ) / (kS ha2 meq
3
), where tc is the charac-
teristic time of the slow micellar process; see Ref. [149]. τC also serves as a characteristic relaxation
280 Handbook of Surface and Colloid Chemistry
time of adsorption in the kinetic regime CD. The expressions for the other characteristic times,
τF, τBC, and τDE (Figure 4.8) are [150] the following:
meq D1
τF = (regime AB) (4.100)
βkm ha2
1 D
= BC = (1 + uβ)(1 + uβBm ) (regime BC) (4.101)
τBC D1
1 D
= DE = [1 + (u + meq )β][1 + (u + meq )βBm ] (regime DE) (4.102)
τDE D1
Here
D BC and D DE are the effective diffusivities of the micellar solutions in the regimes BC and DE,
2
respectively; u = seq /meq
Bm = Dm /D1; Dm is the mean diffusivity of the micelles
100
80 C12TAB
+ 100 mM NaBr
60
SDS
Deff /D1
+ 100 mM NaCl
40
20
0
0 5 10 15 20 25
β = (Ctot – CMC)/CMC
FIGURE 4.9 Plot of the dimensionless effective diffusivity of the micellar solution, Deff/D1, vs. the dimen-
sionless micelle concentration, β, obtained from dynamic surface tension values measured by the maximum
bubble pressure method (MBPM) (From Christov, N.C. et al., Langmuir, 22, 7528, 2006); D1 is the diffusivity
of the surfactant monomers. The lines are guides to the eye.
Chemical Physics of Colloid Systems and Interfaces 281
kinetic regime cannot be DE. On the other hand, a similar estimate of D BC/D1 from Equation 4.101
gives reasonable values. To demonstrate that, from the experimental values of Deff/D1 in Figure 4.9
we calculated u by means of Equation 4.101, substituting Bm = 0.2. For most of the concentrations
we obtain values 0.4 < u < 2, which seem reasonable. Values u > 2 are obtained at β < 2, which
indicate that at the lowest micellar concentrations we are dealing with a rudimentary kinetic regime
[150,151], rather than with the diffusion regime BC.
A
100 β = 100 1,T
km/ks = 107
1,T , m,T , c,T
10–1
B
C
10–2
Perturbations
m,T
10–3
E D
10–4
c,T
10–5
10–7 10–6 10–5 10–4 10–3 10–2 10–1 101 100 102 103 104 105
Dimensionless rate of surface expansion, θ
FIGURE 4.10 Total perturbations in monomer concentration, ξ1,T, micelle concentration, ξc,T, and mean
aggregation number, ξm,T, plotted vs. the dimensionless rate of surface expansion, θ, at micelle dimension-
less concentration β = 100. The curves are obtained by numerical solution of the linear system of equations
derived. (From Danov, K.D. et al., Colloids Surf. A, 282–283, 143, 2006.)
282 Handbook of Surface and Colloid Chemistry
fast micellar process is equilibrated, while the slow process is negligible. In regime CD, the slow
micellar process governs the adsorption kinetics. In regime DE, the adsorption occurs under dif-
fusion control, because both the fast and slow micellar processes are equilibrated. Note that only
the regimes BC and DE correspond to purely diffusion processes. For the regimes AB and CD, the
rate constants of the fast and slow micellar processes, km and kS, respectively, affect the surfactant
adsorption kinetics, and could be in principle determined from the fit of experimental data. For the
specific experimental examples considered here, the adsorption kinetics corresponds to the diffu-
sion regime BC.
1.
Solubilization as a bulk reaction: Molecular dissolution and diffusion of oil into the aque-
ous phase takes place, with a subsequent uptake of oil molecules by surfactant micelles
[156–161]. This mechanism is operative for oils (like benzene, hexane, etc.), which exhibit
a sufficiently high solubility in pure water. Theoretical models have been developed and
verified against the experiment [157,159–161]. The bulk solubilization includes the fol-
lowing processes. First, oil molecules are dissolved from the surface of an oil drop into
water. Kinetically, this process can be characterized by a mass transfer coefficient. Next,
by molecular diffusion, the oil molecules penetrate in the water phase, where they react
with the micelles. Thus, the concentration of free oil molecules diminishes with the dis-
tance from the oil–water interface. In other words, solubilization takes place in a certain
zone around the droplet [159,160].
2.
Solubilization as a surface reaction: This is the major solubilization mechanism for oils
that are practically insoluble in water [156,158,160,162–170]. The uptake of such oils can-
not happen in the bulk of the aqueous phase. The solubilization can be realized only at
the oil–water interface. The mechanism may include (1) micelle adsorption, (2) uptake of
oil, and (3) desorption of the swollen micelles [168–170]. Correspondingly, the theoreti-
cal description of the process involves the rate constants of the three consecutive steps.
If the empty micelles are long rodlike aggregates, upon solubilization they usually break
to smaller spherical aggregates [168,171]. For some systems (mostly solid solubilizates),
the intermediate stages in the solubilization process may involve penetration of surfactant
solution into the oily phase and formation of a liquid crystalline phase at the interface
[172–176].
In the case of solubilization as surface reaction, the detailed kinetic mechanism could be multiform.
Some authors [156,163] expect that the surfactant arrives at the interface in a monomeric form.
Then, at the phase boundary mixed (or swollen) micellar aggregates are formed, which eventually
Chemical Physics of Colloid Systems and Interfaces 283
desorb. This version of the model seems appropriate for solid solubilizates because hemimicelles
can be formed at their surfaces, even at surfactant concentrations below the bulk CMC [177].
Another concept, presented by Plucinski and Nitsch [165], includes a step of partial fusion of the
micelles with the oil–water interface, followed by a step of separation. Such mechanism could take
place in the case when microemulsion drops, rather than micelles, are responsible for the occur-
rence of solubilization.
Experiments with various surfactant systems [166,170,178] showed that the solubilization rates
for solutions of ionic surfactants are generally much lower than those for nonionic surfactants. This
can be attributed to the electrostatic repulsion between the micelles and the similarly charged sur-
factant adsorption monolayer at the oil–water interface. On the other hand, copolymers have been
found to form micelles, which solubilize various hydrophobic compounds well, even in the absence
of low-molecular-weight surfactants [179–187]. Moreover, appropriately chosen copolymers can act
as very efficient promoters of solubilization [160,168–170].
2Hσ = ΔP (4.103)
Here
H is the local mean curvature of the interface
ΔP is the local jump of the pressure across the interface
If z = z(x, y) is the equation of the interface in Cartesian coordinates, then H can be expressed in
the form [191]
⎡ ⎤
⎢ ∇s z ⎥
2 H = ∇s ⋅ ⎢ 1/ 2 ⎥
(4.104)
⎣(
⎢ 1 + ∇s z
2
) ⎥
⎦
where ∇s is the gradient operator in the plane xy. More general expressions for H can be found in the
literature on differential geometry [191–193]. Equation 4.103, along with Equation 4.104, represents
a second-order partial differential equation which determines the shape of the fluid interface. The
interface is bounded by a three-phase contact line at which the boundary conditions for the dif-
ferential equation are formulated. The latter are the respective necessary conditions for mechanical
equilibrium at the contact lines. When one of the three phases is solid (Figure 4.11a), the boundary
condition takes the form of Young [194] equation:
where
α is the three-phase contact angle
σ12 is the tension of the interface between the fluid phases 1 and 2
σ1s and σ2s are the tensions of the two fluid–solid interfaces
284 Handbook of Surface and Colloid Chemistry
(Solid)
α
(1) (2)
σ
(a)
Ψc
(1)
(2)
σ12
(b)
FIGURE 4.11 Sketch of fluid particle (1) attached to the interface between phases (2) and (3). (a) Fluid
particle attached to solid interface; α is the contact angle; σ is the interfacial tension of the boundary between
the two fluid phases. (b) Fluid particle attached to a fluid interface; σ12, σ13, and σ23 are the interfacial tensions
between the respective phases; ψc is the slope angle of the outer meniscus at the contact line.
Insofar as the values of the three σ’s are determined by the intermolecular forces, contact angle α
is the material characteristics of a given three-phase system. However, when the solid is not smooth
and chemically homogeneous, then the contact angle exhibits hysteresis, that is, α has no defined
equilibrium value [6,195]. Contact angle hysteresis can be observed even with molecularly smooth
and homogeneous interfaces under dynamic conditions [196].
When all the three neighboring phases are fluids, then the boundary condition takes the form of
the Neumann [197] vectorial triangle:
(see Figure 4.11b); here vik is a unit vector, which is simultaneously normal to the contact line and
tangential to the boundary between phases i and k. The Laplace, Young, and Neumann equations
can be derived as conditions for minimum of the free energy of the system [37,191,198]; the effect of
the line tension can also be taken into account in Equations 4.105 and 4.106 [198].
In the special case of spherical interface H = 1/R, with R being the sphere radius, and Equation
4.103 takes its most popular form, 2σ/R = ΔP. In the case of axisymmetric meniscus (z-is the axis
of symmetry, Figure 4.11), the Laplace equation reduces to either of the following two equivalent
forms [190,199]:
1 d ⎡ rzʹ ⎤ ΔP
= , z = z(r ) (4.107)
r dr ⎢⎣ (1 + zʹ2 )1/ 2 ⎥⎦ σ
r ʹʹ 1 ΔP
− + = , r = r ( z) (4.108)
(1 + r ʹ )
2 3/2
r (1 + rʹ )
2 1/ 2
σ
Chemical Physics of Colloid Systems and Interfaces 285
Two equivalent parametric forms of Laplace equation are often used for calculations [190,199]:
d sin ϕ sin ϕ ΔP dz
+ = , tan ϕ = (4.109)
dr r σ dr
or
dϕ ΔP sin ϕ dr dz
= − , = cos ϕ, = sin ϕ (4.110)
ds σ r ds ds
Here
φ is the meniscus running slope angle (Figure 4.11a)
s is the arc length along the generatrix of the meniscus
Equation 4.110 is especially convenient for numerical integration, whereas Equation 4.109 may
create numerical problems at the points with tan φ = ±∞, like the particle equator in Figure 4.11a.
A generalized form of Equation 4.109, with account for the interfacial (membrane) bending elastic
modulus, kc,
⎛ d sin ϕ sin ϕ ⎞ kc d ⎧ d ⎡1 d ⎤⎫
σ⎜
dr
+
r ⎟ = ΔP + r cos ϕ dr ⎨r cos ϕ dr ⎢ r dr (r sin ϕ) ⎥ ⎬ (4.111)
⎝ ⎠ ⎩ ⎣ ⎦⎭
serves for description of the axisymmetric configurations of real and model cell membranes
[37,200,201]. The Laplace equation can be generalized to also account for the interfacial bending
moment (spontaneous curvature), shear elasticity, etc.; for review, see Refs. [37,200]. The latter
effects are physically important for systems or phenomena like capillary waves [202], phospholipid
and protein membranes [203–206], emulsions [207], and microemulsions [208].
h III II
Ψc hc
0 R2
2R1
FIGURE 4.12 Capillary menisci formed around two coaxial cylinders of radii R1 and R2. (I) Meniscus
meeting the axis of revolution; (II) meniscus decaying at infinity; (III) meniscus confined between the two
cylinders. h denotes the capillary raise of the liquid in the inner cylinder; hc is the elevation of meniscus II at
the contact line r = R2.
286 Handbook of Surface and Colloid Chemistry
Δρgb2
1 (4.112)
σ
Here
g is the gravity acceleration
Δρ is the difference in the mass densities of the lower and the upper fluid
b is a characteristic radius of the meniscus curvature
For example, if Equation 4.112 is satisfied with b = R1 (see Figure 4.12), the raise, h, of the liquid in
the capillary is determined by means of the equation [6]
2σcosα
h= (4.113)
ΔρgR1
When the gravity effect is not negligible, the capillary pressure, ΔP, becomes dependent on the
z-coordinate:
2σ
ΔP = + Δρgz (4.114)
b
Here b is the radius of curvature at the particle apex, where the two principal curvatures are equal
(e.g., the bottom of the bubble in Figure 4.11a). Unfortunately, Equation 4.107, along with Equation
4.114, has no closed analytical solution. The meniscus shape can be exactly determined by numeri-
cal integration of Equation 4.110. Alternatively, various approximate expressions are available
[199,209,210]. For example, if the meniscus slope is small, z2 1, Equation 4.107 reduces to a
linear differential equation of Bessel type, whose solution reads
1/2
2[I 0 (qr ) − 1] ⎛ Δρg ⎞
z(r ) = q≡⎜ ⎟ (4.115)
bq 2 ⎝ σ ⎠
where I0(x) is the modified Bessel function of the first kind and zeroth order [211,212]. Equation
4.115 describes the shape of the lower surface of the lens in Figure 4.11b; similar expression can also
be derived for the upper lens surface.
ΔP = Δρgz (4.116)
As mentioned earlier, Equation 4.107, along with Equation 4.116, has no closed analytical solu-
tion. On the other hand, the region far from the contact line has always a small slope, z2 1.
Chemical Physics of Colloid Systems and Interfaces 287
In this region Equation 4.107 can be linearized, and then in analogy with Equation 4.115
we derive
where
A is a constant of integration
K0(x) is the modified Bessel function of the second kind and zeroth order [211,212]
The numerical integration of Equation 4.110 can be carried out by using the boundary condition
[199] z′/z = −qK1(qr)/K0(qr) for some appropriately fixed r ≪ q−1 (see Equation 4.117). Alternatively,
approximate analytical solutions of the problem are available [199,210,213]. In particular, Derjaguin
[214] derived an asymptotic formula for the elevation of the contact line at the outer surface of a
thin cylinder,
⎡ qR γ (1 + cos ψ c ) ⎤
hc = − R1 sin ψ c ln ⎢ 1 e † ⎥ , (qR1 ) 1 (4.118)
2
⎣ 4 ⎦
where
R1 is the radius of the contact line
ψc is the meniscus slope angle at the contact line (Figure 4.12)
q is defined by Equation 4.115
γe = 1.781072418… is the constant of Euler–Masceroni [212]
4.3.1.2.3 Meniscus Confined between Two Cylinders (0 < R1 < r < R2 < ∞)
This is the case with the Plateau borders in real foams and emulsions, and with the model films in the
Scheludko–Exerowa cell [215,216]; such is the configuration of the capillary bridges (Figure 4.13a)
and of the fluid particles pressed between two surfaces (Figure 4.13b). When the gravitational defor-
mation of the meniscus cannot be neglected, the interfacial shape can be determined by numeri-
cal integration of Equation 4.110, or by iteration procedure [217]. When the meniscus deformation
caused by gravity is negligible, analytical solution can be found as described in the following section.
z z
r r
0 r0 rc 0 rc r0
P2 P1 P2 P1
(a) (b)
FIGURE 4.13 Concave (a) and convex (b) capillary bridges between two parallel plates. P1 and P2 denote
the pressures inside and outside the capillary bridge, r 0 is the radius of its section with the midplane; rc is the
radius of the three-phase contact lines.
288 Handbook of Surface and Colloid Chemistry
z z
(nodoid) (unduloid)
r r
0 ra rb 0 ra rb
(a) (b)
FIGURE 4.14 Typical shape of nodoid (a) and unduloid (b) plateau curves. Note that the curves are confined
between two cylinders of radii ra and rb.
To determine the shape of the menisci depicted in Figure 4.13a and b, we integrate Equation
4.109 from r0 to r to derive
dz
=
(
k1 r 2 − r02 + r0 ) P −P
, k1 ≡ 1 2 , r1 ≡
1 − k1r0
(4.119)
dr ± ⎡ r 2 − r 2 r 2 − r 2 ⎤1/ 2 k 2σ k1
⎣ ( 0 1 )(⎦ 1 )
The pressures in phases 1 and 2, P1 and P2, and r0 are shown in Figure 4.13. Equation 4.119
describes curves, which after Plateau [189,190,218–220] are called “nodoid” and “unduloid” (see
Figure 4.14). The nodoid (unlike the unduloid) has points with horizontal tangent, where dz/dr = 0.
Then with the help of Equation 4.119, we can deduce that the meniscus generatrix is a part of nodoid
if k1r0 ∈ (−∞, 0) ∪ (1, +∞), while the meniscus generatrix is a part of unduloid if k1r0 ∈ (0, 1).
In the special case, when k1r0 = 1, the meniscus is spherical. In the other special case, k1r0 = 0,
the meniscus has the shape of catenoid, that is,
⎡ 2 ⎤
r ⎛r⎞
z = ±r0 ln ⎢ + ⎜ ⎟ − 1 ⎥ , (k1 = 0) (4.120)
⎢ r0 ⎝ r0 ⎠ ⎥
⎣ ⎦
The meniscus is concave and has a “neck” (Figure 4.13a) when k1r0 ∈ (−∞, 1/2); in particular, the
generatrix is nodoid for k1r0 ∈ (−∞, 0), catenoid for k1r0 = 0, and unduloid for k1r0 ∈ (0, 1/2). For the
configuration depicted in Figure 4.13a, we have r1 > r0 (in Figure 4.14 ra = r0, rb = r1) and Equation
4.119 can be integrated to yield (see tables of integrals):
⎧ ⎡ 1 ⎤⎫
z(r ) = ± ⎨r0 F (φ1, q1 ) + r1 sgn k1 ⎢ E (φ1, q1 ) − (r 2 − r02 )(r12 − r 2 ) ⎥ ⎬ (r0 ≤ r ≤ r1 ) (4.121)
⎩ ⎣ rr1 ⎦⎭
where
sgnx denotes the sign of x
q1 = (1 − r02 /r12 )1/ 2
sin φ1 = q1−1 (1 − r02 /r 2 )1/ 2
F(ϕ, q) and E(ϕ, q) are the standard symbols for elliptic integrals of the first and the second kind
[211,212]
Chemical Physics of Colloid Systems and Interfaces 289
A convenient method for computation of F(ϕ, q) and E(ϕ, q) is the method of the arithmetic–
geometric mean (see Ref. [211], Chapter 17.6).
The meniscus is convex (Figure 4.13b) when k1r0 ∈ (1/2, +∞); in particular, the generatrix is
unduloid for k1r0 ∈ (1/2, 1), circumference for k1r0 = 1, and nodoid for k1r0 ∈ (1, +∞). For the con-
figuration depicted in Figure 4.13b, we have r0 > r1 (in Figure 4.14 ra = r1, rb = r0) and Equation 4.119
can be integrated to yield (see tables of integrals):
⎡⎛ 1⎞ ⎤
z(r ) = ∓ ⎢⎜ r0 − ⎟ F (φ2 , q2 ) − r0 E (φ2 , q2 ) ⎥ , (r1 ≤ r ≤ r0 ) (4.122)
⎣⎝ k1 ⎠ ⎦
where
q2 = (1 − r12 /r02 )1/ 2
sinφ2 = q2−1 (1 − r 2 /r02 )1/ 2
Additional information about the shapes, stability, and nucleation of capillary bridges, and for the
capillary-bridge forces between particles, can be found in Chapter 11 of Ref. [37].
Small capillary bridges, called “pendular rings” [221], give rise to cohesion between the particles
in the wet sand and to adhesion of particles to a flat plate [222]. In their study on the enhancement of
rheology of three-phase (solid–oil–water) dispersions, Koos and Willenbacher [223] identified two
states with different structures: (1) the pendular state, where the solid particles are interconnected
with concave capillary bridges, and (2) the capillary state, where the particles are interconnected
by convex capillary bridges; see Figure 4.13. In the former case, the bridging fluid wets the particles
well (contact angle θ < 90°), whereas in the latter case the bridging fluid does not wet the particles
well (contact angle θ > 90°). However in both cases the capillary bridging phenomenon leads to a
considerable enhancement of the rheological response of the dispersion at a minor volume fraction
of the second fluid [223].
2σ
(μ ) − (μ )
β
i
R
β
i
R =∞
( ) − (μ )
= μiα
R
α
i
R =∞
= Vi α
R
(4.123)
where
μ is chemical potential
Vi is partial volume
the superscripts denote phase
the subscripts denote component
Equation 4.123 is derived under the following assumptions. When β is a gaseous phase, it is assumed
that the partial volume of each component in the gas is much larger than its partial volume in the liq-
uid α; this is fulfilled far enough from the critical point [227]. When phase β is liquid, it is assumed
that Pβ(R) = Pβ(R = ∞), where P denotes pressure.
290 Handbook of Surface and Colloid Chemistry
Piβ ( R) ⎛ 2σVi α ⎞
= exp ⎜ ⎟ (4.124)
Pi (∞)
β
⎝ RkT ⎠
where Piβ ( R) and Piβ (∞) denote, respectively, the equilibrium vapor pressure of component i in the
droplet of radius R and in a large liquid phase of the same composition. Equation 4.124 shows that
the equilibrium vapor pressure of a droplet increases with the decrease of the droplet size. (For a
bubble, instead of a droplet, R must be changed to −R in the right-hand side of Equation 4.124
and the tendency becomes the opposite.) Equation 4.124 implies that in an aerosol of polydisperse
droplets the larger droplets will grow and the smaller droplets will diminish down to complete
disappearance.
The small droplets are “protected” against disappearance when phase α contains a nonvolatile
component. Then instead of Equation 4.124 we have
Piβ ( R) 1 − X ( R) ⎛ 2σVi α ⎞
= exp ⎜ ⎟ (4.125)
Pi (∞) 1 − X (∞)
β
⎝ RkT ⎠
where X denotes the molar fraction of the nonvolatile component in phase α; for X(R) = X(∞)
Equation 4.125 reduces to Equation 4.124. Setting the left-hand side of Equation 4.125 equal to 1,
we can determine the value X(R) needed for a liquid droplet of radius R, surrounded by the gas phase
β, to coexist at equilibrium with a large (R = ∞) liquid phase α of composition X(∞).
When both phases α and β are liquid, Equation 4.123 yields
Xiβ ( R) ⎛ 2σVi α ⎞
= exp ⎜ ⎟ (4.126)
Xiβ (∞) ⎝ RkT ⎠
where
Xiβ ( R) denotes the equilibrium molar fraction of component i in phase β coexisting with a droplet
of radius R
Xiβ (∞) denotes the value of Xiβ ( R) for R → ∞, that is, for phase β coexisting with a large phase α
of the same composition as the droplet
In the case of oil-in-water emulsion, Xiβ can be the concentration of the oil dissolved in the water. In
particular, Equation 4.126 predicts that the large emulsion droplets will grow and the small droplets
will diminish. This phenomenon is called Ostwald ripening (see Section 4.3.1.4). If the droplets
(phase α) contain a component, which is insoluble in phase β, the small droplets will be protected
against complete disappearance; a counterpart of Equation 4.125 can be derived:
Xiβ ( R) 1 − X ( R) ⎛ 2σVi α ⎞
= exp ⎜ ⎟ (4.127)
X i (∞ ) 1 − X (∞ )
β
⎝ RkT ⎠
where X denotes the equilibrium concentration in phase α of the component which is insoluble in
phase β. Setting the left-hand side of Equation 4.127 equal to 1, we can determine the value X(R)
needed for an emulsion droplet of radius R, surrounded by the continuous phase β, to coexist at
equilibrium with a large (R = ∞) liquid phase α of composition X(∞).
Chemical Physics of Colloid Systems and Interfaces 291
4π d 3
R (t ) = 4πDRV1[cm (t ) − ceq ( R)] (4.128)
3 dt
where
t is time
D is the diffusivity of component 1 in the continuous phase
V1 is the volume per molecule of component 1
cm is the number-volume concentration of component 1 in the continuous medium far away from
the droplets surfaces
ceq(R) is the respective equilibrium concentration of the same component for a droplet of radius
R as predicted by the Gibbs–Thomson equation
Note that Equation 4.128 is rigorous only for a diluted emulsion, in which the concentration of
dissolved component 1 levels off at a constant value, c = cm, around the middle of the space between
each two droplets. Some authors [231] also add in the right-hand side of Equation 4.128 terms
accounting for the convective mass transfer (in the case of moving droplets) and thermal contribu-
tion to the growth rate.
Because the theory is usually applied to droplets of diameter not smaller than micrometer (which
are observable by optical microscope), the Gibbs–Thomson equation, Equation 4.126, can be linear-
ized to yield [233]
⎛ b⎞ 2σV1
ceq ( R) ≈ c∞ ⎜ 1 + ⎟ , b ≡ (4.129)
⎝ R ⎠ kT
with c∞ being the value of ceq for flat interface. With σ = 50 mN/m, V1 = 100 Å3, and T = 25°C we
estimate b = 2.5 nm. The latter value justifies the linearization of Gibbs–Thomson equation for
droplets of micrometer size.
Let f(R, t) be the size distribution function of the emulsion droplets such that f(R, t)dR is the num-
ber of particles per unit volume in the size range from R to (R + dR). The balance of the number of
particles in the system reads
⎛ dR ⎞
df dR = (jdt ) |R −( jdt ) |R + dR , ⎜ j ≡ f (4.130)
⎝ dt ⎟⎠
The term in the left-hand side of Equation 4.130 expresses the change of the number of droplets
whose radius belongs to the interval [R, R + dR] during a time period dt; the two terms in the
292 Handbook of Surface and Colloid Chemistry
right-hand side represent the number of the incoming and outgoing droplets in the size interval
[R, R + dR] during time period dt. Dividing both sides of Equation 4.130 by (dR dt), we obtain the
so-called continuity equation in the space of sizes [229–233]:
∂f ∂j
+ = 0 (4.131)
∂t ∂R
One more equation is needed to determine cm. In a closed system this can be the total mass balance
of component 1:
∞
d ⎡⎢ 4π ⎤
cm (t ) + dR R 3 f ( R, t ) ⎥ = 0 (4.132)
dt ⎢
⎣
∫
3
0
⎥
⎦
The first and the second terms in the brackets express the amount of component 1 contained in
the continuous phase and in the droplets, respectively. This expression is appropriate for diluted
emulsions when cm is not negligible compared to the integral in the brackets.
Alternatively, in opened systems and in concentrated emulsions we can use a mean field approxi-
mation based on Equation 4.129 to obtain the following equation for cm:
⎡
cm (t ) = c∞ ⎢1 +
b ⎤ ∫ dRRf (R, t ) (4.133)
⎥ , Rm (t ) ≡
R0
⎣ Rm (t ) ⎦
∞
∫ dR f (R, t )
R0
⎛ 1 1 ⎞
j ( R, t ) = Q ⎜ − 2 ⎟ f ( R, t ), Q ≡ Dbc∞V1 (4.134)
⎝ RRm R ⎠
In calculations, we use the set of Equations 4.128, 4.131, and 4.132 or 4.133 to determine the
distribution f(R, t) at known distribution f(R, 0) at the initial moment t = 0. In other words, the theory
predicts the evolution of the system at a given initial state. From a computational viewpoint it is
convenient to calculate f(R, t) in a finite interval R0 ≤ R < Rmax (see Figure 4.15). The problem can be
solved numerically by discretization: the interval R0 ≤ R < Rmax is subdivided into small portions of
length δ, the integrals are transformed into sums, and the problem is reduced to solving a linear set
of equations for the unknown functions f k(t) ≡ f(Rk, t), where Rk = R0 + kδ, k = 1, 2, ….
In practice, the emulsions are formed in the presence of surfactants. At concentrations above the
CMC the swollen micelles can serve as carriers of oil between the emulsion droplets of different size.
In other words, surfactant micelles can play the role of mediators of the Ostwald ripening. Micelle-
mediated Ostwald ripening has been observed in solutions of nonionic surfactants [234–236]. In
contrast, it was found that the micelles do not mediate the Ostwald ripening in undecane-in-water
emulsions at the presence of an ionic surfactant (SDS) [237]. It seems that the surface charge due to
Chemical Physics of Colloid Systems and Interfaces 293
f(R, t)
R0 Rmax R
δ
FIGURE 4.15 Sketch of the droplet size distribution function, f(R, t) vs. the droplet radius R at a given
moment t. δ is the length of the mesh used when solving the problem by discretization.
the adsorption of ionic surfactant (and the resulting double layer repulsion) prevents the contact of
micelles with the oil drops, which is a necessary condition for micelle-mediated Ostwald ripening.
P0
σlsin α z(x)
σl P0 h/2 σl
α σf γ
Plateau P α0
l τ Pl + П α0
border α rc r
rc1 –h/2 σl
σl
P0
FIGURE 4.16 The detailed and membrane models of a thin liquid film (on the left- and right-hand side,
respectively).
294 Handbook of Surface and Colloid Chemistry
between two identical fluid particles (drops, bubbles) of internal pressure P0. The more complex
case of nonsymmetrical and curved films is reviewed elsewhere [238–240].
Two different, but supplementary, approaches (models) are used in the macroscopic description
of a thin liquid film. The first of them, the “membrane approach,” treats the film as a membrane of
zero thickness and one tension, γ, acting tangentially to the membrane (see the right-hand side of
Figure 4.16). In the “detailed approach”, the film is modeled as a homogeneous liquid layer of thick-
ness h and surface tension σf. The pressure P0 in the fluid particles is larger than the pressure, Pl, of
the liquid in the Plateau border. The difference
Pc = P0 − Pl (4.135)
represents the capillary pressure of the liquid meniscus. By making the balance of the forces acting
on a plate of unit width along the y-axis and height h placed normally to the film at −h/2 < z < h/2
(Figure 4.16), we derive the Rusanov [241] equation:
γ = 2σ f + Pch (4.136)
Equation 4.136 expresses a condition for equivalence between the membrane and detailed models
with respect to the lateral force. To derive the normal force balance we consider a parcel of unit area
from the film surface in the detailed approach. Because the pressure in the outer phase P0 is larger
than the pressure inside the liquid, Pl, the mechanical equilibrium at the film surface is ensured by
the action of an additional disjoining pressure, Π(h), representing the surface force per unit area of
the film surfaces [242]
Π(h) = P0 − Pl = Pc (4.137)
(see Figure 4.16). Note that Equation 4.137 is satisfied only at equilibrium; at nonequilibrium con-
ditions the viscous force can also contribute to the force balance per unit film area. In general,
the disjoining pressure, Π, depends on the film thickness, h. A typical Π(h)-isotherm is depicted
in Figure 4.17 (for details see Section 4.4). We see that the equilibrium condition, Π = Pc, can be
satisfied at three points shown in Figure 4.17. Point 1 corresponds to a film, which is stabilized by
the double layer repulsion; sometimes such a film is called the “primary film” or “common black
film.” Point 3 corresponds to unstable equilibrium and cannot be observed experimentally. Point 2
corresponds to a very thin film, which is stabilized by the short range repulsion; such a film is called
the “secondary film” or “Newton black film.” Transitions from common to Newton black films are
often observed with foam films [243–246]. Note that Π > 0 means repulsion between the film sur-
faces, whereas Π < 0 corresponds to attraction.
k
dγ = −s f dT − ∑ Γ dμ (4.138)
i =1
i i
where
γ is the film tension
T is temperature
sf is excess entropy per unit area of the film
Γi and μi are the adsorption and the chemical potential of the ith component, respectively
Chemical Physics of Colloid Systems and Interfaces 295
Electrostatic barrier
2 3 1 П = Pc
h2 h1 h
Secondary minimum
Primary minimum
FIGURE 4.17 Sketch of a disjoining pressure isotherm of the DLVO type, Π vs. h. The intersection points
of the Π(h)-isotherm with the line Π = Pc correspond to equilibrium films: h = h1 (primary film), h = h2
(secondary film). Point 3 corresponds to unstable equilibrium.
The Gibbs–Duhem equations of the liquid phase (l) and the outer phase (o) read
k
dPχ = s dT +
χ
ν ∑ n dμ ,
i =1
χ
i i χ = l, o (4.139)
where
sνχ and niχ are entropy and number of molecules per unit volume
Pχ is pressure (χ = l, o)
The combination of Equations 4.127 and 4.131 provides an expression for dPc. Let us multiply this
expression by h and subtract the result from the Gibbs–Duhem equation of the film, Equation 4.138.
The result reads:
k
dγ = −s dT + h dPc − ∑ Γ dμ (4.140)
i =1
i i
where
s = s f + sνo − sνl h,
( ) i = Γi + nio − nil h, i = 1,…, k (4.141)
Γ ( )
An alternative derivation of the same equations is possible [248,249]. Imagine two equidistant
planes separated at a distance h. The volume confined between the two planes is thought to be filled
with the bulk liquid phase (l). Taking surface excesses with respect to the bulk phases we can derive
Equations 4.140 and 4.141 with s and Γ i being the excess surface entropy and adsorption ascribed to
the surfaces of this liquid layer [248,249]. A comparison between Equations 4.138 and 4.140 shows
296 Handbook of Surface and Colloid Chemistry
that there is one additional differential in Equation 4.140. It corresponds to one supplementary
degree of freedom connected with the choice of the parameter h. To specify the model, we need an
additional equation to determine h. For example, let this equation be
1 = 0 (4.142)
Γ
Equation 4.142 requires h to be the thickness of a liquid layer from phase (l), containing the same
amount of component 1 as the real film. This thickness is called the thermodynamic thickness of the
film [249]. It can be on the order of the real film thickness if component 1 is chosen in an appropriate
way, say the solvent in the film phase.
From Equations 4.137, 4.140, and 4.142, we obtain [248]
k
dγ = −s dT + h dΠ − ∑ Γ dμ (4.143)
i =2
i i
A corollary of Equation 4.143 is the Frumkin [250] equation
⎛ ∂γ ⎞
⎜ ∂Π ⎟ = h (4.144)
⎝ ⎠T ,μ2 ,…,μk
Equation 4.144 predicts a rather weak dependence of the film tension γ on the disjoining pressure,
Π, for equilibrium thin films (small h). By means of Equations 4.136 and 4.137, Equation 4.143 can
be transformed to read [249]
k
2 dσ f = −s dT − Π dh − ∑ Γ dμ i i (4.145)
i =2
From Equation 4.145, we can derive the following useful relations [248]
⎛ ∂σ f ⎞
2⎜ ⎟ = −Π (4.146)
⎝ ∂h ⎠T ,μ2 ,…,μk
∞
1
σ (h ) = σ +
f l
Π(h)dh (4.147)
2 ∫
h
with σl being the surface tension of the bulk liquid. Equation 4.147 allows calculation of the film
surface tension when the disjoining pressure isotherm is known.
Note that the thermodynamic equations mentioned earlier are, in fact, corollaries from the
Gibbs–Duhem equation of the membrane approach Equation 4.138. There is an equivalent and
complementary approach, which treats the two film surfaces as separate surface phases with their
own fundamental equations [241,251,252]; thus for a flat symmetric film we postulate
k
dU f = T dS f + 2σ f dA + ∑ μ dN i i
f
− ΠA dh (4.148)
i =1
where
A is area
Uf, S f, and Nif are, respectively, excess internal energy, entropy, and number of molecules
ascribed to the film surfaces
Chemical Physics of Colloid Systems and Interfaces 297
Compared with the fundamental equation of a simple surface phase [5], Equation 4.148 contains
an additional term, ΠAdh, which takes into account the dependence of the film surface energy on
the film thickness. Equation 4.148 provides an alternative thermodynamic definition of disjoining
pressure:
1 ⎛ ∂U f ⎞
Π=− ⎜ ⎟ (4.149)
A ⎝ ∂h ⎠
4.3.2.3 The Transition Zone between Thin Film and Plateau Border
4.3.2.3.1 Macroscopic Description
The thin liquid films formed in foams or emulsions exist in permanent contact with the bulk liquid
in the Plateau border, encircling the film. From a macroscopic viewpoint, the boundary between
film and Plateau border is treated as a three-phase contact line, the line at which the two surfaces
of the Plateau border (the two concave menisci sketched in Figure 4.16) intersect at the plane of the
film (see the right-hand side of Figure 4.16). The angle, α 0, subtended between the two meniscus
surfaces represents the thin film contact angle. The force balance at each point of the contact line is
given by Equation 4.106 with σ12 = γ and σ13 = σ23 = σl. The effect of the line tension, κ, can also be
taken into account. For example, in the case of symmetrical flat film with circular contact line, like
those depicted in Figure 4.16, we can write [252]
κ
γ+ = 2σl cos α 0 (4.150)
rc
dU L = T dS L + 2 κ
dL + ∑ μ dN
i i
L
+ τ dh (4.151)
i
Here
UL , SL , and NiL are linear excesses
is the line tension in the detailed approach
κ
1 ⎛ ∂U L ⎞
τ= ⎜ ⎟ (4.152)
L ⎝ ∂h ⎠
is a 1D counterpart of the disjoining pressure (see Equation 4.149). The quantity τ, called the
transversal tension, takes into account the interaction between the two contact lines. The general
force balance at each point of the contact line can be presented in the form of the following vecto-
rial sum [238]
σ if + σ li + σ iκ + τ i = 0, i = 1, 2 (4.153)
298 Handbook of Surface and Colloid Chemistry
rc2
σf2
σl2 Plateau
border
Fluid 1
FIGURE 4.18 The force balance in each point of the two contact lines representing the boundary between a
spherical film and the Plateau border (see Equation 4.153).
κ
σf + = σl cos α (4.154)
rc1
τ = σl sin α (4.155)
Note that, in general α ≠ α 0 (see Figure 4.16). Besides, both α 0 and α can depend on the radius of
the contact line due to line tension effects. In the case of straight contact line from Equations 4.147
and 4.154, we derive [252]
∞
σf 1
cos α r = 1+ l
c1 =∞
=
σl 2σ ∫ Π(h)dh (4.156)
h
Because cos α ≤ 1, the surface tension of the film must be less than the bulk solution surface tension,
σf < σl, and the integral term in Equation 4.156 must be negative in order for a nonzero contact angle
to be formed. Hence, the contact angle, α, and the transversal tension, τ (see Equation 4.155), are
integral effects of the long-range attractive surface forces acting in the transition zone between the
film and Plateau border, where h > h1 (see Figure 4.17).
In the case of a fluid particle attached to a surface (Figure 4.19) the integral of the pressure
Pl = P0 − Δρgz over the particle surface equals the buoyancy force, Fb, which at equilibrium is
counterbalanced by the disjoining pressure and transversal tension forces [238,253]:
Fb is negligible for bubbles of diameter smaller than ca 300 μm. Then the forces due to τ and Π
counterbalance each other. Hence, at equilibrium the role of the repulsive disjoining pressure is to
Chemical Physics of Colloid Systems and Interfaces 299
Pl Pl
rc
0 r
Pl τ Pl + П τ Pl
FIGURE 4.19 Sketch of the forces exerted on a fluid particle (bubble, drop, vesicle) attached to a solid
surface: Π is disjoining pressure, τ is transversal tension, Pl is the pressure in the outer liquid phase.
keep the film thickness uniform, whereas the role of the attractive transversal tension is to keep the
bubble (droplet) attached to the surface. In other words, the particle sticks to the surface at its con-
tact line where the long-range attraction prevails (see Figure 4.17), whereas the repulsion predomi-
nates inside the film, where Π = Pc > 0. Note that this conclusion is valid not only for particle–wall
attachment, but also for particle–particle interaction. For zero contact angle τ is also zero (Equation
4.155) and the particle will rebound from the surface (the other particle), unless some additional
external force keeps it attached.
d 1
(σ sin ϕ) + σ(r )sin ϕ(r ) = Pc − Π(h(r )) (4.158)
dr r
d 1 dz
− (σ cos ϕ) + σ(r )sin ϕ(r ) = Pc , tan ϕ(r ) = (4.159)
dz r dr
where φ(r) and h(r) = 2z(r) are the running meniscus slope angle and thickness of the gap, r espectively.
Equations 4.158 and 4.159 allow calculation of the three unknown functions, z(r), φ(r), and σ(r),
provided that the disjoining pressure, Π(h), is known from the microscopic theory. By eliminating
Pc between Equations 4.158 and 4.159 we can derive [255]
dσ
= −Π(h(r )) cos ϕ(r ) (4.160)
dz
This result shows that the hydrostatic equilibrium in the transition region is ensured by simultane-
ous variation of σ and Π. Equation 4.160 represents a generalization of Equation 4.146 for a film of
300 Handbook of Surface and Colloid Chemistry
z(r)
P0
P1
(r)
α h/2 α
Plateau r
border 0 rc rB
α –h/2 α
σf
P1
P0
σl
FIGURE 4.20 Liquid film between two attached fluid particles (bubbles, drops, vesicles). The solid lines
represent the actual interfaces, whereas the dashed lines show the extrapolated interfaces in the transition zone
between the film and the Plateau border.
uneven thickness and axial symmetry. Generalization of Equations 4.158 through 4.160 for the case
of more complicated geometry is also available [238,239].
For the Plateau border we have z ≫ h, Π → 0, σ → σl = constant, and both Equations 4.158 and
4.159 reduce to Equation 4.109 with ΔP = Pc. The macroscopic contact angle, α, is defined as the
angle at which the extrapolated meniscus, obeying Equation 4.109, meets the extrapolated film sur-
face (see the dashed line in Figure 4.20). The real surface, shown by solid line in Figure 4.20, differs
from this extrapolated (idealized) profile, because of the interactions between the two film surfaces,
which is taken into account in Equation 4.158, but not in Equation 4.109. To compensate for the dif-
ference between the real and idealized system, the line and transversal tensions are ascribed to the
contact line in the macroscopic approach. In particular, the line tension makes up for the differences
in surface tension and running slope angle [255]:
rB
κ ⎡⎛ σ sin 2 ϕ ⎞real ⎛ σ sin 2 ϕ ⎞idealized ⎤
⎥ dr (4.161)
rc ∫
= ⎜ ⎢
⎢⎝ r cos ϕ ⎠
0 ⎣
⎟ −⎜
⎝ r cos ϕ ⎠
⎟
⎥
⎦
rB
1
[(Π )id − Π(r )]r dr (4.162)
τ=
rc∫ 0
where
The superscripts “real” and “idealized” in Equation 4.161 mean that the quantities in the respective
parentheses must be calculated for the real and idealized meniscus profiles; the latter coincide for
Chemical Physics of Colloid Systems and Interfaces 301
micrometer-sized particles, emulsion drops, and biological cells in thinning free foam films or
wetting films. The interference pattern around the entrapped particles allows us to reconstruct the
meniscus shape, to determine the contact angles, and to calculate the particle-to-interface adhesion
energy [280,281].
A conceptually different method, called gel trapping technique, was developed by Paunov
[282] for determining the three-phase contact angle of solid colloid particles at an air–water or
oil–water interface. The method is applicable for particle diameters ranging from several hundred
nanometers to several hundred micrometers. This technique is based on spreading of the particles
on a liquid interface with a subsequent gelling of the water phase with a nonadsorbing polysaccha-
ride. The particle monolayer trapped on the surface of the gel is then replicated and lifted up with
poly(dimethylsiloxane) (PDMS) elastomer, which allows the particles embedded within the PDMS
surface to be imaged with high resolution by using a scanning electron microscope (SEM), which
gives information on the particle contact angle at the air–water or the oil–water interface [282]. This
method has found applications for determining the contact angles of various inorganic [283,284]
and organic [285,286] particles at liquid interfaces.
ψ1 –ψ2 ψ1 –ψ2
FIGURE 4.21 Flotation (a, c, e) and immersion (b, d, f) lateral capillary forces between two particles
attached to fluid interface: (a) and (b) two similar particles; (c) a light and a heavy particle; (d) a hydrophilic
and a hydrophobic particle; (e) small floating particles that do not deform the interface; (f) small particles
captured in a thin liquid film deforming the interfaces due to the wetting effects.
Chemical Physics of Colloid Systems and Interfaces 303
A force of capillary attraction also appears when the particles (instead of being freely float-
ing) are partially immersed in a liquid layer on a substrate [290−292] (see Figure 4.21b). The
deformation of the liquid surface in this case is related to the wetting properties of the particle
surface, that is, to the position of the contact line and the magnitude of the contact angle, rather
than to gravity.
To distinguish between the capillary forces in the case of floating particles and in the case of
partially immersed particles on a substrate, the former are called lateral flotation forces and the lat-
ter, lateral immersion forces [289,292]. These two kinds of forces exhibit similar dependence on the
interparticle separation but very different dependencies on the particle radius and surface tension
of the liquid (see Refs. [37,293] for comprehensive reviews). The flotation and immersion forces can
be both attractive (Figure 4.21a and b) and repulsive (Figure 4.21c and d). This is determined by the
signs of the meniscus slope angles ψ1 and ψ2 at the two contact lines: the capillary force is attractive
when sin ψ1 sin ψ2 > 0 and repulsive when sin ψ1 sin ψ2 < 0. In the case of flotation forces ψ > 0
for light particles (including bubbles) and ψ < 0 for heavy particles. In the case of immersion forces
between particles protruding from an aqueous layer, ψ > 0 for hydrophilic particles and ψ < 0 for
hydrophobic particles. When ψ = 0 there is no meniscus deformation and, hence, there is no capil-
lary interaction between the particles. This can happen when the weight of the particles is too small
to create significant surface deformation (Figure 4.21e).
The immersion force appears not only between particles in wetting films (Figure 4.21b and
d), but also in symmetric fluid films (Figure 4.21f). The theory provides the following asymp-
totic expression for calculating the lateral capillary force between two particles of radii R1 and R2
separated by a center-to-center distance L [37,288−293]:
where
σ is the liquid–fluid interfacial tension
r1 and r 2 are the radii of the two contact lines
Qk = rk sin ψk (k = 1, 2) is the “capillary charge” of the particle [289,292]; in addition
Δρg
q2 = (in thick film)
σ
Δρq − Πʹ
2
q = (in thick films) (4.164)
σ
Here
Δρ is the difference between the mass densities of the two fluids
Π′ is the derivative of the disjoining pressure with respect to the film thickness
K1(x) is the modified Bessel function of the first order
The asymptotic form of Equation 4.163 for qL ≪ 1 (q−1 = 2.7 mm for water),
2πσQ1Q2
F= rk L q −1 (4.165)
L
looks like a 2D analogue of Coulomb’s law, which explains the name “capillary charge” of Q1
and Q2. Note that the immersion and flotation forces exhibit the same functional dependence
304 Handbook of Surface and Colloid Chemistry
on the interparticle distance, see Equations 4.163 and 4.164. On the other hand, their different
physical origin results in different magnitudes of the “capillary charges” of these two kinds of
capillary force. In this aspect they resemble the electrostatic and gravitational forces, which
obey the same power law, but differ in the physical meaning and magnitude of the force con-
stants (charges, masses). In the special case when R1 = R2 = R and rk ≪ L ≪ q−1, we can
derive [292,293]
⎛ R6 ⎞
F ∝⎜ ⎟ K1 (qL ) for flotation force
⎝ σ ⎠
Hence, the flotation force decreases, while the immersion force increases, when the interfacial ten-
sion σ increases. Besides, the flotation force decreases much more strongly with the decrease of R
than the immersion force. Thus Fflotation is negligible for R < 10 μm, whereas Fimmersion can be signifi-
cant even when R = 10 nm. This is demonstrated in Figure 4.22 where the two types of capillary
interactions are compared for a wide range of particle sizes. The values of the parameters used are:
particle mass density ρp = 1.05 g/cm3, surface tension σ = 72 mN/m, contact angle α = 30°, interpar-
ticle distance L = 2R, and thickness of the nondisturbed planar film l0 = R. The drastic difference
in the magnitudes of the two types of capillary forces is due to the different deformation of the
water–air interface. The small floating particles are too light to create substantial deformation of the
liquid surface, and the lateral capillary forces are negligible (Figure 4.21e). In the case of immer-
sion forces the particles are restricted in the vertical direction by the solid substrate. Therefore,
as the film becomes thinner, the liquid surface deformation increases, thus giving rise to a strong
interparticle attraction.
As seen in Figure 4.22, the immersion force can be significant between particles whose radii
are larger than few nanometers. It has been found to promote the growth of 2D crystals from col-
loid particles [294−297], viruses, and globular proteins [298−304]. Such 2D crystals have found
various applications: in nanolithography [305], microcontact printing [306], as nanostructured
materials in photo-electrochemical cells [307], in photocatalytic films [308], photo- and electro-
luminescent semiconductor materials [309], as samples for electron microscopy of proteins and
viruses [310], as immunosensors [311], etc. (for reviews see Refs. [37,312]).
1012
L = 2R
n
1010 tio
te rac
n in
108 rsio
me
Im
n
–∆W/kT
tio
106
ac
ter
in
104
ion
tat
102
Flo
kT
100
10–6 10–5 10–4 10–3 10–2 10–1
R (cm)
FIGURE 4.22 Plot of the capillary interaction energy in kT units, ΔW/kT, vs. the radius, R, of two similar
particles separated at a center-to-center distance L = 2R.
Chemical Physics of Colloid Systems and Interfaces 305
(a) (b)
FIGURE 4.23 Inclusions (say, membrane proteins) in a lipid bilayer: the thickness of the inclusion can be
greater (a) or smaller (b) than the thickness of the (nondisturbed) lipid bilayer. In both cases, the overlap of the
deformations around the inclusions leads to an attraction between them (see Refs. [37,204].)
In the case of interactions between inclusions in lipid bilayers (Figure 4.23), the elasticity of
the bilayer interior must also be taken into account. The calculated energy of capillary interac-
tion between integral membrane proteins turns out to be of the order of several kT [204]. Hence,
this interaction can be a possible explanation for the observed aggregation of membrane proteins
[204,313−316]. The lateral capillary forces have also been calculated for the case of particles
captured in a spherical (rather than planar) thin liquid film or vesicle [316].
Lateral capillary forces between vertical cylinders or between spherical particles have been
measured by means of sensitive electromechanical balance [317], piezo-transducer balance [318],
and torsion microbalance [319]. Good agreement between theory and experiment has been estab-
lished [318,319].
As already mentioned, the weight of micrometer-sized and sub micrometer-sized floating par-
ticles is not sufficient to deform the fluid interface and to bring about capillary force between the
particles (Figure 4.21e). However, the situation changes if the contact line at the particle surface has
undulated or irregular shape (Figure 4.24a). This may happen when the particle surface is rough,
angular, or heterogeneous. In such cases, the contact line sticks to an edge or to the boundary
between two domains of the heterogeneous surface. The undulated contact line induces undula-
tions in the surrounding fluid interface [312,320−324]. Let z = ζ(x, y) be the equation describing the
interfacial shape around such isolated particle. Using polar coordinates (r, φ) in the xy-plane, we can
express the interfacial shape as a Fourier expansion:
∞
ζ(r , ϕ) = ∑r −m
[ Am cos(mϕ) + Bmsin(mϕ)] (4.167)
m =1
where
r is the distance from the particle centre
Am and Bm are coefficients
In analogy with electrical theory, Equation 4.167 can be interpreted as a multipole expansion:
the terms with m = 1, 2, 3,…, play the role of capillary “dipoles,” “quadrupoles,” “hexapoles,”
and multipoles [312,320−324]. The term with m = 0 (capillary “charge”) is missing because
there is no axisymmetric contribution to the deformation (negligible particle weight). Moreover,
the dipolar term with m = 2 is also absent because it is annihilated by a spontaneous rotation
of the floating particle around a horizontal axis [321]. Therefore, the leading term becomes the
quadrupolar one, with m = 2. The interaction between capillary quadrupoles has been investi-
gated theoretically [321−324]. This interaction is nonmonotonic: attractive at long distances,
but repulsive at short distances. Expressions for the rheological properties (surface dilatational
and shear elasticity and yield stress) of Langmuir monolayers from angular particles have been
derived [37,322,323].
306 Handbook of Surface and Colloid Chemistry
Capillary multipoles
Finite menisci
(due to irregular contact line)
hc
L rp
L
(a) (b)
FIGURE 4.24 Special types of immersion capillary forces: (a) the contact line attachment to an irregular
edge on the particle surface produces undulations in the surrounding fluid interface, which give rise to lateral
capillary force between the particles. (b) When the size of particles, entrapped in a liquid film, is much greater
than the nonperturbed film thickness, the meniscus surfaces meet at a finite distance, rp; in this case, the capil-
lary interaction begins at L ≤ 2rp.
Note that Equation 4.167 is approximate and holds for interparticle distances, which are much
smaller than the characteristic capillary length, that is, qr ≪ 1. The general form of the multipolar
expansion, Equation 4.167, for arbitrary interparticle distances reads [321−324]:
∞
ζ(r , ϕ) = A0 K 0 (qr ) + ∑ A K (qr )cos[m(ϕ − ϕ
m m 0, m )] (4.167a)
m =1
where
Am and φ 0,m are constants of integration
Km is the modified Bessel function of the second kind and mth order
The first term with m = 0 in the right-hand side of Equation 4.167a accounts for the contribution of
the “capillary charges” (or “capillary monopoles”). Analytical expressions for the force and energy
of interaction between two capillary multipoles of arbitrary order have been derived [324].
“Mesoscale” capillary multipoles have been experimentally realized by Bowden et al. [325,326],
by appropriate hydrophobization or hydrophilization of the sides of floating plates. Interactions
between capillary quadrupoles have been observed between floating particles, which have the shape
of curved disks [327]. Loudet et al. [328−330] investigated experimentally and theoretically the
capillary forces between adsorbed ellipsoidal particles and found that they behave as capillary
quadrupoles. These authors noted that from a purely geometrical viewpoint, the condition of a
constant contact angle cannot be met for anisotropic particles if the interface remains flat, which
explains the reason for the quadrupolar interfacial deformation. Lateral capillary forces between
ellipsoidal, cylindrical (rodlike), and other anisotropic particles have also been investigated by van
Nierop et al. [331], Lehle et al. [332], Stebe et al. [333−337], and Yunker et al. [338] Gravitation-
like instabilities due to the long-range attractive capillary forces between floating particles have also
been studied [339].
At last, let us consider another type of capillary interactions—between particles surrounded by
finite menisci. Such interactions appear when micrometer-sized or submicrometer-sized particles are
captured in a liquid film of much smaller thickness (Figure 4.24b) [340−343]. If such particles are
approaching each other, the interaction begins when the menisci around the two particles o verlap,
L < 2rp in Figure 4.24b. The capillary force in this case is nonmonotonic: initially the attractive
force increases with the increase of interparticle distance, then it reaches a maximum and further
decays [343]. In addition, there are hysteresis effects: the force is different on approach and separa-
tion at distances around L = 2rp [343].
Chemical Physics of Colloid Systems and Interfaces 307
Here
Q2 and r 2 are the particle capillary charge and contact line radius, respectively
H characterizes the position of the contact line on the wall with respect to the nondisturbed hori-
zontal liquid surface (Figure 4.21)
x is the particle–wall distance
q is defined by Equation 4.164 (thick films)
The first term in the right-hand side of Equation 4.168 expresses the gravity force pushing the
particle to slide down over the inclined meniscus on the wall; the second term originates from
the pressure difference across the meniscus on the wall; the third term expresses the so-called
capillary image force, that is, the particle is repelled by its mirror image with respect to the wall
surface [37,344].
Static [345] and dynamic [346] measurements with particles near walls have been carried out.
In the static measurements the equilibrium distance of the particle from the wall (the distance at
which F = 0) has been measured and a good agreement with the theory has been established [345].
1 2
Water
FIGURE 4.25 Experimental setup for studying the capillary interaction between a floating particle (1) and
a vertical hydrophobic plate (2) separated at a distance, x. The edge of the plate is at a distance, H, lower than
the level of the horizontal liquid surface far from the plate; (3) and (4) are micrometric table and screw; see
Refs. [345,346].
308 Handbook of Surface and Colloid Chemistry
In the dynamic experiments [346] knowing the capillary force F (from Equation 4.168) and
measuring the particle velocity, x , we can determine the drag force, Fd:
where
R2, m, and x are the particle radius, mass, and acceleration, respectively
η is the viscosity of the liquid
fd is the drag coefficient
If the particle were in the bulk liquid, fd would be equal to 1 and Fd would be given by the Stokes
formula. In general, fd differs from unity because the particle is attached to the interface. The exper-
iment [346] results in fd varying between 0.68 and 0.54 for particle contact angle varying from 49°
to 82°; the data are in good quantitative agreement with the hydrodynamic theory of the drag coef-
ficient [347]. In other words, the less the depth of particle immersion, the less the drag coefficient,
as could be expected. However, if the floating particle is heavy enough, it deforms the surrounding
liquid surface; the deformation travels together with the particle, thus increasing fd several times
[346]. The addition of surfactant strongly increases fd. The latter effect can be used to measure the
surface viscosity of adsorption monolayers from low molecular weight surfactants [348], which is
not accessible to the standard methods for measurement of surface viscosity.
In the case of protein adsorption layers, the surface elasticity is so strong that the particle
(Figure 4.25) is arrested in the adsorption film. Nevertheless, with heavier particles and at larger
meniscus slopes, it is possible to break the protein adsorption layer. Based on such experiments, a
method for determining surface elasticity and yield stress has been developed [349].
∞
β23Q 2 β23Q 2
Fz = f , W = Fz ds = w (4.170)
4ε 2 ( R + s ) 2 z ∫
s
4ε 2 ( R + s )
Here
R is the particle radius
s is the distance between the particle surface and the fluid interface (inset in Figure 4.26)
Q = 4πR2σpn is the total charge at the boundary particle–nonpolar fluid
fz and w are dimensionless force and energy coefficients, respectively, which, in general, depend
on the parameters s/R, β12, and β23, where βij = (εi − εj)/(εi + εj); i, j = 1, 2, 3; ε1, ε2, and ε3 are
the dielectric constants of the respective phases
At long distances, s/R > 1, we have fz ≈ w ≈ 1, and then Equation 4.170 reduces to the expressions
for the force and energy of interaction between a point charge Q with the interface between phases
2 and 3. This is the known image charge effect. Expressions that allow us to calculate fz and w for
shorter distances (s/R < 1) are derived in Ref. [350].
Chemical Physics of Colloid Systems and Interfaces 309
R = 1000 nm Phase 1
5 750 nm Attraction
10
104 300 nm
200 nm s Phase 2
103 O Phase 3
100 nm
102 75 nm
50 nm
101 30 nm
100
10–2 10–1 100 101 102 103 104 105
Dimensionless distance, s/R
FIGURE 4.26 Plot of the interaction energy W (scaled with kT) vs. the dimensionless distance, s/R, between
a charged glass particle (phase 1) and a planar interface; phase 2 is tetradecane; phase 3 is water. The curves
correspond to different particle radii, R, denoted in the figure.
In Figure 4.26, numerical results for the particle–interface interaction energy, W, scaled by the
thermal energy kT, are plotted versus the relative distance, s/R, for various values of the particle
radius, R. The other parameter values correspond to the following choice of the phases: phase 1
(the particle) is glass, phase 2 is tetradecane, and phase 3 is water. The curves in Figure 4.26 describe
a strong and long-range attraction between the particle and the interface. The interaction energy, W,
becomes comparable, or smaller than the thermal energy kT for particle radius R < 30 nm. On the
other hand, for R > 30 nm W strongly increases with the particle size (in Equation 4.170 Q2 ~ R4 at
fixed surface charge density, σpn) and reaches W ≈ 105 kT for R = 1 μm at close contact. In addition,
the range of interaction also increases, reaching s/R ≈ 105 for R = 1 μm. In general, this is a strong and
long-range interaction [350]. For example, water drops could attract charged hydrophobic particles
dispersed in the oily phase, which would favor the formation of reverse particle-stabilized emulsions.
γsinψ ψ rc ψ
2rc α
R θ
0
FED FED
Water L
FIGURE 4.27 Sketch of two electrically charged particles attached to an oil–water interface. FED is the elec-
trodipping force, due to the image-charge effect, that pushes the particles into water and deforms the fluid
interface around the particles. FER is the direct electric repulsion between the two like-charged particles. FEC is
the electrocapillary attraction, related to deformations in the fluid interface created by the electric field.
310 Handbook of Surface and Colloid Chemistry
while FER and FEC are interaction forces between two (or more) particles. The presence of electric
field leads to deformations in the fluid interface around the particles, which lead to the appearance of
FEC. The three forces, FED, FER, and FEC, are separately considered in the f ollowing section.
⎛ 4π ⎞
FED = ⎜ ⎟ (σpn R)2 (1− cos α) f (θ, ε pn ) (4.171)
⎝ εn ⎠
Here
εn is the dielectric constant of the nonpolar fluid (oil, air)
σpn is the surface charge density at the boundary particle–nonpolar fluid
εpn = εp/εn is the ratio of the respective two dielectric constants
α is a central angle, while θ = α + ψ is the contact angle (see Figure 4.27)
Soybean
oil
Soybean
oil
ψ = 1.5˚ ψ =26˚
Water Water
FIGURE 4.28 Side-view photographs of hydrophobized spherical glass particles at the boundary water–
soybean oil (no added surfactants). (a) Noncharged particle of radius R = 235 μm: the meniscus slope angle due
to gravity is relatively small, ψ = 1.5°. (b) Charged particle of radius R = 274 μm: the experimental meniscus
slope angle is ψ = 26° owing to the electrodipping force; if this electric force were missing, the gravitational
slope angle would be only ψ = 1.9°.
Chemical Physics of Colloid Systems and Interfaces 311
We could accurately calculate the dimensionless function f(θ, εpn) by means of the relation
f(θ, εpn) = f R (θ, εpn)/(1 − cos θ), where the function f R (θ, εpn) is tabulated in Table 4.3 of Ref. [355] on
the basis of the solution of the electrostatic boundary problem. The tabulated values can be used for
a convenient computer calculation of f R (θ, εpn) with the help of a four-point interpolation formula,
Equation D.1 in Ref. [355]. From the experimental FED and Equation 4.171, we could determine the
surface charge density, σpn, at the particle–oil and particle–air interface. Values of σpn in the range
from 20 to 70 μC/m2 have been obtained [351,354].
3 pd2 ⎛R ⎞
FER =
2ε n L4 ⎜ L 1 ⎟ (4.172)
⎝ ⎠
L is the center-to-center distance between the two particles; pd = 4πσpn DR3 sin3α is the effective
particle dipole moment; as before, R is the particle radius and σpn is the electric charge density at
the particle–nonpolar fluid interface; D = D(α, εpn) is a known dimensionless function, which can
Pd L Pd
Water
FIGURE 4.29 Two particles attached to the boundary water–nonpolar fluid and separated at a center-to-
center distance L. In the nonpolar fluid (oil, air), the electric field of each particle in isolation is asymptotically
identical to the field of a dipole of moment pd. This field is created by charges at the particle–nonpolar fluid
interface.
312 Handbook of Surface and Colloid Chemistry
be calculated by means of Table 4.1 and Equation D.1 in Ref. [355]; εpn ≡ εp/εn is the ratio of the
dielectric constants of the two phases. Equation 4.172 shows that FER asymptotically decays as 1/L 4
like the force between two point dipoles. However, at shorter distances, the finite size of the particle
is expected to lead to a Coulombic repulsion, FER ~ 1/L2; see Refs. [356–358].
Monolayers from electrically charged micron-sized silica particles, spread on the air–water inter-
face, were investigated and surface pressure versus area isotherms were measured by Langmuir
trough and the monolayers’ morphology was monitored by microscope [363]. The experiments
showed that Π ~ L –3 at large L, where Π is the surface pressure and L is the mean interparticle
distance. A theoretical cell model was developed, which predicts not only the aforementioned
asymptotic law but also the whole Π(L) dependence. The model presumes a periodic distribution of
the surface charge density, which induces a corresponding electric field in the air phase. Then, the
Maxwell pressure tensor of the electric field in the air phase was calculated and integrated accord-
ing to the Bakker’s formula [189] to determine the surface pressure. Thus, all collective effects from
the electrostatic interparticle interactions were taken into account, as well as the effects from the
particle finite size.
The effects of applied vertical external electric field on the electrostatic forces acting on a col-
loid particle at a horizontal liquid interface have also been investigated. By varying the strength
of the electric field, it is possible to control the distances between the particles in nondensely
packed 2D colloid crystals formed at liquid interfaces [367–370]. Theoretical expressions for the
forces between floating uncharged [371] and charged [372] dielectric particles in the presence of
external electric field were derived. The particles are located on the boundary water–nonpolar
fluid (air, oil). The effects of the dielectric constants and contact angle (particle wettability)
on the vertical electrodipping force, FED, acting on each particle, and on the horizontal force
between two particles, FER, were investigated. The external field polarizes the uncharged par-
ticles at the fluid interface. The vertical electric force on the particle can be directed upward or
downward. The horizontal interparticle repulsion is dipolar and contact angle dependent. At
given contact angle (for uncharged particles) and external electric field E 0 (for charged particles),
the dipole moment is zero and the repulsion becomes short-range octupolar [371,372]. This mini-
mal electrostatic repulsion could be weaker than the electrocapillary capillary attraction; see the
next section.
FIGURE 4.30 Profile of the oil (tetradecane)–water interface near the contact line of a charged glass par-
ticle, like that in Figure 4.28b: plot of experimental data from Ref. [354]; see Figure 4.27 for the notations. The
dash-dot line shows the gravitational; profile calculated under the assumption that the particle is not charged.
The difference between the real and the gravitational profiles represents the effect of electric field on the
meniscus shape. The fact that the real (experimental) profile is insensitive to the concentration of NaCl in the
water phase indicates that the electric charges are located at the particle–oil interface, so that the interfacial
deformation is due to electric field in the oily phase.
for two identical floating particles with contact radius rc, which are separated at a center-to-center
distance L (see Figure 4.27):
δ = tan ψ, where ψ is the meniscus slope angle for each particle in isolation (Figure 4.27). In
Equation 4.172a, the first term in the brackets is FER in Equation 4.172, whereas the next terms,
which are proportional to the meniscus deformation angle δ, give FEC. Because for micrometer and
submicrometer particles δ is a small quantity, it turns out that for uniform distribution of the surface
charges, the electrocapillary attraction is weaker than the electrostatic repulsion at interparticle dis-
tances, at which the dipolar approximation is applicable, so that the net force, Fx, is repulsive [376].
The final conclusion from the theoretical analysis is that the direct electrostatic repulsion dominates
over the capillary attraction when the surface charge is uniformly distributed; no matter whether the
surface charge is on the polar–liquid or nonpolar–fluid side of the particle.
Electric field–induced attraction that prevails over the electrostatic repulsion was estab-
lished (both experimentally and theoretically) in the case of not-too-small floating particles, for
which the interfacial deformation due to gravity is not negligible [377,378]. If the surface charge
is anisotropically distributed (this may happen at low surface charge density), the electric field
produces a saddle-shaped deformation in the liquid interface near the particle, which is equiva-
lent to a “capillary quadrupole.” The interaction of the latter with the axisymmetric gravitational
deformation around the other particle (which is equivalent to a “capillary charge”) gives rise to a
capillary force that decays ∝ 1/L3, that is, slower than FER ∝ 1/L 4. In such a case, we are dealing
with a hybrid attraction between a gravity-induced “capillary charge” and an electric field–induced
“capillary quadrupole” [378,379].
314 Handbook of Surface and Colloid Chemistry
where, as before, Π denotes disjoining pressure. Derjaguin [381] derived an approximate formula,
which expresses the energy of interaction between two spherical particles of radii R1 and R2 through
integral of f(h):
∞
2πR1R2
U (h0 ) =
R1 + R2 ∫ f (h)dh (4.174)
h0
Here, h0 is the shortest distance between the surfaces of the two particles (see Figure 4.31). In the deriva-
tion of Equation 4.174 it is assumed that the interaction between two parcels from the particle surfaces,
separated at the distance h, is approximately the same as that between two similar parcels in a plane-
parallel film. This assumption is correct when the range of action of the surface forces and the distance
h0 are small compared to the curvature radii R1 and R2. It has been established, both experimentally [36]
and theoretically [382], that Equation 4.174 provides a good approximation in the range of its validity.
Equation 4.174 can be generalized for smooth surfaces of arbitrary shape (not necessarily
spheres). For that purpose, the surfaces of the two particles are approximated with paraboloids
in the vicinity of the point of closest approach (h = h 0). Let the principle curvatures at this point
be c1 and c1 for the first particle, and c2 and c2 for the second particle. Then the generalization of
Equation 4.174 reads [380]:
∞
2π
U (h0 ) =
C ∫ f (h)dh (4.175)
h0
R2
R1
h0
FIGURE 4.31 Two spherical particles of radii R1 and R2; the shortest and the running surface-to-surface
distances are denoted by h 0 and h, respectively.
Chemical Physics of Colloid Systems and Interfaces 315
where ω is the angle subtended between the directions of the principle curvatures of the two
approaching surfaces. For two spheres, we have c1 = c1ʹ = 1/R1, c2 = cʹ2 = 1/R2, and Equation 4.175
reduces to Equation 4.174.
For two cylinders of radii r1 and r 2 crossed at angle ω we have c1 = c2 = 0; c1ʹ = 1/r1, c2ʹ = 1/r2 and
Equation 4.175 yields
∞
2π r1r2
U (h0 ) =
sin ω ∫ f (h)dh (4.176)
h0
Equation 4.176 is often used in connection to the experiments with the surface force apparatus
(SFA) [36,383], in which the interacting surfaces are two crossed cylindrical mica sheets. The diver-
gence in Equation 4.176 for ω = 0 reflects the fact that the axes of the two infinitely long cylinders
are parallel for ω = 0 and thus the area of the interaction zone becomes infinite.
The Derjaguin’s formula is applicable to any type of force law (attractive, repulsive, oscillatory)
if only (1) the range of the forces, and (2) the surface-to-surface distance are much smaller than the
surface curvature radii. This formula is applicable to any kind of surface force, irrespective of its
physical origin: van der Waals, electrostatic, steric, oscillatory-structural, depletion, etc. It reduces
the two-particle interaction problem to the simpler problem for interactions in plane-parallel films.
αij
uij (r ) = − (4.177)
r6
where
uij is the potential energy of interaction
r is the distance between the two molecules
αij is a constant characterizing the interaction
In fact, the van der Waals forces represent an averaged dipole–dipole interaction, which is a superposi-
tion of three main terms: (1) orientation interaction: interaction between two permanent dipoles [384];
(2) induction interaction: interaction between one permanent dipole and one induced dipole [385];
(3) dispersion interaction: interaction between two induced dipoles [386]. The theory yields [36]:
1 ⎡ pi2 p2j 3α α h ν ν ⎤
αij = ⎢ + pi2α 0 j + p2j α 0i + 0i 0 j p i j ⎥ (4.178)
( )
(4πε0 )2 ⎣ 3kT 2(νi + ν j ) ⎦
where
pi and αi0 are molecular dipole moment and electronic polarizability, respectively;
hp is the Planck constant
νi is the orbiting frequency of the electron in the Bohr atom
For van der Waals interactions between molecules in a gas phase, the orientation interaction can
yield from 0% (nonpolar molecules) up to 70% (molecules of large permanent dipole moment, like
H2O) of the value of αij; the contribution of the induction interaction in αij is usually low, about
5%–10%; the contribution of the dispersion interaction might be between 24% (water) and 100%
(nonpolar hydrocarbons); for numerical data, see Ref. [36].
According to the microscopic theory by Hamaker [387], the van der Waals interaction between
two macroscopic bodies can be found by integration of Equation 4.177 over all couples of molecules,
316 Handbook of Surface and Colloid Chemistry
followed by subtraction of the interaction energy at infinite separation between the bodies. The
result depends on the geometry of the system. For a plane-parallel film from component 3 located
between two semi-infinite phases composed from components 1 and 2, the van der Waals interac-
tion energy per unit area and the respective disjoining pressure, stemming from Equation 4.166, are
[387]
AH ∂f A
fvw = − , Π vw = − vw = − H 3 (4.179)
12πh 2
∂h 6πh
where
h is the thickness of the film
AH is the compound Hamaker constant [14]
If Aii and Ajj are known, we can calculate Aij by using the Hamaker approximation
In fact, Equation 4.181 is applicable to the dispersion contribution in the van der Waals
interaction [36].
When components 1 and 2 are identical, AH is positive (see Equation 4.180); therefore, the van
der Waals interaction between identical bodies, in any medium, is always attractive. Besides, two
dense bodies (even if nonidentical) will attract each other when placed in medium 3 of low density
(gas, vacuum). When the phase in the middle (component 3) has intermediate Hamaker constant
between those of bodies 1 and 2, AH can be negative and the van der Waals disjoining pressure can
be repulsive (positive). Such is the case of an aqueous film between mercury and gas [388].
Lifshitz et al. [389,390] developed an alternative approach to the calculation of the Hamaker
constant AH in condensed phases, called the macroscopic theory. The latter is not limited by the
assumption for pair-wise additivity of the van der Waals interaction (see also Refs. [36,380,391]).
The Lifshitz theory treats each phase as a continuous medium characterized by a given uniform
dielectric permittivity, which is dependent on the frequency, ν, of the propagating electromagnetic
waves. For the symmetric configuration of two identical phases “i” interacting across a medium “j,”
the macroscopic theory provides the expression [36]
2 2
(ν =0) (ν >0) 3 ⎛ ε − εj ⎞ 3hpν e ni2 − n2j
( )
AH ≡ Aiji = A
iji +A
iji = kT ⎜ i ⎟ + (4.182)
4 ⎝ εi + ε j ⎠ 16 2 ni2 + n2j 3 / 2
( )
where
εi and εj are the dielectric constants of phases i and j, respectively;
ni and nj are the respective refractive indices for visible light
hp is the Planck constant
νe is the main electronic absorption frequency which is ≈ 3.0 × 1015 Hz for water and the most
organic liquids [36]
Chemical Physics of Colloid Systems and Interfaces 317
The first term in the right-hand side of Equation 4.182, Aiji( ν=0 ), is the so-called zero-frequency term,
expressing the contribution of the orientation and induction interactions. Indeed, these two con-
tributions to the van der Waals force represent electrostatic effects. Equation 4.182 shows that the
zero-frequency term can never exceed (3/4)kT ≈ 3 × 10−21 J. The last term in Equation 4.182, Aijiν>0 ,
( )
accounts for the dispersion interaction. If the two phases, i and j, have comparable densities (as for
emulsion systems, say oil–water–oil), then Aiji( ν>0 ) and Aiji( ν=0 ) are comparable by magnitude. If one
of the phases, i or j, has a low density (gas, vacuum), we obtain Aiji( ν > 0 ) Aiji( ν = 0 ). In the latter case,
the Hamaker microscopic approach may give comparable Aiji( ν>0 ) and Aiji( ν=0 ) in contradiction to the
Lifshitz macroscopic theory, which is more accurate for condensed phases.
A geometrical configuration, which is important for disperse systems, is the case of two spheres
of radii R1 and R2 interacting across a medium (component 3). Hamaker [387] has derived the
following expression for the van der Waals interaction energy between two spheres:
AH ⎛ y y x 2 + xy + x ⎞
U (h0 ) = − ⎜ 2 + 2 + 2 ln 2 ⎟ (4.183)
12 ⎝ x + xy + x x + xy + x + y x + xy + x + y ⎠
where
h0
x=
2 R1
R
y = 2 ≤1 (4.184)
R1
AH y 2πR1R2 AH
U (h0 ) ≈ − =− (4.185)
12 (1 + y) x R1 + R2 12πh0
Equation 4.185 can be also derived by combining Equation 4.179 with the Derjaguin approxima-
tion (Equation 4.174). It is worthwhile noting that the logarithmic term in Equation 4.183 can be
neglected only if x ≪ 1. For example, even when x = 5 × 10−3, the contribution of the logarithmic
term amounts to about 10% of the result (for y = 1); consequently, for larger values of x this term
must be retained.
Another geometrical configuration, which corresponds to two colliding deformable emulsion
droplets, is sketched in Figure 4.32. In this case the interaction energy is given by the expression [392]
AH ⎡ 3 Rs ⎛ h ⎞ r 2 2r 2 ⎤
U (h, r ) = − ⎢ + + 2 ln ⎜ ⎟ + h2 − R h ⎥ (h, r Rs ) (4.186)
12 ⎣ 4 h ⎝ Rs ⎠ s ⎦
where
h and r are the thickness and the radius of the flat film formed between the two deformed drops,
respectively,
Rs is the radius of the spherical part of the drop surface (see Figure 4.32).
Equation 4.186 is a truncated series expansion; the exact formula, which is more voluminous, can be
found in Ref. [392]. Expressions for U for other geometrical configurations are also available [37,391].
318 Handbook of Surface and Colloid Chemistry
2r
Rs
FIGURE 4.32 Thin film of radius r and thickness h formed between two attached fluid particles; the spheri-
cal part of the particle surface has radius Rs.
The asymptotic behavior of the dispersion interaction at large intermolecular separations does
not obey Equation 4.177; instead uij ∝ 1/r 7 due to the electromagnetic retardation effect established
by Casimir and Polder [393]. Several different expressions have been proposed to account for this
effect in the Hamaker constant [391].
The orientation and induction interactions are electrostatic effects, so they are not subjected
to electromagnetic retardation. Instead, they are subject to Debye screening due to the presence
of electrolyte ions in the liquid phases. Thus for the interaction across an electrolyte solution, the
screened Hamaker constant is given by the expression [36,394]
AH = 2 κhA0e −2 κ h + Ad (4.187)
where
A0 denotes the contribution of the (non-screened) orientation and induction interactions to the
Hamaker constant
Ad is the contribution of the dispersion interaction
κ is the Debye screening parameter: κ = κcI1/2 (see Equation 4.34)
Equation 4.187 is accurate to within 15% for κh > 2; see Ref. [36].
planes is very large, the number concentration of both counterions and coions would be equal to
its bulk value, n0, in the middle of the film. However, at finite separation, h, between the surfaces
the two EDLs overlap and the counterion and coion concentrations in the middle of the film, n10
and n20, are no longer equal. Because the solution inside the film is supposed to be in electro-
chemical (Donnan) equilibrium with the bulk electrolyte solution of concentration n0, we can
write [395] n10 n20 = n02, or alternatively
n0 n
n10 = , n20 = n0 m , m ≡ 20 (4.188)
m n10
As pointed out by Langmuir [396], the electrostatic disjoining pressure, Πel, can be identified with
the excess osmotic pressure in the middle of the film:
Equation 4.189 demonstrates that for two identically charged surfaces, Πel is always positive, that is,
corresponds to repulsion between the surfaces. In general, we have 0 < m ≤ 1 because the coions are
repelled from the film due to the interaction with the film surfaces. To find the exact dependence of
Πel on the film thickness, h, we have to solve the Poisson–Boltzmann equation for the distribution of
the electrostatic potential inside the film. The solution provides the following connection between
Πel and h for symmetric electrolytes [380,397]:
cot θ
cos ϕ = (fixed surface potential Φ s ) (4.191)
sinh( ZΦ s / 2)
⎛ ZΦ ∞ ⎞
tan ϕ = (tan θ)sinh ⎜ ⎟ (fixed surface charge σs ) (4.192)
⎝ 2 ⎠
2
1 ⎛ Zeσs ⎞ eψ s
cosh( ZΦ ∞ ) = 1 + ⎜ ⎟ , Φs ≡ (4.193)
2 ⎝ εε0 kTκ ⎠ kT
Here
Φs is the dimensionless surface potential
Φ∞ is the value of Φs for h → ∞
Equation 4.190 expresses the dependence Πel(h) in a parametric form: Πel(θ), h(θ). Fixed surface
potential or charge means that Φs or σs does not depend on the film thickness h. The latter is impor-
tant to be specified when integrating Π(h) or f(h) (in accordance with Equations 4.173 or 4.176) to
calculate the interaction energy.
In principle, it is possible neither the surface potential nor the surface charge to be constant
[398]. In such case a condition for charge regulation is applied, which represents the condition for
dynamic equilibrium with respect to the counterion exchange between the Stern and diffuse parts
320 Handbook of Surface and Colloid Chemistry
of the EDL (i.e., condition for constant electrochemical potentials of the ionic species). As discussed
in Section 4.2.1.2.3, the Stern layer itself can be considered as a Langmuir adsorption layer of coun-
terions. We can relate the maximum possible surface charge density (due to all the surface ionizable
groups) to Γ1 in Equation 4.47: σmax = ZeΓ1. Likewise, the effective surface charge density, σs, which
is smaller by magnitude than σmax (because some ionizable groups are blocked by adsorbed counter-
ions) can be expressed as σs = Ze(Γ1 − Γ2). Then, with the help of Equation 4.44, the Stern isotherm
(Equation 4.47) can be represented in the form
σmax − σs −1
= ⎡⎣1 + ( K 2 I )−1 exp( ZΦ s ) ⎤⎦ (4.194)
σmax
The product ZΦs is always positive. At high surface potential, ZΦs → ∞, from Equation 4.194 we
obtain σs → σmax, that is, there is no blocking of surface ionizable by adsorbed counterions.
When the film thickness is large enough (κh ≥ 1) the difference between the regimes of constant
potential, constant charge, and charge regulation becomes negligible, that is, the usage of each of
them leads to the same results for Πel(h) [14].
When the dimensionless electrostatic potential in the middle of the film
e 1
Φm = ψm = − ln m (4.195)
kT 2Z
is small enough (the film thickness, h, is large enough), we could use the superposition approxima-
tion, that is, we could assume that Φm ≈ 2Φ1(h/2), where Φ1 is the dimensionless electric p otential
at a distance h/2 from the surface (of the film) when the other surface is removed at infinity.
Because
⎛h⎞ ⎛ ZΦ s ⎞
ZΦ1 ⎜ ⎟ = 4e − κh / 4 tanh ⎜ ⎟ (4.196)
⎝2⎠ ⎝ 4 ⎠
from Equations 4.189, 4.195, and 4.196, we obtain a useful asymptotic formula [399]
2
⎛ ZΦ s ⎞ − κh
Π el ≈ n0 kTZ 2Φ 2m ≈ 64n0 kT ⎜ tanh e (4.197)
⎝ 4 ⎟⎠
It should be noted that if Φs is large enough, the hyperbolic tangent in Equation 4.197 is identically
1, and Πel (as well as fel) becomes independent of the surface potential (or charge). Equation 4.197
can be generalized for the case of 2:1 electrolyte (bivalent counterion) and 1:2 electrolyte (bivalent
coion) [400]:
2
⎛ v ⎞
Π el = 432n( 2 )kT ⎜ tanh i: j ⎟ e − κh (4.198)
⎝ 4 ⎠
where n(2) is the concentration of the bivalent ions, the subscript “i:j” takes value “2:1” or “1:2,”, and
⎡ 3 ⎤ ⎡ 2e Φ s + 1 ⎤
v2:1 = ln ⎢ ⎥ , v1:2 = ln ⎢ 3 ⎥ (4.199)
⎣ 1 + 2 e −Φ s ⎦ ⎣ ⎦
Chemical Physics of Colloid Systems and Interfaces 321
Π ψel = (4.200)
2π sinh 2 κh
When the two surface potentials have opposite signs, that is, when ψ s1ψ s 2 < 0,Π ψel is negative for all
h and corresponds to electrostatic attraction (see Figure 4.33a). This result could have been antici-
pated, because two charges of opposite sign attract each other. More interesting is the case, when
ψs1ψs2 > 0, but ψs1 ≠ ψs2. In the latter case, the two surfaces repel each other for h > h 0, whereas
they attract each other for h < h 0 (Figure 4.33a); h 0 is determined by the equation κh 0 = ln(ψs2/ψs1);
ψs2 > ψs1. In addition, the electrostatic repulsion has a maximum value of
εε0 κ2 2 1 ψ
Π ψel (max) = ψ s1 at hmax = arccosh s 2 , ψ s 2 > ψ s1 (4.201)
2π κ ψ s1
Similar electrostatic disjoining pressure isotherm has been used to interpret the experimental data
for aqueous films on mercury [388]. It is worthwhile noting that Π ψel (max) depends only on ψs1, that
is, the maximum repulsion is determined by the potential of the surface of lower charge.
If σs1 = constant, and σs2 = constant, then instead of Equation 4.200 we have [380]
∏ψel ∏σel
+ +
+ + + +
∏ψmax +
+
+ +
+
+
1
ψs1ψs2 > 0 1 σs1σs2 > 0
h0 hmin
0 h 0 h
h0 hmax
2 ψs1ψs2 < 0 σs1σs2 < 0
2
+ – +
+ – + –
+
+
–
–
∏σmin +
+
–
–
(a) (b)
FIGURE 4.33 Electrostatic disjoining pressure at (a) fixed surface potential, Π ψel, and (b) fixed surface charge
density, Π σel, both of them plotted vs. the film thickness h. ψs1 and ψs2 are the potentials of the two surfaces;
σs1 and σs2 are the respective surface charge densities.
322 Handbook of Surface and Colloid Chemistry
When σ1σ2 > 0, Equation 4.202 yields Π σel > 0 for every h (see Figure 4.33b). However, when
σ1σ2 < 0, Π σel is repulsive for small thickness, h < h 0 and attractive for larger separations, h > h 0; h 0
is determined by the equation κh 0 = ln(−σs2/σs1); |σs2| > |σs1|. The electrostatic disjoining pressure in
this case has a minimum value
1 1 ⎛ σ ⎞
Π σel (min) = σs1σs 2 , at hmin = arccosh ⎜ − s 2 ⎟ (4.203)
εε0 κ ⎝ σs1 ⎠
Finally, it should be noted that all curves depicted in Figure 4.24 decay exponentially at h → ∞.
An asymptotic expression for Z:Z electrolytes, which generalizes Equation 4.197, holds [380,399]:
⎛ Zeψ sk ⎞
Π el (h) = 64n0 kTγ1γ 2e − κ h , γ k ≡ tanh ⎜ ⎟ , k = 1, 2 (4.204)
⎝ 4kT ⎠
Equation 4.204 is valid for both low and high surface potentials, only if exp(−κh) ≪ 1.
128πR1R2
U el (h0 ) = n0 kTγ1γ 2e − κ h (4.205)
κ2 ( R1 + R2 )
Equation 4.205 is valid for any values of the surface potentials ψs1 and ψs2 but only for exp(κh) ≫ 1.
Complementary expressions, which are valid for every h ≪ R1, R2, but for small surface potentials,
can be derived by integrating Equations 4.200 and 4.202, instead of Equation 4.204. In this way, for
ψs1 = constant and ψs2 = constant, we can derive [402]:
πεε0 R1R2
U elψ (h0 ) = ⎡(ψ s1 + ψ s 2 )2 ln(1 + e − κh0 ) + (ψ s1 − ψ s 2 )2 ln(1 − e − κh0 ) ⎤ (4.206)
R1 + R2 ⎣ ⎦
or, alternatively, for σs1 = constant and σs2 = constant we obtain [403]:
−πR1R2
U elσ (h0 ) = ⎡(σs1 + σs 2 )2 ln(1 − e − κh0 ) + (σs1 − σs 2 )2 ln(1 + e − κh0 ) ⎤ (4.207)
εε0 κ2 ( R1 + R2 ) ⎣ ⎦
The range of validity of the different approximations involved in the derivations of Equations 4.205
through 4.207 is discussed in the book by Russel et al. [404].
As mentioned earlier, Equations 4.205 through 4.207 hold for h 0 ≪ R. In the opposite case, when
h 0 is comparable to or larger than the particle radius R, we can use the equation [14]:
4πεε0ψ 2s R 2 − κ h0
U el (h0 ) = e (4.208)
2 R + h0
stemming from the theory of Debye and Hückel [405] for two identical particles. Equation 4.208
was derived by using the superposition approximation (valid for weak overlap of the two EDLss)
and the linearized Poisson–Boltzmann equation. A simple approximate formula, representing in
Chemical Physics of Colloid Systems and Interfaces 323
fact interpolation between Equations 4.208 and 4.206 (the latter for R1 = R2 = R), has been derived
by McCartney and Levine [406]:
R + h0 ⎛ Re − κ h0 ⎞
U elψ (h0 ) = 4πεε0 Rψ 2s ln ⎜ 1 + ⎟ (4.209)
2 R + h0 ⎝ R + h0 ⎠
Equation 4.209 has the advantage to give a good approximation for every h 0 provided that the
Poisson–Boltzmann equation can be linearized. Similar expressions for the energy of electrostatic
interaction between two deformed droplets or bubbles (Figure 4.32) can be derived [392].
4.4.4 DLVO Theory
The first quantitative theory of interactions in thin liquid films and dispersions is the DLVO theory
called after the names of the authors Derjaguin and Landau [407] and Verwey and Overbeek [399].
In this theory, the total interaction is supposed to be a superposition of van der Waals and double
layer interactions. In other words, the total disjoining pressure and the total interaction energy are
presented in the form:
Π = Π vw + Π el , U = U vw + U el (4.210)
A typical curve, Π versus h, exhibits a maximum representing a barrier against coagulation and two
minima, called primary and secondary minimum (see Figure 4.17); the U versus h curve has a simi-
lar shape. The primary minimum appears if strong short-range repulsive forces (e.g., steric forces)
are present. With small particles, the depth of the secondary minimum is usually small (Umin < kT).
If the particles cannot overcome the barrier, coagulation (flocculation) does not take place, and the
dispersion is stable due to the electrostatic repulsion, which gives rise to the barrier. With larger
colloidal particles (R > 0.1 μm), the secondary minimum could be deep enough to cause coagulation
and even formation of ordered structures of particles [408].
By addition of electrolyte or by decreasing the surface potential of the particles, we can suppress
the electrostatic repulsion and thus decrease the height of the barrier. According to DLVO theory,
the critical condition determining the onset of rapid coagulation is
dU
U (hmax ) = 0, = 0 (4.211)
dh hmax
For a Z:Z electrolyte, substituting κ 2 = (2Z2 e2 n 0)/(ε0 εkT) into Equation 4.212, we obtain:
1 ⎛ Zeψ s ⎞
n0 (critical) ∝ tanh 4 ⎜ ⎟ (4.213)
⎝ 4kT
6
Z ⎠
When ψs is high enough, the hyperbolic tangent equals 1 and Equation 4.213 yields n0 (critical) ∝
Z −6 which is, in fact, the empirical rule established earlier by Schulze [409] and Hardy [410].
324 Handbook of Surface and Colloid Chemistry
k
kTκ2
Π osm = kT ∑i =1
ni −
24π
(4.214)
The first term in the right-hand side of the Equation 4.214 corresponds to an ideal solution, whereas
the second term takes into account the effect of electrostatic interactions between the ions (the same
effect is accounted for thermodynamically by the activity coefficient, see Equation 4.31).
The expression for Πel in the DLVO theory (Equation 4.189) obviously corresponds to an ideal
solution, the contribution of the ionic correlations being neglected. Hence, in a more general theory
instead of Equation 4.210, we could write:
Π = Π vw + Π el + Π cor (4.215)
where Πcor is the contribution of the ionic correlations to the disjoining pressure. The theory of Πcor
takes into account the following effects: (1) the different ionic concentration (and hence the differ-
ent Debye screening) in the film compared to that in the bulk solution; (2) the energy of deforma-
tion of the counterion atmosphere due to the image forces; and (3) the energy of the long-range
correlations between charge density fluctuations in the two opposite EDLs. For calculating Πcor,
both numerical solutions [411,412] and analytical expressions [413–415] have been obtained. For
example, in the case when the electrolyte is symmetrical (Z:Z) and exp(−κh) ≪ 1 we can use the
asymptotic formula [413]
Z 2e2 κ
Π cor = Π el (ln 2 + 2 IC ) + O(e − κh ) (4.216)
16πεε0 kT
1 2 − 2z3 + z 1
IC = (1 + J ) ln 2 + − (1 − J ) ln( z + z 2 )
2 2 z(2 z 2 − 1)2 2
z2 − 1 z −1
− [1 + J + 4(2 z 2 − 1)−3 ]arctan
z z +1
1/ 2
2
2z2 − 3 ⎡ ⎛ eσ ⎞ ⎤
J≡ , z ≡ ⎢1 + ⎜ s
⎟ ⎥
(2 z − 1)
2 3
⎢⎣ ⎝ 2εε0 kTκ ⎠ ⎥⎦
Chemical Physics of Colloid Systems and Interfaces 325
The results for the case of symmetric electrolytes are the following. Πcor is negative and corresponds
to attraction, which can be comparable by magnitude with Πvw. In the case of 1:1 electrolyte, Πcor is
usually a small correction to Πel. In the case of 2:2 electrolyte, however, the situation can be quite
different: the attractive forces, Πcor + Πvw, prevails over Πel and the total disjoining pressure, Π,
becomes negative. The effect of Πcor is even larger in the presence of ions of higher valence. Short-
range net attractive ion-correlation forces have been measured by Marra [416,417] and Kjellander
et al. [418,419] between highly charged anionic bilayer surfaces in CaCl2 solutions. These forces are
believed to be responsible for the strong adhesion of some surfaces (clay and bilayer membranes) in
the presence of divalent counterions [36,418,420]. In Ref. [421], the attraction mechanism and the
structure of counterionic correlations are discussed in the limit of strong coupling based on numeri-
cal and analytical investigations and for various geometries (planar, spherical, and cylindrical) of
charged objects.
The theory predicts ion-correlation attraction not only across water films with overlapping EDLs,
but also across oily films intervening between two water phases. In the latter case, Πcor is not zero
because the ions belonging to the two outer double layers interact across the thin dielectric (oil) film.
The theory for such a film [422] predicts that Πcor is negative (attractive) and strongly dependent on
the dielectric permittivity of the oil film; Πcor can be comparable by magnitude with Πvw; Πel = 0 in
this case.
Brush
L
(a) (b) (c)
Loop
Tail
Train
(d) (e)
FIGURE 4.34 Polymeric chains adsorbed at an interface: (a) terminally anchored polymer chain of mean
end-to-end distance L; (b) a brush of anchored chains; (c) adsorbed (but not anchored) polymer coils; (d) con-
figuration with a loop, trains and tails; (e) bridging of two surfaces by adsorbed polymer chains.
326 Handbook of Surface and Colloid Chemistry
anchored (grafted) to the liquid interface by their hydrophobic moieties. When the surface con-
centration of adsorbed molecules is high enough, the hydrophilic chains are called to form a brush
(Figure 4.34b). The coils of macromolecules, like proteins, can also adsorb at a liquid surface
(Figure 4.34c). Sometimes, the configurations of the adsorbed polymers are very different from the
statistical coil: loops, trains, and tails can be distinguished (Figure 4.34d).
The osmotic pressure of either dilute or concentrated polymer solutions can be expressed in the
form [426]:
Posm 1 1 1
= + nv + n2 w + (4.217)
nkT N 2 3
Here
N is the number of segments in the polymer chain
n is the number segment density
v and w account for the pair and triplet interactions, respectively, between segments
In fact, v and w are counterparts of the second and third virial coefficients in the theory of imper-
fect gases [11]; v and w can be calculated if information about the polymer chain and the solvent is
available [404]:
vm
w1/ 2 = , v = w1/ 2 (1 − 2χ) (4.218)
NA
where
v (m3/kg) is the specific volume per segment
m (kg/mol) is the molecular weight per segment
NA is the Avogadro number
χ is the Flory parameter
The latter depends on both the temperature and the energy of solvent–segment interaction.
Then, v can be zero (see Equation 4.218) for some special temperature, called the theta temperature.
The solvent at the theta temperature is known as the theta solvent or ideal solvent. The theta tem-
perature for polymer solutions is a counterpart of the Boil temperature for imperfect gases: this
is the temperature at which the intermolecular (intersegment) attraction and repulsion are exactly
counterbalanced. In a good solvent, however, the repulsion due mainly to the excluded volume effect
dominates the attraction and v > 0. In contrast, in a poor solvent the intersegment attraction prevails,
so v < 0.
L ≈ L0 ≡ l N (4.219)
In a good solvent L > L 0, whereas in a poor solvent L < L 0. In addition, L depends on the surface
concentration, Γ, of the adsorbed chains, that is, L is different for an isolated molecule and for a
Chemical Physics of Colloid Systems and Interfaces 327
brush (see Figure 4.34a and b). The mean field approach [404] applied to polymer solutions provides
the following equation for calculating L
⎛ 1 2 ⎞ −1 1
L3 − ⎜ 1 + Γ ⎟ L = 6 v (4.220)
⎝ 9 ⎠
where L, Γ
, and v are the dimensionless values of L, Γ, and v defined as follows:
3/2
L = ΓN w , v = vΓN (4.221)
L = , Γ
l N l l
⎡ π2 L20 ⎛ 8π L20 ⎞ ⎤
f (h) = ΓkT ⎢ − ln ⎜ 2 ⎟⎥
for h < L0 3 (4.222)
⎢⎣ 3 h ⎝ 3 h ⎠ ⎦⎥
2
⎛ 3h2 ⎞
f (h) = 4ΓkT exp ⎜ − 2 ⎟ for h > L0 3 (4.223)
⎝ 2 L0 ⎠
where L 0 is the end-to-end distance as defined by Equation 4.219. The boundary between the power-
law regime (f ∝ 1/h2) and the exponential decay regime is at h = L0 3 ≈ 1.7L0, the latter being
slightly less than 2L 0, which is the intuitively expected onset of the steric overlap. The first term in
the right-hand side of Equation 4.222 comes from the osmotic repulsion between the brushes, which
opposes the approach of the two surfaces; the second term is negative and accounts effectively for
the decrease of the elastic energy of the initially extended chains when the thickness of each of the
two brushes, pressed against each other, decreases.
In the case of good solvent, the disjoining pressure Π = −df/dh can be calculated by means of
Alexander–de Gennes theory as [429,430]:
3/ 4
⎡ 2L 9/ 4 ⎛ ⎞ ⎤
Π(h) = kTΓ 3/2
⎢⎛⎜ g ⎞⎟ − ⎜ h ⎟ ⎥ for h < 2 Lg , Lg = N (Γl 5 )1/ 3 (4.224)
⎢⎝ h ⎠
⎣ ⎝ 2 Lg ⎠ ⎥⎦
where Lg is the thickness of a brush in a good solvent [431]. The positive and the negative terms in
the right-hand side of Equation 4.224 correspond to osmotic repulsion and elastic attraction. The
validity of Alexander–de Gennes theory was experimentally confirmed by Taunton et al. [432] who
measured the forces between two brush layers grafted on the surfaces of two crossed mica cylinders.
In the case of adsorbed molecules, like those in Figure 4.34c, which are not anchored to the
surface, the measured surface forces depend significantly on the rate of approaching of the two
surfaces [433,434]. The latter effect can be attributed to the comparatively low rate of exchange of
polymer between the adsorption layer and the bulk solution. This leads to a hysteresis of the surface
328 Handbook of Surface and Colloid Chemistry
force: different interaction on approach and separation of the two surfaces [36]. In addition, we can
observe two regimes of steric repulsion: (1) weaker repulsion at larger separations due to the overlap
of the tails (Figure 4.34d) and (2) stronger repulsion at smaller separations indicating overlap of the
loops [435].
where
Πos is the disjoining pressure component due to the oscillatory structural forces
ns(h) is the particle number density in the subsurface layer as a function of the distance between
the walls, h
Figure 4.35 illustrates the variation of ns with h and the resulting disjoining pressure, Πos. We see
that in the limit of very small separations, as the last layer of particles is eventually squeezed out,
ns → 0 and
In other words, at small separations Πos is negative (attractive). Equation 4.226 holds for both solva-
tion forces and colloid structural forces. In the latter case, Equation 4.226 represents the osmotic
pressure of the colloid particles and the resulting attractive force is known as the depletion force
(Section 4.4.5.3.3).
The wall induces structuring in the neighboring fluid only if the magnitude of the surface rough-
ness is negligible in comparison with the particle diameter, d. Indeed, when surface irregularities
are present, the oscillations are smeared out and oscillatory structural force does not appear. If the
film surfaces are fluid, the role of the surface roughness is played by the interfacial fluctuation cap-
illary waves, whose amplitude (between 1 and 5 Å) is comparable with the diameter of the solvent
molecules. For this reason, oscillatory solvation forces (due to structuring of solvent molecules) are
observed only with liquid films, which are confined between smooth solid surfaces [36]. In order
for structural forces to be observed in foam or emulsion films, the diameter of the colloidal particles
must be much larger than the amplitude of the surface corrugations. The period of the oscillations
is always about the particle diameter [36,441].
Chemical Physics of Colloid Systems and Interfaces 329
a b c d e f g
h=0 h=d h = 2d h = 3d
(a)
∏os
b
Repulsive
d
f
0
g
e
c
Attractive
a
0 1 2 3 4 5 h/d
(b)
FIGURE 4.35 (a) Sketch of the consecutive stages of the thinning of a liquid film containing spherical
particles; (b) plot of the related oscillatory structural component of disjoining pressure, Πos, vs. the film
thickness h; see Ref. [36] for details.
The theories developed for calculating the oscillatory force are based on modeling by means of
the integral equations of statistical mechanics [449–453] or numerical simulations [454–457]. As a
rule, these approaches are related to complicated theoretical expressions or numerical procedures, in
contrast with the Derjaguin–Landau–Verwey–Overbeek (DLVO) theory, one of its main advantages
being its simplicity [36]. To overcome this difficulty, some relatively simple semiempirical expres-
sions have been proposed [458,459] on the basis of fits of theoretical results for hard-sphere fluids.
The following semiempirical formula for the oscillatory structural component of disjoining pres-
sure reads was proposed in Ref. [458]:
⎛ 2πh ⎞ ⎛ d3 h ⎞
Π os (h) = P0 cos ⎜ ⎟ exp ⎜ 2 − ⎟ for h > d
⎝ d1 ⎠ ⎝ d1 d2 d2 ⎠ (4.227)
where
d is the diameter of the hard spheres
d1 and d2 are the period and the decay length of the oscillations which are related to the particle
volume fraction, ϕ, as follows [458]
d1 2 d2 0.4866
= + 0.237Δφ + 0.633(Δφ)2 ; = − 0.420 (4.228)
d 3 d Δφ
Here
Δϕ = ϕmax − ϕ with fmax = π / (3 2 ) being the value of ϕ at close packing
P0 is the particle osmotic pressure determined by means of the Carnahan–Starling formula [460]
330 Handbook of Surface and Colloid Chemistry
1 + φ + φ2 − φ3 6φ
P0 = nkT , n = 3 (4.229)
(1 − φ)3
πd
where n is the particle number density. For h < d, when the particles are expelled from the slit into
the neighboring bulk suspension, Equation 4.227 describes the depletion attraction. On the other
hand, for h > d the structural disjoining pressure oscillates around P0 as defined by Equation 4.229
in agreement with the results of Kjellander and Sarman [451]. The finite discontinuity of Πos at
h = d is not surprising as, at this point, the interaction is switched over from oscillatory to depletion
regime. It should be noted that in oscillatory regime, the concentration dependence of Πos is domi-
nated by the decay length d2 in the exponent (see Equations 4.227 and 4.228). Roughly speaking, for
a given distance h, the oscillatory disjoining pressure Πos increases five times when ϕ is increased
with 10% [458]. The comparison with available numerical data showed that Equation 4.227 is accu-
rate everywhere except in the region of the first (the highest) oscillatory maximum.
A semiempirical expression for Πos(H), which is accurate in the whole region 0 ≤ H < ∞, includ-
ing the region of the first maximum, was proposed by Trokhymchuk et al. [459]:
Here, Π0, Π1, ω, φ2, κ, and δ are universal functions of particle volume fraction, ϕ, tabulated in
Ref. [459]. Equation 4.230 compares very well with existing computer simulation data [459].
The interactions between the micelles in a nonionic surfactant solution can be adequately
described as interactions in a hard-sphere fluid. Experiments with foam films formed from aque-
ous solutions of two nonionic surfactants, Brij 35 and Tween 20, which contain spherical micelles
of diameters in the range 7–9 nm, have been carried out [461]. From the measured contact angles,
the micelle aggregation number and volume fraction have been determined. In addition, from the
measured disjoining pressure isotherms the micelle diameter has been found. In other words, the
liquid-film measurements could give information about the micelles, which is analogous to that
obtainable by dynamic and static light scattering. As an illustration, Figure 4.36 shows the com-
parison of theory and experiment for the nonionic surfactant Tween 20. The experimental Πos(h)
dependence is obtained by using the porous-plate cell by Mysels and Jones, known also as thin-film
pressure balance [462]. The points on the horizontal axis correspond to the thickness of the meta-
stable states of the film measured by the Scheludko–Exerowa capillary cell [215,216]. The solid line
is calculated by means of Equation 4.230 for particle (micelle) diameter determined by light scat-
tering and micelle volume fraction determined from the contact angle of the thin liquid film [461].
The short-range repulsion at h ≈ 10 nm (Figure 4.36) corresponds to the steric repulsion between the
hydrophilic headgroups of the surfactant molecules. Excellent agreement between Equation 4.230
and experimental data obtained by colloidal-probe atomic force microscopy (CP-AFM) for micellar
solutions of Brij 35 has also been reported [463].
The predictions of different quantitative criteria for stability–instability transitions were inves-
tigated [461], having in mind that the oscillatory forces exhibit both maxima, which play the role
of barriers to coagulation, and minima that could produce flocculation or coalescence in colloidal
dispersions (emulsions, foams, suspensions). The interplay of the oscillatory force with the van der
Waals surface force was taken into account. Two different kinetic criteria were considered, which
give similar and physically reasonable results about the stability–instability transitions. Diagrams
were constructed, which show the values of the micelle volume fraction, for which the oscillatory
barriers can prevent the particles from coming into close contact, or for which a strong floccula-
tion in the depletion minimum or a weak flocculation in the first oscillatory minimum could be
observed [461].
Chemical Physics of Colloid Systems and Interfaces 331
1000
500
–500
10 20 30 40 50 60
Film thickness, h (nm)
FIGURE 4.36 Plot of disjoining pressure, Π, vs. film thickness, h: comparison of experimental data for a
foam film from Ref. [461] (thin-film pressure balance) with the theoretical curve (the solid line) calculated
by means of Equation 4.230. The film is formed from 200 mM aqueous solution of the nonionic surfactant
Tween 20. The volume fraction of the micelles (ϕ = 0.334) is determined from the film contact angle; the
micelle diameter (d = 7.2 nm) is determined by dynamic light scattering. The points on the horizontal axis
denote the respective values of h for the stratification steps measured by a thin-film pressure balance.
d/2
D/2 h0 D/2
FIGURE 4.37 Overlap of the depletion zones around two particles of diameter D separated at a surface-to-
surface distance h 0; the smaller particles have diameter d.
π
S (h0 ) = (2 D + d + h0 )(d − h0 ) for 0 ≤ h0 ≤ d
4
S (h0 ) = 0 for d ≤ h0 (4.232)
Here
h 0 is the shortest distance between the surfaces of the larger particles
n is the number density of the smaller particles
By integrating Equation 4.233, we can derive an expression for the depletion interaction energy
between the two larger particles, Udep(h 0). For D ≫ d, this expression reads:
U dep (h0 ) 3 D
≈ − φ 3 (d − h0 )2 0 ≤ h0 ≤ d (4.233)
kT 2 d
where ϕ = πnd3/6 is the volume fraction of the small particles. The maximum value of Udep at h 0 = 0
is Udep(0)/kT ≈ −3ϕD/(2d). For example, if D/d = 50 and ϕ = 0.1, then Udep(0) = 7.5kT. This depletion
attraction turns out to be large enough to cause flocculation in dispersions. De Hek and Vrij [443]
studied systematically the flocculation of sterically stabilized silica suspensions in cyclohexane by
polystyrene molecules. Patel and Russel [465] investigated the phase separation and rheology of
aqueous polystyrene latex suspensions in the presence of polymer (Dextran T-500).
The stability of dispersions is often determined by the competition between electrostatic repul-
sion and depletion attraction [466]. Interplay of steric repulsion and depletion attraction was studied
theoretically by van Lent et al. [467] for the case of polymer solution between two surfaces coated
with anchored polymer layers. Joanny et al. [468] and Russel et al. [404] re-examined the theory
of depletion interaction by taking into account the internal degrees of freedom of the polymer
molecules. Their analysis confirmed the earlier results of Asakura and Oosawa [442].
The depletion interaction is always present when a film is formed from micellar surfactant
solution; the micelles play the role of the smaller particles. At higher micellar concentrations, the
volume exclusion interaction becomes more complicated: it follows the oscillatory curve depicted
in Figure 4.34. In this case only, the first minimum (that at h → 0) corresponds to the conventional
depletion force.
Chemical Physics of Colloid Systems and Interfaces 333
protein solutions, surfactant–polymer mixtures, and ABA amphiphilic block copolymers, where A and
B denote, respectively, hydrophilic and hydrophobic parts of the molecule [503–511].
In the case of liquid films that contain charged colloid particles (micelles), the oscillatory
period, Δh, is considerably greater than the particle diameter [437,438]. In this case, the theoreti-
cal prediction of Δh demands the use of density functional theory calculations and/or Monte Carlo
simulations [456,512]. However, the theory, simulations, and experiments showed that a simple
inverse-cubic-root relation, Δh = cm−1/ 3, exists between Δh and the bulk number concentration of
micelles (particles), cm [437,438,498,512–514].
The validity of the semiempirical Δh = cm−1/ 3 law is limited at low and high particle concentrations,
characterized by the effective particle volume fraction (particle + counterion atmosphere) [502].
The decrease of the effective particle volume fraction can be experimentally accomplished not only
by dilution, but also by addition of electrolyte that leads to shrinking of the counterion a tmosphere
[500]. The inverse-cubic-root law, Δh = cm−1/ 3, which can be interpreted as an osmotic pressure balance
between the film and the bulk [513], is fulfilled in a wide range of particle/micelle concentrations
that coincide with the range where stratification (step-wise thinning) of free liquid films formed from
particle suspension and micellar solution is observed [139,437–439,493–496,513,514], and where
the surface force measured by CP-AFM [463,490,498–502] or SFA [488,489] exhibits oscillations.
Because the validity of the Δh = cm−1/ 3 law has been proven in numerous studies, it can be used for
determining the aggregation number, Nagg, of ionic surfactant micelles [513]. Indeed, cm = (cs – CMC)/
Nagg, where cs and CMC are the total input surfactant concentration and the CMC expressed as number of
molecules per unit volume. The combination of the latter expression with Δh = cm−1/ 3 law yields [513,514]:
Here
dh is the hydrodynamic diameter of the micelle
k is the Boltzmann constant
T is the absolute temperature
uel(r) is the energy of electrostatic interaction of two micelles in the solution
ψ(r) is the distribution of the electrostatic potential around a given ionic micelle in the solution
r is radial coordinate;
L B ≡ e2/(4πε0 εkT) is the Bjerrum length (L B = 0.72 nm for water at 25°C); ε0 is the permittivity
of vacuum; ε is the dielectric constant of the solvent (water); and e is the elementary charge
Chemical Physics of Colloid Systems and Interfaces 335
TABLE 4.7
Measured Period of the Structural Force,a Δh, and Micelle Aggregation Number,
Nagg, Calculated from Equation 4.235
Experimental Δh (nm) Aggregation Number Nagg
cs (mM) from Ref. [513] from Equation 4.235 Nagg from Literature
Sodium dodecyl sulfate (SDS)
30 15.3 48 50 [515], 55 [516], 59 [517]
40 14.7 61 60 [518], 62 [519], 64 [517]
50 13.7 65 64 [515,517,520], 65 [521]
100 10.6 65 64 [515,517,520], 65 [521]
Cetyl trimethylammonium bromide (CTAB)
10 25.8 95 92 [522], 95 [523,524], 98 [525]
20 21.8 119 —
30 19.9 137 100 [524]
40 18.0 136 —
50 16.6 135 139 [526], 140 [524]
Cetyl pyridinium chloride (CPC)
10 21.2 52 45–90 [527], 56 [528]
20 18.7 75 78 [513]
30 16.6 80 82 [525]
40 15.8 93 —
50 14.6 93 87 [529]
Equation 4.237 reduces the two-particle problem to the single-particle problem; see Refs. [513,514]
for details.
It was found [513] that the relationship deff = cm−1/ 3 = Δh is satisfied in the whole concentration
range where stratifying films are observed; deff is calculated from Equation 4.236, whereas Δh is
experimentally determined from the stratification steps of foam films. In contrast, for deff < cm−1/ 3 the
foam films do not stratify and the oscillations of disjoining pressure vanish. This may happen at low
micelle concentrations, or at sufficiently high salt concentrations [513]. Thus, the relation between
deff and cm can be used as a criterion for the existence of oscillatory structural force with charged
colloidal particles.
of interactions between phospholipid bilayers by Parsegian et al. [536,537]. Israelachvili et al.
[538–540] and Pashley [541–543] examined the validity of the DLVO theory [399,407] at small film
thickness by an SFA in experiments with films from aqueous electrolyte solutions confined between
two curved mica surfaces, bare or covered by adsorbed layers. At electrolyte concentrations below
10−4 M, they observed the typical DLVO maximum. However, at electrolyte concentrations higher
than 10−3 M they did not observe the expected DLVO maximum and primary minimum [540].
Instead a strong short-range repulsion was detected, which can be empirically described by expo-
nential law [36]:
where the decay length λ0 ≈ 0.6–1.1 nm for 1:1 electrolytes and f0 depends on the hydration of the
surfaces but is usually about 3–30 mJ/m2. Similar repulsion was detected between silica sheets
[544,545] and dihexadecyl phosphate monolayers deposited on a solid surface [546].
The conventional electrostatic (double layer) repulsion is suppressed if the solution’s ionic
strength is increased [399,407]. In contrast, the hydration repulsion is detected at higher ionic
strengths [540], at which it is the main stabilizing factor in liquid films and colloidal dispersions.
Such strong repulsion at high salt concentrations was observed between apoferritin molecules in
solutions [547,548] and between the adsorption layers of this and other proteins on solid surfaces
and colloidal particles [549,550]. In general, the hydration force plays an important role for the sta-
bility of proteins in physiological media. Hydration forces have also been observed between DNA
molecules in aqueous solutions [551,552]. Effects of monovalent anions of the Hofmeister series and
other solutes on the hydration repulsion between phospholipid bilayers have been experimentally
investigated [553–556]. The hydration repulsion affects the stability of emulsions [557]; the rheology
of concentrated suspensions [558]; the interactions of biological cells [559]; and the fusion rate of
vesicles in the cellular inter-organelle traffic [560]. Additional information can be found in several
review articles [561–566].
The physical importance of the hydration force is that it stabilizes dispersions at high elec-
trolyte concentrations preventing coagulation in the primary minimum of the DLVO curve
(Figure 4.17). For example, Healy et al. [567] found that even high electrolyte concentrations
cannot cause coagulation of amphoteric latex particles due to binding of strongly hydrated Li+
ions at the particle surfaces. If the Li+ ions are replaced by weakly hydrated Cs+ ions, the hydra-
tion repulsion becomes negligible, compared with the van der Waals attraction, and the particles
coagulate as predicted by the DLVO theory. Hence, the hydration repulsion can be regulated by
ion exchange.
The aforementioned studies indicate that hydration repulsion is observed in (at least) two types
of systems. (1) charged interfaces at relatively high electrolyte concentrations, where electrostatic
and osmotic effects related to the presence of bound and mobile counterions are expected to play
an essential role and (2) electroneutral surfaces with zwitterionic surface groups, like phospholipid
bilayers, where the water structuring near the polar surface and surface charge discreteness could
be the main sources of the observed repulsion. Correspondingly, for the theoretical explanation of
the hydration repulsion different models have been proposed, which could be adequate for different
systems. The most important theoretical models are as follows:
1.
Water-structuring models. In these models, the short-range repulsive interaction is attrib-
uted to alignment of water dipoles in the vicinity of a hydrophilic surface, where the
range of the surface force is determined by the orientation correlation length of the sol-
vent molecules [568–570]. Due to the strong orientation of water molecules near polar
surfaces, we could expect that there are fewer configurations available to maintain the
Chemical Physics of Colloid Systems and Interfaces 337
bulk water structure, which represents a loss of entropy that leads to a repulsive force
[571]. The existence of such effects has been confirmed by molecular dynamics (MD)
simulations [572,573].
2.
Image-charge models. These models take into account the discreteness of surface charges,
which induces orientation in the adjacent water dipoles [574–577]. Dipoles due to zwit-
terionic surface groups, for example, phospholipid headgroups [578], have been also taken
into consideration in models of the electrostatic interaction between planar dipole lattices
[579–583].
3.
Dielectric-saturation models attribute the hydration repulsion to the presence of a layer
with lower dielectric constant, ε, in the vicinity of the interfaces. Models with a stepwise
[584,585] and continuous [586] variation of ε have been proposed.
4.
Excluded-volume models take into account the fact that the finite size of the ions leads to a
lower counterion concentration near a charged surface, and to a weaker Debye screening of
the electrostatic field (in comparison with the point-ion model), which results in a stronger
repulsion between two charged surfaces at short separations [587,588].
5.
Coion expulsion model. This model [589] assumes that at sufficiently small thicknesses all
coions are pressed out of the film so that it contains only counterions dissociated from the
ionized surface groups. Under such conditions, the screening of the electric field of the film
surface weakens, which considerably enhances the electrostatic repulsion in comparison
with that predicted by the DLVO theory. Such reduced screening of the electric field could
exist only in a narrow range of film thicknesses, which practically coincides with the range
where the hydration repulsion is observed.
Let us consider in more details the coion expulsion model, also called “reduced screening model”
[589]. This model was developed to explain the strong short-range repulsion detected in foam
films; see Figure 4.38. It was found [589] that the excluded volume model [588] cannot explain the
observed large deviations from the DLVO theory. Quantitative data interpretation was obtained by
8
Coion expulsion model
7
Disjoining pressure, ∏ (kPa)
6
CsCl NaCl LiCl
5
1 mM SDS
4 100 mM salt
1
DLVO
0
9.6 9.8 10.0 10.2 10.4 10.6 10.8
Total film thickness, ht (nm)
FIGURE 4.38 Plot of the disjoining pressure Π vs. the total thickness ht of foam films formed from 1 mM
aqueous solutions of SDS in the presence of 100 mM electrolyte: LiCl, NaCl and CsCl. At greater thick-
nesses, the Π(ht) dependence obeys the DLVO theory, whereas at the small thickness the steep parts of the
curves are in agreement with the coion-expulsion model. The distances between the experimental curves mea-
sured with different electrolytes are due to the different sizes of the hydrated Cs+, Na+, and Li+ counterions.
(From Kralchevsky, P.A. et al., Curr. Opin. Colloid Interface Sci., 16, 517, 2011.)
338 Handbook of Surface and Colloid Chemistry
assuming that all coions have been pressed out of the thin film (see Figure 4.1). In such case, the
Poisson–Boltzmann equation acquires the form
d 2Φ 1 2 Φ
= κ e (4.239)
dx 2 2
where
Φ is the dimensionless electric potential
κ is the Debye screening length
e|ψ| 2e2 a2 ∞
Φ≡ κ2 ≡ (4.240)
kT εε0 kT
The notations are the same as in Equation 4.237; in particular, ψ(x) is the dimensional electrostatic
potential, x is a coordinate perpendicular to the surfaces of the plane-parallel film, and a2∞ = γ±c2∞ is
the activity of counterions in the bulk solution, which is in contact with the film; see Equation 4.30.
The first integral of Equation 4.239 reads:
dΦ
= κ(eΦ − eΦ m )1/ 2 (4.241)
dx
Here, Φm is the value of Φ in the middle of the film. Integrating Equation 4.241 between the middle
of the film and the film surface, one can derive [589]
exp Φ m
exp Φ s = (4.242)
cos2 [( κh / 4) exp(Φ m / 2)]
where Φs is the value of Φ at the film surface, whereas h is the thickness of the aqueous core of the
foam film. The right-hand side of Equation 4.242 has singularities for those h values, for which the
cosine in the denominator is equal to zero. For this reason, the region of physical applicability of
Equation 4.10 (and of the RS model) corresponds to h values, for which the argument of the cosine
is between 0 and π/2:
2π ⎛ Φ ⎞
0<h< exp ⎜ − m ⎟ (4.243)
κ ⎝ 2 ⎠
In the experiments in Ref. [589], 100 mM electrolyte is present, which leads to κ−1 ≈ 1 nm, and
the midplane potential is Φm ≈ 0.7, so that exp(−Φm /2) ≈ 0.705. Then, Equation 4.243 reduces to
0 < h < 4.4 nm. This range of h values includes the range of thicknesses, where the hydration force
is operative. In the experiments in Ref. [589], the hydration repulsion appears in the interval 0 < h
< 3.71 nm irrespective of the kind of counterion (Li+, Na+, or Cs+). Note that in Figure 4.38 the data
are plotted versus the total film thickness, ht, which includes not only the water core, but also the
two surfactant adsorption layers at the film surfaces.
As already mentioned, the electrostatic component of disjoining pressure can be defined as the
excess osmotic pressure in the film midplane with respect to the bulk solution, see Equation 4.189.
Hence, if all coions are expelled from the film, the expression for the disjoining pressure acquires
the form [589]:
For a given surface electric potential, Φs, Equations 4.242 and 4.244 determine the Π(h) dependence
in a parametric form: h = h(Φm) and Π = Π(Φm). As seen in Figure 4.38, excellent agreement between
theory and experiment has been achieved for reasonable parameter values [589].
fhydrophobic = −2 γe − h / λ0 (4.245)
where typically γ = 10–50 mJ/m2, and λ0 = 1–2 nm in the range 0 < h < 10 nm. Larger decay length,
λ0 = 12–16 nm, was reported by Christenson et al. [595] for the range 20 < h < 90 nm. This long-
range attraction could entirely dominate the van der Waals forces. Ducker et al. [596] measured the
force between hydrophobic and hydrophilic silica particles and air bubbles by means of an atomic
force microscope.
It was found experimentally that 1:1 and 2:2 electrolytes reduce considerably the long-range part
of the hydrophobic attraction [594,595]. The results suggest that this reduction could be due to ion
adsorption or ion exchange at the surfaces rather than to the presence of electrolyte in the solution
itself. Therefore, the physical implication (which might seem trivial) is that the hydrophobic attrac-
tion across aqueous films can be suppressed by making the surfaces more hydrophilic. Besides,
some special polar solutes are found to suppress the hydrophobic interaction at molecular level in
the bulk solution, for example, urea, (NH2)2CO, dissolved in water can cause proteins to unfold.
The polar solutes are believed to destroy the hydrogen-bond structuring in water; therefore they are
sometimes called chaotropic agents [36].
There is no generally accepted explanation of the hydrophobic surface force. One of the pos-
sible explanations is that the hydrogen bonding in water (and other associated liquids) could be the
main underlying factor [36,597]. The related qualitative picture of the hydrophobic interaction is
the following. If there were no thermal motion, the water molecules would form an ice-like tetrahe-
dral network with four nearest neighbors per molecule (instead of 12 neighbors at close packing),
because this configuration is favored by the formation of hydrogen bonds. However, due to the
thermal motion a water molecule forms only about 3–3.5 transient hydrogen bonds with its neigh-
bors in the liquid [598] with lifetime of a hydrogen bond being about 10−11 s. When a water molecule
is brought in contact with a non-hydrogen-bonding molecule or surface, the number of its possible
favorable configurations is decreased. This effect also reduces the number of advantageous configu-
rations of the neighbors of the subsurface water molecules and some ordering propagates in the depth
of the liquid. This ordering might be initiated by the orientation of the water dipoles at a water–air
or water–hydrocarbon interface with the oxygen atom being oriented toward the hydrophobic phase
[599–602]. Such ordering in the vicinity of the hydrophobic wall is entropically unfavorable. When
two hydrophobic surfaces approach each other, the entropically unfavored water is ejected into the
bulk, thereby reducing the total free energy of the system. The resulting attraction could in principle
explain the hydrophobic forces. The existing phenomenological theory [597] has been generalized to
the case of asymmetric films [603], and has been applied to interpret experimental data for breakage
of emulsion and foam films at low surfactant and high electrolyte concentrations [604,605].
Another hypothesis for the physical origin of the hydrophobic force considers a pos-
sible role of formation of gaseous capillary bridges between the two hydrophobic surfaces
340 Handbook of Surface and Colloid Chemistry
(see Figure 4.13a) [36,606,607]. In this case, the hydrophobic force would be a kind of capillary-
bridge force; see Chapter 11 in Ref. [37]. Such bridges could appear spontaneously, by nucleation
(spontaneous dewetting), when the distance between the two surfaces becomes smaller than
a certain threshold value, of the order of several hundred nanometers. Gaseous bridges could
appear even if there is no dissolved gas in the water phase; the pressure inside a bridge can be as
low as the equilibrium vapor pressure of water (e.g., P0 = 2337 Pa at 20°C, which is only 2.3% of
the atmospheric pressure) owing to the high interfacial curvature of the nodoid-shaped bridges;
see Section 4.3.1.2.3 and Ref. [37].
For example, at air–water–solid contact angle θ = 90°–110° the maximal length of a nodoid-
shaped capillary bridge, hmax, can be estimated from the analytical asymptotic formula [37,312]:
−2σ cos θ
hmax = (90° < θ < 110°) (4.246)
P − P0
Substituting P = 1 atm for the outer pressure, P0 = 2337 Pa for the inner pressure (the equilibrium
vapor pressure of water at 20°C), σ = 72.75 mN/m for the surface tension of water, and θ = 94°,
from Equation 4.246 we calculate hmax = 103 nm for a vapor-filled bridge between two parallel
hydrophobic plates in water. This value of hmax is close to the distances at which the experimentally
observed long-range hydrophobic attraction begins to operate. A number of studies [608–616] pro-
vide evidence in support of the capillary-bridge origin of the long-range hydrophobic surface force.
In particular, the observation of “steps” in the experimental data was interpreted as an indication for
separate acts of bridge nucleation [612].
As discussed in Ref. [617], at present, the accumulated experimental data indicate for the exis-
tence of three different force-law regimes of the hydrophobic surface force. At film thickness
h ≤ 1−1.5 nm, a pure short-range hydrophobic force is operative, which is probably related to
water structuring effects and hydrogen bonds at the water–hydrophobic interface. At intermediate
distances, 1.5 < h < 15 nm, a long-range hydrophobic force is acting, which is possibly due to an
enhanced Hamaker constant associated with the “proton-hopping” polarizability of water. Finally,
at h > 15 nm a super-long-range attraction is observed, which could be due to gaseous capillary
bridges (bridging cavities) or to the electrostatic patch-charge attraction [618].
3π2 (kT )2
Π und (h) = (4.247)
64kt h3
where kt is the bending elastic modulus of the bilayer as a whole. The experiment [621] and the
theory [37,204] show that kt is of the order of 10−19 J for lipid bilayers. The undulation force has been
measured, and the dependence Πund ∝ h−3 was confirmed experimentally [622–624].
Chemical Physics of Colloid Systems and Interfaces 341
h h
(a) (b)
(c)
FIGURE 4.39 Surface forces due to configurational confinement of thermally exited modes into a narrow
region of space between two approaching interfaces: (a) bending mode of membrane fluctuations giving rise to
the undulation force; (b) squeezing mode of membrane fluctuations producing the peristaltic force; (c) fluctuat-
ing protrusion of adsorbed amphiphilic molecules engendering the protrusion surface force.
2(kT )2
Π per (h) ≈ (4.248)
π 2 ks h 5
The experiment [626] gives values of ks varying between 135 and 500 mN/m, which depend on
temperature and composition of the lipid membrane.
where
λ is the characteristic protrusion length; λ = 0.14 nm at 25°C for surfactants with paraffin chain
Γ denotes the number of protrusion sites per unit area
342 Handbook of Surface and Colloid Chemistry
Note that Πprotr decays exponentially for h ≫ λ, but Πprotr ∝ h−1 for h < λ, that is, Πprotr is divergent at
h → 0. The respective interaction free energy (per unit film area) is
∞
∂ρ
+ div(ρv) = 0 (4.251)
∂t
∂
(ρv) + div(ρvv − P − Pb ) = 0 (4.252)
∂t
where
ρ is the mass density
v is the local mass average velocity
P is the hydrodynamic stress tensor
Pb is the body-force tensor which accounts for the action of body forces such as gravity, electro-
static forces (the Maxwell tensor), etc.
In a fluid at rest, and in the absence of body forces, the only contact force given by the hydrody-
namic stress tensor is the scalar thermodynamic pressure, p, and P can be written as P = −pI, where
I is the unit tensor in space. For a fluid in motion, the viscous forces become operative and
P = − pI + T (4.253)
where T is the viscous stress tensor. From the definition of the stress tensor (Equation 4.253), it
follows that the resultant hydrodynamic force, F, exerted by the surrounding fluid on the particle
surface, S, and the torque, M, applied to it are given by the expressions [631,633]
F = P ⋅ n dS, M = r0 × P ⋅ n dS (4.254)
∫
S
∫
S
where
r0 is the position vector of a point of S with respect to an arbitrarily chosen coordinate origin
n is the vector of the running unit normal to the surface S
Chemical Physics of Colloid Systems and Interfaces 343
In the presence of body forces, the total force, Ftot, and torque, Mtot, acting on the particle surface are
The dependence of the viscous stress on the velocity gradient in the fluid is a constitutive law, which
is usually called the bulk rheological equation. The general linear relation between the viscous
stress tensor, T, and the rate of strain tensor,
1
D = [∇v + (∇v)T ] (4.256)
2
⎡ 1 ⎤
T = ζ(div v)I + 2η ⎢ D − (div v)I ⎥ (4.257)
⎣ 3 ⎦
The latter equation is usually referred to as the Newtonian model or Newton’s law of viscosity. In
Equation 4.257, ζ is the dilatational bulk viscosity and η is the shear bulk viscosity. The usual liq-
uids comply well with the Newtonian model. On the other hand, some concentrated macromolecu-
lar solutions, colloidal dispersions, gels, etc., may exhibit non-Newtonian behavior; their properties
are considered in detail in some recent review articles and books [636–641]. From Equations 4.252
and 4.257, one obtains the Navier–Stokes equation [642,643]:
dv ⎛ 1 ⎞
ρ = −∇p + ⎜ ζ + η ⎟ ∇(∇ ⋅ v) + η∇ 2 v + f , ( f ≡ ∇ ⋅ Pb ) (4.258)
dt ⎝ 3 ⎠
for homogeneous Newtonian fluids, for which the dilatational and shear viscosities, ζ and η, do not
depend on the spatial coordinates. In Equation 4.258, the material derivative d/dt can be presented
as a sum of a local time derivative and a convective term:
d ∂
= + (v ⋅∇) (4.259)
dt ∂t
If the density, ρ, is constant, the equation of mass conservation (Equation 4.251) and the Navier–
Stokes equation 4.258 reduce to
dv
div v = 0, ρ = −∇p + η∇ 2 v + f (4.260)
dt
For low shear stresses in the dispersions, the characteristic velocity, Vz, of the relative particle motion
is small enough in order for the Reynolds number, Re = ρVzL/η, to be a small parameter, where L
is a characteristic length scale. In this case, the inertia terms in Equations 4.258 and 4.260 can be
neglected. Then, the system of equations becomes linear and the different types of hydrodynamic
motion become additive [404,644,645]; for example, the motion in the liquid flow can be presented
as a superposition of elementary translation and rotational motions.
The basic equations can be further simplified in the framework of the lubrication approxima-
tion, which can be applied to the case when the Reynolds number is small and when the distances
344 Handbook of Surface and Colloid Chemistry
S2: z = h2(t, r)
Film phase
0 r
S1: z = h1(t, r)
Droplet phase
FIGURE 4.40 Sketch of a plane-parallel film formed between two identical fluid particles.
between the particle surfaces are much smaller than their radii of curvature (Figure 4.40) [646,647].
There are two ways to take into account the molecular interactions between the two particles across
the liquid film intervening between them: (1) the body force approach and (2) the disjoining pres-
sure approach. The former approach treats the molecular forces as components of the body force,
f (Equation 4.258); consequently, they give contributions to the normal and tangential stress bound-
ary conditions [648,649]. In the case (2), the molecular interactions are incorporated only in the
normal stress boundary conditions at the particle surfaces. When the body force can be expressed
as a gradient of potential, f = ∇U (that is Pb = UI), the two approaches are equivalent [650].
If two particles are interacting across an electrolyte solution, the equations of continuity and the
momentum balance, Equation 4.260, in lubrication approximation read [651,652]
N
∂vz ∂ 2v
∇ II ⋅ v II + = 0, η 2II = ∇ II ⋅ p + kT ∑z c ∇ Φ i i II
∂z ∂z i =1
(4.261)
N
∂p ∂Φ
+ kT ∑ zi ci =0
∂z i =1
∂z
where
vII and ∇II are the projection of the velocity and the gradient operator on the plane xy; the z-axis
is (approximately) perpendicular to the film surfaces S1 and S2 (see Figure 4.40)
ci = ci(r, z, t) is the ion concentration (i = 1, 2, …, N)
Φ is the dimensionless electric potential (see Sections 4.2.1.2 and 4.2.2)
It turns out that in lubrication approximation, the dependence of the ionic concentrations on the z
coordinate comes through the electric potential Φ(r, z, t): we obtain a counterpart of the Boltzmann
equation ci = ci,n(r, z, t)exp(−ziΦ), where ci,n refers to an imaginary situation of “switched off” elec-
tric charges (Φ ≡ 0). The kinematic boundary condition for the film surfaces has the form:
∂h j
+ u j .∇ IIh j = (vz ) j at S j ( j = 1, 2) (4.262)
∂t
where
ui is the velocity projection in the plane xy at the corresponding film surface
Si, which is close to the interfacial velocity
(vz)i is the z component of the velocity at the surface Si
Chemical Physics of Colloid Systems and Interfaces 345
The general solution of Equations 4.261 and 4.262 could be written as:
N
p = pn + kT ∑ (c − c
i =1
i i,n ) (4.263)
( z − h1 )( z − h2 ) h −z z − h1
v II = ∇ II pn + 2 u1 + u2
2η h h
(4.264)
2 N
kTh ⎡ h2 − z z − h1 ⎤
+
4η ∑ ⎢ m2,i ( z) − h m2,i (h1 ) − h m2,i (h2 ) ⎥ ∇ IIci,n
i =1 ⎣ ⎦
Here h = h2 − h1 is the local film thickness; the meaning of pn(x, y, t) is analogous to that of
ci,n(x, y, t); the functions, mk,i(z), account for the distribution of the ith ionic species in the EDL:
m0,i ≡ exp(− zi Φ ) − 1
z
2 (4.265)
mk ,i ( z ) ≡ (k = 1, 2, 3, i = 1, 2,…, N )
mk −1,i (z)dz
h ∫
0
The equation determining the local thickness, h, of a film with fluid surfaces (or, alternatively, determin-
ing the pressure distribution at the surfaces of the gap between two solid particles of known shape) is
∂h ⎡h ⎤ 1
+ ∇ II ⋅ ⎢ (u1 + u 2 ) ⎥ = ∇ II ⋅ (h3∇ II p)
∂t ⎣2 ⎦ 12η
N
(4.266)
kT ⎧⎪ ⎫⎪
+ ∇II ⋅ ⎨h3 ∑ [ m2,i (h1 ) + m2,i (h2 ) − m3,i (h2 ) + m3,i (h1 )]∇ IIci,n ⎬
8η ⎩⎪ i =1 ⎭⎪
The problem for the interactions upon central collisions of two axisymmetric particles (bubbles,
droplets, or solid spheres) at small surface-to-surface distances was first solved by Reynolds [646]
and Taylor [653,654] for solid surfaces and by Ivanov et al. [655,656] for films of uneven thickness.
Equation 4.266 is referred to as the general equation for films with deformable surfaces [655,656] (see
also the more recent reviews [240,657,658]). The asymptotic analysis [659–661] of the dependence
of the drag and torque coefficient of a sphere, which is translating and rotating in the neighborhood
of a solid plate, is also based on Equation 4.266 applied to the special case of stationary conditions.
Using Equation 4.255, one can obtain expressions for the components of the total force exerted
on the particle surface, S, in the lubrication approximation:
N
⎡ ⎤
Ftot , z = ⎢ pn + kT
⎢
S ⎣
∫ ∑ (c
i =1
is − ci,n ) + Π nel − p∞ ⎥ dS (4.267)
⎥⎦
⎛ ∂v 2kT ∂Φ ⎞
Ftot ,II = − ⎜ η II + 2 ∇ IIΦ ⎟ dS (4.268)
S
⎝ ∂z∫ κ c ∂z ⎠
where
p∞ is the pressure at infinity in the meniscus region (Figure 4.40)
Πnel ≡ Π − Πel accounts for the contribution of nonelectrostatic (non–double-layer) forces to the
disjoining pressure (see Section 4.4)
346 Handbook of Surface and Colloid Chemistry
The normal and the lateral force resultants, Fz and FII, are the hydrodynamic resistance and shear
force, respectively.
4.5.2.1 Taylor and Reynolds Equations, and Influence of the Particle Shape
In the case of two axisymmetric particles moving along the z-axis toward each other with velocity
Vz = −dh/dt, Equation 4.266 can be integrated; and from Equation 4.267, the resistance force can be
calculated. The latter turns out to be proportional to the velocity and bulk viscosity and depends on
the shape in a complex way. For particles with tangentially immobile surfaces and without surface
electric charge (u1 = u2 = 0, Φ = 0) Charles and Mason [664] have derived
∞
r3
Fz = 6πηVz ∫ dr (4.269)
0 h3
where r is the radial coordinate in a cylindrical coordinate system. In the case of two particles
of different radii, R1 and R2, film radius R, and uniform film thickness h (see Figure 4.41), from
Equation 4.269 the following expression can be derived [665,666]:
3 R2 ⎛ R2 R4 ⎞ 2 R1R2
Fz = πηVz * ⎜ 1 + + 2 2 ⎟ , R* ≡ (4.270)
2 h ⎜⎝ hR* h R* ⎟⎠ R1 + R2
R2
R1
FIGURE 4.41 Sketch of a film between two nonidentical fluid particles of radii R1 and R2. The film thickness
and radius are denoted by h and R.
Chemical Physics of Colloid Systems and Interfaces 347
This geometrical configuration has proved to be very close to the real one in the presence of electro-
static disjoining pressure [256]. The Charles–Mason formula (Equation 4.269) and Equation 4.267
have been used to calculate the velocity of film thinning for a large number of cases, summarized
by Hartland [667] in tables for more than 50 cases (2D and 3D small drops, fully deformed large
drops subjected to large forces, 2D hexagonal drops, etc.).
Setting R = 0 in Equation 4.270, we can derive a generalized version of the Taylor formula
[653,654] for the velocity of approach of two nondeformable spheres under the action of an external
(nonviscous) force, Fz [666]:
2hFz
VTa =
3πηR 2 (4.271)
*
When a solid sphere of radius Rc approaches a flat solid surface, we may use the Taylor formula with
R* = 2Rc when the gap between the two surfaces is small compared to Rc. In fact Equation 4.271
does not appear in any of the G.I. Taylor’s publications but it was published in the article by Hardy
and Bircumshaw [653] (see Ref. [654]).
In the case when two plane-parallel ellipsoidal discs of tangentially immobile surfaces are mov-
ing against each other under the action of an external force, Ftot,z, from Equations 4.266 and 4.267,
we can derive the Reynolds equation [646] for the velocity of film thinning:
Fz h3 (a 2 + b2 )
VRe = (4.272)
3πηa 3b3
where a and b are the principal radii of curvature. If there is a contribution of the disjoining pres-
sure, Π, the Reynolds equation for a flat axisymmetrical film (a = b = R) between two fluid particles
of capillary pressure Pc can be written in the form [216]:
2 Fz h3 2( Pc − Π )h3
VRe = = (4.273)
3πηR 4
3ηR 2
From Equations 4.270 and 4.273, the ratio between the Reynolds velocity and the velocity of film
thinning for a given force is obtained. In Figure 4.42, this ratio is plotted as a function of the film
3.0
2.8
2.6
2.4
2.2
VRe/Vz
2.0
1.8
1.6
1.4
1.2
1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
h/hi
FIGURE 4.42 Plot of VRe/Vz vs. h/hi for two fluid particles (Equation 4.270) which are deformed because of
the viscous friction in the transition zone between the film and the bulk phase (see Figure 4.41).
348 Handbook of Surface and Colloid Chemistry
104
2 R/R* = 0.01
R/R* = 0.02
103 R/R* = 0.05
R/R* = 0.1
2
VTa/Vz
102
2
101
2
100
10–4 1.5 2 3 4 5 6
10–3 1.5 2 3 4 5 6 10–2
h/R*
FIGURE 4.43 Plot of VTa /Vz vs. h/R* for various values of the dimensionless film radius, R/R*. VTa corre-
sponds to two nondeformed (spherical) particles (Equation 4.271), whereas Vz is the velocity of approach of
two deformed particles (Equation 4.270).
thickness, h, divided by inversion thickness, hi = R2/R* [657]. We see that the influence of the
viscous friction in the zone encircling the film (this influence is not accounted for in Equation
4.273) decreases the velocity of thinning about three times for the larger distances, whereas for
the small distances this influence vanishes. From Equations 4.270 and 4.271, the ratio between
the Taylor velocity (corresponding to nondeformable spheres) and the approaching velocity of two
deformable particles can be calculated. The dependence of this ratio on the distance between the
particles for different film radii is illustrated in Figure 4.43. We see that an increase of the film
radius, R, and a decrease of the distance, h, lead to a decrease in the velocity. The existence of a
film between the particles can decrease the velocity of particle approach, Vz, by several orders of
magnitude.
R1 R1 ω v
O2 O1 U1
F
ω O1
R1
v
r
R2 R2 R2
O1 O2
O2
U2
(a) (b) (c)
FIGURE 4.44 Types of hydrodynamic interactions between two spherical particles: (a) motion along and
rotation around the line of centers; (b) motion along and rotation around an axis perpendicular to the line of
centers; (c) the first particle moves under the action of an applied external force, F, whereas the second particle
is subjected to the hydrodynamic disturbance created by the motion of the first particle.
moving toward each other. A combination of these results permits evaluation of the total forces and
torques acting on the particles.
The first particle then moves along an axis perpendicular to the center line and rotates around
this axis, whereas the second particle is immobile; see Figure 4.44b (this is a typical 3D hydrody-
namic problem). The contribution of this asymmetric motion of the spheres to the resistance tensors
was determined by Davis [678] and O’Neill and Majumdar [679].
The first particle moves with linear velocity, U1, under the action of an applied external force, F,
whereas the second particle is subjected to the hydrodynamic disturbances (created by the motion
of the first particle) and moves with a linear velocity, U2 (see Figure 4.44c). As a rule, this is a 3D
hydrodynamic problem. For this case, Batchelor [683] and Batchelor and Wen [684] have derived
the following expressions for the instantaneous translational velocities of the two particles in an
otherwise quiescent and unbounded fluid:
F ⎡ rr ⎛ rr ⎞ ⎤
U1 = ⋅ ⎢ A11 (r ) 2 + B11 (r ) ⎜ I − 2 ⎟ ⎥ (4.274)
6πηR1 ⎣ r ⎝ r ⎠⎦
F ⎡ rr ⎛ rr ⎞ ⎤
U2 = ⋅ A12 (r ) 2 + B12 (r ) ⎜ I − 2 ⎟ ⎥ (4.275)
6πη( R1 + R2 ) ⎢⎣ r ⎝ r ⎠⎦
where r is the vector connecting the particle centers and r = |r|. Expressions for the mobility
functions Aij and Bij (i, j = 1, 2) at large values of the dimensionless distance s = 2r/(R1 + R2) and
comparable particle radii λ = R2/R1 = O(1) have been derived by Jeffrey and Onishi [685] and Davis
and Hill [682]. The derived far-field expansions are
In the case of a small heavy sphere falling through a suspension of large particles (fixed in space),
we have λ ≫ 1; the respective expansions, corresponding to Equation 4.276, were obtained by
Fuentes et al. [686]. In the opposite case, when λ ≪ 1, the suspension of small background spheres
will reduce the mean velocity of a large heavy particle (as compared with its Stokes velocity [687])
because the suspension behaves as an effective fluid of larger viscosity as predicted by the Einstein
viscosity formula [683,686].
S2
h(r) h
0 r
S1 2Rd 2R
(a) (b) (c)
ζ(r) 2Rsp h2
θ
2R 2RNBF
(d) (e) (f )
FIGURE 4.45 Main stages of formation and evolution of a thin liquid film between two bubbles or drops:
(a) mutual approach of slightly deformed surfaces; (b) at a given separation, the curvature at the center inverts
its sign and a dimple arises; (c) the dimple disappears, and eventually an almost plane-parallel film forms;
(d) due to thermal fluctuations or other disturbances the film either ruptures or transforms into a thinner
Newton black film (e), which expands until reaching the final equilibrium state (f).
Chemical Physics of Colloid Systems and Interfaces 351
Under the action of an outer driving force, the fluid particles approach each other. The hydrody-
namic interaction is stronger at the front zones and leads to a weak deformation of the interfaces
in this front region. In this case, the usual hydrodynamic capillary number, Ca = ηVz/σ, which is
a small parameter for nondeformable surfaces, should be modified to read Ca = ηVzR*/σh, where
the distance, h, between the interfaces is taken into account. The shape of the gap between two
drops for different characteristic times was calculated numerically by many authors [691–711].
Experimental investigation of these effects for symmetric and asymmetric drainage of foam films
were carried out by Joye et al. [700,701]. In some special cases, the deformation of the fluid
particle can be very fast: for example, the bursting of a small air bubble at an air–water inter-
face is accompanied by a complex motion resulting in the production of a high-speed liquid jet
(see Boulton-Stone and Blake [711]).
When a certain small separation, hi, the inversion thickness, is reached, the sign of the curvature
in the contact of the fluid particles (drops, bubbles) changes. A concave lens–shaped formation called
a dimple is formed (see Frankel and Mysels [712]). This stage is also observed for asymmetric films
[701]. A number of theoretical studies have described the development of a dimple at the initial stage
of film thinning [691–711,713]. The inversion thickness can be calculated from a simple equation in
which the van der Waals interaction is explicitly taken into account (see Section 4.4.2) [240,656,691]
Fz (σ1 + σ2 ) ⎛ AH R* ⎞
hi = ⎜⎜ 1 − ⎟⎟ (4.277)
4πσ1σ2 ⎝ 12 Fz hi ⎠
where
σ1 and σ2 are the interfacial tensions of the phase boundaries S1 and S2
in this case Fz is the external force (of nonviscous and non–van der Waals origin) experienced by
the approaching particles
AH is the Hamaker constant
In the case, when the van der Waals force is negligible, Equation 4.277 reduces to hi = Fz(σ1 + σ2)/
(4πσ1σ2) [240,656]. Danov et al. [665] have shown that in the case of Brownian flocculation of iden-
tical small droplets, hi obeys the following transcendental equation:
⎧ zi z 2 γ ( z )
−1
⎫
kT ⎪ ⎛ i⎞ ⎡ U ( z) − U ( zi ) ⎤ dz ⎪
hi = exp ⎢
2πσzi ∫
⎨ ⎜ ⎟
⎪⎩ 0 ⎝ z ⎠ γ( zi ) ⎣ kT ⎥ ⎬ (4.278)
⎦ zi ⎪
⎭
where
kT is the thermal energy
γ(z) = Fz /Vz is the hydrodynamic resistance given by Equation 4.270
U is the potential energy due to the surface forces (see Equation 4.175)
z is the distance between the droplet mass centers
These authors pointed out that with an increase of droplet size the role of the Brownian force in
the film formation decreases, but for micrometer-sized liquid droplets the Brownian force is still by
several orders of magnitude greater than the buoyancy force due to gravity. If the driving force is
large enough, so that it is able to overcome the energy barrier created by the electrostatic repulsion
and/or the increase of the surface area during the droplet deformation, then film with a dimple will
be formed. On the contrary, at low electrolyte concentration (i.e., strong electrostatic repulsion) such
a dimple might not appear. Parallel experiments [714] on the formation and thinning of emulsion
films of macroscopic and microscopic areas, prepared in the Scheludko–Exerowa cell [215,216]
352 Handbook of Surface and Colloid Chemistry
and in a miniaturized cell, show that the patterns and the time scales of the film evolution in these
two cases are significantly different. There is no dimple formation in the case of thin liquid films of
small diameters [714].
In the case of predominant van der Waals attraction, instead of a dimple, a reverse bell-shape
deformation, called a pimple, appears and the film quickly ruptures [691,698,707,710,713]. The
thickness, hp, at which the pimple appears, can be calculated from the relationship [691]:
1/ 2
⎛ AH R* ⎞
hp = ⎜⎜ ⎟⎟ (4.279)
⎝ 12 Fz ⎠
The pimple formation thickness depends significantly on the radius, R*. If a drop of tangentially
immobile surfaces and radius Rd is driven by the buoyancy force, then we have:
4 3
Fz = πRd Δρg (4.280)
3
where
Δρ is the density difference
g is the gravity acceleration
For the collision of this drop with another immobile one, we have hp2 = AH /(16πΔρgRd2 ). We see that
hp is inversely proportional to the drop radius. For typical values of the Hamaker constant AH = 4 ×
10−20 J, density difference Δρ = 0.12 g/cm3, and Rd = 10 μm, the thickness of pimple formation is
hp = 82.3 nm. Note that this thickness is quite large. The pimple formation can be interpreted as the
onset of instability without fluctuations (stability analysis of the film intervening between the drops
has been carried out elsewhere [62]).
As already mentioned, if the van der Waals force (or other attractive force) is not predominant,
first a dimple forms in the thinning liquid films. Usually the dimple exists for a short period of time;
initially it grows, but as a result of the swift outflow of liquid it decreases and eventually disappears.
The resulting plane-parallel film thins at almost constant radius R. When the electrostatic repulsion is
strong, a thicker primary film forms (see point 1 in Figure 4.17). From the viewpoint of conventional
DLVO theory, this film must be metastable. Indeed, the experiments with microscopic foam films,
stabilized with sodium octyl sulfate or SDS in the presence of different amount of electrolyte [715],
show that a black spot may suddenly form and a transition to secondary (Newton black) film may
occur (see point 2 in Figure 4.17). The rate of thinning depends not only on the capillary pressure (the
driving force) but also very strongly on the surfactant concentration (for details, see Section 4.5.3.2).
The appearance of a secondary film (or film rupture, if the secondary film is not stable) is
preceded by corrugation of the film surfaces due to thermally excited fluctuations or outer dis-
turbances. When the derivative of the disjoining pressure, ∂Π/∂h, is positive, the amplitude of the
fluctuations (ζ in Figure 4.45d) spontaneously grows. As already mentioned, the instability leads to
rupture of the film or to formation of black spots. The theory of film stability was developed by de
Vries [716], Vrij [717], Felderhof [648], Sche and Fijnaut [649], Ivanov et al. [718], Gumerman and
Homsy [719], Malhotra and Wasan [720], Maldarelli and Jain [650], and Valkovska et al. [721]. On
the basis of the lubrication approximation for tangentially immobile surfaces, Ivanov et al. [718]
and Valkovska et al. [721] derived a general expression for the critical film thickness, hcr, by using
long-waves stability analysis:
1/ 4 ⎛ 2 2 htr ⎞
⎛ σh2 ⎞ k R h 3Π ʹ
hcr = htr ⎜ tr ⎟ exp ⎜ − cr 3 dh ⎟ (4.281)
⎝ kT ⎠ ⎜ 32hcr
⎝
∫
hcr
Pc − Π ⎟
⎠
Chemical Physics of Colloid Systems and Interfaces 353
htr
(1/σ) (h3Πʹ / ( Pc − Π ))dh (4.282)
2
k =
cr
∫ hcr
htr
h6 / ( Pc − Π )dh
∫ hcr
In Equation 4.282, htr is the so-called transitional thickness [717,718,721] at which the increase of
free energy due to the increased film area and the decrease of free energy due to the van der Waals
interaction in the thinner part (Figure 4.45d) compensate each other. At htr the most rapidly grow-
ing fluctuation (the critical wave) becomes unstable. The transitional thickness obeys the following
equation [718,721]:
Figures 4.46 and 4.47 show the critical thicknesses of rupture, hcr, for foam and emulsion films,
respectively, plotted versus the film radius [722]. In both cases the film phase is the aqueous phase,
which contains 4.3 × 10−4 M SDS + added NaCl. The emulsion film is formed between two toluene
drops. Curve 1 is the prediction of a simpler theory, which identifies the critical thickness with the
transitional one [720]. Curve 2 is the theoretical prediction of Equations 4.281 through 4.283 (no
adjustable parameters); in Equation 4.182 for the Hamaker constant the electromagnetic retarda-
tion effect has also been taken into account [404]. In addition, Figure 4.48 shows the experimental
dependence of the critical thickness versus the concentration of surfactant (dodecanol) for aniline
films. Figures 4.46 through 4.48 demonstrate that when the film area increases and/or the electrolyte
concentration decreases the critical film thickness becomes larger. Figure 4.49 shows the critical
thickness of foam film rupture for three concentrations of SDS in the presence of 0.3 M NaCl [605].
The dashed and dash-dotted lines, for 1 and 10 μM SDS, respectively, are computed assuming only
the van der Waals attraction (no adjustable parameter). The deviation of the predicted values of hcr
70
(1)
60
Film thickness, nm
50
(2)
40
30
20
0 100 200 300 400 500 600
Film radius, µm
FIGURE 4.46 Dependence of the critical thickness, hcr, on the radius, R, of foam films. (The experimental
points are data from Manev, E.D. et al., J. Colloid Interface Sci., 97, 591, 1984.) The films are formed from a
solution of 4.3 × 10−4 M SDS + 0.25 M NaCl. Curve 1 is the prediction of the simplified theory [720], whereas
Curve 2 is calculated using Equations 4.281 through 4.283; no adjustable parameters.
354 Handbook of Surface and Colloid Chemistry
55
50 (1)
45
Film thickness, nm
(2)
40
35
30
25
20
0 100 200 300 400 500
Film radius, µm
FIGURE 4.47 Critical thickness, hcr, vs. radius, R, of emulsion films, toluene–water–toluene. The experi-
mental points are data from Ref. [722]; the films are formed from a solution of 4.3 × 10−4 M SDS + 0.1 M NaCl.
Curve 1 is the prediction of the simplified theory [720], whereas Curve 2 is calculated using Equations 5.281
through 5.283; no adjustable parameters.
500
Critical thickness, hcr, Å
400
300
10–3 10–2 10–1
Surfactant concentration, c0, M
FIGURE 4.48 Dependence of the critical thickness, hcr, of aniline films on the concentration of dodecanol,
c0 [723]. (Modified from Ivanov, I.B., Pure Appl. Chem., 52, 1241, 1980.)
Chemical Physics of Colloid Systems and Interfaces 355
30
25
Added 0.3 M NaCl
60 80 100 120 140
Film radius (µm)
FIGURE 4.49 Critical thickness, hcr, vs the film radius at a 0.3 M fixed concentration of NaCl for three
SDS concentrations: 0.5; 1; 10 μM. The dashed and dash-dotted lines, for 1 and 10 μM SDS, respectively, are
computed using an absence of the hydrophobic attraction. The solid lines are fits of the experimental points
with λ 0 = 15.8 nm.
from the measured is because of the hydrophobic interaction (Section 4.4.5.4.2). The solid lines
represent fits with the decay length of the hydrophobic interactions λ0 = 15.8 nm using Equations
4.281 through 4.283.
The surface corrugations do not necessarily lead to film rupture. Instead, black spots (secondary
films of very low thickness; h2 in Figure 4.17) can be formed. The typical thickness of plane-
parallel films at stage c (Figure 4.45c) is about 200 nm, while the characteristic thickness h2 of the
Newton black film (Figure 4.45e and f) is about 5–10 nm. The black spots either coalesce or grow in
diameter, forming an equilibrium secondary (Newton black) film with a thickness h2 and radius Rsp.
These spots grow until they cover the whole film area.
After the entire film area is occupied by the Newton black film, the film radius increases until
it reaches its equilibrium value, R = RNBF (Figure 4.45f). Finally, the equilibrium contact angle is
established. For more details about this last stage of film thinning, see part IV.C of Ref. [240].
hin
dh 3πηR* ⎡ ⎛ hin ⎞ R 2 ⎛
2
h ⎞ R4 ⎛ h2 ⎞⎤
⎢ ln ⎜ ⎟ + 1 − f ⎟ + 2 2 ⎜ 1 − 2f
τ=
∫
hf
Vz
= ⎜
2 Fz ⎢⎣ ⎝ h f ⎠ h f R* ⎝ hin ⎠ 2h f R ⎝ hin
*
⎟ ⎥ (4.284)
⎠ ⎥⎦
The final thickness, hf, may coincide with the critical thickness of film rupture. Equation 4.284
is derived for tangentially immobile interfaces from Equation 4.270 at a fixed driving force
(no disjoining pressure).
In the case of gravity-driven coalescence of a droplet with its homophase, the driving force is
given by Equation 4.280 and the mean drop radius is R* = 2Rd. Then from Equations 4.280 and
356 Handbook of Surface and Colloid Chemistry
4.284 we can deduce the droplet lifetime in the so-called Taylor regime, corresponding to nonde-
formed droplets (R = 0):
We see that τTa depends logarithmically on the ratio of the initial and final thickness. Moreover, in
the Taylor regime the lifetime, τ, decreases with the increase of the driving force, Fz, and the drop
radius, Rd. The latter fact is confirmed by the experimental data of Dickinson et al. [726].
In the case of deformed drops (R ≠ 0), the drainage time, τ, is determined by Equation 4.284,
and in such a case the fluid particles approach each other in the Reynolds regime [657,723].
The dependence of τ on Rd in Equation 4.284 is very complex, because the driving force, Fz, and the
film radius, R, depend on Rd. The film radius can be estimated from the balance of the driving and
capillary force [657,723]:
Fz Rd
R2 = (4.286)
2πσ
In this regime, the lifetime, τ, increases with an increase of the driving force, Fz. This is exactly the
opposite trend compared to the results for the Taylor regime (see Equation 4.285). The result can
be rationalized in view of Reynolds equation (Equation 4.273). In the numerator of this equation,
Fz ∝ Rd3, whereas in the denominator R 4 ∝ Rd8 (see Equation 4.286); as a result, the drainage rate
becomes proportional to Rd−5, that is, Vz decreases as the droplet radius increases.
The numerical results from Equations 4.284 through 4.286 for the lifetime or drainage time, τ,
versus the droplet radius, Rd, are plotted in Figure 4.50 for parameter values typical for emulsion
systems: Δρ = 0.2 g/cm3, η = 1 cP, hf = 5 nm, and hin = Rd /10. The various curves in Figure 4.50
correspond to different values of the surface tension, σ, shown in the Figure 4.50. The left branches
of the curves correspond to the Taylor regime (nondeformed droplets), whereas the right branches
correspond to the Reynolds regime (formation of film between the droplets). The presence of a
deep minimum on the τ versus Rd curve was first pointed out by Ivanov [727,728]. The theoretical
4
σ = 30 dyn/cm
3
1
Taylor
0
0 20 40 60 80 100 120 140 160 180 200
Droplet radius, Rd (µm)
FIGURE 4.50 Calculated lifetime, τ, of drops approaching a fluid interface in Taylor regime (the solid line)
and in Reynolds regime (the other lines) as a function of the droplet radius, Rd.
Chemical Physics of Colloid Systems and Interfaces 357
160
140 Reynolds
regime
60
40
20
0
0 100 200 300 400 500 600 700 800 900 1000 1100
Droplet radius (µm)
FIGURE 4.51 Stability of oil drops pressed by buoyancy against a large oil–water interface. Measured
lifetime, (the points), is plotted vs droplet radius in a system consisting of soybean oil and aqueous solution of
4⋅10−4 wt% BSA + 0.15 M NaCl (pH = 6.4). The solid line is drawn in accordance to Equation 4.284.
dependencies in Figure 4.50 agree well with experimental data [652,729–731] for the lifetime of oil
droplets pressed by the buoyancy force against a large oil–water interface in a system containing
protein BSA (Figure 4.51).
∂ck
+ div(ck v + jk ) = rk (k = 1, 2,…, N ) (4.287)
∂t
where
ck and jk are bulk concentration and flux, respectively, of the kth species—note that jk includes
the molecular diffusive flux, the flux driven by external forces (e.g., electrodiffusion
[651,662,663]) and the thermodiffusion flux [662]
rk is the rate of production due to chemical reactions, including surfactant micellization or
micelle decay
358 Handbook of Surface and Colloid Chemistry
The surface mass balance equation for the adsorption, Γk, has the form [651,662,663]:
∂Γ k
+ ∇ s ⋅ (Γ k v s + jsk ) = rks + n ⋅ 〈 jk 〉 (4.288)
∂t
where
n is the unit normal to the interface directed from phase 1 to phase 2
〈 〉 denotes the difference between the values of a given physical quantity at the two sides of the
interface
∇s is the surface gradient operator [733]
vs is the local material surface velocity
jsk is the 2D flux of the kth component along the interface
rks accounts for the rate of production of the kth component due to interfacial chemical reactions
and could include conformational changes of adsorbed proteins
Equation 4.288 provides a boundary condition for the normally resolved flux, jk. From another
viewpoint, Equation 4.288 represents a 2D analogue of Equation 4.287. The interfacial flux, jk , can
s
also contain contributions from the interfacial molecular diffusion, electrodiffusion, and thermo-
diffusion. A simple derivation of the time-dependent convective-diffusion equation for surfactant
transport along a deforming interface is given by Brenner and Leal [734–737], Davis et al. [669],
and Stone [738]. If the molecules are charged, the bulk and surfaces electrodiffusion fluxes can be
expressed in the form [651,739,740]:
Dk ,0 K b (φk ) ∂μ k Ds ∂μ sk
Dk = , Dks = k ,0 K s (Γ k ) (4.290)
kT (1 − φk ) ∂ ln φk kT ∂ ln Γ k
where μk and sk are the bulk and surface chemical potentials, respectively. The dimensionless bulk
friction coefficient, Kb, accounts for the change in the hydrodynamic friction between the fluid and
the particles (created by the hydrodynamic interactions between the particles). The dimensionless
surface mobility coefficient, Ks, accounts for the variation of the friction of a molecule in the adsorp-
tion layer. Feng [741] has determined the surface diffusion coefficient, the dilatational elasticity, and
the viscosity of a surfactant adsorption layer by theoretical analysis of experimental data. Stebe and
Maldarelli [742,743] studied theoretically the surface diffusion driven by large adsorption gradients.
The determination of bulk and surface diffusion coefficients from experimental data for the drain-
age of nitrobenzene films stabilized by different concentrations of dodecanol was reported [739].
Note that the adsorption isotherms, relating the surface concentration, Γk, with the subsurface
value of the bulk concentration, ck (see Section 4.2.2.1), or the respective kinetic Equation 4.86 for
adsorption under barrier control (see Section 4.2.2.5), should also be employed in the computations
based on Equations 4.287 through 4.290 in order for a complete set of equations to be obtained.
Another boundary condition is the equation of the interfacial momentum balance
[634,635,641,663]:
∇ s ⋅ σ = n ⋅ 〈P + Pb 〉 (4.291)
Chemical Physics of Colloid Systems and Interfaces 359
where σ is the interfacial stress tensor, which is a 2D counterpart of the bulk stress tensor, P.
Moreover, a 2D analogue of Equations 4.253, 4.256, and 4.257, called the Boussinesq–Scriven
constitutive law, can be postulated for a fluid interface [240,641,663,732,744–748]:
T
σ = σa I s + ( ηdl − ηsh ) ( ∇ s ⋅ v s ) I s + ηsh ⎡⎢( ∇ s v s ) ⋅ I s + I s ⋅ ( ∇ s v s ) ⎥⎤ (4.292)
⎣ ⎦
where
ηdl and ηsh are the interfacial dilatational and shear viscosities, respectively
Is is the unit surface idemfactor [733]
σa is the scalar adsorption part of the surface tension (see Section 4.2.1.2.2)
In view of the term σaIs in Equation 4.292, the Marangoni effects are hidden in the left-hand side of
the boundary condition (Equation 4.291) through the surface gradient of σa:
N
Ek ET ⎛ ∂σa ⎞ ⎛ ∂σa ⎞
∇ sσa = − ∑Γ ∇sΓ k − ∇ sT , E k = − ⎜ ⎟ , ET = − ⎜ ⎟ (4.293)
k =1 k T ⎝ ∂ ln Γ k ⎠T ,Γ j ≠k ⎝ ∂ ln T ⎠Γk
where
Ek is the Gibbs elasticity for the kth surfactant species (see Equation 4.6)
ET represents the thermal analogue of the Gibbs elasticity
The thermocapillary migration of liquid drops or bubbles and the influence of ET on their motion
are investigated in a number of works [749–751].
In fact, Equation 4.292 describes an interface as a 2D Newtonian fluid. On the other hand,
a number of non-Newtonian interfacial rheological models have been described in the literature
[752–773]. Tambe and Sharma [756] modeled the hydrodynamics of thin liquid films bounded by
viscoelastic interfaces, which obey a generalized Maxwell model for the interfacial stress tensor.
These authors [757,758] also presented a constitutive equation to describe the rheological properties
of fluid interfaces containing colloidal particles. A new constitutive equation for the total stress was
proposed by Horozov et al. [759], Danov et al. [760], and Ivanov et al. [761] who applied a local
approach to the interfacial dilatation of adsorption layers.
The interfacial rheology of protein adsorption layers has been intensively studied in relation
to the properties of foams and emulsions stabilized by proteins and their mixtures with lipids or
surfactants. Detailed information on the investigated systems, experimental techniques, and theo-
retical models can be found in Refs. [762–769]. The shear rheology of the adsorption layers of many
proteins follows the viscoelastic thixotropic model [770–772], in which the surface shear elasticity
and viscosity depend on the surface shear rate. The surface rheology of saponin adsorption layers
has been investigated in Ref. [773].
If the temperature is not constant, the bulk heat transfer equation complements the system and
involves Equations 4.251, 4.252, and 4.287. The heat transfer equation is a special case of the energy
balance equation. It should be noted that more than 20 various forms of the overall differential
energy balance for multicomponent systems are available in the literature [631,634]. The corre-
sponding boundary condition can be obtained as an interfacial energy balance [663,748]. Based
on the derivation of the bulk [774,775] and interfacial [760,775,776] entropy inequalities (using the
Onsager theory), various constitutive equations for the thermodynamic mass, heat, and stress fluxes
have been obtained.
360 Handbook of Surface and Colloid Chemistry
where
h is the closest surface-to-surface distance between the two drops
ηd is the viscosity of the disperse phase (the liquid in the droplets)
In the limiting case of solid particles, we have ηd → ∞, and Equation 4.294 reduces to the Taylor
equation (Equation 4.271). Note that in the case of close approach of two drops (ξ ≫ 1), the velocity
Vz is proportional to h. This implies that the two drops can come into contact (h = 0) in a finite
period of time (τ < ∞) under the action of a given force, Fz, because the integral in Equation 4.284
is convergent for hf = 0. This is in contrast to the case of immobile interfaces (ξ ≪ 1), when Vz ∝ h
and τ → ∞ for hf → 0.
In the other limiting case, that of two nondeformed gas bubbles (ηd → 0) in pure liquid, Equation
4.294 cannot be used; instead, Vz can be calculated from the expression [782,783]
Fz
Vz = (4.295)
πηRd [ln( Rd /h + 1) + 2.5407]
Note that in this case Vz ∝ (ln h)−1, and the integral in Equation 4.284 is convergent for hf → 0. In
other words, the theory predicts that the lifetime, τ, of a doublet of two colliding spherical bubbles
in pure liquid is finite. Of course, the real lifetime of a doublet of bubbles or drops is affected by
the surface forces for h < 100 nm, which should be accounted for in Fz and which may lead to the
formation of thin film in the zone of contact [207,392].
In the case of droplets with equal radii, Rd, in a pure liquid (without surfactant), two asymptotic
expressions are derived for small interdroplet distances [783]:
Fz 3π2ηd Rd ⎛ η2 ⎞ ⎛ Rd ⎞ η Rd
= + ⎜ 1 − d2 ⎟ ln ⎜ + 1 ⎟ + 2.5407 at d < 1 (4.296)
πηRdVz 16η h ⎝ 3η ⎠ ⎝ h ⎠ η h
3/2
Fz 3R 9π2η ⎛ Rd ⎞ ηd h
= d− at 1 (4.297)
πηRdVz 2h 64ηd ⎜⎝ h ⎟⎠ η Rd
Note that for ηd = 0 Equation 4.296 reduces to Equation 4.295. The second term in the right-hand
side of Equation 4.297 represents a correction to the Taylor equation (Equation 4.271).
Chemical Physics of Colloid Systems and Interfaces 361
Next, let us consider the case of deformed fluid particles (Figure 4.32). A number of theoretical
studies [784–787] have been devoted to the thinning of plane-parallel liquid films of pure liquid
phases (no surfactant additives). Ivanov and Traykov [786] derived the following exact expressions
for the velocity of thinning of an emulsion film:
1/ 3 1/ 3
⎛ 32ΔP 2 ⎞ Vz 1 ⎛ ρ η h4 F ⎞
Vz = ⎜ 4 ⎟
h5 / 3 , = , εe ≡ ⎜ d d 3 z4 ⎟ (4.298)
⎝ ρd ηd R ⎠ VRe εe ⎝ 108πη R ⎠
where
ρd is the density of the disperse phase
VRe is the Reynolds velocity defined by Equation 4.273
εe is the so-called emulsion parameter
Substituting typical parameter values in Equations 4.294 and 4.298 we can check that at a given
constant force the velocity of thinning of an emulsion film is smaller than the velocity of approach
of two nondeformed droplets and much larger than VRe. It is interesting to note that the velocity of
thinning as predicted by Equation 4.298 does not depend on the viscosity of the continuous phase,
η, and its dependence on the drop viscosity, ηd, is rather weak. There are experimental observations
confirming this prediction (see Ref. [34], p. 381).
The presence of surfactant adsorption monolayers decreases the mobility of the droplet (bub-
ble) surfaces. This is due to the Marangoni effect (see Equation 4.293). From a general view-
point, we may expect that the interfacial mobility will decrease with the increase of surfactant
concentration until eventually the interfaces become immobile at high surfactant concentrations
(see Section 4.5.2); therefore, a pronounced effect of surfactant concentration on the velocity of
film drainage should be expected. This effect really exists (see Equation 4.299), but in the case
of emulsions it is present only when the surfactant is predominantly soluble in the continuous
phase.
Traykov and Ivanov [787] established (both theoretically and experimentally) the interesting
effect that when the surfactant is dissolved in the disperse phase (i.e., in the emulsion droplets), the
droplets approach each other just as in the case of pure liquid phases, that is, Equation 4.298 holds.
Qualitatively, this effect can be attributed to the fact that the convection-driven surface tension gradi-
ents are rapidly damped by the influx of surfactant from the drop interior; in this way, the Marangoni
effect is suppressed. Indeed, during the film drainage the surfactant is carried away toward the film
border, and a nonequilibrium distribution depicted in Figure 4.52a appears. Because, however, the
mass transport is proportional to the perturbation, the larger the deviation from equilibrium, the
stronger the flux tending to eliminate the perturbation (the surfactant flux is denoted by thick arrows
(a) (b)
FIGURE 4.52 Damping of convection-driven surface tension gradients by influx of surfactant from the drop
interior. (a) Since the mass transport is proportional to the perturbation, the larger the perturbation, the stron-
ger the flux tending to eliminate it. (b) Uniform surfactant distribution is finally reached.
362 Handbook of Surface and Colloid Chemistry
in Figure 4.52b). In this way, any surface concentration gradient (and the related Marangoni effect)
disappears. The emulsion films in this case behave as if surfactant is absent.
In the opposite case, when the surfactant is soluble in the continuous phase, the Marangoni
effect becomes operative and the rate of film thinning becomes dependent on the surface (Gibbs)
elasticity (see Equation 4.293). Moreover, the convection-driven local depletion of the surfactant
monolayers in the central area of the film surfaces gives rise to fluxes of bulk and surface diffu-
sion of surfactant molecules. The exact solution of the problem [651,655,689,739,740,787] gives
the following expression for the rate of thinning of symmetrical planar films (of both foam and
emulsion type):
Vz 1 1 6ηDs 3ηD
= 1+ , = + (4.299)
VRe εe + ε f εf hEG Γ(∂σ /∂c)
where
as usual, D and Ds are the bulk and interfacial collective diffusion coefficients (see Equation 4.290)
EG is the Gibbs elasticity
εf is the so-called foam parameter [723]
In the special case of foam film, one substitutes εe = 0 in Equation 4.299. Note that the diffusive
surfactant transport, which tends to restore the uniform adsorption monolayers, damps the sur-
face tension gradients (which oppose the film drainage) and thus accelerates the film thinning.
However, at large surfactant concentrations, the surface elasticity, EG, prevails, εf increases, and,
consequently, the thinning rate decreases down to the Reynolds velocity, Vz → VRe (see Equation
4.299). Similar expressions for the rate of film thinning, which are appropriate for various ranges
of values of the interfacial parameters, can be found in the literature [240,655,656,703,788,789]. A
table describing the typical ranges of variation of the interfacial properties (Γ, EG, D, Ds, ∂σ/∂c, etc.)
for emulsion and foam systems can be found in Ref. [240], Table 4.2 therein. For h < 100 nm, the
influence of the disjoining pressure should be taken into account (see Equation 4.273). In some stud-
ies [240,666,756,790–793], the effect of the interfacial viscosity on the rate of thinning and the life-
time of plane-parallel films is investigated; this effect is found to decrease when the film thickness,
h, becomes smaller and/or the film radius, R, becomes larger.
Note that Equation 4.299 does not hold in the limiting case of foam films (εe = 0) at low s urfactant
concentration, εf → 0. The following expression is available for this special case [723]:
Vz (1 + 1/ε f )
= (4.300)
VRe [1 + 4h2 / (3R 2ε f )]
The merit of this equation is that it gives as limiting cases both Vz /VRe for foam films without
surfactant, εf → 0, and Equation 4.299 with εe = 0 (note that in the framework of the lubrication
approximation, used to derive Equation 4.299, the terms ∝ h2/R2 are being neglected). Equation 4.300
has also some shortcomings, which are discussed in Ref. [723].
Another case, which is not described by the above equations mentioned earlier, is the approach
of two nondeformed (spherical) bubbles in the presence of surfactant. The velocity of approach in
this case can be described by means of the expression [656,666,728,740]:
−1
Vz h ⎧⎪ ⎡ h(1 + b) ⎤ ⎡ hs ⎤ ⎫⎪
= s ⎨⎢ + 1⎥ ln ⎢ + 1⎥ − 1⎬ (4.301)
VTa 2h ⎪⎩ ⎣ hs ⎦ ⎣ h(1 + b) ⎦ ⎭⎪
Chemical Physics of Colloid Systems and Interfaces 363
where the parameters b and hs account for the influence of bulk and surface diffusivity of surfac-
tants, respectively. From Equation 4.290 these parameters are calculated to be [740]
A generalization of Equation 4.301 to the more complicated case of two nondeformed (spheri-
cal) emulsion droplets with account for the influence of surface viscosity and the solubility
of surfactants in both phases has been published in Ref. [783]. The terms related to the sur-
face viscosities K and E, and the surface elasticity Nel are scaled with the drop radius, Rd , as
follows [783]:
ηdl η E R
K≡ , E ≡ sh , N el ≡ G d (4.303)
ηRd ηRd ηDs
Thus, with the increase of the drop radius the surface viscosity becomes less important and the sup-
pression of surface mobility by the Marangoni effect increases. Figure 4.53 shows the dependence
of the drag force coefficient, f = Fz /(πηVz Rd), on the dimensionless distance between the droplets,
h/Rd. For surfactant concentrations above the CMC, the surfactant relaxation time is so small that
the interfacial tension changes insignificantly during the motion of the drops. In this case, the drag
force coefficient depends on the bulk and surface viscosities (Figure 4.53a). One sees that with the
increase of K and E the drag force changes from the values for two bubbles, f0 (see Equation 4.295),
to those for tangentially immobile drop surfaces, f im (see Equation 4.297). If the effect of bulk and
surface viscosities is negligible, then f depends on the surface elasticity, Nel (Figure 4.53b). Note
that the effect of surface elasticity is inversely proportional to the surface diffusion coefficient
(Equation 4.303). A faster surface diffusion suppresses the gradients of the surface tension and
decreases Nel.
Returning to the parameter values, we note that usually εe ≪ εf and εe ≪ 1. Then, comparing the
expressions for Vz /VRe as given by Equations 4.298 and 4.299, we conclude that the rate of thinning
is much greater when the surfactant is dissolved in the droplets (the disperse phase) in comparison
with the case when the surfactant is dissolved in the continuous phase. This prediction of the theory
was verified experimentally by measuring the number of films that rupture during a given period
of time [794], as well as the rate of thinning. When the surfactant was dissolved in the drop phase,
the average lifetime was the same for all surfactant concentrations (Figure 4.54a), in agreement
with Equation 4.298. For the emulsion film with the same, but inverted, liquid phases (the former
continuous phase becomes disperse phase, and vice versa), that is, the surfactant is in the film
phase, the average lifetime is about 70 times longer—compare curves in Figure 4.54a with curve
B in Figure 4.54b. The theoretical conclusions have been also checked and proved in experimental
measurements with nitroethane droplets dispersed in an aqueous solution of the cationic surfactant
hexadecyl trimethyl ammonium chloride (HTAC) [725].
105
fim
K=E
101
104
102
101 f0
105
fim Nel
105
Drag force coefficient, f
104 104
103
103 102
101
102
101 f0
FIGURE 4.53 Dependence of the drag force coefficient, f, on the dimensionless distance, h/Rd. (a) For sur-
factant concentrations above the CMC at different surface viscosities. (b) For different values of the surface
elasticity, the effects of surface viscosities and the viscosity of drop phase are neglected.
8 to 18 are expected to form oil-in-water (O/W) emulsions. Different formulae for calculating the
HLB numbers are available; for example, the Davies expression [797] reads
where nc is the number of –CH2– groups in the lipophilic part of the molecule. Shinoda and Friberg
[798] proved that the HLB number is not a property of the surfactant molecules only, but also
depends strongly on the temperature (for nonionic surfactants), on the type and concentration of
added electrolytes, on the type of oil phase, etc. They proposed using the phase inversion tempera-
ture (PIT) instead of HLB for characterization of the emulsion stability.
Davis [799] summarized the concepts about HLB, PIT, and Windsor’s ternary phase diagrams
for the case of microemulsions and reported topological ordered models connected with the
Helfrich membrane bending energy. Because the curvature of surfactant lamellas plays a major
role in determining the patterns of phase behavior in microemulsions, it is important to reveal how
the optimal microemulsion state is affected by the surface forces determining the curvature energy
Chemical Physics of Colloid Systems and Interfaces 365
2.0
2 mM
0.1 mM
1.5 0 mM
0.5
0.0
0 5 10 15 20
(a) t (s)
0.25
A
B
0.20
∆N/(N∆t) (s–1)
0.15
0.10
0.05
0.00
0 10 20 30 40 50
(b) t (s)
FIGURE 4.54 Histograms for the lifetimes of emulsion films: ΔN/N is the relative number of films that
have ruptured during a time interval Δt. (a) Surfactant in the drops: benzene films between water drops
containing surfactant sodium octylsulfonate of concentration: 0 M, 0.1 mM, and 2 mM; (b) Surfactant in the
film: (A) benzene film with 0.1 M of lauryl alcohol dissolved in the film, (B) water film with 2 mM of sodium
octylsulfonate inside. (From Traykov, T.T. et al., Int. J. Multiphase Flow, 3, 485, 1977.)
[239,800,801]. It is hoped that lattice models [802,803] and membrane curvature models [804,805]
will lead to predictive formulae for the microemulsion design.
Ivanov et al. [723,727,728,806] have proposed a semiquantitative theoretical approach that
provides a straightforward explanation of the Bancroft rule for emulsions. This approach is
based on the idea of Davies and Rideal [34] that both types of emulsions are formed during the
homogenization process, but only the one with lower coalescence rate survives. If the initial drop
concentration for both emulsions is the same, the coalescence rates for the two emulsions—(Rate)1
for emulsion 1 and (Rate)2 for emulsion 2 (Figure 4.55)—will be proportional to the respective
coalescence rate constants, kc,1 and kc,2 (see Section 4.6), and inversely proportional to the film
lifetimes, τ1 and τ2:
(Rate)1 kc,1 τ2 V1
≈ ≈ ≈ (4.305)
(Rate)2 kc,2 τ1 V2
366 Handbook of Surface and Colloid Chemistry
Phase I Phase II
Phase II
Phase I
Emulsion 1 Emulsion 2
FIGURE 4.55 The two possible types of emulsions obtained just after the homogenization; the surfactant is
soluble into Phase I.
Here V1 and V2 denote the respective velocities of film thinning. After some estimates based on
Equations 4.273, 4.284, 4.298, and 4.299, we can express the ratio in Equation 4.305 in the form:
1/ 3 1/ 3
(Rate)1 1/ 3 ⎛ hcr3 ,1 ⎞ ⎛ ηd ⎞ Pc − Π1
≈ 486ρd Ds3
( ) ⎜ 2 ⎟ ⎜ 2⎟ (4.306)
(Rate)2 ⎝ hcr ,2 ⎠ ⎝ R ⎠ EG ( Pc − Π 2 )2 / 3
where hcr,1 and hcr,2 denote the critical thickness of film rupture for the two emulsions in Figure 4.55.
Many conclusions can be drawn, regarding the type of emulsion to be formed:
1. If the disjoining pressures, Π1 and Π2, are zero, the ratio in Equation 4.306 will be very
small. Hence, emulsion 1 (surfactant soluble in the continuous phase) will coalesce much
more slowly and it will survive. This underlines the crucial importance of the surfactant
location (which is connected with its solubility), thus providing a theoretical foundation for
Bancroft’s rule. The emulsion behavior in this case will be controlled almost entirely by
the hydrodynamic factors (kinetic stability).
2. The disjoining pressure, Π, plays an important role. It can substantially change and even
reverse the behavior of the system if it is comparable by magnitude with the capillary pres-
sure, Pc. For example, if (Pc − Π2) → 0 at a finite value of Pc − Π1 (which may happen, e.g., for
an O/W emulsion with oil-soluble surfactant), the ratio in Equation 4.306 may become much
larger than unity, which means that emulsion 2 will become thermodynamically stable. In
some cases, the stabilizing disjoining pressure is large enough for emulsions with a very
high volume fraction of the disperse phase (above 95% in some cases) to be formed [807].
3. The Gibbs elasticity, EG, favors the formation of emulsion 1, because it slows down the
film thinning. On the other hand, increased surface diffusivity, Ds, decreases this effect,
because it helps the interfacial tension gradients to relax, thus facilitating the formation of
emulsion 2.
4. The film radius, R, increases and the capillary pressure, Pc, decreases with the drop radius,
Rd. Therefore, larger drops will tend to form emulsion 1, although the effect is not very
pronounced.
Chemical Physics of Colloid Systems and Interfaces 367
5. The difference in critical thicknesses of the two emulsions only slightly affects the rate
ratio in Equation 4.306, although the value of hcr itself is important.
6. The viscosity of the continuous phase, η, has no effect on the rate ratio, which depends only
slightly on the viscosity of the drop phase, ηd. This is in agreement with the experimental
observations (see Ref. [34]).
7. The interfacial tension, σ, affects the rate ratio directly only through the capillary pres-
sure, Pc = 2σ/Rd. The electrolyte primarily affects the electrostatic disjoining pressure, Π,
which decreases as the salt content increases, thus destabilizing the O/W emulsion. It can
also influence the stability by changing the surfactant adsorption (including the case of
nonionic surfactants).
8. The temperature strongly affects the solubility and surface activity of nonionic surfac-
tants [3]. It is well known that at higher temperature nonionic surfactants become more oil
soluble, which favors the W/O emulsion. Thus, solubility may change the type of emulsion
formed at the PIT. The surface activity has numerous implications and the most important
is the change of the Gibbs elasticity, EG, and the interfacial tension, σ.
9. Surface-active additives (cosurfactants, demulsifiers, etc.), such as fatty alcohols in the
case of ionic surfactants, may affect the emulsifier partitioning between the phases and its
adsorption, thereby changing the Gibbs elasticity and the interfacial tension. The surface-
active additives may also change the surface charge (mainly by increasing the spacing
among the emulsifier ionic headgroups), thus decreasing the repulsive electrostatic dis-
joining pressure and favoring the W/O emulsion. Polymeric surfactants and adsorbed
proteins increase the steric repulsion between the film surfaces. They may favor either
O/W or W/O emulsions, depending on their conformation at the interface and their sur-
face activity.
10. The interfacial bending moment, B0, can also affect the type of the emulsion, although
this is not directly visible from Equation 4.306. (Note that B0 = −4kcH0, where H0 is the
so-called spontaneous curvature and kc is the interfacial curvature elastic modulus [200].
Typically, B0 is of the order of 5 × 10−11 N.) Usually, for O/W emulsions, B0 opposes the
flattening of the droplet surfaces in the zone of collision (Figure 4.32), but for W/O emul-
sions favors the flattening [207]. This effect might be quantified by the expression for the
curvature contribution in the energy of droplet–droplet interaction [207]:
2
−2πR 2 B0 ⎛ R ⎞
Wc = , ⎜ ⎟ 1 (4.307)
Rd ⎝ Rd ⎠
It turns out that Wc > 0 for the droplet collisions in an O/W emulsion, while Wc < 0 for a W/O
emulsion [207]. Consequently, the interfacial bending moment stabilizes the O/W emulsions but
destabilizes the W/O ones. There is supporting experimental evidence [808] for microemulsions,
that is, for droplets of rather small size. Moreover, the effect of the bending moment can be important
even for micrometer-sized droplets [207]. This is because the bent area increases faster (R2 ∝ Rd2)
than the bending energy per unit area which decreases (Wc/R2 ∝ 1/Rd) when the droplet radius, Rd,
increases (see Equation 4.307).
For micron-sized emulsion droplets the capillary pressure can be so high that a film may
not appear between the drops. In such case, instead of Equation 4.306, we can use analogous
expression for nondeformed (spherical) drops [783,809]. Figure 4.56 shows the calculated ratio
of the coalescence rates of emulsion 1 and emulsion 2 for two water drops (with dissolved
sodium alkyl sulfates, C n SO4Na) in hexadecane at concentrations close to the CMC. The degree
of surface coverage was 0.7 and all parameters are obtained from the fits of surface tension
isotherms [783]. The increase of the surfactant chainlength makes the difference between two
368 Handbook of Surface and Colloid Chemistry
1.0
0.9
C12SO4Na
0.8 C11SO4Na
(Rate)1/(Rate)2
C10SO4Na
0.7
C9SO4Na
0.6 C8SO4Na
C7SO4Na
0.5
0.4
0.3
FIGURE 4.56 Ratio of the coalescence rates of emulsion 1 and emulsion 2 vs. the dimensionless distance,
h/Rd, for water drops stabilized by sodium alkyl sulfates (CnSO4Na) in hexadecane at concentrations close to
the CMC.
systems insignificant. In the calculations, the effect of disjoining pressure has not been taken
into account. Thus, at concentrations close to the CMC, the two systems have a similar hydrody-
namic behavior. The emulsion stability is controlled by the considerable difference between the
values of the disjoining pressure in the cases where the surfactants are in the continuous and in
the disperse phases.
4.5.3.4 Demulsification
It has been known for a long time [34] that one way to destroy an emulsion is to add a surfactant,
which is soluble in the drop phase—this method is termed chemical demulsification. To under-
stand the underlying process, let us consider two colliding emulsion droplets with film formed in
the zone of collision (see Figures 4.32 and 4.57). As discussed earlier, when the liquid is flowing
out of the film, the viscous drag exerted on the film surfaces (from the side of the film interior)
carries away the adsorbed emulsifier toward the film periphery. Thus, a nonuniform surface dis-
tribution of the emulsifier (shown in Figure 4.57a by empty circles) is established. If demulsifier
(the closed circles in Figure 4.57b) is present in the drop phase, it will occupy the interfacial area
freed by the e mulsifier. The result will be a saturation of the adsorption layer, as shown in Figure
4.57b. If the demulsifier is sufficiently surface active, its molecules will be able to decrease sub-
stantially, and even to eliminate completely the interfacial tension gradients, thus changing the
emulsion to type 2 (see Figure 4.55 and Section 4.5.3.2). This leads to a strong increase in the
(a) (b)
FIGURE 4.57 (a) Nonuniform surface distribution of an emulsifier due to drag from the draining film.
(b) Suppression of the surface tension gradients by a demulsifier added in the drop phase.
Chemical Physics of Colloid Systems and Interfaces 369
rate of film thinning, rapid drop coalescence, and emulsion destruction [727,728]. The mecha-
nism mentioned earlier suggests that the demulsifier has to possess the following properties:
1. It must be soluble in the drop phase or in both phases, but in the latter case its solubility in
the drop phase must be much higher.
2. Its diffusivity and concentration must be large enough to provide a sufficiently large demul-
sifier flux toward the surfaces and thus eliminate the gradients of the interfacial tension.
3. Its surface activity must be comparable and even higher than that of the emulsifier; oth-
erwise, even though it may adsorb, it will not be able to suppress the interfacial tension
gradients.
(b)
FIGURE 4.58 Spontaneous cyclic dimpling caused by surfactant diffusion from the aqueous film toward the
two adjacent oil phases. (a) Schematic presentation of the process. (b) Photograph of a large dimple just before
flowing out; the interference fringes in reflected light allow determination of the dimple shape.
Combining the general equation of films with deformable interfaces (Equation 4.266), the
mass balance (Equations 4.287 and 4.288), and the boundary condition for the interfacial stresses
(Equation 4.292), we can derive [814–817]:
∂h 1 ∂ ⎪⎧ 3 ∂ ⎡ σ ∂ ⎛ ∂h ⎞ ⎤ ⎪⎫ 1 ∂ ⎛ jhr 2 ⎞
+ ⎨rh r + Π( h ) ⎥ ⎬ = 2r ∂r ⎜ Γ ⎟ (4.308)
∂t 3ηr ∂r ⎪⎩ ∂r ⎢⎣ r ∂r ⎜⎝ ∂r ⎟⎠ ⎦ ⎪⎭ ⎝ ⎠
where
j is the diffusion flux in the drop phase
r is radial coordinate
h(r, t) is the film thickness
σ is surface tension
Γ is adsorption
Π is disjoining pressure
Chemical Physics of Colloid Systems and Interfaces 371
9
Initial
8 3s
9s
7 17 s
6 29 s
5
h/h0 4
3
2
1
0
0.0 0.2 0.4 0.6 0.8 1.0
r/R
FIGURE 4.59 Comparison between the theory of cyclic dimpling (the lines) and the experimental data (the
points) for the dimple shape, h(r), determined from the interference fringes (see Figure 4.58b); emulsifier is
anionic surfactant sodium nonylphenol polyoxyethylene-25 sulfate and the oil phase is styrene.
The comparison between the numerical calculations based on Equation 4.308 and the experimental
data for the cyclic dimpling with the anionic surfactant sodium nonylphenol polyoxyethylene-25
sulfate shows a very good agreement (Figure 4.59). The experimental points are obtained from the
interference fringes (see Figure 4.58). The shape in the initial moment, t = 0, serves as an initial
condition for determining h(r, t) by solving Equation 4.308. The curves for t = 3, 9, 17, and 29 s
represent theoretical predictions. The scaling parameters along the h- and r-axes in Figure 4.59 are
h 0 = 350 nm and R = 320 μm, with the latter the film radius; the only adjustable parameter is the
diffusion flux, j.
Direction of diffusion
Oil phase +
Surfactant
(a)
(b)
FIGURE 4.60 Osmotic swelling of an aqueous film formed between two oil droplets. (a) The surfactant
dissolved in the oil is transferred by diffusion toward the film, where it forms micelles, the osmotic effects
of which increase the local pressure. (b) Photograph of a typical pattern from a circular film with channels.
counterbalances the outer capillary pressure and arrests further thinning of the film. Moreover, the
excess osmotic pressure in the film gives rise to a convective outflow of solution: this is the physical
origin of the observed channels (Figure 4.60b).
Experimental data [728,811] show that the occurrence of the above phenomenon is the same for
initial surfactant concentration in the water varying from 1 up to 500 times the CMC, if only some
amount of surfactant is also initially dissolved in the oil. This fact implies that the value of the sur-
factant chemical potential inside the oil phase is much greater than that in the aqueous phase, with
the latter closer to its value at the CMC in the investigated range of concentrations.
cases, in the last case the mass transfer is found to destabilize the films. Experimentally, the dif-
fusion transfer of alcohols, acetic acid, and acetone was studied [818,819]. The observed desta-
bilization of the films can be attributed to the appearance of Marangoni instability [812], which
manifests itself through the growth of capillary waves at the interfaces, which eventually can lead
to film rupture.
The Marangoni instabilities can appear not only in thin films, but also in the simpler case of a
single interface. In this case, the Marangoni instability may bring about spontaneous emulsifica-
tion. This effect has been theoretically investigated by Sterling and Scriven [820], whose work
stimulated numerous theoretical and experimental studies on spontaneous emulsification. Lin and
Brenner [821] examined the role of the heat and mass transfer in an attempt to check the hypothesis
of Holly [822] that the Marangoni instability can cause the rupture of tear films. Their analysis was
extended by Castillo and Velarde [823], who accounted for the tight coupling of the heat and mass
transfer and showed that it drastically reduces the threshold for Marangoni convection. Instability
driven by diffusion flux of dissolved oil molecules across an asymmetric liquid film (oil–water–
air film) has been theoretically investigated [813]. It was found that even small decrements of the
water–air surface tension, caused by the adsorbed oil, are sufficient to trigger the instability.
That is the reason why we will now consider expressions for F and M for various types of elemen-
tary motions.
374 Handbook of Surface and Colloid Chemistry
⎛ 8 16 h ⎞ ⎛ 2 Rd ⎞ ⎛ h ⎞
fy = ⎜ + ⎟ ln ⎜ ⎟ + 0.58461 + O ⎜ R ⎟ (4.312)
⎝ 15 375 Rd ⎠ ⎝ h ⎠ ⎝ d ⎠
⎛ 1 43 h ⎞ ⎛ 2 Rd ⎞ ⎛ h ⎞
mx = ⎜ + ⎟ ln ⎜ ⎟ − 0.26227 + O ⎜ R ⎟ (4.313)
⎝ 10 250 Rd ⎠ ⎝ h ⎠ ⎝ d ⎠
where f y and mx are dimensionless drag force and torque coefficients, respectively.
Second, we consider the case of pure rotation: a solid spherical particle of radius Rd is situated at
a surface-to-surface distance, h, from a planar wall and rotates with angular velocity, ω, around the
x-axis in an otherwise quiescent fluid. The corresponding force and torque resultants are [659–661]
2 ⎛ Rd ⎞ ⎛ h ⎞ 2 ⎛ Rd ⎞ ⎛ h ⎞
fy = ln ⎟ − 0.2526 + O ⎜ R ⎟ , m x = 5 ln ⎜ h ⎟ + 0.3817 + O ⎜ R ⎟ (4.315)
15 ⎜⎝ h ⎠ ⎝ d ⎠ ⎝ ⎠ ⎝ d ⎠
From Equations 4.311 through 4.315, it follows that the force and the torque depend weakly
(logarithmically) on the distance, h, as compared to the Taylor or Reynolds laws (Equations 4.271
and 4.272).
Chemical Physics of Colloid Systems and Interfaces 375
101
7 R/Rd = 0.00
6
5
–6
10 2 3 4 6 8 10–5 2 3 4 6 8 10–5
h/Rd
FIGURE 4.61 Deformed fluid particle (the inset) moving tangentially to an immobile solid surface: plot of
the dimensionless drag coefficient, f y, vs. the dimensionless film thickness, h/Rd, for three values of the dimen-
sionless film radius, R/Rd (see Equation 5.317).
As discussed in Sections 4.5.2.1 and 4.5.3.2, a fluid particle in the presence of high surfactant
concentration can be treated as a deformable particle of tangentially immobile surfaces. Such a par-
ticle deforms when pressed against a solid wall (see the inset in Figure 4.61). To describe the drag
due to the film intervening between the deformed particle and the wall, we may use the expression
derived by Reynolds [646] for the drag force exerted on a planar solid ellipsoidal disc, which is
parallel to a solid wall and is moving along the y-axis at a distance h from the wall:
h
Fx = 0, Fy = −πηU (4.316)
ab
Here, a and b are the semiaxes of the ellipse; for a circular disc (or film), we have a = b = R. By
combining Equations 4.311 and 4.312 with Equation 4.316, we can derive an expression for the net
drag force experienced by the deformed particle (the inset in Figure 4.61) when it moves along the
y-axis with a linear velocity U:
R2 ⎛ 8 16 h ⎞ ⎛ 2 Rd ⎞ ⎛ h ⎞
Fy = −6πηURd f y , fy = + + ln ⎟ + 0.58461+ O ⎜ R ⎟ (4.317)
6hRd ⎜⎝ 15 375 Rd ⎟⎠ ⎜⎝ h ⎠ ⎝ d ⎠
Here
h and R denote the film thickness and radius
Rd is the curvature radius of the spherical part of the particle surface
The dependence of the dimensionless drag coefficient, f y, on the distance h for different values of
the ratio R/Rd is illustrated in Figure 4.61. The increase of R/Rd and the decrease of h/Rd may lead
to an increase of the drag force, f y, by an order of magnitude. That is the reason the film between
a deformed particle and a wall can be responsible for the major part of the energy dissipation.
Moreover, the formation of doublets and flocks of droplets separated by liquid films seems to be of
major importance for the rheological behavior of emulsions.
formulation of the same problem was proposed by Liron and Barta [836]. The case of a small drop-
let moving in the restricted space between two parallel solid surfaces was solved by Shapira and
Haber [837,838]. These authors used the Lorentz reflection method to obtain analytical solutions for
the drag force and the shape of a small droplet moving in Couette flow or with constant translational
velocity.
The more complicated case, corresponding to a viscous fluid particle approaching the bound-
ary between two pure fluid phases (all interfaces deformable), was investigated by Yang and Leal
[839,840], who succeeded in obtaining analytical results.
Next, we proceed with the case when surfactant is present and the Marangoni effect becomes
operative. Classical experiments carried out by Lebedev [841] and Silvey [842] show that the mea-
sured velocity of sedimentation, U, of small fluid droplets in a viscous liquid (pure liquid phases
assumed) does not obey the Hadamar [843] and Rybczynski [844] equation:
3ηd + 2η
F = 2πηURd (4.318)
ηd + η
where F is the drag force. The limiting case ηd → 0 corresponds to bubbles, whereas in the other
limit, ηd → ∞, Equation 4.318 describes solid particles. Note that Equation 4.318 is derived for
the motion of a spherical fluid particle (drop or bubble) of viscosity ηd in a liquid of viscosity η in
the absence of any surfactant. The explanation of the contradiction between theory and experi-
ment [841,842] turned out to be very simple: even liquids that are pure from the viewpoint of the
spectral analysis may contain some surface-active impurities, whose bulk concentration might be
vanishingly low, but which can provide a dense adsorption layer at the restricted area of the fluid
particle surface. Then, the effects of Gibbs elasticity and interfacial viscosity substantially affect
the drag coefficient of the fluid particle. The role of the latter two effects was investigated by Levich
[662], Edwards et al. [663], and He et al. [845] for the motion of an emulsion droplet covered with
a monolayer of insoluble surfactant (adsorption and/or desorption not present). These authors used
the Boussinesq–Scriven constitutive law of a viscous fluid interface (Equation 4.292), and estab-
lished that only the dilatational interfacial viscosity, ηdl, but not the shear interfacial viscosity, ηsh,
influences the drag force. If the surfactant is soluble in both phases and the process of adsorption is
diffusion controlled (see Section 4.2.2.1), the generalization of Equation 4.318 is [783]
2η 2
−1
⎡ ⎛ η RE ⎞ ⎤
F = 2πηURd ⎢3 − ⎜ 1 + d + dl + d G ⎟ ⎥ (4.319)
⎢⎣ ⎝ η ηRd 3ηDs 2 + 2( Rd D /ha Ds ) + ( Rd Dd /hd,a Ds ) ⎠ ⎥⎦
where
Dd is the surfactant diffusion coefficient in the drop phase
c and cd are the concentrations of surfactant in the continuous and drop phases, respectively
ha = ∂Γ/∂c and hd,a = ∂Γ/∂cd are the slopes of adsorption isotherms with respect to the surfactant
concentration
In the limiting case without surfactant, Equation 4.319 is reduced to the Hadamar [843] and
Rybczynski [844] equation (Equation 4.318).
A recently developed experimental technique [846–850] gives the possibility to measure pre-
cisely the instantaneous velocity of rising bubbles in a solution as a function of time and the distance
to the starting point. The sensitivity of this technique allows one to determine trace amounts of
impurities in water.
Danov et al. [347,851–854] investigated theoretically the hydrodynamic interaction of a fluid
particle with a fluid interface in the presence of surfactant. The numerical results of these authors
Chemical Physics of Colloid Systems and Interfaces 377
reveal that there is a strong influence of both shear and dilatational interfacial viscosities on the
motion of the fluid particle when the particle–interface distance, h, is approximately equal to or
smaller than the particle radius, Rd. For example, in the presence of an external force acting paral-
lel to the interface (along the y-axis), the stationary motion of the spherical particle close to the
viscous interface is a superposition of a translation along the y-axis with velocity Vy and a rotation
(around the x-axis) with an angular velocity, ωx (see the inset in Figure 4.62a). The numerical results
of Danov et al. [853,854] for Vy and ωx normalized by the Stokes velocity, VStokes = F/(6πηRd), are
plotted in Figure 4.62a and b versus h/Rd for four different types of interfaces: (1) solid particle and
solid wall (see Equations 4.311 through 4.313); (2) fluid particle and fluid interface for K = E = 100
(for the definition of K and E see Equation 4.303); (3) the same system as (2) but for K = E = 10;
(4) the same system as (2) but for K = E = 1. (For the definition of the interfacial viscosities, ηdl and
ηsh, see Equation 4.292). As seen in Figure 4.62a, the velocity of the sphere, Vy, is less than VStokes
for the solid (1) and the highly viscous (2) interfaces, and Vy noticeably decreases when the distance
h decreases. However, in case (4), corresponding to low surface viscosities, the effect is quite dif-
ferent: Vy/VStokes is greater than unity (the sphere moves faster near the interface than in the bulk),
and its dependence on h is rather weak. The result about the angular velocity, ωx, is also intriguing
(Figure 4.62b). The stationary rotation of a sphere close to a solid (1) or highly viscous (2) interface
1.2
(4)
1.1
1.0
0.9
0.8
Vy/VStokes
(3)
0.7
z
0.6 y
(2)
0.5 h
0.4 ωx Vy
0.3 (1)
Rd
0.2
0.001 0.251 0.501 0.751 1.001
(a) h/Rd
0.04 (1)
0.02 (2)
0.00
ωxRd/VStokes
(3)
–0.02
–0.04
–0.06
(4)
–0.08
–0.10
0.001 0.251 0.501 0.751 1.001
(b) h/Rd
FIGURE 4.62 Spherical particle moving tangentially to a viscous interface: plots of the stationary
dimensionless linear (Vy /VStokes) (a) and angular (ωxRd /VStokes) (b) velocities vs. the dimensionless thickness,
h/Rd. The curves correspond to various surface viscosities: (1) K = E = ∞ (solid surfaces); (2) K = E = 100;
(3) K = E = 10, and (4) K = E = 1 (see Equation 4.303).
378 Handbook of Surface and Colloid Chemistry
is in positive direction, that is, ωx > 0. For the intermediate interfacial viscosity (3), the sphere practi-
cally does not rotate, whereas, for the interfaces of low viscosity (4), the drop rotates in the opposite
direction, that is, ωx < 0. The inversion of the sign of ωx is due to the fact that the friction of the
particle with the bulk fluid below it (see the inset in Figure 4.62a) becomes stronger than the friction
with the interface above the particle.
Finally, we consider the case of a solid particle attached to a liquid–fluid interface. This con-
figuration is depicted in Figure 4.21e; note that the position of the particle along the normal to
the interface is determined by the value of the three-phase contact angle. Stoos and Leal [855]
investigated the case when such an attached particle is subjected to a flow directed normally to
the interface. These authors determined the critical capillary number, beyond which the captured
particle is removed from the interface by the flow.
Danov et al. [347] examined the case of an attached particle moving along a liquid–gas interface
under the action of an applied force directed tangentially to the interface. The effects of the contact
angle (the depth of immersion), as well as the effect of adsorbed surfactant on the drag force, were
investigated. These authors also calculated the surface diffusion coefficient of a Brownian particle
attached to the liquid surface. Let Dp and Dp0 be the particle surface diffusion coefficient in the
0.9
K=E=1
0.8
0.7
0.6
Dp/Dp0
0.5 K=E=5
0.4
K = E = 10
0.3
0.2
20 30 40 50 60 70 80 90
Contact angle, θ (deg)
(a)
1.0
0.9 θ = 30˚
θ = 50˚
0.8 θ = 70˚
θ = 90˚
0.7
Dp/Dp0
0.6
0.5
0.4
0.3
0.2
0 1 2 3 4 5 6 7 8 9 10
K=E
(b)
FIGURE 4.63 Effect of adsorbed surfactant on the surface diffusivity, Dp, of a Brownian particle attached
to a fluid interface: (a) plot of Dp /Dp0 vs. particle contact angle, θ, for various surface viscosities (see Equation
4.303); (b) plot of Dp /Dp0 vs. the dimensionless surface viscosity, K = E, for various θ.
Chemical Physics of Colloid Systems and Interfaces 379
presence and absence of surfactant, respectively. In Figure 4.63a, we plot the results for Dp/Dp0
versus the solid–liquid–gas contact angle, θ, for three different values of the parameters K and E
characterizing the surface viscosities (see Equation 4.303): (1) K = E = 1; (2) K = E = 5, and (3) K =
E = 10. The relatively small slope of the curves in Figure 4.63a indicates that Dp/Dp0 depends less
significantly on the contact angle, θ, than on the surface viscosity characterized by K and E. Note,
however, that Dp0 itself depends markedly on θ: the absolute value of Dp0 is smaller for the smaller
values of θ (for deeper immersion of the particle in the liquid phase). Figure 4.63b presents the cal-
culated dependence of Dp/Dp0 on the surface viscosity characterized by K and E (K = E is used in
the calculations) for various fixed values of the contact angle, θ. Apparently, the particle mobility
decreases faster for the smaller values of K and then tends to zero insofar as the fluid surface “solidi-
fies” for the higher values of the surface viscosities. The experimental data from measurements of
the drag coefficient of spherical particles attached to fluid interfaces [346] showed very good agree-
ment with the predictions of the theory [347].
The role of surface viscosity and elasticity on the motion of a solid particle trapped in a thin
film, at an interface, or at a membrane of a spherical vesicle has been recently investigated in Refs.
[856,857]. The theoretical results [856,857] have been applied to process the experimental data
for the drag coefficient of polystyrene latex particles moving throughout the membrane of a giant
lipid vesicle [858–864]. Thus, the interfacial viscosity of membranes has been determined. The
motion of particles with different shapes trapped in thin liquid films and at Langmuir monolay-
ers is studied intensively both theoretically and experimentally because of biological and medical
applications [865–887].
kT
D= (4.320)
6πηm R p
where
Rp is the particle radius
ηm is the viscosity of the liquid medium
This relation was later generalized by Kubo [891] for the cases when the hydrodynamic resistance
becomes important. The further development in this field is reviewed by Davis [824].
The particle–particle interactions lead to a dependence of the viscosity, η, of a colloidal disper-
sion on the particle volume fraction, ϕ. Einstein [892] showed that for a suspension of spherical
particles in the dilute limit:
Later Taylor [893] generalized Equation 4.321 for emulsion systems taking into account the viscous
dissipation of energy due to the flow inside the droplets. Oldroyd [894] took into account the effect
of surface viscosity and generalized the theory of Taylor [893] to diluted monodisperse emulsions
whose droplets have viscous interfaces. Taylor [895], Fröhlich and Sack [896], and Oldroyd [897]
applied asymptotic analysis to derive the next term in Equation 4.321 with respect to the capillary
number. Thus, the effect of droplet interfacial tension was included. This generalization may be
important at high shear rates. Another important generalization is the derivation of appropriate
expressions for the viscosity of suspensions containing particles with different shapes [644,645]. A
third direction of generalization of Equation 4.321 is to calculate the next term in the series with
respect to the volume fraction, ϕ. Batchelor [898] took into account the long-range hydrodynamic
interaction between the particles to derive:
From a mathematical viewpoint, Equation 4.322 is an exact result; however, from a physical
viewpoint, Equation 4.322 is not entirely adequate to the real dispersions, as not only the long-
range hydrodynamic interactions are operative in colloids. A number of empirical expressions
have been proposed in which the coefficient multiplying ϕ2 varies between 5 and 14 [899]. The
development of new powerful numerical methods helped for a better understanding of the rhe-
ology of emulsions [900–908]. The simple shear and Brownian flow of dispersions of elastic
capsules, rough spheres, and liquid droplets were studied in Refs. [901,905,907,908]. The effect
of insoluble surfactants and the drop deformation on the hydrodynamic interactions and on the
rheology of dilute emulsions are the subject of investigation in Refs. [902,904,906]. Loewenberg
and Hinch [900,903] discussed the basic ideas of the numerical simulations of concentrated
emulsion flows. These works are aimed at giving a theoretical interpretation of various experi-
mental results for dilute and concentrated dispersions. When the Peclet number is not small, the
convective term in the diffusion equation (Equations 4.287 and 4.288) cannot be neglected and
the respective problem has no analytical solution. Thus, a complex numerical investigation has
to be applied [909,910].
The formulae of Einstein [890,892], Taylor [893], and Oldroyd [894] have been generalized for
dilute emulsions of mobile surfaces with account for the Gibbs elasticity and the bulk and surface
diffusion and viscosity [911]:
η ⎛ 3 ⎞
= 1 + ⎜ 1 + 〈ε m 〉 ⎟ φ + O(φ2 ), 〈ε m 〉 ≡
∑R ε
3
d m
(4.323)
ηm ⎝ 2 ⎠ ∑R 3
d
where 〈εm〉 is the average value of the interfacial mobility parameter, εm, for all droplets in the con-
trol volume. The mobility parameter of individual drops, εm, and the effective surfactant diffusion
coefficient, Deff, are [911]
Rd D Rd Dd
Deff ≡ Ds + + (4.325)
2ha 3hd,a
(see Equation 4.319 and the following section). If the droplet size distribution in the emulsion and
the interfacial rheological parameters are known, then the average value 〈εm〉 can be estimated.
Chemical Physics of Colloid Systems and Interfaces 381
For monodisperse emulsions, the average value, 〈εm〉, and the interfacial mobility parameter, εm,
are equal. In the special case of completely mobile interfaces, that is, RdEG /(ηmDeff ) → 0 and (3ηdl +
2ηsh)/(Rd ηm) → 0, the mobility parameter, εm, does not depend on the droplet size, and from Equation
4.324 and 4.325, the Taylor [891] formula is obtained. It is important to note that the Taylor formula
takes into account only the bulk properties of the phases (characterized by ηd /ηm); in such a case εm
is independent of Rd and the Taylor equation is also applicable to polydisperse emulsions. If only
the Marangoni effect is neglected (EG → 0), then Equations 4.324 and 4.325 become equivalent to
the Oldroyd [894] formula, which is originally derived only for monodisperse emulsions.
For higher values of the particle volume fraction, the rheological behavior of the colloidal disper-
sions becomes rather complex. We will consider qualitatively the observed phenomena, and next we
will review available semiempirical expressions.
For a simple shear (Couette) flow, the relation between the applied stress, τ, and the resulting
shear rate, γ , can be expressed in the form:
τ = ηγ (4.326)
(e.g., when a liquid is sheared between two plates parallel to the xy plane, we have γ = ∂vx /∂z.)
A typical plot of γ versus τ is shown in Figure 4.64a. For low and high shear rates, we observe
Newtonian behavior (η = constant), whereas in the intermediate region a transition from the lower
shear rate viscosity, η0, to the higher shear rate viscosity, η∞, takes place. This is also visualized in
Figure 4.64b, where the viscosity of the colloidal dispersion, η, is plotted versus the shear rate, γ ;
note that in the intermediate zone η has a minimum value [636,640].
Note also that both η0 and η∞ depend on the particle volume fraction, ϕ. De Kruif et al. [899]
proposed the semiempirical expansions:
η0
= 1 + 2.5φ + (4 ± 2)φ2 + (42 ± 10)φ3 + (4.327)
ηm
η
γ• η0
η∞
cotα2 = η∞
α2 cotα1 = η0
α1
0 τ 0 γ•
(a) (b)
τc Rp3
kT
0.3
0.2
0.1
FIGURE 4.64 Qualitative presentation of basic relations in rheology of suspensions: (a) rate of strain, γ , vs.
applied stress, τ (see Equation 4.326); (b) average viscosity of a suspension, η, vs. rate of strain, γ ; (c) dimen-
sionless parameter τc (Equation 4.330) vs. particle volume fraction ϕ.
382 Handbook of Surface and Colloid Chemistry
η∞
= 1 + 2.5φ + (4 ± 2)φ2 + (25 ± 7)φ3 + (4.328)
ηm
as well as two empirical expressions which can be used in the whole range of values of ϕ:
−2 −2
η0 ⎛ φ ⎞ η∞ ⎛ φ ⎞
= ⎜1 − , = ⎜1 − (4.329)
η m ⎝ 0.63 ⎟⎠ η m ⎝ 0.71 ⎟⎠
In regard to the dependence of η on the shear stress, τ, Russel et al. [404] reported that for the inter-
mediate values of τ, corresponding to non-Newtonian behavior (Figure 4.64a and b), the experimen-
tal data correlate reasonably well with the expression:
η − η∞ 1
= (4.330)
η − η0 1 + (τ /τc )n
with 1 ≤ n ≤ 2, where τc is the value of τ for which η = (η 0 + η∞)/2. In its own turn, τc depends
on the particle volume fraction ϕ (see Figure 4.64c). We see that τc increases with the volume
fraction, ϕ, in dilute dispersions then passes through a maximum and finally decreases down
to zero; note that τc → 0 corresponds to η → η∞. The peak at ϕ ≈ 0.5 is the only indication that
the hard-sphere disorder–order transition either occurs or is rheologically significant in these
systems [404].
The restoring force for a dispersion to return to a random, isotropic situation at rest is either
Brownian (thermal fluctuations) or osmotic [912]. The former is most important for submicrometer
particles and the latter for larger particles. Changing the flow conditions changes the structure, and
this leads to thixotropic effects, which are especially strong in flocculated systems.
Krieger and Dougherty [913] applied the theory of corresponding states to obtain the following
expression for the viscosity of hard-sphere dispersions:
−[ η]φmax
η ⎛ φ ⎞
= ⎜1 − ⎟ (4.331)
ηm ⎝ φmax ⎠
where
[η] is the dimensionless intrinsic viscosity, which has a theoretical value of 2.5 for monodisperse
rigid spheres
ϕmax is the maximum packing volume fraction for which the viscosity η diverges
The value of ϕmax depends on the type of packing of the particles [636] (Table 4.8). The maxi-
mum packing fraction, ϕmax, is very sensitive to particle-size distribution and particle shape [914].
Broader particle-size distributions have greater values of ϕmax. On the other hand, nonspherical
particles lead to poorer space-filling and hence lower ϕmax. Table 4.9 presents the values of [η]
and ϕmax obtained by fitting the results of a number of experiments on dispersions of asymmetric
particles using Equation 4.331. The trend of [η] to increase and of ϕmax to decrease with increas-
ing asymmetry is clearly seen, but the product, [η]ϕmax, is almost constant; [η]ϕmax is about 2 for
spheres and about 1.4 for fibers. This fact can be utilized to estimate the viscosity of a wide variety
of dispersions.
A number of rheological experiments with foams and emulsions are summarized in the reviews
by Prud’home and Khan [915] and Tadros [916]. These experiments demonstrate the influence
Chemical Physics of Colloid Systems and Interfaces 383
TABLE 4.8
Maximum Packing Volume Fraction, ϕmax, for Various Arrangements of
Monodisperse Spheres
Arrangement ϕmax
Simple cubic 0.52
Minimum thermodynamically stable configuration 0.548
Hexagonally packed sheets just touching 0.605
Random close packing 0.637
Body-centered cubic packing 0.68
Face-centered cubic/hexagonal close packed 0.74
TABLE 4.9
Values of [η] and ϕmax for a Number of Dispersions Obtained by Fitting
Experimental Data by Means of Equation 4.331
System [η] ϕmax [η]ϕmax References
Spheres (submicron) 2.7 0.71 1.92 De Kruif et al. [899]
Spheres (40 μm) 3.28 0.61 2.00 Giesekus [926]
Ground gypsum 3.25 0.69 2.24 Turian and Yuan [927]
Titanium dioxide 4.0 0.55 2.75 Turian and Yuan [927]
Glass rods (30 × 700 μm) 9.25 0.268 2.48 Clarke [928]
Quartz grains (53–76 μm) 4.8 0.371 2.15 Clarke [928]
Glass fibers
Axial ratio-7 3.8 0.374 1.42 Giesekus [926]
Axial ratio-14 4.03 0.26 1.31 Giesekus [926]
Axial ratio-21 6.0 0.233 1.40 Giesekus [926]
of films between the droplets (or bubbles) on the shear viscosity of the dispersion as a whole.
Unfortunately, there is no consistent theoretical explanation of this effect accounting for the dif-
ferent hydrodynamic resistance of the films between the deformed fluid particles as compared to
the nondeformed spherical particles (see Sections 4.5.2 and 4.5.3). In the case of emulsions and
foams, the deformed droplets or bubbles have a polyhedral shape, and maximum packing fraction
can be ϕmax ≈ 0.9 and even higher. For this case, a special geometrical rheological theory has been
developed [663,917,918].
Wessel and Ball [919] and Kanai et al. [920] studied in detail the effects of shear rate on the frac-
tal structure of flocculated emulsion drops. They showed that the size of the flocs usually decreases
with the increase of the shear stress; often the flocs are split to single particles at high shear rates.
As a result, the viscosity decreases rapidly with the increase of shear rate.
Interesting effects are observed when dispersion contains both larger and smaller particles; the
latter are usually polymer coils, spherical or cylindrical surfactant micelles, or microemulsion drop-
lets. The presence of the smaller particles may induce clustering of the larger particles due to the
depletion attraction (see Section 4.4.5.3.3); such effects are described in the works on surfactant-
flocculated and polymer-flocculated emulsions [921–924]. Other effects can be observed in disper-
sions representing mixtures of liquid and solid particles. Yuhua et al. [925] have established that if
the size of the solid particles is larger than three times the size of the emulsion drops, the emulsion
384 Handbook of Surface and Colloid Chemistry
can be treated as a continuous medium (of its own average viscosity), in which the solid particles are
dispersed; such treatment is not possible when the solid particles are smaller.
Rheological properties of foams (elasticity, plasticity, and viscosity) play an important role in
foam production, transportation, and applications. In the absence of external stress, the bubbles in
foams are symmetrical and the tensions of the formed foam films are balanced inside the foam and
close to the walls of the vessel [929]. At low external shear stresses, the bubbles deform and the
deformations of the thin liquid films between them create elastic shear stresses. At a sufficiently
large applied shear stress, the foam begins to flow. This stress is called the yield stress, τ0. Then,
Equation 4.326 has to be replaced with the Bingham plastic model [930]:
τ = τ0 + ηγ (4.332)
Experiments show that in steadily sheared foams and concentrated emulsions, the viscosity coef-
ficient η depends on the rate of shear strain, and in most cases the Herschel–Bulkley equation [931]
is applicable:
Here
K is the consistency
n is the power-law index; n < 1 for shear thinning, whereas n > 1 for shear thickening
Systematic studies of the foam rheology [932–939] show that the power-law index varies between
0.25 and 0.5 depending on the elasticity of the individual air–solution surfaces. If the elasticity is
lower than 10 mN/m, then n is close to 0.25, whereas for large surface elasticity (>100 mN/m) n
increases to 0.5.
4.6.1 Irreversible Coagulation
The kinetic theory of fast irreversible coagulation was developed by von Smoluchowsky [940,941].
Later, the theory was extended to the case of slow and reversible coagulation. In any case of coagu-
lation (flocculation), the general set of kinetic equations reads:
k −1
dnk 1
∞
= ∑ a if,k −i ni nk −i − nk ∑a k, i
f n + qk
i (k = 1, 2,…) (4.334)
dt 2 i =1 i =1
Here
t is time
n1 denotes the number of single particles per unit volume
nk is the number of aggregates of k particles (k = 2, 3, …) per unit volume
a if, j (i, j = 1, 2, 3, …) are rate constants of flocculation (coagulation; see Figure 4.65)
qk denotes the flux of aggregates of size k which are products of other processes, different from
the flocculation itself (say, the reverse process of aggregate disassembly or the droplet coales-
cence in emulsions; see Equations 4.346 and 4.350)
In the special case of irreversible coagulation without coalescence, we have qk ≡ 0. The first term in
the right-hand side of Equation 4.334 is the rate of formation of k aggregates by the merging of two
smaller aggregates, whereas the second term expresses the rate of loss of k aggregates due to their
incorporation into larger aggregates. The total concentration of aggregates (as kinetically indepen-
dent units), n, and the total concentration of the constituent particles (including those in aggregated
form), ntot, can be expressed as
∞ ∞
n= ∑ nk , ntot = ∑ kn k (4.335)
k =1 k =1
1,2
af
+
a2,3
f
+
FIGURE 4.65 Elementary acts of flocculation according to the Smoluchowski scheme; a if, j (i, j = 1, 2, 3, …)
are rate constants of flocculation.
386 Handbook of Surface and Colloid Chemistry
The Einstein approach (see Equation 4.320), combined with the Rybczynski–Hadamar equa-
tion (Equation 4.318), leads to the following expression for the relative diffusivity of two isolated
Brownian droplets:
kT ηd + η ⎛ 1 1 ⎞
Di(,0j) = ⎜ + ⎟ (perikinetic coagulationn) (4.337)
2πη 3ηd + 2η ⎝ Ri Rj ⎠
The limiting case ηd → 0 corresponds to two bubbles, whereas in the other limit (ηd → ∞) Equation
4.337 describes two solid particles or two fluid particles of tangentially immobile surfaces.
When the particle relative motion is driven by a body force or by the thermocapillary migration
(rather than by self-diffusion), Equation 4.337 is no longer valid. Instead, in Equation 4.336 we have
to formally substitute the following expression for Di(,0j) (see Rogers and Davis [943]):
1
Di(,0j) = ( Ri + R j ) vi − v j (orthokinetic coagulation) (4.338)
4
Here vj denotes the velocity of a flock of aggregation number j. Physically, Equation 4.338 accounts
for the fact that some particle (usually a larger one) moves faster than the remaining particles and
can “capture” them upon collision. This type of coagulation is called orthokinetic to distinguish it
from the self-diffusion-driven perikinetic coagulation described by Equation 4.337. In the case of
gravity-driven flocculation, we can identify vj with the velocity U in Equation 4.318, where F is to
be set equal to the gravitational force exerted on the particle; for a solid particle or a fluid particle of
tangentially immobile surface, this yields v j = 2gΔρR 2j / (9η) with g the acceleration due to gravity
and Δρ the density difference between the two phases.
In the case of orthokinetic coagulation of liquid drops driven by the thermocapillary migration,
the particle velocity vj is given by the expression (see Young et al. [944]):
2 R j ET λ
vj = ∇(ln T ) (thermocapillary velocity) (4.339)
(3ηd + 2η)(λ d + 2λ)
where the thermal conductivity of the continuous and disperse phases are denoted by λ and λd,
respectively. The interfacial thermal elasticity, ET, is defined by Equation 4.293.
The collision efficiency, Ei,j, in Equation 4.336 accounts for the interactions (of both hydrody-
namic and intermolecular origin) between two colliding particles. The inverse of Ei,j is often called the
stability ratio or the Fuchs factor [945] and can be expressed in the following general form [14,696]:
⎡ U (s) ⎤
∞
1 β(s) 2h
Wi, j = =2 exp ⎢ i, j ⎥ ds, s ≡
Ei, j
0
∫
( s + 2) 2
⎣ kT ⎦ Ri + R j
−1
⎛ 3η + 2η ⎞ Fz
β ≡ ⎜ 2πηR* d ⎟ (4.340a)
⎝ ηd + η ⎠ Vz
where
h is the closest surface-to-surface distance between the two particles
R* is defined by Equation 4.270
Ui,j(s) is the energy of (nonhydrodynamic) interactions between the particles (see Section 4.4)
β(s) accounts for the hydrodynamic interactions
Fz /Vz is the particle friction coefficient
Thus, β → 1 for s → ∞, insofar as for large separations the particles obey the Rybczynski–Hadamar
equation (Equation 4.318). In the opposite limit, s ≪ 1, that is, close approach of the two particles,
Chemical Physics of Colloid Systems and Interfaces 387
Fz /Vz can be calculated from Equations 4.271, 4.294 through 4.297, or 4.301, depending on the
specific case. In particular, for s ≪ 1 we have β ∝ 1/s for two solid particles (or fluid particles of
tangentially immobile surfaces), β ∝ s−1/2 for two liquid droplets, and β ∝ ln s for two gas bubbles.
We see that for two solid particles (β ∝ 1/s), the integral in Equation 4.340a may be divergent. To
overcome this problem, one usually accepts that for the smallest separations Ui,j is dominated by
the van der Waals interaction, as given by Equation 4.185, that is, Ui,j → −∞, and, consequently, the
integrand in Equation 4.340a tends to zero for s → 0.
Note that the value of Wi,j is determined mainly by the values of the integrand in the vicinity of the
electrostatic maximum (barrier) of Ui,j (see Figure 4.17), insofar as Ui,j enters Equation 4.340a as an expo-
nent. By using the method of the saddle point, Derjaguin [14] estimated the integral in Equation 4.340a:
1/ 2
1 ⎡ 8πkT ⎤ β(sm ) ⎡ U (s ) ⎤
Wi, j ≡ ≈⎢ ⎥ exp ⎢ i, j m ⎥ (4.340b)
Ei, j ⎣ −Uiʹʹ, j (sm ) ⎦ ( sm + 2 )2 ⎣ kT ⎦
where sm denotes the value of s corresponding to the maximum. We see that the larger the barrier,
Ui,j(sm), the smaller the collision efficiency, Ei,j, and the slower the coagulation.
Note also that for imaginary particles, which experience neither long-range surface forces
(Ui,j = 0) nor hydrodynamic interactions (β = 1), Equation 4.340a yields a collision efficiency Ei,j = 1
and Equation 4.336 reduces to the Smoluchowski [940,941] expression for the rate constant of the
fast irreversible coagulation. In this particular case, Equation 4.334 represents an infinite set of
nonlinear differential equations. If all flocculation rate constants are the same and equal to af, the
problem has a unique exact solution [940,941]:
n0 (a f n0t / 2)k −1
n= , nk = n0 (k = 1, 2,…) (4.341)
1 + a f n0t / 2 (1 + a f n0t / 2)k +1
It is supposed that the total average concentration of the constituent particles (in both singlet and
aggregated form), ntot, does not change and is equal to the initial number of particles, n0. Unlike ntot,
the concentration of the aggregates, n, decreases with time, while their size increases. Differentiating
Equation 4.341 we obtain:
dn a dV a f φ
= − f n2 , = φ0 , V ≡ 0 (4.342)
dt 2 dt 2 n
where
V is the average volume per aggregate
ϕ 0 is the initial volume fraction of the constituent particles
Combining Equations 4.336 and 4.342, we obtain the following result for perikinetic (Brownian)
coagulation:
V t R02
= 1+ , t Br = (4.343)
V0 t Br 3φ0 D0 E0
where
V0 = 4πR03 /3 is the volume of a constituent particle
t Br is the characteristic time of the coagulation process in this case
E 0 is an average collision efficiency
D 0 is an average diffusion coefficient
388 Handbook of Surface and Colloid Chemistry
In contrast, V is not a linear function of time for orthokinetic coagulation. When the flocculation is
driven by a body force, that is, in case of sedimentation or centrifugation, we obtain [942]:
−3
V ⎛ t ⎞ 2 R0
= ⎜1 − , t bf = (4.344)
V0 ⎝ 3tbf ⎠
⎟ 3φ0 vbf E0
where
t bf is the characteristic time in this case
vbf is an average velocity of aggregate motion
As discussed earlier, when the body force is gravitational, we have vbf = 2gΔρR02 / (9η).
When the orthokinetic coagulation is driven by the thermocapillary migration, the counterpart
of Equation 4.340 reads [942]
V ⎛ t ⎞ 2 R0
= exp ⎜ ⎟ , t tm = 3φ v E (4.345)
V0 ⎝ t tm ⎠ 0 tm 0
where
vtm is an average velocity of thermocapillary migration
ttm is the respective characteristic time
Note that D0 ∝ R0−1, vbf ∝ R02, and vtm ∝ R0 (see Equations 4.320 and 4.339). Then, from Equations
4.343 through 4.345, it follows that the three different characteristic times exhibit different depen-
dencies on particle radius: t Br ∝ R03 , t bf ∝ R0−1, while ttm is independent of R0. Thus, the Brownian
coagulation is faster for the smaller particles, the body force–induced coagulation is more rapid for
the larger particles, whereas the thermocapillary driven coagulation is not so sensitive to the particle
size [946].
The Smoluchowski scheme based on Equations 4.341 and 4.342 has found numerous a pplications.
An example for biochemical application is the study [947,948] of the kinetics of flocculation of latex
particles caused by human gamma globulin in the presence of specific “key-lock” interactions. The
infinite set of Smoluchowski equations (Equation 4.334) was solved by Bak and Heilmann [949] in
the particular case when the aggregates cannot grow larger than a given size; an explicit analytical
solution was obtained by these authors.
4.6.2 Reversible Coagulation
In the case of reversible coagulation, the flocs can disaggregate because the primary minimum
(Figure 4.17) is not deep enough [14]. For example, an aggregate composed of i + j particles can be
split on two aggregates containing i and j particles. We denote the rate constant of this reverse pro-
cess by ari, j (Figure 4.66a). It is assumed that both the straight process of flocculation (Figure 4.65)
and the reverse process (Figure 4.66a) take place. The kinetics of aggregation in this more general
case is described by the Smoluchowski set of equations, Equation 4.334, where we have to substitute:
k −1
1
∞ ∞
q1 = ∑ ar1,i ni +1, qk = ∑ ark ,i ni + k − nk ∑a i ,k −i
r (k = 2,, 3,…) (4.346)
i =1 i =1
2 i =1
In Equation 4.346 qk equals the rate of formation of k aggregates in the process of disassembly
of larger aggregates minus the rate of decay of the k aggregates. As before, the total number of
Chemical Physics of Colloid Systems and Interfaces 389
ari,j
+
ack,i
k particles i particles
(b)
FIGURE 4.66 Elementary acts of aggregate splitting (a) and droplet coalescence within an aggregate (b); ari, j
and ack ,i (i, j, k = 1, 2, 3, …) are the rate constants of the respective processes.
constituent particles, ntot, does not change. However, the total number of the aggregates, n, can either
increase or decrease depending on whether the straight or the reverse process prevails. Summing up
all Equations in 4.334 and using Equation 4.346, we derive the following equation for n:
dn 1
∞ ∞
=
dt 2 ∑ ∑(ai =1 j =1
i, j
r n i+ j )
− a if, j ni n j (4.347)
Martinov and Muller [950] reported a general expression for the rate constants of the reverse process:
Di(,0j) Ei, j 1
ari, j = (4.348)
Z i, j ( Ri + R j )2
where Zi,j is the so-called irreversible factor, which can be presented in the form
1 ⎡ U (s) ⎤
Zi, j = (s + 2)2 exp ⎢ − i, j ⎥ ds (4.349)
8 ∫
Ui , j <0
⎣ kT ⎦
The integration in Equation 4.349 is carried out over the region around the primary minimum,
where Ui,j takes negative values (see Figure 4.17). In other words, Zi,j is determined by the values
of Ui,j in the region of the primary minimum, whereas Ei,j is determined by the values of Ui,j in
the region of the electrostatic maximum (see Equations 4.340b and 4.349). When the minimum is
deeper, Zi,j is larger and the rate constant in Equation 4.348 is smaller. In addition, as seen from
Equations 4.340b and 4.348, the increase of the height of the barrier also decreases the rate of the
reverse process. The physical interpretation of this fact is that to detach from an aggregate a particle
has to first go out from the well and then to “jump” over the barrier (Figure 4.17).
To illustrate the effect of the reverse process on the rate of flocculation, we solved numeri-
cally the set of Equations 4.334, 4.346, and 4.347. To simplify the problem, we used the following
assumptions: (1) the von Smoluchowski assumption that all rate constants of the straight process are
equal to af ; (2) aggregates containing more than M particles cannot decay; (3) all rate constants of
the reverse process are equal to ar; and (4) at the initial moment, only single constituent particles
of concentration n0 are available. In Figure 4.67, we plot the calculated curves of n0/n versus the
dimensionless time, τ = afn 0 t/2, for a fixed value, M = 4, and various values of the ratio of the rate
390 Handbook of Surface and Colloid Chemistry
n0/n
3
b = 0.0
b = 1.0
2 b = 2.0
b = 3.0
b = 5.0
1
0 2 4 6 8 10 12 14 16 18 20
Dimensionless time, τ
FIGURE 4.67 Reversible coagulation: theoretical plot of the inverse dimensionless aggregate concentration,
n 0/n, vs. the dimensionless time, τ = afn 0 t/2, in the case of M = 4 and various values of the dimensionless ratio,
b = 2ar/(n 0 af), of the rate constants of the reverse and straight process, ar and af.
constants of the straight and the reverse process, b = 2ar/(n 0 af). Note that n is defined by Equation
4.335. We see that in an initial time interval all curves in Figure 4.67 touch the von Smoluchowski
distribution (corresponding to b = 0), but after this period we observe a reduction in the rate of floc-
culation, which is larger for the curves with larger values of b (larger rate constants of the reverse
process). These S-shaped curves are typical for the case of reversible coagulation, which is also
confirmed by the experiment [14,951].
∞ ∞ k −1
q1 = ∑ aci,1ni , qk = ∑ aci,k ni − nk ∑a k ,i
c (k = 2, 3,…) (4.350)
i =2 i = k +1 i =1
where ack ,i is the rate constant of transformation (by coalescence) of an aggregate containing k drop-
lets into an aggregate containing i droplets (see Figure 4.66b). The newly formed aggregate is fur-
ther involved in the flocculation scheme, which thus accounts for the fact that the flocculation and
coalescence processes are interdependent. In this scheme, the total coalescence rate, aci,tot, and the
total number of droplets, ntot, obey the following equation [955]:
i −1
dntot
∞
=− aci,tot ni ,
∑ aci,tot = ∑ (i − k)a i,k
c (i = 2, 3,…) (4.351)
dt i =2 k =1
Chemical Physics of Colloid Systems and Interfaces 391
To determine the rate constants of coalescence, ack ,i, Danov et al. [665] examined the effects of drop-
let interactions and Brownian motion on the coalescence rate in dilute emulsions of micrometer- and
submicrometer-sized droplets. The processes of film formation, thinning, and rupture were included
as consecutive stages in the scheme of coalescence. Expressions for the interaction energy due to the
various DLVO and non-DLVO surface forces between two deformed droplets were obtained [392]
(see also Section 4.4).
Average models for the total number of droplets are also available [956,957]. The average
model of van den Tempel [956] assumes linear structure of the aggregates. The coalescence rate
is supposed to be proportional to the number of contacts within an aggregate. To simplify the
problem, van den Tempel has used several assumptions, one of them is that the concentration of
the single droplets, n1, obeys the Smoluchowski distribution (Equation 4.341) for k = 1. The aver-
age model of Borwankar et al. [957] is similar to that of van den Tempel but is physically more
adequate. The assumptions used by the latter authors [957] make their solution more applicable
to cases in which the flocculation (rather than the coalescence) is slow and is the rate determining
stage. This is confirmed by the curves shown in Figure 4.68 which are calculated for the same rate
of coalescence, but for three different rates of flocculation. For relatively high rates of flocculation
(Figure 4.68a), the predictions of the three theories differ. For the intermediate rates of floccula-
tion (Figure 4.68b), the prediction of the model by Borwankar et al. [957] is close to that of the
more detailed model by Danov et al. [955]. For very low values of the flocculation rate constant,
af, for which the coalescence is not the rate-determining stage, all three theories [955–957] give
numerically close results (Figure 4.68c). Details about the coupling of coalescence and floccula-
tion in dilute oil-in-water emulsions, experimental investigations, and numerical modeling can be
found in Refs. [958–966].
1
2 1
100 100 2
3
3
10–1 2
10–1
10–2
ntot/n0
ntot/n0
2
10–3 10–2
10–4 2
10–3
10–5
2
0 5,000 10,000 15,000 20,000 25,000 0 5,000 10,000 15,000 20,000 25,000
(a) Time (s) (b) Time (s)
1
100 2
9
3
8
7
6
ntot/n0
3
2.5
2
0 5,000 10,000 15,000 20,000 25,000
(c) Time (s)
FIGURE 4.68 Relative change in the total number of drops, ntot, vs. time, t; initial number of primary
drops n 0 = 1012 cm−3; coalescence rate constant kc2,1 †= 10 −3 s−1. Curve 1: numerical solution of Equation
4.351. Curve 2: output of the model of Borwankar et al. [957]. Curve 3: output of the model of van den
Tempel [956]. The values of the flocculation rate constant are (a) af = 10 –11 cm3/s; (b) af = 10 –13 cm3/s;
(c) af = 10 –16 cm3/s.
392 Handbook of Surface and Colloid Chemistry
ACKNOWLEDGMENTS
The authors gratefully acknowledge the support from the FP7 project Beyond-Everest, and from
COST Action CM1101.
REFERENCES
1. Jungermann, E., Cationic Surfactants, Marcel Dekker, New York, 1970.
2. Lucassen-Reynders, E.H., Anionic Surfactants—Physical Chemistry of Surfactant Action, Marcel
Dekker, New York, 1981.
3. Schick, M.J., Nonionic Surfactants: Physical Chemistry, Marcel Dekker, New York, 1986.
4. Gibbs, J.W., The Scientific Papers of J.W. Gibbs, Vol. 1, Dover, New York, 1961.
5. Ono, S. and Kondo, S., Molecular theory of surface tension in liquids, in Handbuch der Physik, Vol. 3/10,
Flügge, S. (Ed.), Springer, Berlin, Germany, 1960, pp. 134–280.
6. Adamson, A.W. and Gast, A.P., Physical Chemistry of Surfaces, 6th edn., Wiley, New York, 1997.
7. Freundlich, H., Colloid and Capillary Chemistry, Methuen, London, U.K., 1926.
8. Langmuir, I., J. Am. Chem. Soc., 40, 1361, 1918.
9. Volmer, M., Z. Physikal. Chem., 115, 253, 1925.
10. Frumkin, A., Z. Physikal. Chem., 116, 466, 1925.
11. Hill, T.L., An Introduction to Statistical Thermodynamics, Addison-Wesley, Reading, MA, 1962.
12. Lucassen-Reynders, E.H., J. Phys. Chem., 70, 1777, 1966.
13. Borwankar, R.P. and Wasan, D.T., Chem. Eng. Sci., 43, 1323, 1988.
14. Derjaguin, B.V., Theory of Stability of Colloids and Thin Liquid Films, Plenum Press, Consultants
Bureau, New York, 1989.
15. Shchukin, E.D., Pertsov, A.V., and Amelina, E.A., Colloid Chemistry, Moscow University Press, Moscow,
Russia, 1982 (Russian); Elsevier, 2001 (English).
16. Danov, K.D. and Kralchevsky, P.A., Colloid J., 74, 172, 2012.
17. Rosen, M.J. and Aronson, S., Colloids Surf., 3, 201, 1981.
18. Zeldowitch, J., Acta Physicochim. (USSR), 1, 961, 1934.
19. Halsey, G. and Taylor, H.S., J. Chem. Phys., 15, 624, 1947.
20. Gurkov, T.G., Kralchevsky, P.A., and Nagayama, K., Colloid Polym. Sci., 274, 227, 1996.
21. Butler, J.A.V., Proc. Roy. Soc. Ser. A, 135, 348, 1932.
22. Fainerman, V.B. and Miller, R., Langmuir, 12, 6011, 1996.
23. Vaughn, M.W. and Slattery, J. C., J. Colloid Interface Sci., 195, 1, 1997.
24. Makievski, A.V., Fainerman, V.B., Bree, M., Wüstneck, R., Krägel, J., and Miller, R., J. Phys. Chem. B,
102, 417, 1998.
25. Landau, L.D. and Lifshitz, E.M., Statistical Physics, Part 1, Pergamon, Oxford, U.K., 1980.
26. Kralchevsky, P.A., Danov, K.D., Broze, G., and Mehreteab, A., Langmuir, 15, 2351, 1999.
27. Hachisu, S., J. Colloid Interface Sci., 33, 445, 1970.
28. Kalinin, V.V. and Radke, C.J., Colloids Surf. A, 114, 337, 1996.
29. Warszyński, P., Barzyk, W., Lunkenheimer, K., and Fruhner, H., J. Phys. Chem. B, 102, 10948, 1998.
30. Prosser, A.J. and Frances, E.I., Colloids Surf. A, 178, 1, 2001.
31. Kirkwood, J.G. and Oppenheim, I., Chemical Thermodynamics, McGraw-Hill, New York, 1961.
32. Robinson, R.A. and Stokes, R.H., Electrolyte Solutions, Butterworths, London, U.K., 1959.
33. Gouy, L.G., J. Phys., 9, 457, 1910.
34. Davies, J. and Rideal, E., Interfacial Phenomena, Academic Press, New York, 1963.
35. Grahame, D.C., Chem. Rev., 41, 441, 1947.
36. Israelachvili, J.N., Intermolecular and Surface Forces, Academic Press, London, U.K., 2011.
37. Kralchevsky, P.A. and Nagayama, K., Particles at Fluid Interfaces and Membranes, Elsevier, Amsterdam,
the Netherlands, 2001.
38. Matijevič, E. and Pethica, B.A., Trans. Faraday Soc., 54, 1382, 1958.
39. van Voorst Vader, F., Trans. Faraday Soc., 56, 1067, 1960.
40. Tajima, K., Bull. Chem. Soc. Jpn., 44, 1767, 1971.
41. Stern, O., Ztschr. Elektrochem., 30, 508, 1924.
42. Tajima, K., Muramatsu, M., and Sasaki, T., Bull. Chem. Soc. Jpn., 43, 1991, 1970.
43. Tajima, K., Bull. Chem. Soc. Jpn., 43, 3063, 1970.
44. Kolev, V.L., Danov, K.D., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Langmuir, 18, 9106, 2002.
Chemical Physics of Colloid Systems and Interfaces 393
45. Cross, A.W. and Jayson, G.G., J. Colloid Interface Sci., 162, 45, 1994.
46. Johnson, S.B., Drummond, C.J., Scales, P.J., and Nishimura, S., Langmuir, 11, 2367, 1995.
47. Alargova, R.G., Danov, K.D., Petkov, J.T., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Langmuir,
13, 5544, 1997.
48. Rathman, J.F. and Scamehorn, J.F., J. Phys. Chem., 88, 5807, 1984.
49. Berr, S.S., Coleman, M.J., Marriot, J., and Johnson Jr., J.S., J. Phys. Chem., 90, 6492, 1986.
50. Rosen, M.J., Surfactants and Interfacial Phenomena, Wiley, New York, 1989.
51. Clint, J., Surfactant Aggregation, Chapman & Hall, London, U.K., 1992.
52. Alargova, R.G., Danov, K.D., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Langmuir, 14, 4036,
1998.
53. Dimov, N.K., Kolev, V.L., Kralchevsky, P.A., Lyutov, L.G., Brose, G., and Mehreteab, A., J. Colloid
Interface Sci., 256, 23, 2002.
54. Kralchevsky, P.A., Danov, K.D., Kolev, V.L., Broze, G., and Mehreteab, A., Langmuir, 19, 5004, 2003.
55. Danov, K.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., J. Colloid Interface Sci., 300,
809, 2006.
56. Lunkenheimer, K., Barzyk, W., Hirte, R., and Rudert, R., Langmuir, 19, 6140, 2003.
57. Christov, N.C., Danov, K.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., Langmuir, 22,
7528, 2006.
58. Danov, K.D., Kralchevska, S.D., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Langmuir, 19, 5019,
2003.
59. Danov, K.D., Kralchevska, S.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., Langmuir,
20, 5445, 2004.
60. Valkovska, D.S., Shearman, G.C., Bain, C.D., Darton, R.C., and Eastoe, J., Langmuir, 20, 4436, 2004.
61. Day, J.P.R., Campbell, R.A., Russell, O.P., and Bain, C.D., J. Phys. Chem. C, 111, 8757, 2007.
62. Valkovska, D.S., Danov, K.D., and Ivanov, I.B., Colloids Surf. A, 175, 179, 2000.
63. Danov, K.D., Kralchevsky, P.A., and Ivanov, I.B., Dynamic processes in surfactant stabilized emulsions,
in Encyclopedic Handbook of Emulsion Technology, Sjöblom, J. (Ed.), Marcel Dekker, New York, 2001,
Chapter 26 pp. 621–659.
64. Dukhin, S.S., Kretzschmar, G., and Miller, R., Dynamics of Adsorption at Liquid Interfaces, Elsevier,
Amsterdam, the Netherlands, 1995.
65. Eastoe, J. and Dalton, J.S., Adv. Colloid Interface Sci., 85, 103, 2000.
66. Rayleigh, L., Proc. Roy. Soc. (Lond.), 29, 71, 1879.
67. Bohr, N., Philos. Trans. Roy. Soc. (Lond.) A, 209, 281, 1909.
68. Defay, R. and Pétré, G., Dynamic surface tension, in Surface and Colloid Science, Vol. 3, Matijevič, E.
(Ed.), Wiley, New York, 1971, p. 27.
69. Miller, R. and Kretzschmar, G., Adv. Colloid Interface Sci., 37, 97, 1991.
70. Wantke, K.-D., Lunkenheimer, K., and Hempt, C., J. Colloid Interface Sci., 159, 28, 1993.
71. Chang, C.-H. and Franses, E.I., J. Colloid Interface Sci., 164, 107, 1994.
72. Johnson, D.O. and Stebe, K.J., J. Colloid Interface Sci., 182, 525, 1996.
73. Horozov, T. and Arnaudov, L., J. Colloid Interface Sci., 219, 99, 1999.
74. Horozov, T. and Arnaudov, L., J. Colloid Interface Sci., 222, 146, 2000.
75. van den Tempel, M. and Lucassen-Reynders, E.H., Adv. Colloid Interface Sci., 18, 281, 1983.
76. Langevin, D., Colloids Surf., 43, 121, 1990.
77. Lemaire, C. and Langevin, D., Colloids Surf., 65, 101, 1992.
78. Grigorev, D.O., Krotov, V.V., and Noskov, B.A., Colloid J., 56, 562, 1994.
79. Mysels, K.J., Colloids Surf., 43, 241, 1990.
80. Kralchevsky, P.A., Radkov, Y.S., and Denkov, N.D., J. Colloid Interface Sci., 161, 361, 1993.
81. Fainerman, V.B., Miller, R., and Joos, P., Colloid Polym. Sci., 272, 731, 1994.
82. Fainerman, V.B. and Miller, R., J. Colloid Interface Sci., 176, 118, 1995.
83. Horozov, T.S., Dushkin, C.D., Danov, K.D., Arnaudov, L.N., Velev, O.D., Mehreteab, A., and Broze, G.,
Colloids Surf. A, 113, 117, 1996.
84. Mishchuk, N.A., Dukhin, S.S., Fainerman, V.B., Kovalchuk, V.I., and Miller, R., Colloids Surf. A, 192,
157, 2001.
85. van den Bogaert, R. and Joos, P., J. Phys. Chem., 83, 17, 1979.
86. Möbius, D. and Miller, R. (Eds.), Drops and Bubbles in Interfacial Research, Elsevier, Amsterdam, the
Netehrlands, 1998.
87. Jho, C. and Burke, R., J. Colloid Interface Sci., 95, 61, 1983.
88. Joos, P. and van Hunsel, J., Colloid Polym. Sci., 267, 1026, 1989.
394 Handbook of Surface and Colloid Chemistry
89. Fainerman, V.B. and Miller, R., Colloids Surf. A, 97, 255, 1995.
90. Miller, R., Bree, M., and Fainerman, V.B., Colloids Surf. A, 142, 237, 1998.
91. Senkel, O., Miller, R., and Fainerman, V.B., Colloids Surf. A, 143, 517, 1998.
92. Bain, C.D., Manning-Benson, S., and Darton, R.C., J. Colloid Interface Sci., 229, 247, 2000.
93. Rotenberg, Y., Boruvka, L., and Neumann, A.W., J. Colloid Interface Sci., 37, 169, 1983.
94. Makievski, A.V., Loglio, G., Krägel, J., Miller, R., Fainerman, V.B., and Neumann, A.W., J. Phys. Chem.,
103, 9557, 1999.
95. Joos, P., Dynamic Surface Phenomena, VSP BV, AH Zeist, the Netherlands, 1999.
96. Ward, A.F.H. and Tordai, L., J. Chem. Phys., 14, 453, 1946.
97. Miller, R., Colloid Polym. Sci., 259, 375, 1981.
98. McCoy, B.J., Colloid Polym. Sci., 261, 535, 1983.
99. Hansen, R.S., J. Chem. Phys., 64, 637, 1960.
100. Filippov, L.K., J. Colloid Interface Sci., 164, 471, 1994.
101. Daniel, R. and Berg, J.C., J. Colloid Intrface Sci., 237, 294, 2001.
102. Sutherland, K.L., Aust. J. Sci. Res., A5, 683, 1952.
103. Arfken, G.B., Weber, H.J, and Harris, F.E., Mathematical Methods for Physicists, Elsevier, Amsterdam,
2013.
104. Korn, G.A. and Korn, T.M., Mathematical Handbook, McGraw-Hill, New York, 1968.
105. Danov, K.D., Kolev, V.L., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Langmuir, 16, 2942, 2000.
106. Dukhin, S.S., Miller, R., and Kretzschmar, G., Colloid Polym. Sci., 261, 335, 1983.
107. Dukhin, S.S. and Miller, R., Colloid Polym. Sci., 272, 548, 1994.
108. MacLeod, C. and Radke, C.J., Langmuir, 10, 3555, 1994.
109. Vlahovska, P.M., Danov, K.D., Mehreteab, A., and Broze, G., J. Colloid Interface Sci., 192, 194, 1997.
110. Danov, K.D., Vlahovska, P.M., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Colloids Surf. A, 156,
389, 1999.
111. Diamant, H. and Andelman, D., J. Phys. Chem., 100, 13732, 1996.
112. Diamant, H., Ariel, G., and Andelman, D., Colloids Surf. A, 183–185, 259, 2001.
113. Dattwani, S.S. and Stebe, K.J., J. Colloid Interface Sci., 219, 282, 1999.
114. Danov, K.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., J. Colloid Interface Sci., 303,
56, 2006.
115. Nayfeh, A.H., Perturbation Methods, Wiley, New York, 1973.
116. Rillaerts, E. and Joos, P., J. Colloid Interface Sci., 88, 1, 1982.
117. Durbut, P., Surface activity, in Handbook of Detergents, Part A, Broze, G. (Ed.), Marcel Dekker, New
York, 1999, Chapter 3 pp. 47–98.
118. Bond, W.N. and Puls, H.O., Philos. Mag., 24, 864, 1937.
119. Doss, K.S.G., Koll. Z., 84, 138, 1938.
120. Blair, C.M., J. Chem. Phys., 16, 113, 1948.
121. Ward, A.F.H., Surface Chemistry, Butterworths, London, U.K., 1949.
122. Dervichian, D.G., Koll. Z., 146, 96, 1956.
123. Hansen, R.S. and Wallace, T., J. Phys. Chem., 63, 1085, 1959.
124. Baret, J.F., J. Phys. Chem., 72, 2755, 1968.
125. Baret, J.F., J. Chem. Phys., 65, 895, 1968.
126. Baret, J.F., J. Colloid Interface Sci., 30, 1, 1969.
127. Borwankar, R.P. and Wasan, D.T., Chem. Eng. Sci., 38, 1637, 1983.
128. Alexandrov, N.A., Marinova, K.G., Gurkov, T.D., Danov, K.D., Kralchevsky, P.A., Stoyanov, S.D.,
Blijdenstein, T.B.J., Arnaudov, L.N., Pelan, E.G., and Lips, A., J. Colloid Interface Sci., 376, 296, 2012.
129. Danov, K.D., Valkovska, D.S., and Kralchevsky, P.A., J. Colloid Interface Sci., 251, 18, 2002.
130. Dong, C., Hsu, C.-T., Chin, C.-Y., and Lin, S.-Y., Langmuir, 16, 4573, 2000.
131. McBain, J.W., Trans. Faraday Soc., 9, 99, 1913.
132. Vincent, B., Adv. Colloid Interface Sci., 203, 51, 2014.
133. Missel, P.J., Mazer, N.A., Benedek, G.B., Young, C.Y., and Carey, M.C., J. Phys. Chem., 84, 1044, 1980.
134. Anachkov, S.E., Kralchevsky, P.A., Danov, K.D., Georgieva, G.S., and Ananthapadmanabhan, K.P.,
J. Colloid Interface Sci., 416, 258, 2014.
135. Kralchevsky, P.A., Danov, K.D., Anachkov, S.E., Georgieva, G.S., and Ananthapadmanabhan, K.P., Curr.
Opin. Colloid Interface Sci., 18, 524, 2013.
136. Danov, K.D., Kralchevsky, P.A., and Ananthapadmanabhan, K.P., Adv. Colloid Interface Sci., 206, 17,
2014.
137. Mitchell, D.J. and Ninham, B.W., J. Phys. Chem., 87, 2996, 1983.
Chemical Physics of Colloid Systems and Interfaces 395
138. Evans, E.A. and Skalak, R., CRC Crit. Rev. Bioengin., 3, 181, 1979.
139. Kralchevsky, P.A., Danov, K.D., and Anachkov, S.E., Colloid J., 76, 255, 2014.
140. Kresheck, G.C., Hamory, E., Davenport, G., and Scheraga, H.A., J. Am. Chem. Soc., 88, 246, 1966.
141. Aniansson, E.A.G. and Wall, S.N., J. Phys. Chem., 78, 1024, 1974.
142. Lucassen, J., Faraday Discuss. Chem. Soc., 59, 76, 1975.
143. Noskov, B.A., Kolloidn. Zh., 52, 509, 1990.
144. Johner, A. and Joanny, J.F., Macromolecules, 23, 5299, 1990.
145. Dushkin, C.D., Ivanov, I.B., and Kralchevsky, P.A., Colloids Surf., 60, 235, 1991.
146. Joos, P. and van Hunsel, J., Colloids Surf., 33, 99, 1988.
147. Li, B., Joos, P., and van Uffelen, M., J. Colloid Interface Sci., 171, 270, 1995.
148. Geeraerts, G. and Joos, P., Colloids Surf. A, 90, 149, 1994.
149. Danov, K.D., Kralchevsky, P.A., Denkov, N.D., Ananthapadmanabhan, K.P., and Lips, A., Adv. Colloid
Interface Sci., 119, 1, 2006.
150. Danov, K.D., Kralchevsky, P.A., Denkov, N.D., Ananthapadmanabhan, K.P., and Lips, A., Adv. Colloid
Interface Sci., 119, 17, 2006.
151. Danov, K.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., Colloids Surf. A, 282–283,
143, 2006.
152. McBain, J.W., Colloidal Science, D.C. Heat, Lexington, MA, 1950.
153. Christian, S.D. and Scamehorn, J.F., Solubilization in Surfactant Aggregates, Marcel Dekker, New York,
1995.
154. Miller, C.A., Micellar systems and microemulsions: solubilization aspects, in Handbook of Surface and
Colloid Chemistry, 1st edn., Birdi, K.S. (Ed.), CRC Press, Boca Raton, FL, 1997, p. 157.
155. Vasilescu, M., Caragheorgheopol, A., and Caldararu, H., Adv. Colloid Interface Sci., 89–90, 169, 2001.
156. Carroll, B.J., J. Colloid Interface Sci., 79, 126, 1981.
157. Kabalnov, A. and Weers, J., Langmuir, 12, 3442, 1996.
158. Weiss, J., Coupland, J.N., Brathwaite, D., and McClements, D.J., Colloids Surf. A, 121, 53, 1997.
159. Todorov, P.D., Kralchevsky, P.A., Denkov, N.D., Broze, G., and Mehreteab, A., J. Colloid Interface Sci.,
245, 371, 2002.
160. Kralchevsky, P.A. and Denkov, N.D., Triblock copolymers as promoters of solubilization of oils in
aqueous surfactant solutions, in Molecular Interfacial Phenomena of Polymers and Biopolymers,
Chen, P. (Ed.), Woodhead Publishing, Cambridge, U.K., 2005, Chapter 15, p. 538.
161. Sailaja, D., Suhasini, K.L., Kumar, S., and Gandhi, K.S., Langmuir, 19, 4014, 2003.
162. Chan, A.F., Fennel Evans, D., and Cussler, E.L., AIChE J., 22, 1006, 1976.
163. Huang, C., Fennel Evans, D., and Cussler, E.L., J. Colloid Interface Sci., 82, 499, 1981.
164. Shaeiwitz, J.A., Chan, A.F.-C., Cussler, E.L. and Fennel Evans, D., J. Colloid Interface Sci., 84, 47, 1981.
165. Plucinski, P. and Nitsch, W., J. Phys. Chem., 97, 8983, 1993.
166. Chen, B.-H., Miller, C.A., and Garrett, P.R., Colloids Surf. A, 128, 129, 1997.
167. Chen, B.-H., Miller, C.A., and Garrett, P.R., Langmuir, 14, 31, 1998.
168. Christov, N.C., Denkov, N.D., Kralchevsky, P.A., Broze, G., and Mehreteab, A., Langmuir, 18, 7880,
2002.
169. Kralchevsky, P.A., Denkov, N.D., Todorov, P.D., Marinov, G.S., Broze, G., and Mehreteab, A., Langmuir,
18, 7887, 2002.
170. Todorov, P.D., Marinov, G.S., Kralchevsky, P.A., Denkov, N.D., Durbut, P., Broze, G., and Mehreteab, A.,
Langmuir, 18, 7896, 2002.
171. Granek, R., Langmuir, 12, 5022, 1996.
172. Lawrence, A.S.C., Discuss. Faraday Soc., 25, 51, 1958.
173. Lawrence, A.S.C., Bingham, A., Capper, C.B., and Hume, K., J. Phys. Chem., 68, 3470, 1964.
174. Stowe, L.R. and Shaeiwitz, J.A., J. Colloid Interface Sci., 90, 495, 1982.
175. Raterman, K.T. and Shaeiwitz, J.A., J. Colloid Interface Sci., 98, 394, 1984.
176. Lim, J.-C. and Miller, C.A., Langmuir, 7, 2021, 1991.
177. Somasundaran, P. and Krishnakumar, S., Colloids Surf. A, 123–124, 491, 1997.
178. Ward, A.J., Kinetics of solubilization in surfactant-based systems, in Solubilization in Surfactant
Aggregates, Christian, S.D. and Scamehorn, J.F. (Eds.), Marcel Dekker, New York, 1995, Chapter 7,
pp. 237–272.
179. Nagarajan, R. and Ganesh, K., J. Colloid Interface Sci., 184, 489, 1996.
180. Lebens, P.J.M. and Keurentjes, J.T.F., Ind. Eng. Chem. Res., 35, 3415, 1996.
181. Xing, L. and Mattice, W.L., Macromolecules, 30, 1711, 1997.
182. Křiž, J., Masař, B., and Doskočilová, D., Macromolecules, 30, 4391, 1997.
396 Handbook of Surface and Colloid Chemistry
183. Marinov, G., Michels, B., and Zana, R., Langmuir, 14, 2639, 1998.
184. Kositza, M.J., Bohne, C., Alexandridis, P.T., Hatton, T.A., and Holzwarth, J.F., Langmuir, 15, 322, 1999.
185. Walderhaug, H., J. Phys. Chem. B, 103, 3352, 1999.
186. Paterson, I.F., Chowdhry, B.Z., Leharne, S.A., Langmuir, 15, 6178, 1999.
187. Bromberg, L. and Temchenko, M., Langmuir, 15, 8627, 1999.
188. Laplace, P.S., Traité de mécanique céleste; Suppléments au Livre X, 1805, 1806.
189. Bakker, G., Kapillatytät und oberflächenspannung, in Handbuch der Experimentalphysik, Band 6,
Akademische Verlagsgesellschaft, Leipzig, Germany, 1928.
190. Princen, H.M., The equilibrium shape of interfaces, drops, and bubbles, in Surface and Colloid Science,
Vol. 2, Matijevic, E. (Ed.), Wiley, New York, 1969, p. 1.
191. Finn, R., Equilibrium Capillary Surfaces, Springer-Verlag, New York, 1986.
192. Weatherburn, C.E., Differential Geometry in Three Dimensions, Cambridge, U.K., 1930.
193. McConnell, A.J., Application of Tensor Analysis, Dover, New York, 1957.
194. Young, T., Philos. Trans. Roy. Soc. (Lond.), 95, 55, 1805.
195. Jonson, R.E. and Dettre, Wettability and contact angles, in Surface and Colloid Science, Vol. 2, Matijevic,
E. (Ed.), Wiley, New York, 1969, p. 85.
196. Starov, V.M., Adv. Colloid Interface Sci., 39, 147, 1992.
197. Neumann, F., Vorlesungen über die Theorie der Capillarität, B.G. Teubner, Leipzig, Germany, 1894.
198. Ivanov, I.B., Kralchevsky, P.A., and Nikolov, A.D., J. Colloid Interface Sci., 112, 97, 1986.
199. Hartland, S. and Hartley, R.W., Axisymmetric Fluid–Liquid Interfaces, Elsevier, Amsterdam, the
Netherlands, 1976.
200. Kralchevsky, P.A., Eriksson, J.C., and Ljunggren, S., Adv. Colloid Interface Sci., 48, 19, 1994.
201. Tachev, K.D., Angarska, J.K., Danov, K.D., and Kralchevsky, P.A., Colloids Surf. B, 19, 61, 2000.
202. Meunier, J. and Lee, L.T., Langmuir, 7, 1855, 1991.
203. Dan, N., Pincus, P., and Safran, S.A., Langmuir, 9, 2768, 1993.
204. Kralchevsky, P.A., Paunov, V.N. Denkov, N.D., and Nagayama, K., J. Chem. Soc. Faraday Trans.,
91, 3415, 1995.
205. Danov, K.D., Kralchevsky, P.A., and Stoyanov, S.D., Langmuir, 26, 143, 2010.
206. Basheva, E.S., Kralchevsky, P.A., Christov, N.C., Danov, K.D., Stoyanov, S.D., Blijdenstein, T.B.J., Kim,
H.-J., Pelan, E.G., and Lips, A., Langmuir, 27, 2382, 2011.
207. Petsev, D.N., Denkov, N.D., and Kralchevsky, P.A., J. Colloid Interface Sci., 176, 201, 1995.
208. De Gennes, P.G. and Taupin, C., J. Phys. Chem., 86, 2294, 1982.
209. Concus, P., J. Fluid Mech., 34, 481, 1968.
210. Kralchevsky, P.A., Ivanov, I.B., and Nikolov, A.D., J. Colloid Interface Sci., 112, 108, 1986.
211. Abramowitz, M. and Stegun, I.A., Handbook of Mathematical Functions, Dover, New York, 1965.
212. Jahnke, E., Emde, F., and Lösch, F., Tables of Higher Functions, McGraw-Hill, New York, 1960.
213. Lo, L.L., J. Fluid Mech., 132, 65, 1983.
214. Derjaguin, B.V., Dokl. Akad. Nauk USSR, 51, 517, 1946.
215. Scheludko, A. and Exerowa, D., Comm. Dept. Chem. Bulg. Acad. Sci., 7, 123, 1959.
216. Sheludko, A., Adv. Colloid Interface Sci., 1, 391, 1967.
217. Dimitrov, A.S., Kralchevsky, P.A., Nikolov, A.D., and Wasan, D.T., Colloids Surf., 47, 299, 1990.
218. J. Plateau, Experimental and theoretical researches on the figures of equilibrium of a liquid mass with-
drawn from the action of gravity, in The Annual Report of the Smithsonian Institution, Washington, DC,
1863, pp. 207–285.
219. J. Plateau, The figures of equilibrium of a liquid mass, in The Annual Report of the Smithsonian
Institution, Washington, DC, 1864, pp. 338–369.
220. J. Plateau, Statique Expérimentale et Théoretique des Liquides Soumis aux Seules Forces Moléculaires,
Gauthier-Villars, Paris, France, 1873.
221. Orr, F.M., Scriven, L.E., and Rivas, A.P., J. Fluid Mech., 67, 723, 1975.
222. McFarlane, J.S. and Tabor, D., Proc. R. Soc. Lond. A, 202, 224, 1950.
223. Koos, E. and Willenbacher, N., Science, 331, 897, 2011.
224. Zettlemoyer, A.C., Nucleation, Marcel Dekker, New York, 1969.
225. Abraham, E.F., Homogeneous Nucleation Theory, Academic Press, New York, 1974.
226. Thomson, W. (Lord Kelvin), Proc. Roy. Soc., 9, 225, 1858; Thomson, W., Philos. Mag., 17, 61, 1859.
227. Lupis, C.H.P., Chemical Thermodynamics of Matherials, North-Holland, New York, 1983.
228. Lifshitz, I.M. and Slyozov, V.V., Zh. Exp. Teor. Fiz., 35, 479, 1958 (in Russian).
229. Wagner, C., Z. Electrochem., 35, 581, 1961.
230. Kalhweit, M., Faraday Discuss. Chem. Soc., 61, 48, 1976.
Chemical Physics of Colloid Systems and Interfaces 397
231. Parbhakar, K., Lewandowski, J., and Dao, L.H., J. Colloid Interface Sci., 174, 142, 1995.
232. Kabalnov, A.S., Pertzov, A.V., and Shchukin, E.D., Colloids Surf., 24, 19, 1987.
233. Kabalnov, A.S. and Shchukin, E.D., Adv. Colloid Interface Sci., 38, 69, 1992.
234. McClements, D.J., Dungan, S.R., German, J.B., and Kinsela, J.E., Food Hydrocolloids, 6, 415, 1992.
235. Weiss, J., Coupland, J.N., and McClements, D.J., J. Phys. Chem., 100, 1066, 1996.
236. Weiss, J., Canceliere, C., and McClements, D.J., Langmuir, 16, 6833, 2000.
237. Kabalnov, A.S., Langmuir, 10, 680, 1994.
238. Ivanov, I.B. and Kralchevsky, P.A., Mechanics and thermodynamics of curved thin liquid films, in Thin
Liquid Films, Ivanov, I.B. (Ed.), Marcel Dekker, New York, 1988, p. 49.
239. Kralchevsky, P.A. and Ivanov, I.B., J. Colloid Interface Sci., 137, 234, 1990.
240. Kralchevsky, P.A., Danov, K.D., and Ivanov, I.B., Thin liquid film physics, in Foams: Theory,
Measurements and Applications, Prud’homme, R.K. (Ed.), Marcel Dekker, New York, 1995, p. 1.
241. Rusanov, A.I., Phase Equilibria and Surface Phenomena, Khimia, Leningrad, Russia, 1967 (Russian);
Phasengleichgewichte und Grenzflächenerscheinungen, Akademie Verlag, Berlin, Germany, 1978
(German).
242. Derjaguin, B.V. and Kussakov, M.M., Acta Physicochem. USSR, 10, 153, 1939.
243. Exerowa, D. and Scheludko, A., Bull. Inst. Chim. Phys. Bulg. Acad. Sci., 4, 175, 1964.
244. Mysels, K.J., J. Phys. Chem., 68, 3441, 1964.
245. Exerowa, D., Commun. Dept. Chem. Bulg. Acad. Sci., 11, 739, 1978.
246. Kruglyakov, P.M., Equilibrium properties of free films and stability of foams and emulsions, in Thin
Liquid Films, Ivanov, I.B. (Ed.), Marcel Dekker, New York, 1988, p. 767.
247. Martynov, G.A. and Derjaguin, B.V., Kolloidn. Zh., 24, 480, 1962.
248. Toshev, B.V. and Ivanov, I.B., Colloid Polym. Sci., 253, 558, 1975.
249. Ivanov, I.B. and Toshev, B.V., Colloid Polym. Sci., 253, 593, 1975.
250. Frumkin, A., Zh. Phys. Khim. USSR, 12, 337, 1938.
251. de Feijter, J.A., Thermodynamics of thin liquid films, in Thin Liquid Films, Ivanov, I.B. (Ed.), Marcel
Dekker, New York, 1988, p. 1.
252. Kralchevsky, P.A. and Ivanov, I.B., Chem. Phys. Lett., 121, 111, 1985.
253. Nikolov, A.D., Kralchevsky, P.A., Ivanov, I.B., and Dimitrov, A.S., AIChE Symp. Ser., 82(252), 82, 1986.
254. de Feijter, J.A. and Vrij, A., J. Electroanal. Chem., 47, 9, 1972.
255. Kralchevsky, P.A. and Ivanov, I.B., Chem. Phys. Lett., 121, 116, 1985.
256. Denkov, N.D., Petsev, D.N., and Danov, K.D., J. Colloid Interface Sci., 176, 189, 1995.
257. Derjaguin, B.V., Acta Physicochim. USSR, 12, 181, 1940.
258. Princen, H.M. and Mason, S.G., J. Colloid Sci., 20, 156, 1965.
259. Prins, A., J. Colloid Interface Sci., 29, 177, 1969.
260. Clint, J.H., Clunie, J.S., Goodman, J.F., and Tate, J.R., Nature (Lond.), 223, 291, 1969.
261. Yamanaka, T., Bull. Chem. Soc. Jpn, 48, 1755, 1975.
262. Princen, H.M., J. Phys. Chem., 72, 3342, 1968.
263. Princen, H.M. and Frankel, S., J. Colloid Interface Sci., 35, 186, 1971.
264. Scheludko, A., Radoev, B., and Kolarov, T., Trans. Faraday Soc., 64, 2213, 1968.
265. Haydon, D.A. and Taylor, J.L., Nature (Lond.), 217, 739, 1968.
266. Kolarov, T. and Zorin, Z.M., Kolloidn. Zh., 42, 899, 1980.
267. Kruglyakov, P.M. and Rovin, Y.G., Physical Chemistry of Black Hydrocarbon Films, Nauka, Moscow,
Russia, 1978 (in Russian).
268. Marinova, K.G., Gurkov, T.D., Velev, O.D., Ivanov, I.B., Campbell, B., and Borwankar, R.P., Colloids
Surf. A, 123, 155, 1997.
269. Françon, M., Progress in Microscopy, Pergamon Press, London, U.K., 1961.
270. Beyer, H., Theorie und Praxis der Interferenzmicroscopie, Academische Verlagessellschaft, Leipzig,
Germany, 1974.
271. Zorin, Z.M., Kolloidn. Zh., 39, 1158, 1977.
272. Zorin, Z., Platikanov, D., Rangelova, N., and Scheludko, A., Measurements of contact angles between
bulk liquids and Newton black films for determining the line tension, in Surface Forces and Liquid
Interfaces, Derjaguin, B.V. (Ed.), Nauka, Moscow, Russia, 1983, p. 200 (in Russian).
273. Nikolov, A.D., Kralchevsky, P.A., and Ivanov, I.B., J. Colloid Interface Sci., 112, 122, 1986.
274. Lobo, L.A., Nikolov, A.D., Dimitrov, A.S., Kralchevsky, P.A., and Wasan, D.T., Langmuir, 6, 995, 1990.
275. Dimitrov, A.S., Nikolov, A.D., Kralchevsky, P.A., and Ivanov, I.B., J. Colloid Interface Sci., 151, 462,
1992.
276. Picard, G., Schneider, J.E., and Fendler, J.H., J. Phys. Chem., 94, 510, 1990.
398 Handbook of Surface and Colloid Chemistry
277. Picard, G., Denicourt, N., and Fendler, J.H., J. Phys. Chem., 95, 3705, 1991.
278. Skinner, F.K., Rotenberg, Y., and Neumann, A.W., J. Colloid Interface Sci., 130, 25, 1989.
279. Dimitrov, A.S., Kralchevsky, P.A., Nikolov, A.D., Noshi, H., and Matsumoto, M., J. Colloid Interface
Sci., 145, 279, 1991.
280. Hadjiiski, A., Dimova, R., Denkov, N.D., Ivanov, I.B., and Borwankar, R., Langmuir, 12, 6665, 1996.
281. Ivanov, I.B., Hadjiiski, A., Denkov, N.D., Gurkov, T.D., Kralchevsky, P.A., and Koyasu, S., Biophys. J.,
75, 545, 1998.
282. Paunov, V.N., Langmuir, 19, 7970, 2003.
283. Cayre, O.J. and Paunov, V.N., Langmuir, 20, 9594, 2004.
284. Ikeda, S., Takahara, Y.K., Ishino, S., Matsumura, M., and Ohtani, B., Chem. Lett., 34(10), 1386, 2005.
285. Paunov, V.N., Cayre, O.J., Noble, P.F., Stoyanov, S.D., Velikov, K.P., and Golding, M., J. Colloid Interface
Sci., 312, 381, 2007.
286. Park, B.J., Pantina, J.P., Furst, E.M., Oettel, M., Reynaert, S., and Vermant, J., Langmuir, 24, 1686, 2008.
287. Nicolson, M.M., Proc. Camb. Philos. Soc., 45, 288, 1949.
288. Chan, D.Y.C., Henry, J.D., and White, L.R., J. Colloid Interface Sci., 79, 410, 1981.
289. Paunov, V.N., Kralchevsky, P.A., Denkov, N.D., Ivanov, I.B., and Nagayama, K., J. Colloid Interface Sci.,
157, 100, 1993.
290. Kralchevsky, P.A., Paunov, V.N., Ivanov, I.B., and Nagayama, K., J. Colloid Interface Sci., 151, 79, 1992.
291. Kralchevsky, P.A., Paunov, V.N., Denkov, N.D., Ivanov, I.B., and Nagayama, K., J. Colloid Interface Sci.,
155, 420, 1993.
292. Kralchevsky, P.A. and Nagayama, K., Langmuir, 10, 23, 1994.
293. Kralchevsky, P.A. and Nagayama, K., Adv. Colloid Interface Sci., 85, 145, 2000.
294. Denkov, N.D., Velev, O.D., Kralchevsky, P.A., Ivanov, I.B., Nagayama, K., and Yoshimura, H., Langmuir,
8, 3183, 1992.
295. Dimitrov, A.S., Dushkin, C.D., Yoshimura, H., and Nagayama, K., Langmuir, 10, 432, 1994.
296. Sasaki, M. and Hane, K., J. Appl. Phys., 80, 5427, 1996.
297. Du, H., Chen, P., Liu, F., Meng, F.-D., Li, T.-J., and Tang, X.-Y., Mater. Chem. Phys., 51, 277, 1977.
298. Price, W.C., Williams, R.C., and Wyckoff, R.W.G., Science, 102, 277, 1945.
299. Cosslett, V.E. and Markham, R., Nature, 161, 250, 1948.
300. Horne, R.W. and Pasquali-Ronchetti, I., J. Ultrastruct. Res., 47, 361, 1974.
301. Harris, J.R., Micron Microsc. Acta, 22, 341, 1991.
302. Yoshimura, H., Matsumoto, M., Endo, S., and Nagayama, K., Ultramicroscopy, 32, 265, 1990.
303. Yamaki, M., Higo, J., and Nagayama, K., Langmuir, 11, 2975, 1995.
304. Nagayama, K., Colloids Surf. A, 109, 363, 1996.
305. Burmeister, F., Schäfle, C., Keilhofer, B., Bechinger, C., Boneberg, J., and Leiderer, P., Adv. Mater., 10,
495, 1998.
306. Xia, Y., Tien, J., Qin, D., and Whitesides, G.M., Langmuir, 12, 4033, 1996.
307. Lindström, H., Rensmo, H., Sodergren, S., Solbrand, A., and Lindquist, S.E., J. Phys. Chem., 100, 3084,
1996.
308. Matsushita, S., Miwa, T., and Fujishima, A., Langmuir, 13, 2582, 1997.
309. Murray, C.B., Kagan, C.R., and Bawendi, M.G., Science, 270, 1335, 1995.
310. Jap, B.K., Zulauf, M., Scheybani, T., Hefti, A., Baumeister, W., Aebi, U., and Engel, A., Ultramicroscopy,
46, 45, 1992.
311. De Rossi, D., Ahluwalia, A., and Mulè, M., IEEE Eng. Med. Biol., 13, 103, 1994.
312. Kralchevsky, P.A. and Denkov, N.D., Curr. Opin. Colloid Interface Sci., 6, 383, 2001.
313. Gil, T., Ipsen, J.H., Mouritsen, O.G., Sabra, M.C., Sperotto, M.M., and Zuckermann, M.J., Biochim.
Biophys. Acta, 1376, 245, 1998.
314. Mansfield, S.L., Gotch, A.J., and Harms, G.S., J. Phys. Chem., B, 103, 2262, 1999.
315. Fisher, L.R. and Malloy, A.R., Annu. Rep. Prog. Chem., Sect. C, 95, 373, 1999.
316. Kralchevsky, P.A., Paunov, V.N., and Nagayama, K., J. Fluid Mech., 299, 105, 1995.
317. Camoin, C., Roussel, J.F., Faure, R., and Blanc, R., Europhys. Lett., 3, 449, 1987.
318. Velev, O.D., Denkov, N.D., Paunov, V.N., Kralchevsky, P.A., and Nagayama, K., Langmuir, 9, 3702, 1993.
319. Dushkin, C.D., Kralchevsky, P.A., Yoshimura, H., and Nagayama, K., Phys. Rev. Lett., 75, 3454, 1995.
320. Lucassen, J., Colloids Surf., 65, 131, 1992.
321. Stamou, D., Duschl, C., and Johannsmann, D., Phys. Rev., E, 62, 5263, 2000.
322. Kralchevsky, P.A., Denkov, N.D., and Danov, K.D., Langmuir, 17, 2001, 7694.
323. Danov, K.D., Kralchevsky, P.A., Naydenov, B.N., and Brenn, G., J. Colloid Interface Sci., 287, 121,
2005.
Chemical Physics of Colloid Systems and Interfaces 399
324. Danov, K.D. and Kralchevsky, P.A., Adv. Colloid Interface Sci., 154, 91, 2010.
325. Bowden, N., Terfort, A., Carbeck, J., and Whitesides, G.M., Science, 276, 233, 1997.
326. Bowden, N., Choi, I.S., Grzybowski, B.A., and Whitesides, G.M., J. Am. Chem. Soc., 121, 5373, 1999.
327. Brown, A.B.D., Smith, C.G., and Rennie, A.R., Phys. Rev. E, 62, 951, 2000.
328. Loudet, J.C., Alsayed, A.M., Zhang, J., and Yodh, A.G., Phys. Rev. Lett., 94, 018301, 2005.
329. Loudet, J.C., Yodh, A.G., and Pouligny, B., Phys. Rev. Lett., 97, 018304, 2006.
330. Loudet, J.C. and Pouligny, B., Europhys. Lett., 85, 28003, 2009.
331. van Nierop, E.A., Stijnman, M.A., and Hilgenfeldt, S., Europhys. Lett., 72, 671, 2005.
332. Lehle, H., Noruzifar, E., and Oettel, M., Eur. Phys. J. E, 26, 151, 2008.
333. Lewandowski, E.P., Bernate, J.A., Searson, P.C., and Stebe, K.J., Langmuir, 24, 9302, 2008.
334. Lewandowski, E.P., Bernate, J.A., Tseng, A., Searson, P.C., and Stebe, K.J., Soft Matter, 5, 886, 2009.
335. Lewandowski, E.P., Cavallaro, M., Botto, L., Bernate, J.C., Garbin, V., and Stebe, K.J., Langmuir, 26,
15142, 2010.
336. Cavallaro, M., Botto, L., Lewandowski, E.P., Wang, M., and Stebe, K.J., Proc. Natl. Acad. Sci. U.S.A.,
108, 20923, 2011.
337. Botto, L., Lewandowski, E.P., Cavallaro, M., and Stebe, K.J., Soft Matter, 8, 9957, 2012.
338. Yunker, P.J., Still, T., Lohr, M.A., and Yodh, A.G., Nature, 476, 308, 2011.
339. Bleibel, J., Dominguez, A., and Oettel, M., Eur. Phys. J.—Special Topics, 222, 3071, 2013.
340. Velikov, K.P., Durst, F., and Velev, O.D. Langmuir, 14, 1148, 1998.
341. Sur, J. and Pak, H.K., J. Korean Phys. Soc., 38, 582, 2001.
342. Danov, K.D., Pouligny, B., Angelova, M.I., and Kralchevsky, P.A., Strong capillary attraction between
spherical inclusions in a multilayered lipid membrane, in Studies in Surface Science and Catalysis, Vol. 132,
Iwasawa, Y., Oyama, N., and Kunieda, H. (Eds.), Elsevier, Amsterdam, the Netehrlands, 2001, p. 519.
343. Danov, K.D., Pouligny, B., and Kralchevsky, P.A., Langmuir, 17, 2001, 6599.
344. Kralchevsky, P.A., Paunov, V.N., Denkov, N.D., and Nagayama, K., J. Colloid Interface Sci., 167, 47,
1994.
345. Velev, O.D., Denkov, N.D., Paunov, V.N., Kralchevsky, P.A., and Nagayama, K., J. Colloid Interface Sci.,
167, 66, 1994.
346. Petkov, J.T., Denkov, N.D., Danov, K.D., Velev, O.D., Aust, R., and Durst, F., J. Colloid Interface Sci.,
172, 147, 1995.
347. Danov, K.D., Aust, R., Durst, F., and Lange, U., J. Colloid Interface Sci., 175, 36, 1995.
348. Petkov, J.T., Danov, K.D., Denkov, N.D., Aust, R., and Durst, F., Langmuir, 12, 2650, 1996.
349. Petkov, J.T., Gurkov, T.D., and Campbell, B.E., Langmuir, 17, 4556, 2001.
350. Danov, K.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., Langmuir, 22, 106, 2006.
351. Danov, K.D., Kralchevsky, P.A., and Boneva, M.P., Langmuir, 20, 6139, 2004.
352. Aveyard, R., Clint, J.H., Nees, D., and Paunov, V.N., Langmuir, 16, 1969, 2000.
353. Nikolaides, M.G., Bausch, A.R., Hsu, M.F., Dinsmore, A.D., Brenner, M.P., Gay, C., and Weitz, D.A.,
Nature, 420, 299, 2002.
354. Danov, K.D., Kralchevsky, P.A., and Boneva, M.P., Langmuir, 22, 2653, 2006.
355. Danov, K.D. and Kralchevsky, P.A., J. Colloid Interface Sci., 298, 213, 2006.
356. Aveyard, R., Binks, B.P., Clint, J.H., Fletcher, P.D.I., Horozov, T.S., Neumann, B., Paunov, V.N. et al.,
Phys. Rev. Lett., 88, 246102, 2002.
357. Horozov, T.S., Aveyard, R., Clint, J.H., and Binks, B.P., Langmuir, 19, 2822, 2003.
358. Horozov, T.S., Aveyard, R., Binks, B.P., and Clint, J.H., Langmuir, 21, 7407, 2005.
359. Horozov, T.S. and Binks, B.P., Colloids Surf. A, 267, 64, 2005.
360. Ray, M.A. and Li, J., Adv. Mater., 19, 2020, 2007.
361. Stancik, E.J., Kouhkan, M., and Fuller, G.G., Langmuir, 20, 90, 2004.
362. Reynaert, S., Moldenaers, P., and Vermant, J., Langmuir, 22, 4936, 2006.
363. Petkov, P.V., Danov, K.D., and Kralchevsky, P.A., Langmuir, 30, 2768, 2014.
364. Domínguez, A., Oettel, M., and Dietrich, S., J. Chem. Phys. 127, 204706, 2007.
365. Labib, M.E. and Williams, R., J. Colloid Interface Sci., 115, 330, 1987.
366. Philipse, A.P. and Vrij, A., J. Colloid Interface Sci., 128, 121, 1989.
367. Aubry, N. and Singh, P., Phys. Rev. E, 77, 056302, 2008.
368. Aubry, N., Singh, P., Janjua, M., and Nudurupati, S., Proc. Natl. Acad. Sci. U.S.A., 105, 3711, 2008.
369. Janjua, M., Nudurupati, S., Singh, P., and Aubry, N., Mech. Res. Commun., 36, 55, 2009.
370. Singh, P., Joseph, D.D., and Aubry, N., Soft Matter, 6, 4310, 2010.
371. Danov, K.D. and Kralchevsky, P.A., J. Colloid Interface Sci., 405, 278, 2013.
372. Danov, K.D. and Kralchevsky, P.A., J. Colloid Interface Sci., 405, 269, 2013.
400 Handbook of Surface and Colloid Chemistry
423. Tadros, Th.F., Steric interactions in thin liquid films, in Thin Liquid Films, Ivanov, I.B. (Ed.), Marcel
Dekker, New York, 1988, p. 331.
424. Patel, S.S. and Tirel, M., Ann. Rev. Phys. Chem., 40, 597, 1989.
425. Ploehn, H.J. and Russel, W.B., Adv. Chem. Eng., 15, 137, 1990.
426. de Gennes, P.G., Scaling Concepts in Polymer Physics, Cornell University Press, Ithaca, NY, 1979,
Chapter 3.
427. Dolan, A.K. and Edwards, S.F., Proc. Roy. Soc. (Lond.) A, 337, 509, 1974.
428. Dolan, A.K. and Edwards, S.F., Proc. Roy. Soc. (Lond.) A, 343, 427, 1975.
429. de Gennes, P.G., C. R. Acad. Sci. (Paris), 300, 839, 1985.
430. de Gennes, P.G., Adv. Colloid Interface Sci., 27, 189, 1987.
431. Alexander, S.J., Physique, 38, 983, 1977.
432. Taunton, H.J., Toprakcioglu, C., Fetters, L.J., and Klein, J., Macromolecules, 23, 571, 1990.
433. Klein, J. and Luckham, P., Nature, 300, 429, 1982; Klein, J. and Luckham, P., Macromolecules, 17, 1041,
1984.
434. Luckham, P.F. and Klein, J., J. Chem. Soc. Faraday Trans., 86, 1363, 1990.
435. Sonntag, H., Ehmka, B., Miller, R., and Knapschinski, L., Adv. Colloid Interface Sci., 16, 381, 1982.
436. Horn, R.G. and Israelachvili, J.N., Chem. Phys. Lett., 71, 192, 1980.
437. Nikolov, A.D., Wasan, D.T., Kralchevsky, P.A., and Ivanov, I.B., Ordered structures in thinning micellar
foam and latex films, in Ordering and Organisation in Ionic Solutions, Ise, N. and Sogami, I. (Eds.),
World Scientific, Singapore, 1988.
438. Nikolov, A.D. and Wasan, D.T., J. Colloid Interface Sci., 133, 1, 1989.
439. Nikolov, A.D., Kralchevsky, P.A., Ivanov, I.B., and Wasan, D.T., J. Colloid Interface Sci., 133, 13, 1989.
440. Nikolov, A.D., Wasan, D.T., Denkov, N.D., Kralchevsky, P.A., and Ivanov, I.B., Prog. Colloid Polym.
Sci., 82, 87, 1990.
441. Wasan, D.T., Nikolov, A.D., Kralchevsky, P.A., and Ivanov, I.B., Colloids Surf., 67, 139, 1992.
442. Asakura, S. and Oosawa, F., J. Chem. Phys., 22, 1255, 1954; Asakura, S. and Oosawa, F., J. Polym. Sci.,
33, 183, 1958.
443. de Hek, H. and Vrij, A., J. Colloid Interface Sci., 84, 409, 1981.
444. Snook, I.K. and van Megen, W., J. Chem. Phys., 72, 2907, 1980.
445. Kjellander, R. and Marčelja, S., Chem. Phys. Lett., 120, 393, 1985.
446. Tarazona, P. and Vicente, L., Mol. Phys., 56, 557, 1985.
447. Evans, R., and Parry, A.O., J. Phys. Condens. Matter, 2, SA15, 1990.
448. Henderson, J.R., Mol. Phys., 59, 89, 1986.
449. Mitchell, D.J., Ninham, B,W., and Pailthorpe, B.A., J. Chem. Soc. Faraday Trans. 2, 74, 1116, 1978.
450. Henderson, D., J. Colloid Interface Sci., 121, 486, 1988.
451. Kjellander, R. and Sarman, S., Chem. Phys. Lett., 149, 102, 1988.
452. Trokhymchuk, A. and Henderson, D., Curr. Opin. Colloid Interface Sci., 20, 33, 2015.
453. Pollard, M.L. and Radke, C.J., J. Chem. Phys., 101, 6979, 1994.
454. Chu, X.L., Nikolov, A.D., and Wasan, D.T., Langmuir, 10, 4403, 1994.
455. Chu, X.L., Nikolov, A.D., and Wasan, D.T., J. Chem. Phys., 103, 6653, 1995.
456. Trokhymchuk, A., Henderson, D., Nikolov, A., and Wasan, D.T., J. Phys. Chem. B, 107, 3927, 2003.
457. Blawzdziewicz, J. and Wajnryb, E., Europhys. Lett., 71, 269, 2005.
458. Kralchevsky, P.A. and Denkov, N.D., Chem. Phys. Lett., 240, 385, 1995.
459. Trokhymchuk, A., Henderson, D., Nikolov, A., and Wasan, D.T., Langmuir, 17, 4940, 2001.
460. Carnahan, N.F. and Starling, K.E., J. Chem. Phys., 51, 635, 1969.
461. Basheva, E.S., Kralchevsky, P.A., Danov, K.D., Ananthapadmanabhan, K.P., and Lips, A., Phys. Chem.
Chem. Phys., 9, 5183, 2007.
462. Mysels, K.J. and Jones, M.N., Discuss. Faraday Soc., 42, 42, 1966.
463. Christov, N.C., Danov, K.D., Zeng, Y., Kralchevsky, P.A., and von Klitzing, R., Langmuir, 26, 915, 2010.
464. Bondy, C., Trans. Faraday Soc., 35, 1093, 1939.
465. Patel, P.D. and Russel, W.B., J. Colloid Interface Sci., 131, 192, 1989.
466. Aronson, M.P., Langmuir, 5, 494, 1989.
467. van Lent, B., Israels, R., Scheutjens, J.M.H.M., and Fleer, G.J., J. Colloid Interface Sci., 137, 380, 1990.
468. Joanny, J.F., Leibler, L., and de Gennes, P.G., J. Polym. Sci., 17, 1073, 1979.
469. Evans, E. and Needham, D., Macromolecules, 21, 1822, 1988.
470. Yang, S., Tan, H., Yan, D., Nies, E., and Shi, A.-C., Phys. Rev. E, 75, 061803, 2007.
471. Melby, P., Prevost, A., Egolf, D.A., and Urbach, J.S., Phys. Rev. E, 76, 051307, 2007.
472. Lang, P.R., J. Chem. Phys., 127, 124906, 2007.
402 Handbook of Surface and Colloid Chemistry
473. July, C., Kleshchanok, D., and Lang, P.R., Eur. Phys. J. E, 35, 60, 2012.
474. Tulpar, A., Tilton, R.D., and Walz, J.Y., Langmuir, 23, 4351, 2007.
475. Lekkerkerker, H.N.W. and Tuinier, R., Colloids and the Depletion Interaction, Springer, Berlin, Germany,
2011.
476. Johonnott, E.S., Philos. Mag., 11, 746, 1906.
477. Perrin, J., Ann. Phys. (Paris), 10, 160, 1918.
478. Bruil, H.G. and Lyklema, J., Nature, 233, 19, 1971.
479. Friberg, S., Linden, St.E., and Saito, H., Nature, 251, 494, 1974.
480. Keuskamp, J.W. and Lyklema, J., ACS Symp. Ser., 8, 191, 1975.
481. Kruglyakov, P.M., Kolloidn. Zh., 36, 160, 1974.
482. Kruglyakov, P.M. and Rovin, Yu.G., Physical Chemistry of Black Hydrocarbon Films, Nauka, Moscow,
Russia, 1978 (in Russian).
483. Denkov, N.D., Yoshimura, H., Nagayama, K., and Kouyama, T., Phys. Rev. Lett., 76, 2354, 1996.
484. Denkov, N.D., Yoshimura, H., and Nagayama, K., Ultramicroscopy, 65, 147, 1996.
485. Kralchevsky, P.A., Nikolov, A.D., Wasan, D.T., and Ivanov, I.B., Langmuir, 6, 1180, 1990.
486. Bergeron, V. and Radke, C.J., Langmuir, 8, 3020, 1992.
487. Bergeron, V., Jimenez-Laguna, A.I., and Radke, C.J., Langmuir, 8, 3027, 1992.
488. Richetti, P. and Kékicheff, P., Phys. Rev. Lett., 68, 1951, 1992.
489. Parker, J.L., Richetti, P., and Kékicheff, P., Phys. Rev. Lett., 68, 1955, 1992.
490. McNamee, C.E., Tsujii, Y., Ohshima, H., and Matsumoto, M., Langmuir, 20, 1953, 2004.
491. Krichevsky, O. and Stavans, J., Phys. Rev. Lett., 74, 2752, 1995.
492. Bergeron, V. and Radke, C.J., Colloid Polym. Sci., 273, 165, 1995.
493. Marinova, K.G., Gurkov, T.D., Dimitrova, T.D., Alargova, R.G., and Smith, D., Langmuir, 14, 2011,
1998.
494. Basheva, E.S., Nikolov, A.D., Kralchevsky, P.A., Ivanov, I.B., and Wasan, D.T., Multi-stepwise drainage
and viscosity of macroscopic films formed from latex suspensions, in Surfactants in Solution, Vol. 12,
Mittal, K.L. and Shah, D.O. (Eds.), Plenum Press, New York, 1991, p. 467.
495. Basheva, E.S., Danov, K.D., and Kralchevsky, P.A., Langmuir, 13, 4342, 1997.
496. Dushkin, C.D., Nagayama, K., Miwa, T., and Kralchevsky, P.A., Langmuir, 9, 3695, 1993.
497. Sethumadhavan, G.N., Nikolov, A.D., and Wasan, D.T., J. Colloid Interface Sci., 240, 105, 2001.
498. Piech, M. and Walz, J.Y., J. Phys. Chem. B, 108, 9177, 2004.
499. Klapp, S.H.L., Zeng, Y., Qu, D., and von Klitzing, R., Phys. Rev. Lett., 100, 118303, 2008.
500. Klapp, S.H.L., Grandner, S., Zeng, Y., and von Klitzing, R., J. Phys.: Condens. Matter, 20, 494232, 2008.
501. Klapp, S.H.L., Grandner, S., Zeng, Y., and von Klitzing, R., Soft Matter, 6, 2330, 2010.
502. Zeng, Y., Grandner, S., Oliveira, C.L.P., Thünemann, A.F., Paris, O., Pedersen, J.S., Klapp, S.H.L., and
von Klitzing, R., Soft Matter, 7, 10899, 2011.
503. Koczo, K., Nikolov, A.D., Wasan, D.T., Borwankar, R.P., and Gonsalves, A., J. Colloid Interface Sci.,
178, 694, 1996.
504. Asnacios, A., Espert, A., Colin, A., and Langevin, D., Phys. Rev. Lett., 78, 4974, 1997.
505. Bergeron, V. and Claesson, P.M., Adv. Colloid Interface Sci., 96, 1, 2002.
506. Kolaric, B., Förster, S., and von Klitzing, R., Progr. Colloid Polym. Sci., 117, 195, 2001.
507. von Klitzing, R. and Kolaric, B., Tenside Surfactants Deterg., 39, 247, 2002.
508. Stubenrauch, C. and von Klitzing, R., J. Phys.: Condens. Matter, 15, R1197, 2003.
509. Beltran, C.M., Guillot, S., and Langevin, D., Macromolecules, 36, 8506, 2003.
510. Beltran, C.M. and Langevin, D., Phys. Rev. Lett., 94, 217803, 2005.
511. Heinig, P., Beltran, C.M., and Langevin, D., Phys. Rev. E, 73, 051607, 2006.
512. Jönsson, B., Broukhno, A., Forsman, J., and Åkesson, T., Langmuir, 19, 9914, 2003.
513. Danov, K.D., Basheva, E.S., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips, A., Adv. Colloid
Interface Sci., 168, 50, 2011.
514. Anachkov, S.E., Danov, K.D., Basheva, E.S., Kralchevsky, P.A., and Ananthapadmanabhan, K.P., Adv.
Colloid Interface Sci., 183–184, 55, 2012.
515. Bales, B.L. and Almgren, M., J. Phys. Chem., 99, 15153, 1995.
516. Turro, N.J. and Yekta, A., J. Am. Chem. Soc., 100, 5951, 1978.
517. Quina, F.H., Nassar, P.M., Bonilha, J.B.S., and Bales, B.L., J. Phys. Chem., 99, 17028, 1995.
518. Gehlen, M.H. and De Schryver, F.C., J. Phys. Chem., 97, 11242, 1993.
519. Mukerjee, P. and Mysels, K.J., Critical micelle concentration of aqueous surfactant systems, National
Standard Reference Data Series 36, National Bureau of Standards, Washington, DC, 1971.
520. Shah, S.S., Saeed, A., and Sharif, Q.M., Colloids Surf. A, 155, 405, 1999.
Chemical Physics of Colloid Systems and Interfaces 403
521. Benrraou, M., Bales, B.L., and Zana, R., J. Phys. Chem. B, 107, 13432, 2003.
522. Weidemaier, K., Tavernier, H.L., and Fayer, M.D., J. Phys. Chem. B, 101, 9352, 1997.
523. Hansson, P., Jönsson, B., Ström, C., and Söderman, O., J. Phys. Chem. B, 104, 3496, 2000.
524. Van Stam, J., Depaemelaere, S., and De Schryver, F.C., J. Chem. Educ., 75, 93, 1998.
525. Mata, J., Varade, D., and Bahadur, P., Thermochim. Acta, 428, 147, 2005.
526. Hassan, P.A., Hodgdon, T.K., Sagasaki, M., Fritz-Popovski, G., and Kaler, E.W., Comp. Rend. Chim., 12,
18, 2009.
527. Mukhim, T. and Ismail, K., J. Surf. Sci. Technol., 21, 113, 2005.
528. Bhat, M.A., Dar, A.A., Amin, A., Rashid, P.I., and Rather, G.M., J. Chem. Thermodyn., 39, 1500, 2007.
529. Amos, D.A., Lynn, S., and Radke, C.J., Langmuir, 14, 2297, 1998.
530. Stanley, H.E. and Teixeira, J., J. Chem. Phys., 73, 3404, 1980.
531. Norrish, K., Discuss. Faraday Soc., 18, 120, 1954.
532. Van Olphen, H., Clays Clay Miner., 2, 418, 1954.
533. Clunie, J.S., Goodman, J.F., and Symons, P.C., Nature (Lond.), 216, 1203, 1967.
534. Jordine, E.St.A., J. Colloid Interface Sci., 45, 435, 1973.
535. Derjaguin, B.V., Churaev, N.V., Croatica Chem. Acta, 50, 187, 1977.
536. LeNeveu, D.M., Rand, R.P., Parsegian, V.A., and Gingell, D., Biophys. J., 18, 209, 1977.
537. Lis, L.J., McAlister, M., Fuller, N., Rand, R.P., and Parsegian, V.A., Biophys. J., 37, 657, 1982.
538. Israelachvili, J.N. and Adams, G.E., J. Chem. Soc. Faraday Trans. 1, 74, 975, 1978.
539. Pashley, R.M. and Israelachvili, J.N., Colloids Surf., 2, 169, 1981.
540. Pashley, R.M. and Israelachvili, J.N., J. Colloid Interface Sci., 101, 511, 1984.
541. Pashley, R.M., J. Colloid Interface Sci., 80, 153, 1981.
542. Pashley, R.M., J. Colloid Interface Sci., 83, 531, 1981.
543. Pashley, R.M., Adv. Colloid Interface Sci., 16, 57, 1982.
544. Peschel, G., Belouschek, P., Müller, M.M., Müller, M.R., and König, R., Colloid Polym. Sci., 260, 444,
1982.
545. Horn, R.G., Smith, D.T., and Haller, W., Chem. Phys. Lett., 162, 404, 1989.
546. Claesson, P., Carmona-Ribeiro, A.M., Kurihara, K., J. Phys. Chem., 93, 917, 1989.
547. Petsev, D.N. and Vekilov, P.G., Phys. Rev. Lett., 84, 1339, 2000.
548. Petsev, D.N., Thomas, B.R., Yau, S.-T., and Vekilov, P.G., Biophys. J., 78, 2060, 2000.
549. Valle-Delgado, J.J., Molina-Bolívar, J.A., Galisteo-González, F., Gálvez-Ruiz, M.J., Feiler, A., and
Rutland, M.W., Langmuir, 21, 9544, 2005.
550. Valle-Delgado, J.J., Molina-Bolívar, J.A., Galisteo-González, F., and Gálvez-Ruiz, M.J., Curr. Opin.
Colloid Interface Sci., 16, 572, 2011.
551. Stanley, C. and Rau, D.C., Curr. Opin. Colloid Interface Sci., 16, 551, 2011.
552. Zhang, N.H., Tan, Z.Q., Li, J.J., Meng, W.L., and Xu, L.W., Curr. Opin. Colloid Interface Sci., 16, 592,
2011.
553. Aroti, A., Leontidis, E., Dubios, M., and Zemb, T., Biophys. J., 93, 1580, 2007.
554. Leontidis, E., Aroti, A., Belloni, L., Dubios, M., and Zemb, T., Biophys. J., 93, 1591, 2007.
555. Sparr, E. and Wennerström, H., Curr. Opin. Colloid Interface Sci., 16, 561, 2011.
556. Demé, B. and Zemb, T., Curr. Opin. Colloid Interface Sci., 16, 584, 2011.
557. Sanfeld, A. and Steinchen, A., Adv. Colloid Interface Sci., 140, 1, 2008.
558. Kaldasch, J., Senge, B., and Laven, J., Rheol. Acta, 48, 665, 2009.
559. Bongrand, P., Intermolecular forces, in Physical Basis of Cell–Cell Adhesion, Bongrand, P. (Ed.), CRC
Press, Boca Raton, FL, 1987, p. 1.
560. Binder, B., Goede, A., Berndt, N., and Holzhütter, H.-G., PLoS One, 4, 1, 2009.
561. Rand, R. and Parsegian, V., Biochim. Biophys. Acta, 988, 351, 1989.
562. Cevc, G., J. Chem. Soc. Faraday Trans., 87, 2733, 1991.
563. Leikin, S., Parsegian, V.A., Rau, D.C., Rand, R.P., Annu. Rev. Phys. Chem., 44, 369, 1993.
564. Israelachvili, J. and Wennerström, H., Nature, 379, 219, 1996.
565. Butt, H.-J., Capella, B., and Kappl, M., Surf. Sci. Rep., 59, 1, 2005.
566. Leite, F.L., Bueno, C.C., Da Róz, A.L., Ziemath, E.C., and Oliveira Jr., O.N., Int. J. Mol. Sci., 13, 12773,
2012.
567. Healy, T.W., Homola, A., James, R.O., and Hunter, R.J., Faraday Discuss. Chem. Soc., 65, 156, 1978.
568. Marčelja, S. and Radič, N., Chem. Phys Lett., 42, 129, 1976.
569. Besseling, N.A.M., Langmuir, 13, 2113, 1997.
570. Forsman, J., Woodward, C.E., and Jönsson, B., Langmuir, 13, 5459, 1997.
571. Attard, P. and Batchelor, M.T., Chem. Phys. Lett., 149, 206, 1988.
404 Handbook of Surface and Colloid Chemistry
572. Faraudo, J., Curr. Opin. Colloid Interface Sci., 16, 557, 2011.
573. Schneck, E. and Netz, R.R., Curr. Opin. Colloid Interface Sci., 16, 607, 2011.
574. Jönsson, B. and Wennerström, H., J. Chem. Soc. Faraday Trans. 2, 79, 19, 1983.
575. Kjellander, R., J. Chem. Soc. Faraday Trans. 2, 80, 1323, 1984.
576. Bratko, D. and Jönsson, B., Chem. Phys. Lett., 128, 449, 1986.
577. Trokhymchuk, A., Henderson, D., and Wasan, D.T., J. Colloid Interface Sci., 210, 320, 1999.
578. Simon, S.A. and McIntosh, T.J., Proc. Natl. Acad. Sci. U.S.A., 86, 9263, 1989.
579. Richmond, P., J. Chem. Soc. Faraday Trans. 2, 70, 1066, 1974.
580. Dzhavakhidze, P.G., Kornyshev, A.A., and Levadny, V.G., Nuovo Cimento, 10D, 627, 1988.
581. Kornyshev, A.A. and Leikin, S., Phys. Rev. A, 40, 6431, 1989.
582. Leikin, S. and Kornyshev, A.A., J. Chem. Phys., 92, 6890, 1990.
583. Attard, P. and Patey, G.N., Phys. Rev. A, 43, 2953, 1991.
584. Henderson, D. and Lozada-Cassou, M., J. Colloid Interface Sci., 114, 180, 1986.
585. Henderson, D. and Lozada-Cassou, M., J. Colloid Interface Sci., 162, 508, 1994.
586. Basu, S. and Sharma, M.M., J. Colloid Interface Sci., 165, 355, 1994.
587. Paunov, V.N., Dimova, R.I., Kralchevsky, P.A., Broze, G., and Mehreteab, A., J. Colloid Interface Sci.,
182, 239, 1996.
588. Paunov, V.N. and Binks, B.P., Langmuir, 15, 2015, 1999.
589. Kralchevsky, P.A., Danov, K.D., and Basheva, E.S., Curr. Opin. Colloid Interface Sci., 16, 517, 2011.
590. Tchaliovska, S., Herder, P., Pugh, R., Stenius, P., and Eriksson, J.C., Langmuir, 6, 1535, 1990.
591. Pashley, R.M., McGuiggan, P.M., Ninham, B.W., and Evans, D.F., Science, 229, 1088, 1985.
592. Rabinovich, Y.I. and Derjaguin, B.V., Colloids Surf., 30, 243, 1988.
593. Parker, J.L., Cho, D.L., and Claesson, P.M., J. Phys. Chem., 93, 6121, 1989.
594. Christenson, H.K., Claesson, P.M., Berg, J., and Herder, P.C., J. Phys. Chem., 93, 1472, 1989.
595. Christenson, H.K., Fang, J., Ninham, B.W., and Parker, J.L., J. Phys. Chem., 94, 8004, 1990.
596. Ducker, W.A., Xu, Z., and Israelachvili, J.N., Langmuir, 10, 3279, 1994.
597. Eriksson, J.C., Ljunggren, S., and Claesson, P.M., J. Chem. Soc. Faraday Trans. 2, 85, 163, 1989.
598. Joesten, M.D. and Schaad, L.J., Hydrogen Bonding, Marcel Dekker, New York, 1974.
599. Stillinger, F.H. and Ben-Naim, A., J. Chem. Phys., 47, 4431, 1967.
600. Conway, B.E., Adv. Colloid Interface Sci., 8, 91, 1977.
601. Kuzmin, V.L. and Rusanov, A.I., Kolloidn. Z., 39, 455, 1977.
602. Dubrovich, N.A., Kolloidn. Z., 57, 275, 1995.
603. Eriksson, J.C., Henriksson, U., and Kumpulainen, A., Colloids Surf. A, 282–283, 79, 2006.
604. Paunov, V.N., Sandler, S.I., and Kaler, E.W., Langmuir, 17, 4126, 2001.
605. Angarska, J.K., Dimitrova, B.S., Danov, K.D., Kralchevsky, P.A., Ananthapadmanabhan, K.P., and Lips,
A., Langmuir, 20, 1799, 2004.
606. Christenson, H.K. and Claesson, P.M., Science, 239, 390, 1988.
607. Parker, J.L., Claesson, P.M., and Attard, P., J. Phys. Chem., 98, 8468, 1994.
608. Carambassis, A., Jonker, L.C., Attard, P., and Rutland, M.W., Phys. Rev. Lett., 80, 5357, 1998.
609. Mahnke, J., Stearnes, J., Hayes, R.A., Fornasiero, D., and Ralston, J., Phys. Chem. Chem. Phys., 1, 2793,
1999.
610. Considine, R.F., Hayes, R.A., and Horn, R.G., Langmuir, 15, 1657, 1999.
611. Considine, R.F. and Drummond, C., Langmuir, 16, 631, 2000.
612. Attard, P., Langmuir, 16, 4455, 2000.
613. Yakubov, G.E., Butt, H.-J., and Vinogradova, O., J. Phys. Chem. B, 104, 3407, 2000.
614. Ederth, T., J. Phys. Chem. B, 104, 9704, 2000.
615. Ishida, N., Sakamoto, M., Miyahara, M., and Higashitani, K., Langmuir, 16, 5681, 2000.
616. Ishida, N., Inoue, T., Miyahara, M., and Higashitani, K., Langmuir, 16, 6377, 2000.
617. Hammer, M.U., Anderson, T.H., Chaimovich, A., Shell, M.S., and Israelachvili, J., Faraday Discuss.,
146, 299, 2010.
618. Popa, I., Gillies, G., Papastavrou, G., and Borkovec, M., J. Phys. Chem. B, 113, 8458, 2009.
619. Helfrich, W., Z. Naturforsch., 33a, 305, 1978.
620. Servuss, R.M. and Helfrich, W., J. Phys. (France), 50, 809, 1989.
621. Fernandez-Puente, L., Bivas, I., Mitov, M.D., and Méléard, P., Europhys. Lett., 28, 181, 1994.
622. Safinya, C.R., Roux, D., Smith, G.S., Sinha, S.K., Dimon, P., Clark, N.A., and Bellocq, A.M., Phys. Rev.
Lett., 57, 2718, 1986.
623. McIntosh, T.J., Magid, A.D., and Simon, S.A., Biochemistry, 28, 7904, 1989.
624. Abillon, O. and Perez, E., J. Phys. (France), 51, 2543, 1990.
Chemical Physics of Colloid Systems and Interfaces 405
625. Israelachvili, J.N. and Wennerström, H., J. Phys. Chem., 96, 520, 1992.
626. Evans, E.A. and Skalak, R., Mechanics and Thermodynamics of Biomembranes, CRC Press, Boca Raton,
FL, 1980.
627. Aniansson, G.A.E., Wall, S.N., Almgren, M., Hoffman, H., Kielmann, I., Ulbricht, W., Zana, R., Lang. J.,
and Tondre, C., J. Phys. Chem., 80, 905, 1976.
628. Aniansson, G.A.E., J. Phys. Chem., 82, 2805, 1978.
629. Dimitrova, T.D., Leal-Calderon, F., Gurkov, T.D., and Campbell, B., Langmuir, 17, 8069, 2001.
630. Danov, K.D., Ivanov, I.B., Ananthapadmanabhan, K.P., Lips, A., Adv. Colloid Interface Sci., 128–130,
185, 2006.
631. Bird, R.B., Stewart, W.E., and Lightfoot, E.N., Transport Phenomena, Wiley, New York, 1960.
632. Germain, P., Mécanique des Milieux Continus, Masson et Cie, Paris, France, 1962.
633. Batchelor, G.K., An Introduction of Fluid Mechanics, Cambridge University Press, London, U.K., 1967.
634. Slattery, J.C., Momentum, Energy, and Mass Transfer in Continua, R.E. Krieger, Huntington, New York,
1978.
635. Landau, L.D. and Lifshitz, E.M., Fluid Mechanics, Pergamon Press, Oxford, U.K., 1984.
636. Barnes, H.A., Hutton, J.F., and Walters, K., An Introduction to Rheology, Elsevier, Amsterdam, the
Netherlands, 1989.
637. Walters, K., Overview of macroscopic viscoelastic flow, in Viscoelasticity and Rheology, Lodge, A.S.,
Renardy, M., and Nohel, J.A. (Eds.), Academic Press, London, U.K., 1985, p. 47.
638. Boger, D.V., Ann. Rev. Fluid Mech., 19, 157, 1987.
639. Barnes, H.A., J. Rheol., 33, 329, 1989.
640. Barnes, H.A., A Handbook of Elementary Rheology, University of Wales, Institute of Non-Newtonian
Fluid Mechanics, Aberystwyth, U.K., 2000.
641. Sagis, L.M.C., Rev. Modern Phys., 83, 1367, 2011.
642. Navier, M., Mém. de l’Acad. d. Sci., 6, 389, 1827.
643. Stokes, G.G., Trans. Camb. Philos. Soc., 8, 287, 1845.
644. Happel, J. and Brenner, H., Low Reynolds Number Hydrodynamics with Special Applications to
Particulate Media, Prentice-Hall, Englewood Cliffs, New York, 1965.
645. Kim, S. and Karrila, S.J., Microhydrodynamics: Principles and Selected Applications, Butterworth-
Heinemann, Boston, MA, 1991.
646. Reynolds, O., Philos. Trans. Roy. Soc. (Lond.), A177, 157, 1886.
647. Lamb, H., Hydrodynamics, Cambridge University Press, London, U.K., 1963.
648. Felderhof, B.U., J. Chem. Phys., 49, 44, 1968.
649. Sche, S. and Fijnaut, H.M., Surf. Sci., 76, 186, 1976.
650. Maldarelli, Ch. and Jain, R.K., The hydrodynamic stability of thin films, in Thin Liquid Films, Ivanov,
I.B. (Ed.), Marcel Dekker, New York, 1988, p. 497.
651. Valkovska, D.S. and Danov, K.D., J. Colloid Interface Sci., 241, 400, 2001.
652. Danov, K.D., Effect of surfactants on drop stability and thin film drainage, in Fluid Mechanics of
Surfactant and Polymer Solutions, Starov, V. and Ivanov, I.B. (Eds.), Springer, New York, 2004, p. 1.
653. Hardy, W. and Bircumshaw, I., Proc. Roy. Soc. (Lond.), A108, 1, 1925.
654. Horn, R.G., Vinogradova, O.I., Mackay, M.E. and Phan-Thien, N., J. Chem. Phys., 112, 6424, 2000.
655. Dimitrov, D.S. and Ivanov, I.B., J. Colloid Interface Sci., 64, 97, 1978.
656. Ivanov, I.B., Dimitrov, D.S., Somasundaran, P., and Jain, R.K., Chem. Eng. Sci., 40, 137, 1985.
657. Ivanov, I.B. and Dimitrov, D.S., Thin film drainage, in Thin Liquid Films, Ivanov, I.B. (Ed.), Marcel
Dekker, New York, 1988, p. 379.
658. Danov, K.D., Kralchevsky, P.A. and Ivanov, I.B., in Handbook of Detergents, Part A: Properties, Broze,
G. (Ed.), Marcel Dekker, New York, 1999, p. 303.
659. O’Neill, M.E. and Stewartson, K., J. Fluid Mech., 27, 705, 1967.
660. Goldman, A.J., Cox, R.G., and Brenner, H., Chem. Eng. Sci., 22, 637, 1967.
661. Goldman, A.J., Cox, R.G., and Brenner, H., Chem. Eng. Sci., 22, 653, 1967.
662. Levich, V.G., Physicochemical Hydrodynamics, Prentice-Hall, Englewood Cliffs, NJ, 1962.
663. Edwards, D.A., Brenner, H., and Wasan, D.T., Interfacial Transport Processes and Rheology, Butterworth-
Heinemann, Boston, MA, 1991.
664. Charles, G.E. and Mason, S.G., J. Colloid Sci., 15, 236, 1960.
665. Danov, K.D., Denkov, N.D., Petsev, D.N., Ivanov, I.B., and Borwankar, R., Langmuir, 9, 1731, 1993.
666. Danov, K.D., Valkovska, D.S., and Ivanov, I.B., J. Colloid Interface Sci., 211, 291, 1999.
667. Hartland, S., Coalescence in dense-packed dispersions, in Thin Liquid Films, Ivanov, I.B. (Ed.), Marcel
Dekker, New York, 1988, p. 663.
406 Handbook of Surface and Colloid Chemistry
668. Hetsroni, G. (Ed.), Handbook of Multiphase System, Hemisphere Publishing, New York, 1982, pp. 1–96.
669. Davis, A.M.J., Kezirian, M.T., and Brenner, H., J. Colloid Interface Sci., 165, 129, 1994.
670. Brenner, H., Chem. Eng. Sci., 18, 1, 1963.
671. Brenner, H., Chem. Eng. Sci., 19, 599, 1964; Brenner, H., Chem. Eng. Sci., 19, 631, 1964.
672. Brenner, H. and O’Neill, M.E., Chem. Eng. Sci., 27, 1421, 1972.
673. Van de Ven, T.G.M., Colloidal Hydrodynamics, Academic Press, London, U.K., 1988.
674. Jeffery, G.B., Proc. Lond. Math. Soc., 14, 327, 1915.
675. Stimson, M. and Jeffery, G.B., Proc. Roy. Soc. (Lond.), A111, 110, 1926.
676. Cooley, M.D.A. and O’Neill, M.E., Mathematika, 16, 37, 1969.
677. Cooley, M.D.A. and O’Neill, M.E., Proc. Camb. Philos. Soc., 66, 407, 1969.
678. Davis, M.H., Chem. Eng. Sci., 24, 1769, 1969.
679. O’Neill, M.E. and Majumdar, S.R., Z. Angew. Math. Phys., 21, 164, 1970; O’Neill, M.E. and Majumdar,
S.R., Z. Angew. Math. Phys., 21, 180, 1970.
680. Davis, M.H., Two Unequal Spheres in a Slow Linear Shear Flow, Report NCAR-TN/STR-64, National
Center for Atmospheric Research, Bolder, CO, 1971.
681. Batchelor, G.K., J. Fluid Mech., 74, 1, 1976.
682. Davis, R.H. and Hill, N.A., J. Fluid Mech., 236, 513, 1992.
683. Batchelor, G.K., J. Fluid Mech., 119, 379, 1982.
684. Batchelor, G.K. and Wen, C.-S., J. Fluid Mech., 124, 495, 1982.
685. Jeffrey, D.J. and Onishi, Y., J. Fluid Mech., 139, 261, 1984.
686. Fuentes, Y.O., Kim, S., and Jeffrey, D.J., Phys. Fluids, 31, 2445, 1988; Fuentes, Y.O., Kim, S., and
Jeffrey, D.J., Fuentes, Y.O., Kim, S., and Jeffrey, D.J., Phys. Fluids, A1, 61, 1989.
687. Stokes, G.G., Trans. Camb. Philos. Soc., 1, 104, 1851.
688. Exerowa, D. and Kruglyakov, P. M., Foam and Foam Films: Theory, Experiment, Application, Elsevier,
New York, 1998.
689. Ivanov, I.B., Radoev, B.P., Traykov, T.T., Dimitrov, D.S., Manev, E.D., and Vassilieff, C.S., in Proceedings
of the International Conference on Colloid Surface Science, Wolfram, E. (Ed.), Akademia Kiado,
Budapest, Hungary, 1975, p. 583.
690. Denkov, N.D., Petsev, D.N., and Danov, K.D., Phys. Rev. Lett., 71, 3226, 1993.
691. Valkovska, D.S., Danov, K.D., and Ivanov, I.B., Colloid Surf. A, 156, 547, 1999.
692. Davis, R.H., Schonberg, J.A., and Rallison, J.M., Phys. Fluids, A1, 77, 1989.
693. Chi, B.K. and Leal, L.G., J. Fluid Mech., 201, 123, 1989.
694. Ascoli, E.P., Dandy, D.S., and Leal, L.G., J. Fluid Mech., 213, 287, 1990.
695. Yiantsios, S.G. and Davis, R.H., J. Fluid Mech., 217, 547, 1990.
696. Zhang, X. and Davis, R.H., J. Fluid Mech., 230, 479, 1991.
697. Chesters, A.K., Trans. Inst. Chem. Eng. A, 69, 259, 1991.
698. Yiantsios, S.G. and Davis, R.H., J. Colloid Interface Sci., 144, 412, 1991.
699. Yiantsios, S.G. and Higgins, B.G., J. Colloid Interface Sci., 147, 341, 1991.
700. Joye, J.-L., Hirasaki, G.J., and Miller, C.A., Langmuir, 8, 3083, 1992.
701. Joye, J.-L., Hirasaki, G.J., and Miller, C.A., Langmuir, 10, 3174, 1994.
702. Abid, S. and Chestrers, A.K., Int. J. Multiphase Flow, 20, 613, 1994.
703. Li, D., J. Colloid Interface Sci., 163, 108, 1994.
704. Saboni, A., Gourdon, C., and Chesters, A.K., J. Colloid Interface Sci., 175, 27, 1995.
705. Rother, M.A., Zinchenko, A.Z., and Davis, R.H., J. Fluid Mech., 346, 117, 1997.
706. Singh, G., Miller, C.A., and Hirasaki, G.J., J. Colloid Interface Sci., 187, 334, 1997.
707. Cristini, V., Blawzdziewicz, J., and Loewenberg, M., J. Fluid Mech., 366, 259, 1998.
708. Bazhlekov, I.B., Chesters, A.K., and van de Vosse, F.N., Int. J. Multiphase Flow, 26, 445, 2000.
709. Bazhlekov, I.B., van de Vosse, F.N., and Chesters, A.K., J. Non-Newtonian Fluid Mech., 93, 181, 2000.
710. Chesters, A.K. and Bazhlekov, I.B., J. Colloid Interface Sci., 230, 229–243, 2000.
711. Boulton-Stone, J.M. and Blake, J.R., J. Fluid Mech., 254, 437, 1993.
712. Frankel, S. and Mysels, K., J. Phys. Chem., 66, 190, 1962.
713. Tabakova, S.S. and Danov, K.D., J. Colloid Interface Sci., 336, 273, 2009.
714. Velev, O.D., Constantinides, G.N., Avraam, D.G., Payatakes, A.C., and Borwankar, R.P., J. Colloid
Interface Sci., 175, 68, 1995.
715. Exerowa, D., Nikolov, A., and Zacharieva, M., J. Colloid Interface Sci., 81, 419, 1981.
716. de Vries, A.J., Rec. Trav. Chim. Pays-Bas., 77, 441, 1958.
717. Vrij, A., Discuss. Faraday Soc., 42, 23, 1966.
718. Ivanov, I.B., Radoev, B.P., Manev, E.D., and Sheludko, A.D., Trans. Faraday Soc., 66, 1262, 1970.
Chemical Physics of Colloid Systems and Interfaces 407
719. Gumerman, R.J. and Homsy, G.M., Chem. Eng. Commun., 2, 27, 1975.
720. Malhotra, A.K. and Wasan, D.T., Chem. Eng. Commun., 48 35, 1986.
721. Valkovska, D. S., Danov, K.D., and Ivanov, I.B., Adv. Colloid Interface Sci., 96, 101, 2002.
722. Manev, E.D., Sazdanova, S.V., and Wasan, D.T., J. Colloid Interface Sci., 97, 591, 1984.
723. Ivanov, I.B., Pure Appl. Chem., 52, 1241, 1980.
724. Aveyard, R., Binks, B.P., Fletcher, P.D.I., and Ye, X., Prog. Colloid Polymer Sci., 89, 114, 1992.
725. Velev, O.D., Gurkov, T.D., Chakarova, Sv.K., Dimitrova, B.I., and Ivanov, I.B., Colloids Surf. A, 83, 43,
1994.
726. Dickinson, E., Murray, B.S., and Stainsby, G., J. Chem. Soc. Faraday Trans., 84, 871, 1988.
727. Ivanov, I.B., Lectures at INTEVEP, Petroleos de Venezuela, Caracas, Venezuela, June 1995.
728. Ivanov, I.B. and Kralchevsky, P.A., Colloid Surf. A, 128, 155, 1997.
729. Basheva, E.S., Gurkov, T.D., Ivanov, I.B., Bantchev, G.B., Campbell, B., and Borwankar, R.P., Langmuir,
15, 6764, 1999.
730. Gurkov, T.D. and Basheva, E.S., Hydrodynamic behavior and stability of approaching deformable drops,
in Encyclopedia of Surface & Colloid Science, Hubbard, A.T. (Ed.), Marcel Dekker, New York, 2002.
731. Gurkov, T.D., Angarska, J.K., Tahcev, K.D., and Gaschler, W., Colloids Surf. A, 382, 174, 2011.
732. Boussinesq, M.J., Ann. Chim. Phys., 29, 349, 1913; Boussinesq, M.J., Ann. Chim. Phys., 29, 357, 1913.
733. Aris, R., Vectors, Tensors, and the Basic Equations of Fluid Mechanics, Prentice Hall, Englewood Cliffs,
NJ, 1962.
734. Brenner, H. and Leal, L.G., J. Colloid Interface Sci., 62, 238, 1977.
735. Brenner, H. and Leal, L.G., J. Colloid Interface Sci., 65, 191, 1978.
736. Brenner, H. and Leal, L.G., AIChE J., 24, 246, 1978.
737. Brenner, H. and Leal, L.G., J. Colloid Interface Sci., 88, 136, 1982.
738. Stone, H.A., Phys. Fluids, A2, 111, 1990.
739. Valkovska, D.S. and Danov, K.D., J. Colloid Interface Sci., 223, 314, 2000.
740. Stoyanov, S.D. and Denkov, N.D., Langmuir, 17, 1150, 2001.
741. Feng, S.-S., J. Colloid Interface Sci., 160, 449, 1993.
742. Stebe, K.J. and Maldarelli, Ch., Phys. Fluids, A3, 3, 1991.
743. Stebe, K.J. and Maldarelli, Ch., J. Colloid Interface Sci., 163, 177, 1994.
744. Scriven, L.E., Chem. Eng. Sci., 12, 98, 1960.
745. Scriven, L.E. and Sternling, C.V., J. Fluid Mech., 19, 321, 1964.
746. Slattery, J.C., Chem. Eng. Sci., 19, 379, 1964; Chem. Eng. Sci., 19, 453, 1964.
747. Slattery, J.C., I&EC Fundam., 6, 108, 1967.
748. Slattery, J.C., Interfacial Transport Phenomena, Springer-Verlag, New York, 1990.
749. Barton, K.D. and Subramanian, R.S., J. Colloid Interface Sci., 133, 214, 1989.
750. Feuillebois, F., J. Colloid Interface Sci., 131, 267, 1989.
751. Merritt, R.M. and Subramanian, R.S., J. Colloid Interface Sci., 131, 514, 1989.
752. Mannheimer, R.J. and Schechter, R.S., J. Colloid Interface Sci., 12, 98, 1969.
753. Pintar, A.J., Israel, A.B., and Wasan, D.T., J. Colloid Interface Sci., 37, 52, 1971.
754. Gardner, J.W. and Schechter, R.S., Colloid Interface Sci., 4, 98, 1976.
755. Li, D. and Slattery, J.C., J. Colloid Interface Sci., 125, 190, 1988.
756. Tambe, D.E. and Sharma, M.M., J. Colloid Interface Sci., 147, 137, 1991.
757. Tambe, D.E. and Sharma, M.M., J. Colloid Interface Sci., 157, 244, 1993.
758. Tambe, D.E. and Sharma, M.M., J. Colloid Interface Sci., 162, 1, 1994.
759. Horozov, T., Danov, K., Kralchevsky, P., Ivanov, I., and Borwankar, R., A local approach in interfacial
rheology: Theory and experiment, in First World Congress on Emulsion, Paris, France, 3-20, 137,
1993.
760. Danov, K. D., Ivanov, I. B., and Kralchevsky, P. A., Interfacial rheology and emulsion stability, in Second
World Congress on Emulsion, Paris, France, 2-2, 152, 1997.
761. Ivanov, I.B., Danov, K.D., Ananthapadmanabhan, K.P. and Lips, A. Adv. Colloid Interface Sci., 114–115,
61, 2005.
762. Murray, B.S. and Dickinson, E., Food Sci. Technol., 2, 131, 1996.
763. Miller, R., Wüstneck, R., Kärgel, J., and Kretzschmar, G., Colloids Surf. A, 111, 75, 1996.
764. Dickinson, E., Colloids Surf. B, 20, 197, 2001.
765. Bos, M.A. and van Vliet, T., Adv. Colloid Interface Sci., 91, 437, 2001.
766. Kärgel, J. and Derkatch, S.R., Curr. Opin. Colloid Interface Sci., 15, 246, 2010.
767. Miller, R., Ferri, J.K., Javadi, A., Kärgel, J., Mucic, N., and Wüstneck, R., Colloid Polym. Sci., 288, 937,
2010.
408 Handbook of Surface and Colloid Chemistry
768. Kärgel, J., Derkatch, S.R., and Miller, R., Adv. Colloid Interface Sci., 144, 38, 2008.
769. Murray, B.S., Curr. Opin. Colloid Interface Sci., 16, 27, 2011.
770. Radulova, G.M., Danov, K.D., Kralchevsky, P.A., Petkov, J.T., and Stoyanov, S.D., Soft Matter, 10, 5777,
2014.
771. Danov, K.D., Radulova, G.M., Kralchevsky, P.A., Golemanov, K., and Stoyanov, S.D., Faraday Discuss.,
158,195, 2012.
772. Danov, K.D., Kralchevsky, P.A., Radulova, G.M., Basheva, E.S., Stoyanov, S.D., and Pelan, E.G., Adv.
Colloid Interface Sci., 2015, doi: 10.1016/j.cis.2014.04.009.
773. Stanimirova, R., Marinova, K., Tcholakova, S., Denkov, N.D., Stoyanov, S., and Pelan, E., Langmuir, 27,
12486, 2011.
774. de Groot, S.R. and Mazur, P., Non-Equilibrium Thermodynamics, Interscience, New York, 1962.
775. Sagis, L.M.C and Öttinger, H.C., Phys. Rev. E, 88, 022149, 2013.
776. Moeckel, G.P., Arch. Rat. Mech. Anal., 57, 255, 1975.
777. Rushton, E. and Davies, G.A., Appl. Sci. Res., 28, 37, 1973.
778. Haber, S., Hetsroni, G., and Solan, A., Int. J. Multiphase Flow, 1, 57, 1973.
779. Reed, L.D. and Morrison, F.A., Int. J. Multiphase Flow, 1, 573, 1974.
780. Hetsroni, G. and Haber, S., Int. J. Multiphase Flow, 4, 1, 1978.
781. Morrison, F.A. and Reed, L.D., Int. J. Multiphase Flow, 4, 433, 1978.
782. Beshkov, V.N., Radoev, B.P., and Ivanov, I.B., Int. J. Multiphase Flow, 4, 563, 1978.
783. Danov, K.D., Stoyanov, S.D., Vitanov, N.K., and Ivanov, I.B., J. Colloid Interface Sci., 368, 342, 2012.
784. Murdoch, P.G. and Leng, D.E., Chem. Eng. Sci., 26, 1881, 1971.
785. Reed, X.B., Riolo, E., and Hartland, S., Int. J. Multiphase Flow, 1, 411, 1974; Reed, X.B., Riolo, E., and
Hartland, S., Int. J. Multiphase Flow, 1, 437, 1974.
786. Ivanov, I.B. and Traykov, T.T., Int. J. Multiphase Flow, 2, 397, 1976.
787. Traykov, T.T. and Ivanov, I.B., Int. J. Multiphase Flow, 3, 471, 1977.
788. Lu, C.-Y.D. and Cates, M.E., Langmuir, 11, 4225, 1995.
789. Jeelany, S.A.K. and Hartland, S., J. Colloid Interface Sci., 164, 296, 1994.
790. Zapryanov, Z., Malhotra, A.K., Aderangi, N., and Wasan, D.T., Int. J. Multiphase Flow, 9, 105, 1983.
791. Malhotra, A.K. and Wasan, D.T., Chem. Eng. Commun., 55, 95, 1987.
792. Malhotra, A.K. and Wasan, D.T., Interfacial rheological properties of adsorbed surfactant films with
applications to emulsion and foam stability, in Thin Liquid Films, Ivanov, I.B. (Ed.), Marcel Dekker, New
York, 1988, p. 829.
793. Singh, G., Hirasaki, G.J., and Miller, C.A., J. Colloid Interface Sci., 184, 92, 1996.
794. Traykov, T.T., Manev, E.D., and Ivanov, I.B., Int. J. Multiphase Flow, 3, 485, 1977.
795. Bancroft, W.D., J. Phys. Chem., 17, 514, 1913.
796. Griffin, J., Soc. Cosmet. Chem., 5, 4, 1954.
797. Davies, J.T., in Proceedings of the Second International Congress on Surface Activity, Vol. 1, Butterworths,
London, U.K., 1957, p. 426.
798. Shinoda, K. and Friberg, S., Emulsion and Solubilization, Wiley, New York, 1986.
799. Davis, H.T., Factors determining emulsion type: HLB and beyond, in Proceedings of First World
Congress on Emulsion, October 19–22, Paris, France, 1993, p. 69.
800. Israelachvili, J., The history, applications and science of emulsion, in Proceedings of First World Congress
on Emulsion, October 19–22, Paris, France, 1993, p. 53.
801. Kralchevsky, P.A., J. Colloid Interface Sci., 137, 217, 1990.
802. Gompper, G. and Schick, M., Phys. Rev., B41, 9148, 1990.
803. Lerczak, J., Schick, M., and Gompper, G., Phys. Rev., 46, 985, 1992.
804. Andelman, D., Cates, M.E., Roux, D., and Safran, S., J. Chem. Phys., 87, 7229, 1987.
805. Chandra, P. and Safran, S., Europhys. Lett., 17, 691, 1992.
806. Danov, K.D., Velev, O.D., Ivanov, I.B., and Borwankar, R.P., Bancroft rule and hydrodynamic stability of
thin films and emulsions, in First World Congress on Emulsion, October 19–22, Paris, France, 1993, p. 125.
807. Kunieda, H., Evans, D.F., Solans, C., and Yoshida, Colloids Surf., 47, 35, 1990.
808. Koper, G.J.M., Sager, W.F.C., Smeets, J., and Bedeaux, D., J. Phys. Chem., 99, 13291, 1995.
809. Ivanov, I.B., Danov, K.D., and Kralchevsky, P.A., Colloids Surf. A, 152, 161, 1999.
810. Velev, O.D., Gurkov, T.D., and Borwankar, R.P., J. Colloid Interface Sci., 159, 497, 1993.
811. Velev, O.D., Gurkov, T.D., Ivanov, I.B., and Borwankar, R.P., Phys. Rev. Lett., 75, 264, 1995.
812. Danov, K., Ivanov, I., Zapryanov, Z., Nakache, E., and Raharimalala, S., Marginal stability of emulsion
thin film, in Proceedings of the Conference of Synergetics, Order and Chaos, Velarde, M. (Ed.), World
Scientific, Singapore, 1988, p. 178.
Chemical Physics of Colloid Systems and Interfaces 409
813. Valkovska, D.S., Kralchevsky, P.A., Danov, K.D., Broze, G., and Mehreteab, A., Langmuir, 16, 8892,
2000.
814. Danov, K.D., Gurkov, T.D., Dimitrova, T.D., and Smith, D., J. Colloid Interface Sci., 188, 313, 1997.
815. Chevaillier, J.P., Klaseboer, E., Masbernat, O., and Gourdon, C., J. Colloid Interface Sci., 299, 472,
2006.
816. Chan, D.Y.C., Klaseboer, E., and Manica, R., Soft Matter, 7, 2235, 2011.
817. Chan, D.Y.C., Klaseboer, E., and Manica, R., Adv. Colloid Interface Sci., 165, 70, 2011.
818. Ivanov, I.B., Chakarova, S.K., and Dimitrova, B.I., Colloids Surf., 22, 311, 1987.
819. Dimitrova, B.I., Ivanov, I.B., and Nakache, E., J. Dispersion Sci. Technol., 9, 321, 1988.
820. Sternling, C.V. and Scriven, L.E., AIChE J., 5, 514, 1959.
821. Lin, S.P. and Brenner, H.J., J. Colloid Interface Sci., 85, 59, 1982.
822. Holly, F.J., On the wetting and drying of epithelial surfaces, in Wetting, Spreading and Adhesion, Padday,
J.F. (Ed.), Academic Press, New York, 1978, p. 439.
823. Castillo, J.L. and Velarde, M.G., J. Colloid Interface Sci., 108, 264, 1985.
824. Davis, R.H., Adv. Colloid Interface Sci., 43, 17, 1993.
825. Uijttewaal, W.S.J., Nijhof, E.-J., and Heethaar, R.M., Phys. Fluids, A5, 819, 1993.
826. Zapryanov, Z. and Tabakova, S., Dynamics of Bubbles, Drops and Rigid Particles, Kluwer Academic
Publishers, London, U.K., 1999.
827. Lorentz, H.A., Abhandl. Theoret. Phys., 1, 23, 1906.
828. Faxen, H., Arkiv. Mat. Astron. Fys., 17, 27, 1923.
829. Wakiya, S., J. Phys. Soc. Jpn, 12, 1130, 1957.
830. Dean, W.R. and O’Neill, M.E., Mathematika, 10, 13, 1963.
831. O’Neill, M.E., Mathematika, 11, 67, 1964.
832. Cooley, M.D.A. and O’Neill, M.E., J. Inst. Math. Appl., 4, 163, 1968.
833. Keh, H.J. and Tseng, C.H., Int. J. Multiphase Flow, 1, 185, 1994.
834. Schonberg, J. and Hinch, E.J., J. Fluid Mech., 203, 517, 1989.
835. Ryskin, G. and Leal, L.G., J. Fluid Mech., 148, 1, 1984; Ryskin, G. and Leal, L.G., J. Fluid Mech., 148,
19, 1984; Ryskin, G. and Leal, L.G., J. Fluid Mech., 148, 37, 1984.
836. Liron, N. and Barta, E., J. Fluid Mech., 238, 579, 1992.
837. Shapira, M. and Haber, S., Int. J. Multiphase Flow, 14, 483, 1988.
838. Shapira, M. and Haber, S., Int. J. Multiphase Flow, 16, 305, 1990.
839. Yang, S.-M. and Leal, L.G., J. Fluid Mech., 149, 275, 1984.
840. Yang, S.-M. and Leal, L.G., Int. J. Multiphase Flow, 16, 597, 1990.
841. Lebedev, A.A., Zhur. Russ. Fiz. Khim., 48, 1916.
842. Silvey, A., Phys. Rev., 7, 106, 1916.
843. Hadamar, J.S., Comp. Rend. Acad. Sci. (Paris), 152, 1735, 1911.
844. Rybczynski, W., Bull. Intern. Acad. Sci. (Cracovie), A, 1911.
845. He, Z., Dagan, Z., and Maldarelli, Ch., J. Fluid Mech., 222, 1, 1991.
846. Malysa, K., Krasowska, M., and Krzan, M., Adv. Colloid Interface Sci., 114−115, 205, 2005.
847. Krzan, M., Zawala, J., and Malysa, K., Colloids Surf. A, 298, 42, 2007.
848. Krzan, M. and Malysa, K., Physicochem. Problems Mineral Process., 43, 43, 2009.
849. Krzan, M. and Malysa, K., Physicochem. Problems Mineral Process., 48, 49, 2012.
850. Zawala, J., Malysa, E., Krzan, M., and Malysa, K., Physicochem. Problems Mineral Process., 50, 143, 2014.
851. Danov, K.D., Aust, R., Durst, F., and Lange, U., Chem. Eng. Sci., 50, 263, 1995.
852. Danov, K.D., Aust, R., Durst, F., and Lange, U., Chem. Eng. Sci., 50, 2943, 1995.
853. Danov, K.D., Aust, R., Durst, F., and Lange, U., Int. J. Multiphase Flow, 21, 1169, 1995.
854. Danov, K.D., Gurkov, T.D., Raszillier, H., and Durst, F., Chem. Eng. Sci., 53, 3413, 1998.
855. Stoos, J.A. and Leal, L.G., J. Fluid Mech., 217, 263, 1990.
856. Danov, K.D., Dimova, R.I., and Pouligny, B., Phys. Fluids, 12, 2711, 2000.
857. Dimova, R.I., Danov, K.D., Pouligny, B., and Ivanov, I.B., J. Colloid Interface Sci., 226, 35, 2000.
858. Angelova, M.I. and Pouligny, B., Pure Appl. Optics, 2, 261, 1993.
859. Pouligny, B., Martinot-Lagarde, G., and Angelova, M.I., Progr. Colloid Polym. Sci., 98, 280, 1995.
860. Dietrich, C., Angelova, M., and Pouligny, B., J. Phys. II France, 7, 1651, 1997.
861. Velikov, K., Dietrich, C., Hadjiski, A., Danov, K., and Pouligny, B., Europhys. Lett., 40(4), 405, 1997.
862. Velikov, K., Danov, K., Angelova, M., Dietrich, C., and Pouligny, B., Colloids Surf. A, 149, 245, 1999.
863. Dimova, R., Dietrich, C., Hadjiisky, A., Danov, K., and Pouligny, B., Eur. Phys. J. B, 12, 589, 1999.
864. Danov, K.D., Pouligny, B., Angelova, M.I., and Kralchevsky, P.A., Studies in Surface Science and
Catalysis, Vol. 132, Elsevier, Amsterdam, the Netherlands, 2001.
410 Handbook of Surface and Colloid Chemistry
865. Barentin, C., Ybert, C., di Meglio, J.M., and Joanny, J.F., J. Fluid Mech., 397, 331, 1999.
866. Maenosono, S., Dushkin, C.D., and Yamaguchi, Y., Colloid Polym. Sci., 227, 993, 1999.
867. Barentin, C., Muller, P., and Ybert, C., Eur. Phys. J. E, 2, 153, 2000.
868. Dimova, R., Pouligny, B., and Dietrich, C., Biophys. J., 79, 340, 2000.
869. Joseph, D.D., Wang, J., Bai, R., Yang, B.H., and Hu, H.H., J. Fluid Mech., 496, 139, 2003.
870. Sickert, M. and Rondelez, F., Phys. Rev. Lett., 90, 126104, 2003.
871. Fischer, T.M., Phys. Rev. Lett., 92, 139603, 2004.
872. Sickert, M. and Rondelez, F., Phys. Rev. Lett., 92, 139604, 2004.
873. Haris, S.S. and Giorgio, T.D., Gene Therapy, 12, 512, 2005.
874. Khattari, Z., Ruschel, Y., Wen, H.Z., Fischer, A. and Fischer, T.M., J. Phys. Chem. B, 109, 3402, 2005.
875. Singh, P. and Joseph, D.D., J. Fluid Mech., 530, 31, 2005.
876. Fischer, T.M., Dhar, P., and Heinig, P., J. Fluid Mech., 558, 451, 2006.
877. Sickert, M., Rondelez, F., and Stone, H.A., Europhys. Lett., 79, 66005, 2007.
878. Chen, W. and Tong, P., Europhys. Lett., 84, 28003, 2008.
879. Peng, Y., Chen, W., Fisher, T.M., Weitz, D.A., and Tong, P., J. Fluid Mech., 618, 243, 2009.
880. Lee, M.H., Reich, D.H., Stebe, K.J., and Leheny, R.L., Langmuir, 26, 2650, 2010.
881. Wilke, N., Vega Merkado, F., and Maggio, B., Langmuir, 26, 11050, 2010.
882. Singh, P., Joseph, D.D., and Aubry, N., Soft Matter, 6, 4310, 2010.
883. Ally, J. and Amirfazli, A., AIP Conf. Proc., 1311, 307, 2010.
884. Abras, D., Pranami, G., and Abbott, N.L., Soft Matter, 8, 2026, 2012.
885. Koynov, K. and Butt, H.-J., AIP Conf. Proc., 1518, 357, 2013.
886. Morse, J., Huang, A., Li, G., Maxey, M.R., and Tang, J.X., Biophys. J., 105, 21, 2013.
887. Memdoza, A.J., Guzman, E., Martines-Pedrero, F., Ritacco, H., Rubio, R.G., Ortega, F., Starov, V.M. and
Miller, R., Adv. Colloid Interface Sci., 206, 303, 2014.
888. Hunter, R.J., Foundation of Colloid Science, Vol. 1, Clarendon Press, Oxford, U.K., 1987.
889. Hunter, R.J., Foundation of Colloid Science, Vol. 2, Clarendon Press, Oxford, U.K., 1989.
890. Einstein, A., Ann. Phys., 19, 289, 1906.
891. Kubo, R., Rep. Prog. Phys., 29, 255, 1966.
892. Einstein, A., Ann. Phys., 34, 591, 1911.
893. Taylor, G.I., Proc. Roy. Soc. A, 138, 41, 1932.
894. Oldroyd, J.G., Proc. Roy. Soc. A, 232, 567, 1955.
895. Taylor, G.I., Proc. Roy. Soc. A, 146, 501, 1934.
896. Fröhlich, H. and Sack, R., Proc. Roy. Soc. A, 185, 415, 1946.
897. Oldroyd, J.G., Proc. Roy. Soc. A, 218, 122, 1953.
898. Batchelor, G.K., J. Fluid Mech., 83, 97, 1977.
899. De Kruif, C.G., Van Iersel, E.M.F., Vrij, A., and Russel, W.B., J. Chem. Phys., 83, 4717, 1985.
900. Loewenberg, M. and Hinch, E.J., J. Fluid Mech., 321, 395, 1996.
901. Da Cunha, F.R. and Hinch, E.J., J. Fluid Mech., 309, 211, 1996.
902. Li, X. and Pozrikidis, C., J. Fluid Mech., 341, 165, 1997.
903. Loewenberg, M., J. Fluids Eng., 120, 824, 1998.
904. Blawzdziewicz, J., Vajnryb, E., and Loewenberg, M., J. Fluid Mech., 395, 29, 1999.
905. Ramirez, J.A., Zinchenko, A., Loewenberg, M., and Davis, R.H., Chem. Eng. Sci., 54, 149, 1999.
906. Blawzdziewicz, J., Vlahovska, P., and Loewenberg, M., Physica A, 276, 50, 2000.
907. Breyannis, G. and Pozrikidis, C., Theor. Comp. Fluid Dynam., 13, 327, 2000.
908. Li, X. and Pozrikidis, C., Int. J. Multiphase Flow, 26, 1247, 2000.
909. Rednikov, A.Y., Ryazantsev, Y.S., and Velarde, M.G., Phys. Fluids, 6, 451, 1994.
910. Velarde, M.G., Philos. Trans. Roy. Soc., Math. Phys. Eng. Sci., 356, 829, 1998.
911. Danov, K.D., J. Colloid Interface Sci., 235, 144, 2001.
912. Barnes, H.A., Rheology of emulsions—A review, in Proceedings of First World Congress on Emulsion,
October 19–22, Paris, France, 1993, p. 267.
913. Krieger, L.M. and Dougherty, T.J., Trans. Soc. Rheol., 3, 137, 1959.
914. Wakeman, R., Powder Tech., 11, 297, 1975.
915. Prud’home, R.K. and Khan, S.A., Experimental results on foam rheology, in Foams: Theory,
Measurements, and Applications, Prud’home, R.K. and Khan, S.A. (Eds.), Marcel Dekker, New York,
1996, p. 217.
916. Tadros, T.F., Fundamental principles of emulsion rheology and their applications, in Proceedings of First
World Congress on Emulsion, October 19–22, Paris, France, 1993, p. 237.
917. Pal, R., Bhattacharya, S.N., and Rhodes, E., Can. J. Chem. Eng., 64, 3, 1986.
Chemical Physics of Colloid Systems and Interfaces 411
918. Edwards, D.A. and Wasan, D.T., Foam rheology: The theory and role of interfacial rheological prop-
erties, in Foams: Theory, Measurements, and Applications, Prud’home, R.K. and Khan, S.A. (Eds.),
Marcel Dekker, New York, 1996, p. 189.
919. Wessel, R. and Ball, R.C., Phys. Rev., A46, 3009, 1992.
920. Kanai, H., Navarrete, R.C., Macosko, C.W., and Scriven, L.E., Rheol. Acta, 31, 333, 1992.
921. Pal, R., Chem. Eng. Commun., 98, 211, 1990.
922. Pal, R., Colloids Surf. A, 71, 173, 1993.
923. Prins, A., Liquid flow in foams as affected by rheological surface properties: A contribution to a better
understanding of the foaming behaviour of liquids, in Hydrodynamics of Dispersed Media, Hulin, J.P.,
Cazabat, A.M., Guyon, E., and Carmona, F. (Eds.), Elsevier/North-Holland, Amsterdam, the Netherlands,
1990, p. 5.
924. Babak, V.G., Colloids Surf. A, 85, 279, 1994.
925. Yuhua, Y., Pal, R., and Masliyah, J., Chem. Eng. Sci., 46, 985, 1991.
926. Giesekus, H., Disperse systems: Dependence of rheological properties on the type of flow with impli-
cations for food rheology, in Physical Properties of Foods, Jowitt, R. et al. (Eds.), Applied Science
Publishers, London, 1983, Chapter 13.
927. Turian, R. and Yuan, T.-F., AIChE J., 23, 232, 1977.
928. Clarke, B., Trans. Inst. Chem. Eng., 45, 251, 1967.
929. Denkov, N.D., Tcholakova, S., Höhler, R., and Cohen-Addad, S., Foam rheology, in Foam Engineering:
Fundamentals and Applications, Stevenson, P. (Ed.), John Wiley & Sons, Chichester, U.K., 2012, p. 91.
930. Bingham, E.C., U.S. Bur. Standards Bull., 13, 309, 1916.
931. Herschel, W.K. and Bulkley, R., Kolloid Z., 39, 291, 1926.
932. Denkov, N.D., Tcholakova, S., Golemanov, K., Ananthapadmanabhan, K.P., and Lips, A., Phys. Rev.
Lett., 100, 138301, 2008.
933. Tcholakova, S., Denkov, N.D., Golemanov, K., Ananthapadmanabhan, K.P., and Lips, A., Phys. Rev. E,
78, 011405, 2008.
934. Golemanov, K., Denkov, N.D., Tcholakova, S., Vethamunthu, M., and Lips, A., Langmuir, 24, 9956, 2008.
935. Denkov, N.D., Tcholakova, S., Golemanov, K., Hu, T., Lips, A., Amer. Inst. Physics Conference
Proceedings, 1027, 902, 2008.
936. Denkov, N.D., Tcholakova, S., Golemanov, K., Ananthapadmanabhan, K.P., and Lips, A., Soft Matter, 5,
3389, 2009.
937. Politova, N., Tcholakova, S., Golemanov, K., Denkov, N.D., Vethamunthu, M., and Ananthapadmanabhan,
K.P., Langmuir, 28, 1115, 2012.
938. Mitrinova, Z., Tcholakova, S., Golemanov, K., Denkov, N.D., Vethamunthu, M., and Ananthapadmanabhan,
K.P., Colloids Surf. A, 438, 186, 2013,
939. Mitrinova, Z., Tcholakova, S., Popova, Z., Denkov, N., Dasgupta, B.R., and Ananthapadmanabhan, K.P.,
Langmuir, 29, 8255, 2013.
940. von Smoluchowsky, M., Phys. Z., 17, 557, 1916.
941. von Smoluchowsky, M., Z. Phys. Chem., 92, 129, 1917.
942. Wang, H. and Davis, R.H., J. Colloid Interface Sci., 159, 108, 1993.
943. Rogers, J.R. and Davis, R.H., Mettal. Trans., A21, 59, 1990.
944. Young, N.O., Goldstein, J.S., and Block, M.J., J. Fluid Mech., 6, 350, 1959.
945. Fuchs, N.A., Z. Phys., 89, 736, 1934.
946. Friedlander, S.K., Smoke, Dust and Haze: Fundamentals of Aerosol Behaviour, Wiley-Interscience, New
York, 1977.
947. Singer, J.M., Vekemans, F.C.A., Lichtenbelt, J.W.Th., Hesselink, F.Th., and Wiersema, P.H., J. Colloid
Interface Sci., 45, 608, 1973.
948. Leckband, D.E., Schmitt, F.-J., Israelachvili, J.N., and Knoll, W., Biochemistry, 33, 4611, 1994.
949. Bak, T.A. and Heilmann, O., J. Phys. A: Math. Gen., 24, 4889, 1991.
950. Martinov, G.A. and Muller, V.M., in Research in Surface Forces, Vol. 4, Plenum Press, Consultants
Bureau, New York, 1975, p. 3.
951. Elminyawi, I.M., Gangopadhyay, S., and Sorensen, C.M., J. Colloid Interface Sci., 144, 315, 1991.
952. Hartland, S. and Gakis, N., Proc. Roy. Soc. (Lond.), A369, 137, 1979.
953. Hartland, S. and Vohra, D.K., J. Colloid Interface Sci., 77, 295, 1980.
954. Lobo, L., Ivanov, I.B., and Wasan, D.T., AIChE J., 39, 322, 1993.
955. Danov, K.D., Ivanov, I.B., Gurkov, T.D., and Borwankar, R.P., J. Colloid Interface Sci., 167, 8, 1994.
956. van den Tempel, M., Recueil, 72, 433, 1953.
957. Borwankar, R.P., Lobo, L.A., and Wasan, D.T., Colloid Surf., 69, 135, 1992.
412 Handbook of Surface and Colloid Chemistry
958. Dukhin, S., Sæther, Ø., and Sjöblom, J., Coupling of coalescence and flocculation in dilute O/W emul-
sions, in Encycloped Handbook of Emulsion Technology, Sjöblom, J. (Ed.), Marcel Dekker, New York,
2001, p. 71.
959. Urbina-Villalba, G. and Garcia-Sucre, M., Mol. Simulat., 27, 75, 2001.
960. Madras, G. and McCoy, B.J., J. Colloid Interface Sci., 246, 356, 2002.
961. Barthelmes, G., Pratsinis, S.E., and Buggisch, H., Chem. Eng. Sci., 58, 2893, 2003.
962. Han, B.B., Akeprathumchai, S., Wickramasinghe, S.R., and Qian, X., AIChE J., 49, 1687, 2003.
963. Urbina-Villalba, G., Garcia-Sucre, M., and Toro-Mendoza, J., Phys. Rev. E, 68, 061408, 2003.
964. Petsev, D.N. (Ed.), Emulsions: Structure, Stability and Interactions, Elsevier, London, U.K., 2004.
965. Urbina-Villalba, G., Int. J. Mol. Sci., 10, 761, 2009.
966. Rahn-Chique, K., Puertas, A.M., Rumero-Cano, M.S., Rojas, C., and Urbina-Villalba, G., Adv. Colloid
Interface Sci., 178, 1, 2012.
5 Subsurface Colloidal Fines,
Behavior, Characterization,
and Their Role in Groundwater
Contamination
Tushar Kanti Sen and Chittaranjan Ray
CONTENTS
5.1 Introduction........................................................................................................................... 413
5.2 Subsurface Colloidal Fines and Their Classification............................................................. 414
5.3 Classification of Subsurface Environment and Importance on Colloid-Contaminant
Transport................................................................................................................................ 415
5.4 Sampling and Characterization............................................................................................. 417
5.4.1 Particle Size............................................................................................................... 420
5.4.2 Surface Charge and Zeta Potential............................................................................ 421
5.4.3 Chemical Characterization of Fines.......................................................................... 427
5.5 Brief Description of the Role of Inorganic Colloidal Fines–Associated Contaminant
Transport in Groundwater Flows and Associated Health Effect........................................... 427
5.6 Brief Description of the Role of Biocolloids in Groundwater Contamination and
Associated Health Effect....................................................................................................... 428
5.7 Conclusions............................................................................................................................ 431
References....................................................................................................................................... 431
5.1 INTRODUCTION
There are several classes of subsurface colloids—abiotic and biotic. The mobilization and migration
of these subsurface colloidal particles take place under different physical and geochemical condi-
tions (Massoudieh and Ginn, 2010; Sen, 2011; Sen and Khilar, 2006). Therefore, subsurface colloi-
dal fines can enhance or retard the mobility and dispersion of various contaminants in groundwater
flows (Sen and Khilar, 2009). There are two categories of colloid-induced subsurface contaminant
transport: (a) colloid-associated contaminant transport and (b) transport of biocolloids. General
research on subsurface colloid transport originated in the early 1930s (Kretzschmar and Schafer,
2005). First findings on the partitioning of aqueous solution constituents onto colloids appeared
in the late 1970s, as summarized by Gustafasson and Gschwend (1987). Subsequently, reports on
the transport facilitation of contaminants via association with subsurface colloidal fines emerged
in the 1980s (Enfield and Bengtsson, 1988; McCarthy and Zachara, 1989). Our knowledge and
understanding of the colloidal fines–associated contaminant transport in subsurface porous media
have increased substantially over the last three decades, which is reviewed by various researchers
from time to time such as Kretzschmar et al. (1999), Elimelech and Ryan (2002), Sen and Khilar
(2006, 2009), and Bin et al. (2011). The focus of this chapter is to review subsurface colloidal fines,
sampling methods and characterization, and their role in groundwater contamination. This has been
413
414 Handbook of Surface and Colloid Chemistry
evidenced by various experimental, modeling, and field studies, which have been briefly compiled
here and also partially updated in Chapter 4 (Sen and Khilar, 2009). Finally, authors here briefly
discussed subsurface inorganic/organic colloid–associated contaminant transport in groundwater
and their associated health effects.
Clay minerals
Metal oxides
Humic acids
Bacteria
Viruses
? Fine Coarse
clay clay Silt Sand
1 nm 1 μm 1 mm
FIGURE 5.1 Diameter range of various types of subsurface colloidal fines. (From Kretzschmar, R. et al.,
Adv. Agron., 66, 121, 1999.)
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 415
oxides and hydroxides, and OM. The clay minerals have fixed charge as well as a variable pH-
dependent charge (Khilar and Fogler, 1998; Sen and Khilar, 2009), whereas oxides and hydroxides
generally have a positive charge at neutral pH, but the charge is highly pH dependent (Cornell and
Schwertmann, 1998). Waterborne viruses include enteroviruses, coxsackievirus, echovirus, rotavi-
rus, norovirus, and hepatitis A and B. Bacteria of concern are chiefly enteropathogenic Escherichia
coli, Salmonella and Shigella spp., Campylobacter jejuni, and Aeromonas hydrophila, among oth-
ers. The main protozoa that have been transmitted by groundwater are Cryptosporidium parvum
and Giardia lamblia (Macler and Merkle, 2000). Bacteria are microscopic (1–6 μm in size) unicellu-
lar organisms with a nucleus; viruses are submicroscopic (20–120 nm in size) biological agents con-
sisting of molecules of nucleic acids and a protein envelope (Nevecherya et al., 2005) and negatively
charged at high pH (pH 7). With this unique feature of viral structure and colloidal physicochemical
properties, the transport of viruses in soil and groundwater can act with a combination of character-
istics ranging from solutes, colloids, and microorganisms. The contamination of subsurface water
by migrating pathogenic bacteria and viruses has caused large outbreaks of waterborne diseases
(Pekdeger et al., 1985; Stevik et al., 2004). The most important pathogenic germs that might possibly
be transported on the subsurface water path are the bacteria Salmonella sp., Shigella sp., Yersinia
enterocolitica, Y. pseudotuberculosis, Leptospira, Dyspepsia coli, enterotoxigenic E. coli, Vibrio
sp., Legionella sp., and the infectious hepatitis virus, polio virus, coxsackie viruses, rotavirus, and
norwalk-like virus (Pekdeger et al., 1985). Bradford et al. (2002) mentioned that these inorganic/
organic colloid particles can be released into the soil solution and groundwater through a variety
of hydrologic, geochemical, and microbiological processes including translocation from the vadose
zone (Nyham et al., 1985), dissolution of minerals and surface coatings (Ryan and Gschwend, 1990),
precipitation from solution (Gschwend and Reynolds, 1987), deflocculation of aggregates (McCarthy
and Zachara, 1989), and microbial-mediated solubilization of humic substances from kerogen and
lignitic materials (Quyang et al., 1996). Scientific interest in the way these colloidal particles are
transported through the intricate pores of subsurface porous environments has been inspired by two
facts affecting public health (Morales et al., 2011): colloids that can be contaminants by themselves
(bacteria, viruses, organics) and colloids that can act as carriers for inorganic/organic contaminants
(Elimelech and Ryan, 2002; Sen and Khilar, 2006, 2009). Basically, subsurface colloidal particles
can be released into soil solution and groundwater through a variety of hydrologic and geochemical
processes and also have been implicated in the transport of metals, radionuclide, and certain ioniz-
able organic pesticides in laboratory and field tests, which has been reviewed by Sen and Khilar
(2009, 2006), Flurry and Qiu (2008), Elimelech and Ryan (2002), Bin et al. (2011), and Massoudieh
and Ginn (2010). It is now generally accepted that in situ release of subsurface soil particles, col-
loidal in nature, may facilitate the contaminant transport in groundwater (Flurry and Qiu, 2008;
Kersting et al., 1999; Ryan and Elimelech, 1996; Sen and Khilar, 2009; Sen et al., 2002).
Evapotranspiration
Precipitation
Unsaturated zone
Soil zone
Recharge to
Water table
water table
Capillary fringe
Saturated zone
below the water table
(Ground water)
FIGURE 5.2 Classification of subsurface environment and characterization. (From USGS Circular 1186,
General facts and concepts about groundwater in sustainability of ground water—Water resources, pubs.usgs.
gov/circ/circ1186/html/gen_facts.html.)
phase (Figure 5.2). In comparison with the vadose zone, chemical conditions will be less vari-
able in time and space because, in general, the influence of dilution and evapotranspiration will
be negligible and the composition of the solid phase in the horizontal and vertical direction will
be less variable. Transport of contaminants in the saturated zone will be mainly horizontal (Sen
and Khilar, 2006) and is important for the transport of inorganic contaminants. The transport of
colloids/biocolloids through saturated and unsaturated porous media is of significant interest, from
the perspective of protecting the groundwater supplies from contamination (e.g., Redman et al.,
2001; Surampalli et al., 1997), assessment of risk from pathogens in groundwater (e.g., Bitton and
Gerba, 1984; Bruins et al., 2000; Goyal et al., 1989; Morris and Foster, 2000; Nola et al., 2001; Rose
et al., 2000; Scandura and Sobsey, 1997; Taylor et al., 2004; Yates and Yates, 1988), and design of
better water treatment systems to remove biocolloids from drinking water supplies (e.g., Gerba and
Goyal, 1985; Macler and Merkle, 2000; Schijven and Hassanizadeh, 2001). Because biocolloids are
living organisms, their transport in the subsurface is more complex than in the case for colloidal
solutes transport (Sen and Khilar, 2006, 2009). Not only are they subject to the same physicochemi-
cal phenomena as are colloids (Sen et al., 2004), but there are also a number of strictly biological
processes that affect their transport. Three key features of the vadose zone play a critical role in
colloidal movement: (1) the presence of air–water interfaces, (2) transients in flow and chemistry,
and (3) soil structure and heterogeneity.
It has been found that the fate and transport of colloids and colloid-associated contaminants in
saturated porous media are controlled by a combination of several basic factors including (1) the
presence of colloidal particles in the subsurface environment; (2) their release, dispersion stability,
migration, and straining/plugging at pore constrictions; and (3) association of contaminants with
colloidal particles/solid matrix (Sen and Khilar, 2006, 2009; Sen et al., 2002, 2004, 2005). Similar
observations by several investigators have been compiled by Sen and Khilar (2009).
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 417
Chemical colloidal and hydrodynamic forces are important to colloidal particle release, deposi-
tion, and straining/plugging in saturated porous media, which are also dependent on geochemi-
cal and hydrodynamic conditions of the subsurface medium (Ryan and Elimelech, 1996; Sen and
Khilar, 2002, 2006). However, differences in the mobilization mechanisms and deposition behavior
of colloids are vastly different for saturated and unsaturated systems: straining, for instance, is
anticipated to be even more important in unsaturated than in saturated systems (Bradford et al.,
2006; Gao et al., 2006). The retention of colloidal fines in saturated porous media is mainly due to
attachment at the solid–liquid interface and the change in surface properties of solid and solution
that has been well documented in literature (Kretzschmar et al., 1999; McCarthy and Zachara, 1989;
Sen and Khilar, 2006, 2009; Sharma et al., 2008). But mechanisms of colloidal retention in unsatu-
rated porous media may include adsorption (attachment/detachment processes) to the solid–water
interface (SWI) (Ryan and Elimelech, 1996; Sen and Khilar, 2006, 2009), the air–water interface
(AWI) (Nicole et al., 2004; Torkzaban et al., 2008; Zhang et al., 2010), the solid–water–air (SWA)
straining (Bradford et al., 2006; Torkzaban et al., 2008), and film straining (Chin and Flurry, 2005;
Kretzschmar and Schafer, 2005; Sen, 2011).
TABLE 5.1
Recommended Protocol for Groundwater Sampling of Colloids and Colloid-Associated
Contaminants
1. Measured depth to water calculates standing water volume.
2. (Optional) Make dipstick observation of any layering in the well (e.g., floating NAPL).
3. Slowly lower inlet/pump to sampling depth.
4. Inflate packer(s) to reduce well volume and isolate inlet from settled particles at well bottom or floating NAPL at the
surface.
5. Initiate pumping at a slow flow rate (100 mL/min). Avoid surging. Observe air bubbles displaced from sample tube to
assess progress of steady pumping until water arrives at the surface.
6. Measure volumetric flow rate and adjust pumping rate as necessary.
7. Use a flow-through cell to monitor temperature, specific conductivity, pH, Eh, and dissolved oxygen. Repeatedly collect
water and assess turbidity.
8. (Optional) Initiate slow pumping at another well(s) on-site. In addition to sampling wells within the contaminated site
of the aquifer, background wells should be sampled. The background well should be located outside the contaminated
zone but constructed and developed in the same manner and screened in the same aquifer as the contaminated well.
9. After turbidity has stabilized, collect water samples for mobile contaminant analysis in suitable flow-through vessels.
Minimize contact with air. Do not filter for groundwater analyses. Preserve as appropriate (e.g., store in cold and dark
for organics).
10. (Optional) Collect colloid samples on membrane filters and store for characterization (SEM/EDX and XRD analysis).
Source: Adapted from Backhus, D.A. et al., Ground Water, 31, 466, 1993.
Different techniques are usually used to collect water sample from the vadose zone such as
free drainage lysimeters, suction cups or porous plates, and fiberglass wicks or rock wool samplers
(Takala and Manninen, 2006). The suitability of fiberglass wicks for colloidal sampling was tested
by Czinany (2004). A zero-tension lysimeter is used to sample colloidal particles and to routinely
monitor colloidal concentrations in field soils (Thompson and Scharf, 1994). Such zero-tension
lysimeters are installed at a certain depth below undisturbed soil cores and these basically con-
sist of a sampling cup with an air inlet tube and a sampling tube. Field flow fractionation (FFF),
a rapid and high-resolution fractionation technique, may be used for the characterization of low
concentration natural colloids (Beckett and Hart, 1993). This sampling technique is based on colloid
interaction with an externally applied force field in a thin, ribbon-like flow channel.
Sampling and analysis of microorganisms in groundwater are especially problematic because tra-
ditional analytical methods (APHA, 1995) are based on culturing of viable cells and many pathogens
of interest are not readily culturable in the laboratory (Maier et al., 2000; Sen and Khilar, 2009). This
problem has been partly solved by the development of nucleic acid-based analysis, which can detect
both viable and nonviable microorganisms (Josephson et al., 1993; McCarthy and McKay, 2004).
The nature of colloidal materials is complex and several properties should be considered to char-
acterize the dispersion of colloids. Different techniques for characterization of colloids and nanofines
(synthetic or natural) have been presented in different articles (Hassellov et al., 2008; Hennebert
et al., 2013; McCarthy and Degueldre, 1993; Schurtenberger and Newman, 1993). When character-
izing colloidal particles, one is usually primarily interested in composition (mineralogy, elemental
composition, and organic components), concentration (mass concentration, number concentration),
size and shape (size distribution), and surface characteristics (specific surface area (SA), surface
charge, and sorption capacity). Some important techniques that can be used to analyze subsurface
colloids are summarized in Table 5.2, which is updated and taken from our earlier publication (Sen
and Khilar, 2009). The elemental and morphological analysis of colloidal particles can be done
either from large single particles (<500 nm) or from agglomerate of the colloidal particles or from a
filter cake with an energy dispersive x-ray spectroscope, SEM/EDS (Takala and Manninen, 2006).
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 419
Mineralogical characteristics of colloidal particles can be found from XRD analysis. Wang and
Keller (2009) collected three bulk agricultural soils and one sediment (denoted as Ag#1, Ag#2, and
Ag#3, and sediment from the topsoil layer of Santa Barbara, CA) and separated natural colloids and
analyzed various properties, which are presented in Table 5.3. From Table 5.3, although natural col-
loids only vary from 2.7% to 13.8% by weight, they have significantly higher organic carbon (OC),
cation exchange capacity (CEC), and BET SA than their bulk soils and sediments and hence have
strong implications in the colloid-associated contaminant transport.
There are only few studies dealing with the release and characterization of colloids from solid
waste. Hennebert et al. (2013) reported potential colloidal release and their various characterizations
TABLE 5.2
Various Methods Used for the Characterization of Colloidal Particles in Subsurface
Systems
Colloid Property Method Information
Particle Inductively coupled plasma-atomic Bulk elemental composition for most elements with atomic
composition emission spectroscopy (ICP-AES) number >10, isotopic analysis with ICP-MS, mass
and inductively coupled plasma- concentration if all constituent elements are analyzed
mass spectroscopy (ICP-MS)
Energy dispersive analysis of x-rays Elemental composition of single particles viewed by SEM
(EDAX) or TEM, analysis of most elements with atomic number
>10, element mapping
X-ray diffraction analysis (XRD) Mineralogical composition
Fourier transform infrared Mineralogical composition and organic functional groups
spectroscopy (FT-IR, FT-IR image)
Thermogravimetric analysis (TGA), Quantitative analysis of minerals
differential thermal analysis (DTA)
Mossbauer spectroscopy Oxidation state and structure of Fe-containing minerals
Morphology and Light microscopy Particle number, morphology
size distribution Transmission electron microscopy Particle morphology, size distribution
(TEM)
Atomic force microscopy (AFM) Particle morphology, size distribution
Static light scattering (SLS) Particle size and shape with the small angle option for
particle size distribution in the range of 0.1–100 μm
Dynamic light scattering (DLS) Average hydrodynamic radius of particles <1 μm, upper
size cut-off for polydisperse samples from multiangle
measurements
Sedimentation analysis Particle size distribution based on sedimentation velocity in
a gravitational field, size range 1–50 μm
Centrifuge techniques Particle size distribution based on sedimentation velocity in
a centrifugal field, analytical and preparative techniques,
wide size ranges from 1 to 1000 nm
Field flow fractionation (FFF) Size distribution based on sedimentation velocity or
diffusion coefficient
Particle surface Electrophoretic mobility Electrophoretic mobility, which is an indicator of surface
characteristics measurement (EM), zeta analyzer charge (zeta potential)
Gas adsorption (BET) Specific surface area, microporosity
Sources: Adapted from Kretzschmar, R. et al., Adv. Agron., 66, 121, 1999; Sen, T.K. and Khilar, K.C., Mobile subsurface
colloids and colloid-mediated transport of contaminants in subsurface soils, in: Birdi, K.S. (Ed.), Handbook of
Surface and Colloid Chemistry, 3rd edn., Taylor & Francis Group/CRC Press, 2009, pp. 107–130.
420 Handbook of Surface and Colloid Chemistry
TABLE 5.3
Measured Properties of Natural Colloids, Bulk Soils, and Sediment
Surface Exchangeable Ca2+ +
Soil Weight (%) OC (%) CEC (cmol/kg) Area (m2/g) Mg2+ (cmol/kg) Fe Content (%)
Ag#1
Bulk 100 1.51 6.20 3.5 5.8 0.19
Colloids 5.3 4.95 40.2 30.8 39.0
Ag#2
Bulk 100 1.50 15.2 9.4 14.1 0.14
Colloids 6.8 4.36 59.0 53.1 57.5
Ag#3
Bulk 100 1.52 15.4 14.1 15.1 0.17
Colloids 13.8 4.50 54.4 63.1 53.6
Sediment
Bulk 100 1.12 5.4 2.0 5.2 0.19
Colloids 2.7 6.02 42.2 36.6 40.7
Source: Adapted from Wang, P. and Keller, A.A., Langmuir, 25(12), 6856, 2009.
Abbreviations: CEC, cation exchange capacity; OC, organic carbon
from solid industrial and municipal waste leachates, contaminated soil, contaminated sediments,
and landfill leachates. Potentially released colloids and their characterization were carried out with
standard leaching, membrane filtration, and elemental determination of filtrate fraction (ICP), and
visible morphological characteristics of colloids were done by the TEM-EDS system. They have
reported that Mn, As, Co, Pb, Sn, and Zn always had a colloidal form and total organic carbon
(TOC), Fe, P, Ba, Cr, Cu, and Ni had a partially colloidal form that existed in various waste leach-
ates and contaminated soil. Koster et al. (2007) show the release of Cu, Pb, Zn, and Cr in colloidal
form from municipal solid waste incinerator’s bottom ash.
FIGURE 5.3 SEM micrograph of an agglomerate of the colloidal particles from mine drainage tunnel.
(From Zanker, H. et al., A Separation and Detection Scheme for Environmental Colloids, Institute of
Radiochemistry, Dresden, Germany, 2000, www.fz-rossendorf.de/FWR/COLL/Coll_2.htm; Takala, M. and
Manninen, P., Sampling and analysis of groundwater colloids—A literature review, Working report 2006-15,
Posiva, Finland, 2006, (29), 43, http://www.posiva.fi/tyoraportit/WR2006-98web.pdf.)
TABLE 5.4
Dependence of Plugging on the Ratio of Size of Fines to Size of Pore Constrictions
Size of Fines/Size of Pore Constrictions Occurrence
≥1 Plugging due to blocking or size exclusion.
0.1–0.6 Plugging due to bridging and multiparticle blocking.
0.04–0.10 Plugging due to surface deposition, bridging, and multiparticle blocking.
0.01–0.04 Surface deposition. Multiparticle blocking may or may not occur.
Source: Khilar, K.C. and Fogler, H.S., Migration of Fines in Porous Media, Kluwer Academic Publishers, Dordrecht, the
Netherlands, 1998, Chapters 1, 3, 9.
surface can exchange its portable cations present on the surface for cations present in the solu-
tion. The amount exchanged at equilibrium is known as the CEC. Likewise, we can define anion
exchange capacity (AEC) of positively charged colloidal fines. The CEC of several clay minerals
has been listed by Grim (1968). For migratory clayey fines such as kaolin and illite, the CEC is of
the order of 3–15 meq/g (Khilar and Fogler, 1998). Further soil colloids are heterogeneous in com-
position and consist of inorganic and organic constituents or a mixture of the two. Therefore, charge
development of soil colloids in geochemical environment is a complex process due to heterogeneity
in the composition and structures of colloidal particles (Bin et al., 2011).
The surfaces of oxide colloidal particles or silicate minerals in contact with aqueous solutions are
charged positively or negatively by adsorption or desorption of H+ ions. An electrical double layer
at solid/liquid interface is formed by adsorbing counterions from the aqueous solution to its surface
(Sen and Khilar, 2009). The charge developing process on the surfaces can be represented as follows
(Cornell and Schwertmann, 1998):
where
MOH denotes a surface site
int int
K a1 and K a2 are the intrinsic surface acidity constants
Schindler and Stumm (1987) summarized the intrinsic acidity constants of various metal oxides/
minerals under solution chemistry. In this system, H+ and OH− are potential determining ions and
the surface charge and surface potential are defined by the pH of the solution. There is a particular
pH at which the surface charge becomes zero, that is, the point of zero charge (pzc) or the isoelectric
point (iep). K aint is correlated to the pzc, which is as follows (Bin et al., 2011):
Here intrinsic acidity constant (K aint) is equal to the apparent acidity constant (Ka) at a surface poten-
tial of zero. Point of zero charge (pzc) can be measured by electrophoresis as the pH at which there is
no motion of the particles in suspending medium under an applied external electric field. Table 5.5
presents the point of zero charge of simple hydro(oxides) and other materials (Kosmulski, 2004;
Sen and Khilar, 2009).
The more relevant and useful method of surface charge characterization of colloidal fines is,
however, the measurement of zeta potential of the particle. The zeta potential of kaolinite has been
studied by several investigators (Herrington et al., 1992; Hussain et al., 1996; Khilar and Fogler,
1983; Kia et al., 1987; Williams and Williams, 1978). The kaolinite face zeta potential varies
from approximately –35 to –25 mV at a pH of 6.5 in the range of sodium chloride concentration
between 10 −4 and 10 −2 M (VanOlphen, 1977). Hussain et al. (1996) measured the zeta potential of
kaolinite, illite, and chlorite clay minerals against pH and found to have negative zeta potential
in the pH range of 2.5–11. The kaolinite is the most negative clay, ranging in zeta potential value
from –24 to –49.5 mV. There have been numerous investigations of the electrochemical proper-
ties of goethite, hematite, and ferrihydrite (Atkinson et al., 1967; Breeuwsma and Lyklema, 1971,
1973; Crawford et al., 1996; Davis and Leckie, 1978; Onoda and Debruyn, 1966; Park, 1967; Puls
and Powell, 1992).
Park (1967) summarized the factors controlling the sign and magnitude of surface charge of
oxide and mineral oxides, especially hydrous metal oxides.
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 423
TABLE 5.5
Point of Zero Charge (pzc) of Simple Hydro(oxides) and Other Materials
Material Description Salt T Method pzc References
Al2O3 γ, >99% 0.1 mol/dm 3 25 pH 9.3 Fein and Brady
NaCl 60 9 (1995)
Al2O3 α 0.01 mol/dm3 25 IEP 8 Guo et al. (1997)
NaCl
Al2O3 α, A16, Alcoa None IEP 8.5 Yang and
Troczynski
(1999)
Al2O3 γ, from 0.001–0.1 mol/ 25 CIP 8.6 Ardizzone et al.
sec-butoxide dm3 KNO3 (2000)
Al2O3 Riedel Haen, purity 0.01 mol/dm3 20 IEP 8 Ramos-Tejada
98% KCl et al. (2002)
Al2O3 α, AKP-30 0.01 mol/dm3 22 IEP 9 Hackley et al.
NaNO3 (2002)
Al2O3 γ, spherical, 0.01 mol/dm3 IEP 9.6 Tang et al. (2002)
NanoTek NaCl
Al2O3 γ, Aldrich, 0–0.1 mol/dm3 CIP 9.1 Wang et al. (2002)
mesoporous NaCl
Al2O3 Fisher 0–0.1 mol/dm3 CIP 8.7 Wang et al. (2002)
NaCl
Al2O3 γ, from nitrate HNO3 + KOH 25 IEP 8.2 Vakros et al.
CIP 8.6 (2002), Bourikas
pH 8–8.4 et al. (2003)
Al2O3 α, sapphire 0.001 mol/dm3 IEP 5 Franks and
KBr, KNO3 Meagher (2003)
Al2O3 γ 0.001–0.1 mol/ 25 IEP 8.4 De Lint et al.
dm3 NaCl (2003)
Al2O3 Alcoa Acousto 7.8 Cordelair and
Greil (2003)
Al2O3 Aldrich, 99.5%, α IEP 9 Sun and Berg
(2003)
Al2O3 Alfa Aesar, 99.9% 0.01 mol/dm3 IEP 9.1 Hu and Dai (2003)
KCl
Al2O3 γ, Merck 0.1 mol/dm3 25 CIP 8.2 Lefevre et al.
NaNO3 (2004)
AlOOH, Six different HCl + KOH 25 IEP 8.5–9 Katatny et al.
Al (OH)3 recipes (2003)
Co3O4 Thermal 0.005–0.3 mol/ CIP 7.2 Faria and Trasatti
decomposition of dm3 KNO3 (2003)
nitrate at 400°C
Cr2O3 From chloride 0.1 mol/dm3 KCl pH 4 Onija et al. (2003)
Cr2O3 Fluka 0.001–0.1 mol/ 22 pH 6.7 Onija and Milonjic
dm3 KCl (2003)
Cr(OH)3 From chloride 0.1 mol/dm3 KCl pH 4.7 Onija et al. (2003)
CuO Tenorite, Merck 0.001 mol/dm3 pH 6.9, 7.6 Gonzalez and
KClO4 Laskowski (1974)
(Continued)
424 Handbook of Surface and Colloid Chemistry
Sources: Taken from Kosmulski, M., J. Colloid Interface Sci., 275, 214, 2004; Sen, T.K. and Khilar, K.C., Mobile
subsurface colloids and colloid-mediated transport of contaminants in subsurface soils, in: Birdi, K.S. (Ed.),
Handbook of Surface and Colloid Chemistry, 3rd edn., Taylor & Francis Group/CRC Press, 2009, pp. 107–130.
More data are available in Kosmulski, 2004.
426 Handbook of Surface and Colloid Chemistry
Crawford et al. (1996) measured the zeta potential of amorphous hydrous iron (III) oxide (HFO)
and amorphous hydrous chromium (III) oxide (HCO) as a function of pH during adsorption and
coprecipitation of Cr3+, Zn2+, and Ni2+. For sand materials, the zeta potential is negative and it
increases as the pH increases (Elimelech et al., 2000).
Hennebert et al. (2013) measured the zeta potential of solid industrial and municipal waste
leachates under different conditions, and it was found that six leachates had a positive zeta poten-
tial (<8 mV) and all others had a negative zeta potential (−15 to −38 mV). A particle with a zeta
potential >− 30 mV is considered as electrostatically stable (Hassellov and Kaegi, 2009). The mea-
sured zeta potentials for the three types of natural colloids (Ag#1, Ag#2, Ag#3) and engineered
zeolite particles under different salt concentrations are presented in Figure 5.4 (Wang and Keller,
2009). Detailed mechanistic explanations are given in Wang and Keller (2009). The zeta potential
of all the natural colloids showed similar trends (Figure 5.4).
Ag#1_K+ 0
0
Ag#3_K+
–10
–10 Zeolite_K+
–40 –40
–5 –4 –3 –2 –5 –4 –3 –2
(a) Log (KCI) (b) Log (CaCl2)
Ag#1_divalent
0 0
Ag#3_divalent
Zeolite_Ca2+
Zeta potential (mV)
Zeta potential (mV)
–10 –10
–20 –20
Ag#1_divalent
–30 Ag#3_divalent
–30
Zeolite_Ca2+
–40 –40
–5 –4 –3 –2 –5 –4 –3 –2
(c) Log (KCI) (d) Log (CaCl2)
Ag#1_divalent Ag#1_K+
Ag#3_divalent Ag#3_K+
0 Zeolite_Ca2+ Zeolite_K+
Zeta potential (mV)
–10
–20
–30
–40
–5 –4 –3 –2
(e) Log (KCI)
FIGURE 5.4 Zeta potentials of the original and cation-exchanged natural colloids (Ag#1 and Ag#3) and
engineered zeolite as a function of log molar concentrations of KCl and CaCl2 (pH 7.0): (a) K+ saturated col-
loids with KCl in the bulk solution, (b) K+ saturated colloids with CaCl2 in the bulk solution, (c) divalent cation
saturated colloids with KCl in the bulk solution, (d) divalent cation saturated colloids with CaCl2 in the bulk
solution, and (e) comparison of K+ and divalent cation saturated colloids with KCl in the bulk solution. (From
Wang, P. and Keller, A.A., Langmuir, 25(12), 6856, 2009.)
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 427
TABLE 5.6
Range in Chemical Composition of Migratory Clays
Clay SiO2 Al2O3 Fe2O3 TiO2 CaO MgO K2O Na2O
Kaolin 40–49 35–40 0–13 0–2 0–0.8 0–0.5 0–0.10 0–0.2
Illite 50–56 18–31 2–5 0–0.8 0–0.2 1–4 — 0–1
Chlorite 31–38 18–20 — — — 35–38 — —
Montmorillonite 45–55 19–50 0–3 — — 0–2 — —
Source: Khilar, K.C. and Fogler, H.S., Migration of Fines in Porous Media, Kluwer Academic Publishers, Dordrecht, the
Netherlands, 1998, Chapters 1, 3, 9.
carry with them the attached contaminates, the phenomena have been termed as colloid-facilitated
transport (McCarthy and Zachara, 1989).
Nanoparticles have found their way for many human use products over the years. They are in
the colloidal or subcolloidal size range. Nowack and Bucheli (2007) present a review of classes of
nanoparticles relevant to the environment and summarize their formation, emission, and occurrence
in the environment. At elevated concentrations, they reported the toxic ecotoxicological effects of
certain nanoparticles.
Silver nanoparticles are widely used for their antimicrobial properties. Toxic effects of soluble
silver on humans are presented in a review by Panyala et al. (2008). In recent years, the regulating
agencies are concerned about the widespread use of inorganic nanoparticles, including the use of
carbon nanotubes (CNTs) in various items used by human. Choi et al. (2010) show that nanopar-
ticles that have a hydrodynamic diameter less than 34 nm and have noncationic surface charge
can translocate rapidly from the lungs to the mediastinal lymph nodes. Nanoparticles with hydro-
dynamic radius less than 6 nm can traffic rapidly from lungs to lymph nodes and blood streams.
However, the kidney may subsequently clear them. Colvin (2003) is one of the early researchers to
alarm about the widespread use of nanoparticles in products for human use. She pointed out that a
few studies have investigated the direct or indirect exposure to nanoparticles and no clear guidelines
exist to quantify their effects.
Lam et al. (2006) conducted a review of potential CNT toxicity. From rodent studies they sum-
marized that regardless of the synthesis methods or metals used, CNTs caused inflammation, epi-
thelioid granulomas (microscopic nodules), fibrosis, and biochemical/toxicological changes in the
lungs. Their review suggests that single-walled CNTs are toxic than quartz (which is considered as
a serious occupational hazard when chronically inhaled).
TABLE 5.7
Selected Waterborne Bacterial and Protozoan Diseases, Their Agents, Symptoms, and
Sources in Water Supply
Bacterial Species or
Disease Genera Symptoms Sources in Water Supply
Amoebiasis Entamoeba hystolytica Diarrhea, bloating, fever Sewage-contaminated water
Typhoid Salmonella typhi Very high temperature, sweating, etc. Fecal-contaminated water
Cryptosporidiosis Cryptosporidium Watery diarrhea, flu-like symptom Unfiltered water containing
parvum animal manure
Giardiasis Giardia lamblia Diarrhea, abdominal discomfort Untreated water, water from
beaver and muskrat ponds
Cholera Vibrio cholerae Watery diarrhea, nausea, Drinking water contaminated
vomiting, etc. with the bacterium
Leptospirosis Bacteria genus Flu-like symptom followed by Water contaminated by animal
Leptospira meningitis, liver damage, etc. urine carrying the bacterium
Legionellosis Bacterial genus Fever, chills, pneumonia, etc. Contaminated water-cooling
Legionella towers
Escherichia coli Escherichia coli Diarrhea and can lead to death in Water contaminated with the
infection people with weak immune response bacterium
Dysentery Bacterial genera Frequent passing of feces with blood Water contaminated with the
Shigella and and/or mucus bacteria
Salmonella
Botulism Clostridium botulinum Dry mouth, blurred vision, slurred Bacterium enters through
speech wounds or gastrointestinal tracts
Campylobacteriosis Campylobacter jejuni Dysentery with high fever Drinking water contaminated
with feces
Source: Kolwzan, B. et al., Introduction to Environmental Microbiology, Oficyna Wydawnicza poletechruki wroclawskieji,
Nigeria, 2006, http://www.oficyna.pwr.wroc.pl.
TABLE 5.8
Selected Waterborne Viral Diseases and Symptom
Disease and/or Agent Symptoms Sources in Water Supply
Poliomyelitis (polio virus) Minor or no symptoms to headache, fever, etc. Feces-contaminated water
Hepatitis A Fatigue, fever, abdominal pain Can manifest in water and food
Severe acute respiratory Fever, lethargy, cough, sore throat Manifests in improperly treated
syndrome water
Coxsackie A Herpangina, hemorrhagic conjunctivitis, meningitis Feces contamination
Enteroviruses Meningitis, encephalitis, hemorrhagic conjunctivitis, Similarity with polio virus
fever
Enteric cytopathic human Meningitis, respiratory system problems, rash, diarrhea, Feces contamination
orphan (ECHO) virus fever
Norwalk virus Diarrhea, fever Feces contamination
Adenoviruses Respiratory problems, diarrhea, eye infection Respiratory droplets, fecal
route in water
Rotaviruses Severe diarrhea Fecal–oral route exposure
Source: Kolwzan, B. et al., Introduction to Environmental Microbiology, Oficyna Wydawnicza poletechruki wroclawskieji,
Nigeria, 2006, http://www.oficyna.pwr.wroc.pl.
430 Handbook of Surface and Colloid Chemistry
pathogenic microorganisms (bacteria and virus), waterborne bacterial infections, and intestinal
viruses that may be transmitted by water and diseases caused by them (Kolwzan et al., 2006).
In rural areas, human and animal waste sources contribute to groundwater contamination. For
example, septic tanks and cesspools used for human waste disposal have traditionally been con-
sidered as source of fecal contamination of groundwater. Additionally, animal waste lagoons and
stockyards are becoming important sources for fecal contamination of shallow groundwater in rural
landscapes and the number of these units and the animals housed in the facilities are becoming
more and more. Yates (1985) states that septic tanks contribute 800 billion gallons of water per year
to subsurface, and this subsurface water is a source of pathogens including bacteria and viruses. She
points out that in regions having more than 40 systems per square mile (1 system per 16 acres), the
Environmental Protection Agency (EPA) estimates high potential for groundwater contamination.
Ahmed et al. (2005) conducted biochemical fingerprinting enterococci and E. coli to relate possible
migration of these organisms from septic tanks to surface water (a creek).
In 1976, cryptosporidiosis was first recognized to infect humans and the symptoms were diar-
rhea that generally ends in 2–3 weeks. Among the most well-known public health concerns involv-
ing biocolloids is the outbreak of cryptosporidium in the water system of the city of Milwaukee,
Wisconsin, in 1993 (MacKenzie et al., 1994). Contamination occurred as the sand filtration system
at the Howard Avenue Water treatment Plant on the shore of Lake Michigan failed to remove the
oocysts of cryptosporidium during rapid sand filtration. Because of a storm even in early spring
season (when cows were out in open areas), the oocysts in manure mixed in runoff water ended up
near the intake of the treatment plant. The filtration system was not operating optimally and that led
to a bypassing of the oocysts during the backwash of the filters. Over a period of 2 weeks (March
23 to April 8, 1993), over 400,000 people became infected and more than 100 weak or immune-
compromised individuals died.
An outbreak of cryptosporidiosis in a disinfected groundwater supply occurred in Warrington, a
town in Northwest England, between November 1992 and February 1993 (Bridgeman et al., 1995).
A total of 47 cases were recorded over this period, and the bypass flow of surface water containing
animal feces to aquifer was attributed to the cause. The bypass flow did not take advantage of filtra-
tion in natural sandstone present in subsurface. Once the source (e.g., well) was taken out of service,
the outbreak quickly subsided. The fecal samples from infected patients showed the presence of
oocysts of cryptosporidium. Chlorination appeared not to have complete disinfection potential in
the outbreak.
The town of North Battleford, Saskatchewan, Canada, suffered a cryptosporidium outbreak
(Sterling et al., 2001). The mechanism for failure was the defect of the sand filtration units and
cryptosporidium oocysts being resistant to chlorine passed through the treatment plant to the dis-
tribution system. The outbreak occurred between late March and early May of 2001. Between 5800
and 7100 people were affected by the outbreak that included many visitors. Confirmed 275 cases
of cryptosporidiosis were found by May 30, 2001. Again, failure of filtration was attributed to the
outbreak. The source water was surface water and the parasites must have entered the river some
point upstream from the intake.
Another serious case of contamination of the water supply of Walkerton, Ontario, Canada, with
O157:H7 strain of E. coli led to the death of 7 people and sickening of 2500 people in year 2000
(Hrudey et al., 2002, 2003). Contaminated surface runoff entered the well possibly due to bypass
flow. The operators were held liable for misconduct.
The resort on the island of South Bass in Lake Erie, Ohio, witnessed an outbreak of gastroenteri-
tis in 2004 (Fong et al., 2007; O’Reilly et al., 2007) that exposed the visitors and the local residents.
Fong et al. (2007) concluded the following:
Massive groundwater contamination on the island was likely caused by transport of microbiological
contaminants from wastewater treatment facilities and septic tanks to the lake and the subsurface,
after extreme precipitation events in May–July 2004. This likely raised the water table, saturated the
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 431
subsurface, and along with very strong Lake Erie currents on 24 July, forced a surge in water levels and
rapid surface water–groundwater interchange throughout the island. Landsat images showed massive
influx of organic material and turbidity surrounding the island before the peak of the outbreak. These
combinations of factors and information can be used to examine vulnerabilities in other coastal sys-
tems. Both wastewater and drinking water issues are now being addressed by the Ohio Environmental
Protection Agency and the Ohio Department of Health.
Understanding the fate and transport of biocolloids in environmental systems is quite important
because they pose a serious water quality risk. Biocolloids may remain infectious and may survive
for long periods of time in natural environmental systems (Chrysikopoulos et al., 2010). In order to
estimate the potential health risk associated with aquifers contaminated by various biocolloids, the
prediction of biocolloid fate and transport is necessary.
Virus adsorption has some correlation with soil moisture; whereas its transport is faster when
water saturation in pores is high. It has been proved that virus adsorption on AWI in unsaturated
porous media is governed by temporal and spatial water saturation fluctuations. Virus adsorption is
stronger on AWIs than on liquid–solid interfaces. However, virus inactivation is enhanced under the
presence of AWIs. Released viruses in water or wastes can be adsorbed on the sand surface, clays,
bacterial cells, natural colloidal dispersions, and sludge particles. Adsorbed viruses survive for
longer time periods than viruses suspended in the aqueous phase. The prevailing mechanisms for
biocolloid adsorption onto clays are not well defined and they vary for varying kinds of biocolloids.
The electrostatic interactions and chemical properties such as ionic strength and pH are dominant
parameters for the adsorption of a virus on mineral surfaces (Chrysikopoulos et al., 2010).
5.7 CONCLUSIONS
Within the subsurface systems such as soils, groundwater aquifers, sediments, or fractured rocks,
colloidal particles/fines (particles with an average diameter between 10 −9 and 10 −10 m and carry
surface charge) are ubiquitous with a pronounced variability concerning morphology and chemical
composition. There are several classes of subsurface colloids—abiotic and biotic. These include
silicate clays, iron and aluminum oxides, mineral precipitates, humic materials, micro emulsions of
nonaqueous phase liquids, viruses, and bacteria. The mobilization and migration of these subsurface
colloidal particles take place under different physical and geochemical conditions. The contamina-
tion of groundwater and soil due to these migrating colloidal particles is well recognized and is
classified into two categories: (a) colloid-associated contaminant transport and (b) transport of bio-
colloids. This chapter has briefly reviewed these subsurface colloidal fines, classifications, sampling
methods, and characterization, and finally their role in groundwater contamination. Groundwater
contamination due to mobilization and migration of these colloidal fines and associated health
effects has also been presented here.
REFERENCES
Ahmed A, Neller N, Katouli M. J. Appl. Microbiol., 2005;98:910.
APHA. Standard Methods, 19th edn. American Public Association, Washington, DC, 1995.
Appel C, Ma LQ, Rhue RD, Kennelley E, Geoderma, 2003;113:77.
Arab D, Pourafshary P, Ayatollahi AH. Int. J. Environ. Sci. Technol., 2014;11:207.
Ardizzone S, Bianchi CL, Galassi CJ, Electroanal. Chem., 2000;490:48.
Atkinson RJ, Posner AM, Quirk JP, J. Phys. Chem., 1967;71:550.
Auset M, Keller AA, Water Resour. Res., 2004;40:3503.
Backhus DA, Ryan JN, Groher DM, MacFarlane JK, Gschwend PM, Ground Water, 1993;31:466.
Beckett R, Hart BT. The use of field-flow fractionation techniques to characterize aquatic particles, colloids and
macromolecules. In: Buffle J, Van Leeuwen HP (Eds.), Environmental Particles, Vol. 2. Lewis Publisher,
London, U.K., 1993, p. 165.
Bin G, Cao X, Dong Y, Luo Y, Ma LQ. Crit. Rev. Environ. Sci. Technol., 2011;41:336.
432 Handbook of Surface and Colloid Chemistry
Bitton G, Gerba CP. Groundwater Pollution Microbiology: The Emerging Issue. Wiley, New York, 1984, p. 713.
Borkovec MQ, Wu GD, Laggner P, Sticher H. Colloids Surf. A, 1993;73:65.
Bourikas K, Vakros J, Kordulis C, Lycourghiotis A. J. Phys. Chem. B, 2003;107:9441.
Bradford SA, Simunik J, Walker SL. Water Resour. Res., 2006;42:W12S12. doi: 10.1029/2005WR004805.
Bradford SA, Yates SR, Bettahar M, Simunek J. Water Resour. Res., 2002;38:1327.
Breeuwsma A, Lyklema J. Discuss. Faraday Soc., 1971;52:324.
Breeuwsma A, Lyklema J. J. Colloid Interface Sci., 1973;43:434.
Bridgeman SA, Robertson RMP, Sayed Q, Speed N, Andrews N, Hunter PR. Epidemiol. Infect.,
1995;115:555–566.
Bruins MR, Kapil S, Ochme FW. Ecotoxicol. Environ. Saf., 2000;47(2):105.
Buffle J, Leppard GG. Environ. Sci. Technol., 1995;29:2169.
Chibowski S, Wisniewska M. Colloids Surf. A, 2002;208:131.
Chin G, Flurry M. Colloids Surf. A, 2005;256:207–216.
Choi HS, Ashitate Y, Lee JH, Kim SH, Matsui A, Insin N, Bawendi MG, Semmler-Behnke M, Frangioni JV,
Tsuda A. Nat. Biotechnol., 2010;28:1300.
Chrysikopoulos CV, Masciopinto C, Mantia RL, Manariotis ID. Environ. Sci. Technol., 2010;44(3):971.
Chu Y, Jin Y, Flurry M, Yates MV. Water Resour. Res., 2001;37:253–263.
Colvin VL. Nat. Biotechnol., 2003;21(10):1166.
Cordelair J, Greil P. J. Colloid Interface Sci., 2003;265:359.
Cornell RM, Schwertmann U. The Iron Oxide, 1st edn. VCH Publishers, New York, 1998.
Crane SR, Moore JR. Water Air Soil Pollut., 1984;22(1):67–83.
Crawford RJ, Harding IH, Mainwaring DE. J. Colloid Interface Sci., 1996;181:561.
Czinany S. Subsurface colloids: Stability, sampling and transport under gravitational and centrifugal accelerations.
Academic dissertation, Department of Crop and Soil Sciences, Washington State University, Pullman, WA, 2004.
Davis JA, Leckie JO. J. Colloid Interface Sci., 1978;67:90.
Degueldre C, Missana T. Colloids Surf. A: Physicochem. Eng. Aspects, 2003;217:33.
De Lint WBS, Benes NE, Lyklema J, Bouwmeester HJM, van der Linde AJ, Wessling M. Langmuir,
2003;19:5861.
D’Antonio RG et al. Ann. Intern. Med., 1985;103:886–888.
Elimelech M, Nagai M, Ko C-H, Ryan J. Environ. Sci. Technol., 2000;34:2143.
Elimelech M, Ryan JN. The role of mineral colloids in the facilitated transport of contaminants in saturated
porous media. In: Huang PM, Bollag J-M, Senesi N (Eds.), Interactions between Soil Particles and
Microorganisms: Impact on the Terrestrial Ecosystem. John Wiley & Sons, New York, 2002.
Enfield CG, Bengtsson G. Groundwater, 1988;26:64.
Faria LAD, Trasatti S. J. Electroanal. Chem., 2003;554:355.
Faur-Brasquet C, Reddad Z, Kadirvelu K, Le Cloirec P. Appl. Surf. Sci., 2002;196:356.
Fein JB, Brady PV. Chem. Geol., 1995;121:11.
Flurry M, Qiu H. Vadose Zone J., 2008;7:682.
Fong T-T, Mansfield LS, Wilson DL, Schwab DJ, Molloy SL, and Joan B. Environ. Health Perspect.,
2007;115(6):856.
Franks GV, Meagher L. Colloids Surf. A, 2003;214:99.
Fuerstenau DW, Shibata J. Int. J. Miner. Process., 1999;57:205.
Gao B, Saiers JE, Ryan J. Water Resour. Res., 2006;42:W01410.
Gerba CP, Goyal SM. Pathogen Removal from Wastewater during Groundwater Recharge Artificial Recharge
of Groundwater. Butterworth Publishers, Boston, MA, 1985, p. 317.
Gonzalez G, Laskowski J. Electronal. Chem. Int. Electrochem., 1974;53:452.
Goyal SM, Amundson DA, Robinson RA, Gerba CP. J. Minn. Acad. Sci., 1989;55(1):58.
Grim RE. Clay Mineralogy. McGraw Hill, New York, Chapter 6, 1968.
Gschwend PM, Reynolds MD. J. Contam. Hydrol., 1987;1:309.
Guo LC, Zhang Y, Uhida N, Uematsu K. J. Eur. Ceram. Soc., 1997;17:345.
Gustafsson O, Gschwend M. Limnol. Oceanoger., 1997;42(3):519.
Hackley VA, Patton J, Lum LSH, Wasche RJ, Naito M, Abe H, Hotta Y, Pendse H. J. Dispers. Sci. Technol.,
2002;23:601.
Hassellov M, Kaegi R. Analysis and characterization of manufactured nanoparticles in aquatic environments.
In: Lead JR, Smith E. (Eds.), Environmental and Human Health Impacts of Nanotechnology. Wiley,
Chichester, U.K., 2009, p. 211.
Hassellov M, Readman JW, Ranville JF, Tiede K. Ecotoxicology, 2008;15(5):344.
Hennebert P, Avellan A, Yan J, Aguerre-Chariol O. Waste Manage., 2013;33(9):1870.
Subsurface Colloidal Fines, Behavior, Characterization, and Their Role 433
CONTENTS
6.1 Introduction........................................................................................................................... 437
6.2 Application of Activated Carbon in Phenolics Treatment..................................................... 437
6.2.1 Activated Carbon Adsorption.................................................................................... 437
6.2.2 Problems Faced in Activated Carbon Adsorption of Phenols—Oligomerization....... 438
6.2.3 Activated Carbon Fibers............................................................................................440
6.2.4 Effect of Functional Groups...................................................................................... 441
6.2.5 Pore Size Distribution and Its Effect on Oligomerization......................................... 443
6.2.6 Sources and Methods of Developing Activated Carbon............................................446
6.2.7 Tailoring of Activated Carbon...................................................................................448
6.3 Experimental Studies............................................................................................................. 450
6.3.1 Development of Novel Activated Carbon for Oligomerization Control.................... 450
6.3.2 Oligomerization Control of Single Solute Isotherm.................................................. 452
6.3.3 Oligomerization Control of Multisolute Adsorption Isotherm.................................. 455
References....................................................................................................................................... 458
6.1 INTRODUCTION
Activated carbons are widely used as an adsorbent in numerous applications and especially in the
phenolics treatment field, due to its excellent adsorption capacity and great flexibility. However, molec-
ular oxygen has been found to be a significant factor in the adsorption of phenolics onto activated
carbon (Grant and King 1990; Nakhla et al. 1994; Tessmer et al. 1997). Molecular oxygen was found
to be able to promote a significant increase in the adsorptive capacity of activated carbon at the cost of
reduction in the recovery of adsorbents. Studies have attributed this phenomenon to the oligomeriza-
tion of phenolics (oxidative coupling) on the surface of activated carbon (Vidic et al. 1993). Analyzing
the solvent extracted products by GC–MS has shown the presence of oligomerized phenols, which are
essentially irreversibly bound to carbon surface (Vidic et al. 1997). As a consequence, the regenera-
tion efficiency of activated carbon after the oligomerization of phenolic compounds is low. The cost of
adsorbents has been the major concern in activated carbon adsorption usage. In this chapter, the theory
as well as factors for controlling oligomerization is outlined. The description of development of novel
activated carbon for oligomerization control is also the main subject of this contribution.
437
438 Handbook of Surface and Colloid Chemistry
surface area and pore volume (Walker 1969). Nowadays, different shapes of activated carbon can be
fabricated based on different utilizing purposes.
Activated carbon is a porous adsorbent having high capacity for removing various small molecu-
lar weight organic compounds from water. The major liquid-phase application of activated carbon
is for removing traces of water organic compounds after other conventional treatment processes
in order to attain regulatory compliance. In the beginning, activated carbon adsorption was only
employed for taste and odor control. But with the great development of analytical technology, many
contaminants can be identified and quantified at very low concentration. Therefore, new water qual-
ity issues of trace compound contamination were arising. Especially synthetic organic compounds
(SOCs) are highly concerned because of their potential toxicity and carcinogenicity. Due to this rea-
son, activated carbon has been used for several decades to treat public water supplies for the removal
of SOCs. Activated carbon adsorption is considered to be the best currently available technology for
removing SOCs. In addition, due to ubiquity in the environment and risk to human health, volatile
organic compounds (VOCs) have received great attention in the field of environmental control. The
adsorption of VOCs onto porous adsorbents such as activated carbons has been suggested as an
innovative treatment technology.
Activated carbons are used extensively in industrial purification and chemical recovery opera-
tions because of their advantage of removing a wide range of organic substances with relatively low
cost and no hazardous by-products. The market for activated carbons is indeed vast. The global con-
sumption of activated carbon was 750,000 t in 2002. The estimated growth of worldwide demand
is 4%–5% per year, with higher growth rates of 5%–6% per year projected for the United States
between 2002 and 2005 (Zhang et al. 2004). Activated carbon is commonly used either in PAC
or granular activated carbon (GAC) form. GAC provides continuous removal and is most efficient
when organic contaminants are frequently present. It can be regenerated to recover the adsorption
capacity and is also a good medium for biological stabilization. PAC has the flexibility of being used
whenever it is needed. Therefore, it is best used for occasional taste and odor problems, episodic
contaminant spills, and seasonal high level of pesticides and herbicides. Activated carbon adsorp-
tion has not been used extensively in wastewater treatment; yet, demands for a better quality of
treated wastewater effluent, including toxicity reduction, have led to an intensive examination and
use of activated carbon adsorption. In wastewater treatment, activated carbon adsorption is used
to remove dissolved organic matter (DOM), the remaining portion of normal biological treatment.
OH H 3C
OH OH HO O
OH OH
HO CH3
+
+
OH
OH H2
C O
O
(1) (2)
OH OH OH
H 3C
CH2 O CH2 O
H2C
OH OH OH OH O O
CH3 CH3 O
+ + + H2C H2C
CH3 O CH2
CH3 CH3
OH HO OH
(3)
the reaction of phenol oxidation included usually dimers and higher polymers (Figure 6.1). The
abundance of the polymer decreased with molecular weight. More importantly, they found the car-
bon–oxygen linkages are less frequently formed. No minerals in the carbon caused the oxidative
coupling reaction. In addition, temperature is always an important factor affecting activated carbon
adsorption. The effect of temperature on the dissolved oxygen-induced oligomerization of phenolics
was also investigated in their study. They found that higher temperature enhances the polymeriza-
tion reaction and hence increases the adsorption capacity.
The oligomerization of phenolics depends strongly not only on the presence of molecular oxygen
but also on the amount of oxygen (Vidic and Suidan 1991). The variation of molecular oxygen
amount can lead to significant differences of adsorption capacity of activated carbon for pheno-
lics. The presence of molecular oxygen increases the adsorption capacity of activated carbon for
phenolics by imparting oligomerization reaction. Then the adsorption isotherms can be denoted
into two types: anoxic (absence of molecular oxygen) isotherms and oxic (presence of molecular
oxygen) isotherms. These suggest that physical adsorption equilibrium on the activated carbon can
be achieved by conducting anoxic isotherms with the exclusion of oxygen from the experimental
container. In addition, the adsorption capacity including chemisorption like oligomerization can be
obtained by conducting oxic isotherms. Therefore, as compared to anoxic isotherm, any increasing
amount of adsorbed phenolics predicted under the oxic condition can be attributed to oligomeriza-
tion. It has been shown that the amount of dissolved oxygen (DO) consumed during the adsorption
of phenolic adsorbates is linearly proportional to the amount of irreversibly adsorbed compound
(Vidic et al. 1993).
Reviews on studies pertaining to oxygen-containing functional groups on the activated carbon
seems to indicate that only basic surface functional groups have little influence on the extent of
oligomerization (Tessmer et al. 1997). The presence of basic surface functional groups can slightly
promote phenolic compounds adsorption under oxic conditions by increasing its effectiveness in pro-
moting adsorption via oxidative coupling reactions. Tessmer et al. (1997) investigated the modifica-
tion of activated carbon surface functional groups with outgassing thermal treatment to enhance the
440 Handbook of Surface and Colloid Chemistry
catalytic ability of activated carbon. It was found that outgassing at 900°C can e ffectively eliminate
the acidic functional groups and encourage the formation of oxygen basic groups. But outgassing
at higher temperature (1200°C) could remove all oxygen complexes. In addition, only oxygenated
basic groups can promote the oligomerization on the surface of activated carbon. Hence, outgas-
sing at proper temperature (900°C) can promote the oligomerization by changing the amount and
composition of oxygen functional groups.
As the oligomerization of phenolics can only occur in the presence of oxygen, researchers inves-
tigated the mechanism of interactions between oxygen and carbons. Zawadzki and Biniak (1988)
suggested that super-oxo ions are formed in a first step of oxygen adsorption by the carbon surface.
These ions undergo further rearrangement on the carbon surface, forming a number of surface
oxygen compounds. This hypothesis was further confirmed by a study conducted by Terzyk (2003)
where they found the oligomerization extent was having direct relationship with the concentration of
surface lactones at neutral pH. The larger is the concentration of lactones, the less are the oligomer-
ization reactions. They postulated that as lactones have a considerable influence on the temperature
dependence of the carbon resistance, it weakens the work function of the electrons. Consequently,
lactones can decrease oligomerization by creating strong bonds with adsorbed molecules.
effects and the low inorganic impurities result in minimal effects on pore development during the
activation process (Parker 1995). The advantage of ACFs is that it is very useful in studying the
role of functional groups of adsorbate by eliminating the interrupting factors of adsorbent func-
tional groups.
For the adsorption application, ACFs offer the fabrication flexibility that GAC does not have,
and thus they are easier to be made into filter media or devices. ACFs can be made in various forms
such as fabrics, nonwovens, paper, and composites, which make them suitable for handling in spe-
cial devices (Suzuki 1994). GAC are widely used in a variety of applications, but a large volume of
adsorption devices are always needed in applications due to the relatively slow adsorption rate of
GAC. Meanwhile, although the size of PAC is smaller than GAC, PAC may result in a strong flow
resistance because of potential compact under flow (Shmidt et al. 1997). ACFs can effectively solve
this issue by not only significantly improving the adsorption rate of adsorbates but also have good
adsorption capacity. This is probably because of the unique porous structure of ACFs. For ACFs,
the surface pores can lead directly into the center of the ACF. But for GAC and PAC, the penetration
of adsorbate into the center of PAC and GAC is controlled by diffusion through micropore layers
that interconnect individual macropores. In water purification systems, ACFs have advantages of
less possibility to produce ash, less pressure drop, and a higher water flow rate. In air purification
systems, ACFs show faster adsorption and desorption rate (Shmidt et al. 1997).
10
2 Methylphenol
2-Chlorophenol
2-Nitrophenol
Phenol
qe (mmol/g)
2-Ethylphenol
10
qe (mmol/g)
10
qe (mmol/g)
FIGURE 6.2 Adsorption of phenolics on (a) F400, (b) ACC-10, and (c) ACC-15. (Reprinted from Lu, Q. and
Sorial, G.A., J. Hazard. Mater., 148, 436, 2007. With permission.)
The trend was similar to F400 with the exception of 2-methylphenol and 2-ethylphenol. Both of
these adsorbates are three dimensional, however, 2-ethylphenol (6.06 Å * 5.62 Å * 1.61 Å) is larger
than 2-methylphenol (5.68 Å * 4.12 Å * 1.63 Å). The limited PSD of ACC-10 as compared to F400
restricted the adsorption of 2-ethylphenol due to its larger size as compared to 2-methylphenol. In
the case of ACC-15 under anoxic conditions, a similar adsorption trend was observed but more
differences in adsorption capacities were also displayed as compared to ACC-10. This could be
attributed to the larger critical pore diameter of ACC-15 as compared to ACC-10.
Activated Carbon Adsorption in Water Treatment 443
More importantly, it can be seen that the oxic adsorptive capacity of phenolic was higher than
that of anoxic ones on F400 due to oligomerization of phenolics on the activated carbon surface
(Figure 6.2). It is known that the first step in this oligomerization reaction is forming phenoxy
radical by losing one proton; the phenoxy radical is highly reactive and can participate in oligomer
formation through carbon–oxygen or carbon–carbon bonding. However, they found that the pres-
ence of an electron attracting group, like the nitro and the chloro group, inhibited the parent phenol
molecule to lose the proton, that is, hydrogen; thus the first step of oligomerization was inhibited to
some degree. In contrast, they reported the presence of an electron-donating group, like the methyl
and the ethyl group decreased the COP and thus aided the parent phenol molecule to form phenoxy
radical.
Figure 6.2 shows that the sequence of phenolic oxic adsorptive capacity was in the same
order as the anoxic isotherm while the relative difference under oxic conditions was less. This
can be explained by the existence of electron withdrawing groups, like chloro and nitro groups.
2-Chlorophenol and 2-nitrophenol are more adsorbable than other compounds under anoxic con-
ditions, while these existed chloro or nitro groups also make 2-chlorophenol and 2-nitrophenol
less oligomerized than the others, which diminished the differences between these compounds.
In case of oxic adsorption on ACC-10, the sequence of adsorptive capacity was found to be phenol
> 2-chlorophenol > 2-nitrophenol ∼ 2-methylphenol > 2-ethylphenol, which is different from the
order shown under anoxic conditions. Obviously, the limited pore diameter is more accessible for
the 2D molecules (phenol) than the 3D ones (2-methylphenol and 2-ethylphenol). Thus, the large
molecular size of phenolics made them less oligomerized on ACC-10.
In reported adsorption studies, the activated carbon PSD was shown to be an important param-
eter affecting the removal of phenolics. Hsieh and Teng (2000) described the adsorption of phenol
onto activated carbons developed with different activation levels. They prepared activated carbons
with different porosities by carbonizing bituminous coal to different extents of burn-off. Larger
surface area and higher surface area were obtained by increasing the extent of carbon burn-off.
However, the burn-off level only showed a little influence on the average pore diameter of activated
carbons. These indicated that pore deepening, rather than widening, occurred predominantly upon
activation in CO2 atmosphere. Meanwhile, they found that mesopores facilitate the adsorption of
the adsorbates in the inner and narrow micropores since the adsorptive capacity was an increasing
function of mesopore volume when surface area and micropore volumes were similar. In addition,
Juang et al. (2001) reported that the ratio of monolayer capacity and volume of micropores for phe-
nol always decreased with increasing micropore volume, indicating that the adsorption of phenols
is not completely restricted to occur within the micropores.
The PSD of activated carbon determines the fraction of its structure that can be accessed by
a molecule of given size and shape. For a given size pore, the size of adsorbates relative to the
pore size of the adsorbents are important in controlling the adsorptive competitive mechanism. A
good understanding of the impact of PSD on oligomerization is required as a basis for developing
activated carbon for oligomerization control. However, very few studies have been conducted to
study the impact of PSD on oligomerization. Therefore, adsorption isotherms under the control of
oxygen can potentially constitute a significant advance in the understanding of the impact of PSD
on oligomerization. In addition, activated carbon fiber (ACF) is very suitable for elucidating the
impact of PSD on oligomerization due to its perfect surface properties. Firstly, ACFs have unique
PSD and all the micropores are directly on the carbon surface (Ryu et al. 2000; Shen et al. 2004).
Secondly, elemental analysis of ACFs indicated that they had <5% surface oxides and contain very
low amounts of inorganic impurities (Pelekani and Snoeyink 1999, 2000). The low concentration of
surface oxides results in limited surface chemistry effects and the low inorganic impurities results
in minimal effects on pore development during the activation process (Pelekani and Snoeyink
1999). These parameters plus the controlled pore structure during activation are a key advantage of
ACFs over granular activated carbon (GAC) and make them ideal for targeting the effects of PSD
on adsorption phenomena (Parker 1995).
Lu and Sorial (2004) conducted an adsorption isotherm of phenolics on two types of ACFs with
different pore sizes and GAC F400 to investigate the impact of PSD of activated carbon on the
adsorption of phenolics (Figures 6.3 and 6.4). In order to study the impact of molecular oxygen on
the adsorptive capacity, two types of isotherms were conducted. One was conducted in the pres-
ence of molecular oxygen (oxic isotherm) and the other in the absence of molecular oxygen (anoxic
isotherm).
The results clearly indicate that there is a significant difference in adsorptive capacity of
2-methylphenol between oxic and anoxic conditions for the three adsorbents studied (Figure 6.3).
The adsorptive capacity of 2-methylphenol under oxic conditions is higher than that under anoxic
conditions. The difference in the adsorptive capacity between oxic and anoxic conditions for ACC-
10 is smaller as compared to the other adsorbents. A similar behavior is seen for 2-ethylphenol
(Figure 6.4). In addition, they compared the adsorptive capacity of phenolics on the three adsor-
bents under oxic and anoxic conditions at two equilibrium liquid-phase concentrations of 0.5 and
5 mM (Table 6.1). They reported the increase in adsorptive capacity under oxic conditions is in
the sequence of F400 > ACC-15 > ACC-10 for all the two adsorbates considered. They found
that F400 with a wide range of PSD cannot hamper the oligomerization of phenolics. But ACC-
10 with a critical pore diameter of 8.0 Å and narrow PSD (microporosity above 90%) was more
effective in hampering the oligomerization of phenolics. It is interesting to note that the molecu-
lar size of the adsorbate played another significant role in the increase of adsorptive c apacity.
The second widest dimension of 2-methylphenol molecules is 4.17 Å. Therefore, it is possible
Activated Carbon Adsorption in Water Treatment 445
1000 1000
Anoxic
Oxic
Myers equation
qe (mg/g)
qe (mg/g)
100 100
10 100 1000 10 100 1000
(a) Ce (mg/L) (b) Ce (mg/L)
1000
qe (mg/g)
100
10 100 1000
(c) Ce (mg/L)
FIGURE 6.3 Adsorption isotherms of o-cresol. (a) F400, (b) ACC-10, and (c) ACC-15. (Reprinted from
Chemosphere, 55, Lu, Q. and Sorial, G.A., Adsorption of phenolics on activated carbon—Impact of pore size
and molecular oxygen, 671–679, Copyright 2002, with permission from Elsevier.)
for 2-methylphenol to form dimers, while the formation of trimers is nearly impossible. In case
of 2-ethylphenol, the second widest dimension is 5.62 Å, thus it is impossible for 2-ethylphenol
to get oligomerized. This is further confirmed in Table 6.1 where 2-ethylphenol did not show
adsorption capacity gain under oxic c onditions. In case of ACC-15, forming dimers and trimers
of 2-methylphenol is possible, while tetramers or pentamers are very difficult or even impossible.
At the same time, they also found that the oligomerization extent is directly related to the critical
oxidation potential (COP) value of phenolics. 2-Methylphenol is more amenable to oligomeriza-
tion than 2-ethylphenol due to its lower COP value (1.04 V). In their study, the effective control
of oligomerization on ACF is further confirmed by the high recovery efficiency of ACF after oxic
adsorption isotherms. F400 showed the lowest recovery efficiency as compared to ACC-10 and
ACC-15, and ACC-15 was lower than ACC-10. ACC-10 showed higher adsorbate extraction effi-
ciency, above 85%, than the other adsorbents. More importantly, the adsorptive capacity of F400
was lower than the two ACFs for both adsorption conditions (oxic and anoxic). Thus, for better
use of activated carbon, it is mandatory to use adsorbents with high adsorptive capacity and also
high regeneration efficiency. Therefore, activated carbon with similar pore structure as ACC-10
(having their pore volume mostly distributed in micropores and low critical pore diameter) needs
to be developed and used for hampering oligomerization in order to make activated carbon more
cost effective.
446 Handbook of Surface and Colloid Chemistry
1000 1000
Anoxic
Oxic
Myers equation
qe (mg/L)
qe (mg/L)
100 100
10 100 1000 10 100 1000
(a) Ce (mg/L) (b) Ce (mg/L)
1000
qe (mg/L)
100
10 100 1000
(c) Ce (mg/L)
FIGURE 6.4 Adsorption isotherms of 2-ethylphenol. (a) F400, (b) ACC-10, and (c) ACC-15. (Reprinted from
Chemosphere, 55, Lu, Q. and Sorial, G.A., Adsorption of phenolics on activated carbon—Impact of pore size
and molecular oxygen, 671–679, Copyright 2002, with permission from Elsevier.)
TABLE 6.1
Comparison of 2-Methylphenol and 2-Ethylphenol
F400 ACC-10 ACC-15
Ce Anoxic qe Oxic qe 𝝙 qe Anoxic qe Oxic qe 𝝙 qe Anoxic qe Oxic qe 𝝙 qe
Adsorbate (mM) (mmol/g) (mmol/g) (%) (mmol/g) (mmol/g) (%) (mmol/g) (mmol/g) (%)
2-Methylphenol 0.5 1.61 2.45 52.3 2.00 2.32 15.7 2.54 3.23 27.0
5 2.17 2.87 32.3 2.42 2.62 8.7 3.20 3.72 16.1
2-Ethylphenol 0.5 1.73 2.44 40.8 2.04 1.93 −5.5 2.90 3.41 17.5
5 2.23 2.73 22.6 2.28 2.28 0.4 3.49 3.83 9.6
Source: Data from Lu, Q. and Sorial, G.A., Chemosphere, 55, 671, 2002.
methods used (Guo and Lua 2000). It is found that the carbon content of char was s ignificantly higher
after carbonization, and the oxygen and hydrogen were lower. In addition, the pore characteristics of
activated carbons are important in determining the particular application of the carbon.
Wood is one of the most important materials for preparing activated carbons, which have par-
ticular porous characteristics and are appropriate for the adsorption of solutes in liquid phase. For
example, quercus agrifolia and eucalyptus woods activated with CO2 yielded a surface area of 1201
and 1190 m2/g, respectively, and a fraction of micropore volume of 0.76 and 0.50. In addition,
rubber wood sawdust and wood flour activated with 1.5 and 2.4 times of H3PO4 obtained an area
of 1496 and 1780 m2/g, respectively, and a fraction of micropore volume of 0.69 and 0.66. These
carbons show higher ratios of mass transfer voids (Wu et al. 2005). The general process to pro-
duce activated carbon is based on carbonizing and activating the original carbonaceous m aterial.
Activated carbons may be produced through physical or chemical activation. For example, physical
activation via pyrolysis to produce a char followed by steam gasification, alternatively, the bio-
mass may be treated with a chemical, for example, zinc chloride followed by thermal treatment via
pyrolysis to produce an activated carbon. Physical activation involves carbonization of the precursor
under inert atmosphere, followed by gasification of the char under oxidizing agents such as steam or
carbon dioxide. In industrial practice, physical activation is carried out most frequently by burning
off some of the raw carbon in an oxidizing environment to create micropores. The usual commercial
choices of activation gas are steam, CO2, air, or their mixtures. Activation normally takes place at
temperatures between 700°C and 1000°C in steam and CO2, and lower temperature in air (Zhang
et al. 2004). Chemical activation is one of the less commonly commercial methods. In this process,
the starting materials are impregnated with a solution of dehydrating agent (such as zinc chloride
[ZnCl2], phosphoric acid [H3PO4], sulfuric acid [H2SO4], or potassium hydroxide [KOH]) to retard
the formation of tars during the carbonization process. They are then washed, dried, and carbonized
in an inert atmosphere and finally activated with CO2 or steam to produce the final activated carbon
(Yamashita and Ouchi 1982; Marsh et al. 1984; Molinasabio et al. 1995). Chemical activation has
intrinsic advantages over physical activation, which includes lower temperature, single step, higher
product yields, higher surface areas, and better microporosity distribution control (Lozano-Castello
et al. 2007). A large number of agricultural by-products such as peach stone, sugar cane bagasse,
coconut shell, almond shell, and hazelnut shell have been successfully converted into activated car-
bons by chemical activation on a laboratory scale. Caturla et al. (1991) studied the preparation of
activated carbons from peach stones with ZnCl2 impregnation to determine the effects of various
parameters such as starting material size, impregnation method, and activation temperature on the
BET surface area, porosity, and density of the resulting carbons. Girgis et al. (1994) prepared acti-
vated carbons from sugar cane bagasse impregnated with 50% inorganic acids. The effectiveness
of such impregnation was in the following order: H3PO4 > H2SO4 > HCl > HNO3. They also found
out that activated carbons prepared at low temperatures were essentially microporous, but at higher
temperatures, products of higher surface areas with developed mesoporosity but low microporos-
ity were obtained. Laine and Calafat prepared a series of activated carbons from coconut shells,
impregnated with various potassium solutions (KCl, KNO3, KOH, K2CO3, and K3PO4) at differ-
ent concentrations (Laine et al. 1989; Laine and Calafat 1991). It was found that all the potassium
compounds studied except KCl were effective catalysts for activation with CO2. Balci et al. (1994)
used NH4Cl impregnated and untreated almond and hazelnut shells for the preparation of activated
carbons. Experimental results showed that treatment with NH4Cl increased the total surface area
and improved adsorptive capacity. Further, Hussein et al. (1996) prepared activated carbons from oil
palm trunk chips by impregnation with ZnCl2. The BET surface area and the porosity of the result-
ing activated carbons were found to be related to the concentrations of the impregnating solutions. It
entails the addition of materials such as zinc salts or phosphoric acid to the carbon precursors could
generate secondary environmental pollution during disposal.
The development of micropores and mesopores for the carbon is crucial because the adsorption
of low molecular weight constituents from gas or liquid streams is responsible for the existence of
448 Handbook of Surface and Colloid Chemistry
these two types of pores. Microporous carbons are generally applied in gas-phase adsorption due
to the small gas molecules, whereas mesoporous carbons are often used in liquid-phase adsorption.
Recent progress of industrial technologies provides new application fields for activated carbons
in super-capacitors and, at the same time, requires the carbons having a desired pore structure.
An increased activation temperature and/or time will increase the surface area and the fraction of
mesopore volume simultaneously. Hu and Vansant (1995) have shown similar trends for the car-
bons prepared from eucalyptus wood and activated in the atmosphere containing CO2 and a small
amount of O2. However, when carbons were activated with KOH, ZnCl2, or H3PO4, the increase of
such chemical doses did not proportionally enhance the fraction of mesopore volume. Accordingly,
the fraction of mesopore volume was usually <0.2, while sometimes it increased up to 0.5. The
adsorption energetic heterogeneity commonly observed on activated carbon is induced by the struc-
tural heterogeneity, which can be characterized by the distribution of the slit-shaped micropores
(MPSD). This MPSD is the intrinsic property of the activated carbon and dictates the adsorption
equilibria and kinetics of different adsorbate molecules on the activated carbon through the adsor-
bate–pore interaction (Wang and Do 1999).
It has long been recognized that physicochemical characteristics affect the performance of acti-
vated carbon. Several research groups have studied thermal treatment, apart from other techniques, to
enhance the activated carbon properties and improve their sorption capacity for specific c ontaminants.
For instance, Kaneko et al. (1989) tailored activated carbon at high temperature by hydrogen. They
concluded that this protocol could remove acidic functional groups from the carbon’s surface and
enhance the sorption capacity for organic compounds such as benzene. Economy et al. (1996) reported
that ACF treated by NH3 at elevated temperature can demonstrate basic characteristics. These ther-
mally treated carbon fibers showed higher sorption capacity for acidic substances like SO2 and HCl.
Mangun et al. (2001) treated activated carbon with ammonia at higher temperature above 400°C. It
was found that a higher temperature and a longer treatment time resulted in an increase of surface
area and pore size. They also proposed that the nature of the nitrogen-containing surface groups intro-
duced by this method differed with the treatment temperature. Ammonia-treated samples presented
higher HCl sorption capacity. In addition, other researchers have used the chemical vapor deposition
of hydrocarbons to modify the physical and chemical properties of activated carbons. Nowack et al.
(2004) observed improved 2-methyl isoborneol (MIB) sorption when activated carbon was thermally
treated with steam or methane plus steam. Less favorable improvements were achieved with nitrogen
or hydrogen. For instance, when GAC was treated with nitrogen between 450°C and 850°C the sorption
capacity slightly increased; however, when the modified carbon sample was subsequently exposed to
air at ambient temperature it lost its enhanced MIB adsorption capacity and the authors attributed this
to the re-incorporation of oxygen. Hydrogen-treated lignite GAC at 1025°C was tested in mini-column
experiments and the results showed that this carbon could remove MIB to below detection about 2.5
times more bed volumes longer than the commercial lignite GAC HD4000 could. Conversely, tailored
carbon made by applying the methane and steam method at 1000°C presented the highest MIB sorp-
tion capacity. This modified carbon sample (13% mass gain above the pyrolysed mass, then 13% net
mass loss for the final product) processed about 3.6 times more water than did HD4000 at initial MIB
breakthrough. They also found MIB adsorption mainly occurred in the pore size range of 5.5–63 Å.
The steam and methane-plus-steam processes have potential to produce a carbon product with high
sorption capacity for MIB and other similar organic compounds such as geosmin.
Various methods have been already proposed in order to get a needed porosity by direct elaboration,
by pore size reduction, or by pore size widening. The main limitation of those techniques leads in the
competition between the simultaneous internal diffusion (heat and mass transfer) and oxidative reac-
tions. This is usually responsible for a radical heterogeneity in textural properties and consequently to
a wide PSD. Many researchers have proposed to overcome this limitation by means of the separation of
the diffusional and the reaction phenomena by doing the former one at a temperature level at which the
second one is ineffective, and the second one at a higher temperature level but under inert atmosphere.
Therefore, at the particle scale and during the first step, the gaseous reagent is uniformly chemisorbed
at the available intra-particular surface of the material. During the second step, the chemical func-
tional groups previously formed are decomposed withdrawing carbon atoms from the material. In this
way, there is no preferential burn-off at the pore mouth but preferential initiation of new micropores.
Moreover, this technique presents the advantage of allowing the applicability of a usually unusable
activating agent like molecular oxygen, which is known to react with carbon with an exothermic effect
leading to combustion of the material. The proposed cycling protocol is based on air as oxidative
reagent, which presents an obvious economical interest. The carbon surface concerned by the sorption
is heterogeneous; it consists of the faces of graphene sheets and of edges of such layers. The edge sites
are much more reactive than the atoms in the interior of the graphene sheets, and chemisorbed foreign
elements as oxygen are predominantly located on the edges at the twist and tilt crystallites boundaries.
Oxygen can be bound in the form of various functional groups that are similar to those known from
organic chemistry. On the other hand, it is well known that essentially all oxygen-containing groups
are removed during high temperature treatment in a nitrogen atmosphere. The gasification step con-
sists in the decomposition of the carbon–oxygen complexes into CO and CO2 and new freshly formed
carbon sites that might be viewed as an unsaturated carbon at the edge of the carbon matrix.
450 Handbook of Surface and Colloid Chemistry
TABLE 6.2
Porous Properties of GAC F400 and Activated Carbon by Using KOH as the
Activating Agent
Activation SBETa Vmicrob Vmeso + Vmacroc VTotald Micropore
Sample Parameters (m2/g) (cm3/g) (cm3/g) (cm3/g) Percentagee (%)
KOH/coal ratio
F400 — 993 0.373 0.242 0.615 60.7
BC-11-700-1-400 1:1 354 0.162 0.064 0.226 71.6
BC-21-700-1-400 2:1 971 0.405 0.154 0.559 72.4
BC-31-700-1-400 3:1 1158 0.434 0.235 0.669 64.9
BC-41-700-1-400 4:1 1247 0.389 0.348 0.737 52.8
BC-51-700-1-400 5:1 920 0.295 0.242 0.537 55.0
Temperature
BC-21-600-1-400 600 845 0.381 0.104 0.485 78.4
BC-21-650-1-400 650 1128 0.507 0.144 0.651 77.9
BC-21-700-1-400 700 971 0.405 0.154 0.559 72.4
BC-21-750-1-400 750 1335 0.529 0.236 0.765 69.2
BC-21-800-1-400 800 1451 0.516 0.319 0.835 61.8
BC-21-850-1-400 850 1285 0.329 0.430 0.759 43.4
Time
BC-21-650-0.5-400 0.5 1085 0.499 0.122 0.621 80.3
BC-21-650-1-400 1 1128 0.507 0.144 0.651 77.9
BC-21-650-2-400 2 1476 0.684 0.157 0.841 81.3
BC-21-650-3-400 3 1107 0.499 0.139 0.638 78.3
BC-21-650-4-400 4 903 0.415 0.109 0.524 79.2
N2 flow rate
BC-21-650-2-100 100 1203 0.534 0.158 0.692 77.1
BC-21-650-2-200 200 1472 0.676 0.179 0.855 79.0
BC-21-650-2-400 400 1476 0.683 0.158 0.841 81.3
BC-21-650-2-600 600 1091 0.480 0.153 0.633 75.8
Source: Modified from Yan, L. and Sorial, G.A., J. Hazard. Mater., 197, 311, 2011.
a S
BET is BET surface area.
b V
micro is the volume of micropore.
c V
meso + Vmacro = VTotal − Vmicro.
d V
Total is the total pore volume.
e Micropore percentage = V
micro/VTotal 100%.
*
samples was larger than F400 (the commercial activated carbon). However, the sample that was
made at a temperature below 650°C has a worse texture structure, resulting in much lower total BET
surface area. The authors attributed this to the pores in the sample as they are not interconnected
because the volatile components need to be released from the uniform structure at higher activa-
tion temperature. In addition, volatile components were released much more at higher activation
temperature, which cause the development of bigger pores in the activated carbon. On the other
hand, the release of volatile components made it easy for the activation agent to penetrate into the
coal particle. Consequently, the mesopore and macropore volume increased at higher temperatures.
It is possible that during activation, most reactive carbon atoms were eliminated after the activa-
tion process, while the carbon microporosity decreased. Therefore, the activation temperature is
a very important parameter in developing activated carbon with a suitable structure required for
452 Handbook of Surface and Colloid Chemistry
controlling oligomerization. Activated samples with a high level of microporosity can be obtained
at 650°C (BC-21-650-1-400, see Table 6.2).
In order to examine the best activation time on the activated carbon, this parameter was varied from
0.5 to 4 h while keeping the other parameters constant. No significant difference of microporosity was
found in activated carbon samples, while considerable BET surface area difference could be found
between each other (Table 6.2). All the activated carbon samples are mainly microporous (around 80%
microporosity). Carbon sample with 2 h activation was found to have highest micropore volume, high-
est micropore percentage, and highest BET surface area. Therefore, an activation time of 2 h could be
considered to be optimum for the preparation of a good activated carbon for oligomerization control.
In the activation process, nitrogen was used as a protection gas in this study. To select the best
nitrogen flow rate, the nitrogen flow rate was varied in the range 100–600 mL/min with the values of
the other parameters kept constant (KOH/coal ratio of 2:1, activation temperature of 650°C, and acti-
vation time of 2 h). The authors found increasing N2 flow rates favored the development of total BET
surface area until a flow rate of 400 mL/min was reached. It can be seen that a further increase in the
N2 flow rate led to a decrease in the surface area (Table 6.2). The highest micropore percentage was
obtained with a nitrogen flow rate of 400 mL/min. This finding is consistent with others reported in the
literature (Jiang and Zhao 2004; Jimenez et al. 2009). This pore opening behavior could be explained
on the basis that the gaseous products (containing CO2, CO, H2O, etc.) were obtained during the activa-
tion. Previous studies reported that the removal of CO, CO2, and H2 would favor the activation process,
while the removal of H2O formed during the pyrolysis reaction work against the activation. In addition,
a nitrogen flow rate too high could result in removing a large amount of H2O from the reaction site
and this would be detrimental to the porous properties of carbon, while a nitrogen flow rate too low
would probably not remove sufficient amounts of CO, CO2, which would also be in detrimental to the
activation. Therefore, a balance between these two removal processes must be reached under a medium
nitrogen flow rate (400 mL/min). Based on these results, it can be seen that the BC-21-650-2-400 sam-
ple, which was prepared with activation condition (KOH/coal = 2/1, activation temperature = 650°C,
activation time = 2 h, and nitrogen flow rate = 400 mL/min), has the best potential for controlling
oligomerization of phenolic compounds. This carbon was referred to as BC-21.
In the study, total acidic and basic groups on carbon surfaces was also performed for the selected
BC-21 and F400 using NaOH or HCl uptake methods. Based on the titration test, the BC-21 was
found to be an acidic carbon where total acidity and total basicity are 0.54 and 0 meq/g, respectively.
In contrast, the total acidity and total basicity for F400 is 0.15 and 0.475 meq/g. Batch equilibrium
method was used for determining the pHpzc of BC-21 selected and F400. The pHpzc was deter-
mined as the pH of the water sample that did not change after contact with the samples. The values
of pHpzc for BC-21 and F400 are 3.4 and 7.1, respectively.
q
Ce = e ∗ exp( Kqep ) (6.1)
H
where
Ce is the liquid-phase concentrations, mM
qe is surface loading, mmol/g
H, K, and p are regression parameters
Activated Carbon Adsorption in Water Treatment 453
10
Oxic
Anoxic
Myers equation
qe (mmol/g)
1
0.1 1 10 0.1 1 10
(a) Ce (mM) (b) Ce (mM)
FIGURE 6.5 Single solute adsorption isotherm of phenol. (a) BC-21 and (b) F400. (Reprinted from J. Hazard.
Mater., 197, Yan, L. and Sorial, G.A., Chemical activation of bituminous coal for hampering oligomerization
of organic contaminants, pp. 311–319, Copyright 2011, with permission from Elsevier.)
10
Oxic
Anoxic
Myers equation
qe (mmol/g)
1
0.1 1 10 0.1 1 10
(a) Ce (mM) (b) Ce (mM)
FIGURE 6.6 Single solute adsorption isotherm of 2-methylphenol. (a) BC-21 and (b) F400. (Reprinted from
J. Hazard. Mater., 197, Yan, L. and Sorial, G.A., Chemical activation of bituminous coal for hampering
oligomerization of organic contaminants, pp. 311–319, Copyright 2011, with permission from Elsevier.)
10
Oxic
Anoxic
Myers equation
qe (mmol/g)
1
0.1 1 10 0.1 1 10
(a) Ce (mM) (b) Ce (mM)
FIGURE 6.7 Single solute adsorption isotherm of 2-ethylphenol. (a) BC-21 and (b) F400. (Reprinted from
J. Hazard. Mater., 197, Yan, L. and Sorial, G.A., Chemical activation of bituminous coal for hampering
oligomerization of organic contaminants, pp. 311–319, Copyright 2011, with permission from Elsevier.)
454 Handbook of Surface and Colloid Chemistry
At low concentration, the Myers equation becomes a linear equation (6.2), that is, following Henry’s
law equation,
dCe 1
= ; qe → 0 (6.2)
dqe H
This criterion is very important when predicting multicomponent adsorption systems by the ideal
adsorption theory (IAST) because of the integration limits in the Gibbs adsorption equation.
Meanwhile, qe (mmol/g) was calculated from mass balance equation (6.3)
(C − Ce )V
qe = O (6.3)
M
where
CO (mM) is initial liquid-phase concentration
M (g) is the mass of adsorbent
V (L) is the adsorbate volume
The adsorbate volume, V, was 0.125 L for the anoxic isotherms and 0.1 L for the oxic isotherms.
Figure 6.5 clearly indicates that the adsorptive capacity of F400 for phenol under oxic condi-
tions increased significantly as compared to anoxic data due to the presence of molecular oxygen.
However, the presence of molecular oxygen in the test environment did not have a tremendous
impact on the adsorptive capacity of BC-21 for phenol. A similar behavior is seen for 2-methylphe-
nol and 2-ethylphenol (see Figures 6.6 and 6.7). Furthermore, the difference in the 2-methylphenol
adsorptive capacity obtained for F400 under oxic and anoxic conditions is more significant than
that for phenol and 2-ethylphenol. However, for BC-21, the adsorption capacity under both oxic and
anoxic conditions is nearly identical. The major differences only occurred at low Ce values, that is,
at high carbon dosage.
In this experiment by Yan and Sorial, anoxic and oxic adsorptive capacity of phenol,
2-methylphenol, and 2-ethylphenol on the two adsorbents were compared at two equilibrium con-
centrations of 0.5 and 5 mM (Table 6.3). These two limits were selected in order to represent the low
and high carbon dosage. It is clear that the difference between anoxic and oxic condition achieved
with F400 is highly significant as compared to BC-21. In addition, they found 2-methylphenol is
easier to have oligomerization phenomenon in comparison to phenol and 2-ethylphenol because the
difference for 2-methylphenol is larger than 2-ethylphenol and phenol. This distinct performance
can be attributed to the differences of critical oxidation potential (COP) of these phenolic com-
pounds. The COP for 2-methylphenol is 1.040 V while that for phenol and 2-ethylphenol is 1.089
and 1.080 V. As shown in Table 6.3, the oxic adsorptive capacity of phenol on F400 compared to the
anoxic capacity increased from 32.2% to 32.7% for an equilibrium concentration of 5 and 0.5 mM,
respectively. Furthermore, in the case of 2-ethylphenol, for the equilibrium adsorbate concentra-
tions of 5 and 0.5 mM, the adsorptive capacities of F400 exhibited under oxic condition are 25.1%
and 33.5% higher than the anoxic adsorptive capacity, respectively. This difference in capacity is
much more pronounced for 2-methylphenol. The oxic adsorption capacity for 2-methylphenol com-
pared to the anoxic adsorptive capacity on F400 increases from 42.3% to 52.5% for an equilibrium
concentration of 5 and 0.5 mM, respectively.
It is also of interest to notice that the BC-21 exhibited very slight differences between the oxic and
anoxic adsorption isotherms for the three phenolic compounds as compared to F400 (see Table 6.3).
This indicates that there is no considerable change between both conditions. In other words, it is
safe to assume that BC-21 has a significant positive impact on hampering the oligomerization of
phenolic compounds on its surface. Especially, in the case of 2-ethylphenol, the capacity difference
Activated Carbon Adsorption in Water Treatment 455
TABLE 6.3
Comparison of Anoxic and Oxic Adsorption Capacity
Phenol 2-Methylphenol 2-Ethylphenol
Ce Anoxic qe Oxic qe 𝚫 qe Anoxic qe Oxic qe 𝚫 qe Anoxic qe Oxic qe 𝚫 qe
(mM) (mmol/g) (mmol/g) (%) (mmol/g) (mmol/g) (%) (mmol/g) (mmol/g) (%)
F400 0.5 1.854 2.461 32.7 1.760 2.683 52.5 2.146 2.865 33.5
5 2.401 3.174 32.2 2.828 4.027 42.3 4.001 5.010 25.1
BC-21 0.5 1.943 2.182 12.3 2.552 2.796 9.54 2.844 2.877 1.16
5 3.195 3.010 6.12 4.185 4.341 3.70 3.537 3.626 2.53
Source: Data from Yan, L. and Sorial, G.A., J. Hazard. Mater., 197, 311, 2011.
on BC-21 under both conditions (low and high carbon dosage) is <2.5%. The main reason for this
distinct performance is that BC-21 had a significant higher micropore percentage than F400. In the
case of BC-21, forming larger stable oligomerized molecular products is much more difficult than
F400. This could be attributed to the molecular dimensions of the adsorbates and the critical pore
size of the adsorbent. For example, F400 has a broad PSD from 10 to 500 Å and oligomerization
would not be constricted by its PSD. In contrast, BC-21 has a critical pore diameter of 10 Å and
a narrow PSD. The second widest dimensions of phenol, 2-methylphenol, and 2-ethylphenol are
4.17, 5.32, and 5.70 Å, respectively. Therefore, it is just possible for phenol to form dimers, while
the formation of trimers is impossible. In the case of 2-methylphenol and 2-ethylphenol, it is even
impossible for them to get oligomerized. On the other hand, Vidic et al. (1997) have postulated that
oxygen-containing basic surface functional groups slightly promote the oligomerization of phenolic
compounds. Therefore, the behavior difference between BC-21 and F400 may be partially due to
the difference of the total basicity. The total basicity (0 meq/g) for BC-21, as determined by the
titration method, was much lower than that of the GAC F400 (0.475 meq/g). This shows that GAC
F400 has more basic surface functional groups than BC-21 that could promote the oligomerization
of phenolic compounds.
It is worth to note that the equilibrium adsorptive capacity of BC-21 for all three phenolic com-
pounds was higher than F400 under anoxic adsorption condition (see Table 6.3). Some studies
have shown that the effect of surface functional group content on phenolic compounds adsorptive
capacity in the absence of molecular oxygen can be negligible (Efremenko and Sheintuch 2006).
Therefore, the results can be mainly attributed to the following physical reasons. Firstly, these
results illustrate that the higher BET surface area of BC-21 can lead to higher adsorption than F400.
Secondly, this phenomenon can also be elucidated by the comparison of the microporosity of the
two adsorbents. The microporosity of F400 is only 60.7%, while for BC-21 it is 81.8%. Therefore,
the adsorption capacity of BC-21 is higher than F400 because micropores are primarily responsible
for the adsorption.
Ideal adsorbed solution theory (IAST) was used in this study because it is the most common
approach used to predict the multicomponent adsorption isotherms onto activated carbon by using
only single solute equilibrium data. The IAST is based on the assumption that the adsorbed mix-
ture forms an ideal solution at a constant spreading pressure. The model can be represented by the
following Equation 6.4:
Cio (. m, T )qe,i
Ce ,i = i=n (6.4)
- i =1
qe,i
where Cio (÷ m, T ) is equilibrium liquid-phase concentration of pure solute i at the same temperature
T and spreading pressure Πm of the mixture with N components:
i=N i=N
qe,i Ce ,i
- = -C = 1 (6.5)
i =1
qio i =1
o
i
Equation 6.5 is the expression of the relationship between mixture equilibrium solid-phase concen-
tration qe,i and single solute equilibrium concentrations qio in the case of an ideal multicomponent
system:
qio
As πi q o dc o
= io io dqio (6.6)
RT ∫
ci dqi
0
In addition, a further relation is established between the spreading pressure and directly measur-
able quantities as Equation 6.6, where As is the external surface area per unit mass of adsorbent
and πi is the spreading pressure of single component i. This method did not consider the chemi-
cal effects during the adsorption process. Therefore, this approach will not accurately predict
adsorption equilibrium of trace organic contaminants under oxic condition due to oligomerization
of adsorbates on the surface of the activated carbon. However, it is very suitable to examine the
effectiveness of BC-21 in hampering oligomerization under oxic conditions in multicomponent
adsorption.
In this study, a binary adsorption of 2-methylphenol/2,4-dimethylphenol was performed on F400
and BC-21. Figure 6.8a and b presents the binary adsorption isotherms of phenol/2-methylphenol
for F400 and BC-21, respectively. Meanwhile, the Myers parameters obtained from the single solute
anoxic isotherm was applied into the IAST model to predict the competitive adsorption behavior.
The evaluation of the predictability of IAST was also done by calculating the sum of squares of
relative errors (SSREs). All the detailed concentration combination and SSRE values of both runs
for both adsorbents are shown in Table 6.4.
Figure 6.8a demonstrates that the presence of molecular oxygen had a significant impact
on the binary adsorptive properties of F400. The obvious differences between oxic and anoxic
binary adsorption isotherms due to oligomerization on the carbon surface was observed for both
2-methylphenol and 2,4-dimethylphenol. On the other hand, the anoxic adsorption isotherms can
be well predicted by IAST. However, the oxic adsorption isotherms for both 2-methylphenol and
2,4-dimethylphenol are obviously higher than oxic adsorption value predicted by IAST. The under
prediction of the IAST is due to the oligomerization behavior that is taking place in the presence of
molecular oxygen, which is not accounted for in IAST.
Activated Carbon Adsorption in Water Treatment 457
10
qe (mmol/g)
0.1
0.01 0.1 1 10 0.1 1 10
(a)
Ce (mM) (b) Ce (mM)
FIGURE 6.8 Binary adsorption isotherms of 2-methylphenol and 2,4-dimethylphenol on (a) F400 and
(b) BC-21. (Reprinted from Yan, L. and Sorial, G.A., Water Air Soil Pollut., 224, 1588, 2013. With permission.)
TABLE 6.4
Summary of the Multicomponent Adsorption System
Binary Adsorption
Initial Concentration (mM) SSRE
Adsorbent Conditions 2-Methylphenol 2,4-Dimethylphenol 2-Methylphenol 2,4-Dimethylphenol
F400 Anoxic 1.75 2.39 0.0655 0.0201
Oxic 1.74 2.38 0.8094 0.4581
Carbonexp Anoxic 2.13 4.10 0.1294 0.029
Oxic 2.12 4.08 0.0257 0.1810
Source: Date from Yan, L. and Sorial, G.A., Water Air Soil Pollut., 224, 1588, 2013.
In contrast, BC-21 (see Figure 6.8b) showed remarkable different binary adsorption behavior
than that documented for F400. For the binary adsorption studied, oligomerization could happen
in various ways, such as 2-methylphenol molecules, 2-methylphenol and 2,4-dimethylphenol, or
2,4-dimethylphenol molecules. Therefore, it seems there are more opportunities than single solute
adsorption. However, it is apparent from Figure 6.8b that the oxic and anoxic isotherm lines are
really close to each other for both adsorbates on Carbonexp. This observation can be attributed to
three reasons. Firstly, regarding multicomponent adsorption, the possibility for either compound to
get oligomerized becomes less as compared to the single solute system because of the competitive
effect that would lead to less available sites. Secondly, the narrow PSD also leads to the oligomer-
ization constriction effect on binary adsorption. Thirdly, BC-21 is a more acidic carbon that is unfa-
vorable for the oligomerization of phenolic compounds. It is seen from Figure 6.8b that the binary
adsorptive behavior of BC-21 agreed with those predicted by the IAST model for both the oxic and
anoxic binary system. Therefore, the good prediction (see Table 6.4) for both the oxic and anoxic
binary adsorption on BC-21 by IAST confirmed again the effectiveness of BC-21 in hampering
oligomerization for binary solute system adsorption.
458 Handbook of Surface and Colloid Chemistry
REFERENCES
Balci, S., T. Dogu, and H. Yucel. 1994. Characterization of activated carbon produced from almond shell and
hazelnut shell. Journal of Chemical Technology and Biotechnology 60(4):419–426.
Caturla, F., M. Molinasabio, and F. Rodriguezreinoso. 1991. Preparation of activated carbon by chemical acti-
vation with ZnCl2. Carbon 29(7):999–1007.
Centeno, T. A., M. Marban, and A. B. Fuertes. 2003. Importance of micropore size distribution on adsorption
at low adsorbate concentrations. Carbon 41(4):843–846.
Cooney, D. O. and Z. P. Xi. 1994. Activated carbon catalyzes reactions of phenolics during liquid-phase adsorp-
tion. AiChE Journal 40(2):361–364.
Donnet, J.-B. 1998. Carbon Fibers. 3rd edn. New York: Marcel Dekker.
Economy, J. and R. Anderson. 1966. A new route to boron nitride. Inorganic Chemistry 5(6):989–992.
Economy, J., M. Daley, and C. Mangun. 1996. Activated carbon fibers—Past, present, and future. Abstracts of
Papers of the American Chemical Society 211:84-Fuel.
Efremenko, I. and M. Sheintuch. 2006. Predicting solute adsorption on activated carbon: Phenol. Langmuir
22(8):3614–3621.
Girgis, B. S., L. B. Khalil, and T. A. M. Tawfik. 1994. Activated carbon from sugar-cane Bagasse by car-
bonization in the presence of inorganic acids. Journal of Chemical Technology and Biotechnology
61(1):87–92.
Grant, T. M. and C. J. King. 1990. Mechanism of irreversible adsorption of phenolic-compounds by activated
carbons. Industrial & Engineering Chemistry Research 29(2):264–271.
Guo, J. and A. C. Lua. 2000. Textural characterization of activated carbons prepared from oil-palm stones
pre-treated with various impregnating agents. Journal of Porous Materials 7(4):491–497.
Hsieh, C. T. and H. S. Teng. 2000. Liquid-phase adsorption of phenol onto activated carbons prepared with
different activation levels. Journal of Colloid and Interface Science 230(1):171–175.
Hu, Z. H. and E. F. Vansant. 1995. Synthesis and characterization of a controlled micropore-size carbonaceous
adsorbent produced from walnut shell. Microporous Materials 3(6):603–612.
Hussein, M. Z., R. S. H. Tarmizi, Z. Zainal, R. Ibrahim, and M. Badri. 1996. Preparation and characterization
of active carbons from oil palm shells. Carbon 34(11):1447–1449.
Jiang, Q. and Y. Zhao. 2004. Effects of activation conditions on BET specific surface area of activated carbon
nanotubes. Microporous and Mesoporous Materials 76(1–3):215–219.
Jimenez, V., J. A. Diaz, P. Sanchez, J. L. Valverde, and A. Romero. 2009. Influence of the activation condi-
tions on the porosity development of herringbone carbon nanofibers. Chemical Engineering Journal
155(3):931–940.
Juang, R. S., R. L. Tseng, and F. C. Wu. 2001. Role of microporosity of activated carbons on their adsorption
abilities for phenols and dyes. Adsorption—Journal of the International Adsorption Society 7(1):65–72.
Jung, M. W., K. H. Ahn, Y. Lee, K. P. Kim, J. S. Rhee, J. T. Park, and K. J. Paeng. 2001. Adsorption char-
acteristics of phenol and chlorophenols on granular activated carbons (GAC). Microchemical Journal
70(2):123–131.
Kaneko, Y., M. Abe, and K. Ogino. 1989. Adsorption characteristics of organic-compounds dissolved in water
on surface-improved activated carbon-fibers. Colloids and Surfaces 37:211–222.
Laine, J. and A. Calafat. 1991. Factors affecting the preparation of activated carbons from coconut shell cata-
lyzed by potassium. Carbon 29(7):949–953.
Laine, J., A. Calafat, and M. Labady. 1989. Preparation and characterization of activated carbons from coconut
shell impregnated with phosphoric-acid. Carbon 27(2):191–195.
Lillo-Rodenas, M. A., D. Cazorla-Amoros, and A. Linares-Solano. 2003. Understanding chemical reactions
between carbons and NaOH and KOH—An insight into the chemical activation mechanism. Carbon
41(2):267–275.
Lillo-Rodenas, M. A., J. Juan-Juan, D. Cazorla-Amoros, and A. Linares-Solano. 2004. About reactions occur-
ring during chemical activation with hydroxides. Carbon 42(7):1371–1375.
Lozano-Castello, D., J. M. Calo, D. Cazorla-Amoros, and A. Linares-Solano. 2007. Carbon activation
with KOH as explored by temperature programmed techniques, and the effects of hydrogen. Carbon
45(13):2529–2536.
Lozano-Castello, D., D. Cazorla-Amoros, A. Linares-Solano, and D. F. Quinn. 2002. Influence of pore size
distribution on methane storage at relatively low pressure: Preparation of activated carbon with optimum
pore size. Carbon 40(7):989–1002.
Lu, Q. L. and G. A. Sorial. 2004. Adsorption of phenolics on activated carbon—Impact of pore size and molec-
ular oxygen. Chemosphere 55(5):671–679.
Activated Carbon Adsorption in Water Treatment 459
Lu, Q. L. and G. A. Sorial. 2007. The effect of functional groups on oligomerization of phenolics on activated
carbon. Journal of Hazardous Materials 148(1–2):436–445.
Magne, P. and P. L. Walker. 1986. Phenol adsorption on activated carbons—Application to the regeneration of
activated carbons polluted with phenol. Carbon 24(2):101–107.
Mangun, C. L., K. R. Benak, J. Economy, and K. L. Foster. 2001. Surface chemistry, pore sizes and adsorption
properties of activated carbon fibers and precursors treated with ammonia. Carbon 39(12):1809–1820.
Marsh, H., D. S. Yan, T. M. Ogrady, and A. Wennerberg. 1984. Formation of active carbons from cokes using
potassium hydroxide. Carbon 22(6):603–611.
Martin-Gullon, I., R. Andrews, M. Jagtoyen, and F. Derbyshire. 2001. PAN-based activated carbon fiber com-
posites for sulfur dioxide conversion: Influence of fiber activation method. Fuel 80(7):969–977.
Martinez, M. L., M. M. Torres, C. A. Guzman, and D. M. Maestri. 2006. Preparation and characteristics of
activated carbon from olive stones and walnut shells. Industrial Crops and Products 23(1):23–28.
Molinasabio, M., F. Rodriguezreinoso, F. Caturla, and M. J. Selles. 1995. Porosity in granular carbons activated
with phosphoric-acid. Carbon 33(8):1105–1113.
Moore, B. C., F. S. Cannon, J. A. Westrick, D. H. Metz, C. A. Shrive, J. DeMarco, and D. J. Hartman. 2001.
Changes in GAC pore structure during full-scale water treatment at Cincinnati: A comparison between
virgin and thermally reactivated GAC. Carbon 39(6):789–807.
Nakhla, G., N. Abuzaid, and S. Farooq. 1994. Activated carbon adsorption of phenolics in oxic systems effect
of Ph and temperature-variations. Water Environment Research 66(6):842–850.
Nowack, K. O., F. S. Cannon, and D. W. Mazyck. 2004. Enhancing activated carbon adsorption of 2-methyliso-
borneol: Methane and steam treatments. Environmental Science & Technology 38(1):276–84.
Otowa, T., Y. Nojima, and T. Miyazaki. 1997. Development of KOH activated high surface area carbon and its
application to drinking water purification. Carbon 35(9):1315–1319.
Parker, G. R. 1995. Optimum isotherm equation and thermodynamic interpretation for aqueous
1,1,2-trichloroethene adsorption isotherms on three adsorbents. Adsorption—Journal of the International
Adsorption Society 1(2):113–132.
Pelekani, C. and V. L. Snoeyink. 1999. Competitive adsorption in natural water: Role of activated carbon pore
size. Water Research 33(5):1209–1219.
Pelekani, C. and V. L. Snoeyink. 2000. Competitive adsorption between atrazine and methylene blue on acti-
vated carbon: The importance of pore size distribution. Carbon 38(10):1423–1436.
Ryu, Z. Y., H. Q. Rong, J. T. Zheng, M. Z. Wang, and B. J. Zhang. 2002. Microstructure and chemical analysis of
PAN-based activated carbon fibers prepared by different activation methods. Carbon 40(7):1144–1147.
Ryu, Z. Y., J. T. Zheng, M. H. Wang, and B. J. Zhang. 2000. Nitrogen adsorption studies of PAN-based acti-
vated carbon fibers prepared by different activation methods. Journal of Colloid and Interface Science
230(2):312–319.
Shen, W. Z., J. T. Zheng, Z. F. Qin, J. G. Wang, and Y. H. Liu. 2004. Preparation of mesoporous activated carbon
fiber by steam activation in the presence of cerium oxide and its adsorption of Congo red and Vitamin
B12 from solution. Journal of Materials Science 39(14):4693–4696.
Shmidt, J. L., A. V. Pimenov, A. I. Lieberman, and H. Y. Cheh. 1997. Kinetics of adsorption with granular,
powdered, and fibrous activated carbon. Separation Science and Technology 32(13):2105–2114.
Singh, B. K. and N. S. Rawat. 1994. Comparative sorption equilibrium studies of toxic phenols on fly-ash and
impregnated fly-ash. Journal of Chemical Technology and Biotechnology 61(4):307–317.
Stohr, B., H. P. Boehm, and R. Schlogl. 1991. Enhancement of the catalytic activity of activated carbons in
oxidation reactions by thermal-treatment with ammonia or hydrogen-cyanide and observation of a super-
oxide species as a possible intermediate. Carbon 29(6):707–720.
Suzuki, M. 1994. Activated carbon-fiber—Fundamentals and applications. Carbon 32(4):577–586.
Teng, H. S. and T. S. Yeh. 1998. Preparation of activated carbons from bituminous coals with zinc chloride
activation. Industrial & Engineering Chemistry Research 37(1):58–65.
Teng, H. S., T. S. Yeh, and L. Y. Hsu. 1998. Preparation of activated carbon from bituminous coal with phos-
phoric acid activation. Carbon 36(9):1387–1395.
Terzyk, A. P. 2003. Further insights into the role of carbon surface functionalities in the mechanism of phenol
adsorption. Journal of Colloid and Interface Science 268(2):301–329.
Tessmer, C. H., R. D. Vidic, and L. J. Uranowski. 1997. Impact of oxygen-containing surface functional groups
on activated carbon adsorption of phenols. Environmental Science & Technology 31(7):1872–1878.
Vidic, R. D. and M. T. Suidan. 1991. Role of dissolved-oxygen on the adsorptive capacity of activated carbon
for synthetic and natural organic-matter. Environmental Science & Technology 25(9):1612–1618.
Vidic, R. D., M. T. Suidan, and R. C. Brenner. 1993. Oxidative coupling of phenols on activated carbon—
Impact on adsorption equilibrium. Environmental Science & Technology 27(10):2079–2085.
460 Handbook of Surface and Colloid Chemistry
Vidic, R. D., C. H. Tessmer, and L. J. Uranowski. 1997. Impact of surface properties of activated carbons on
oxidative coupling of phenolic compounds. Carbon 35(9):1349–1359.
Villar-Rodil, S., F. Suarez-Garcia, J. I. Paredes, A. Martinez-Alonso, and J. M. D. Tascon. 2005. Activated
carbon materials of uniform porosity from polyaramid fibers. Chemistry of Materials 17(24):5893–5908.
Walker, P. L. 1969. Chemical and physical properties of Carbon. In Uoker, F. M. Ed., Khimicheskie i fizicheskie
svoĭstva ugleroda.
Wang, K., and D. D. Do. 1999. Sorption equilibria and kinetics of hydrocarbons onto activated carbon sam-
ples having different micropore size distributions. Adsorption—Journal of the International Adsorption
Society 5(1):25–37.
Wu, F. C., R. L. Tseng, and R. S. Juang. 2005. Preparation of highly microporous carbons from fir wood by
KOH activation for adsorption of dyes and phenols from water. Separation and Purification Technology
47(1–2):10–19.
Yamashita, Y. and K. Ouchi. 1982. Influence of alkali on the carbonization process. 1. Carbonization of
3,5-dimethylphenol-formaldehyde resin with NaOH. Carbon 20(1):41–45.
Yan, L. and G. A. Sorial. 2011. Chemical activation of bituminous coal for hampering oligomerization of
organic contaminants. Journal of Hazardous Materials 197:311–319.
Yan, L. and G. A. Sorial. 2013. Carbon activation for hampering oligomerization of phenolics in multicompo-
nent systems. Water Air and Soil Pollution 224(6):1–11.
Yue, Z. G., C. Mangun, J. Economy, P. Kemme, D. Cropek, and S. Maloney. 2001. Removal of chemical
contaminants from water to below USEPA MCL using fiber glass supported activated carbon filters.
Environmental Science & Technology 35(13):2844–2848.
Zawadzki, J. and S. Biniak. 1988. IR spectral studies of the basic properties of carbon. Polish Journal of
Chemistry 62(1–3):195–202.
Zhang, T. Y., W. P. Walawender, L. T. Fan, M. Fan, D. Daugaard, and R. C. Brown. 2004. Preparation of
activated carbon from forest and agricultural residues through CO2 activation. Chemical Engineering
Journal 105(1–2):53–59.
7 Solvation in
Heterogeneous Media
Sanjib Bagchi
CONTENTS
7.1 Introduction........................................................................................................................... 461
7.2 Self-Organized Assembles.................................................................................................... 462
7.2.1 Aqueous Normal Micelles and Vesicles Formed by Surfactants.............................. 463
7.3 Solvent Polarity and Solvatochromic Probes for Pure Solvents............................................465
7.3.1 Single Parameter Approach.......................................................................................466
7.3.1.1 Z-Scale........................................................................................................466
7.3.1.2 ET (30)-Scale................................................................................................ 467
7.3.2 Multiparameter Approach.........................................................................................468
7.4 Solvation in Microheterogeneous Media............................................................................... 470
7.4.1 Estimation of Micellar Phase Property Free from Aqueous Phase Contribution........ 471
7.4.2 Determination of Polarity Parameters....................................................................... 473
7.4.2.1 Method Using Absorption Probes............................................................... 473
7.4.2.2 Method Using Fluorescence Probes........................................................... 474
7.4.3 Polarity of the Micellar Phase................................................................................... 476
7.5 Concluding Remarks............................................................................................................. 477
Acknowledgments........................................................................................................................... 477
References....................................................................................................................................... 478
7.1 INTRODUCTION
Solute–solvent interactions play a key role in determining the observed kinetic, equilibrium, and
spectroscopic properties of a solute in different media. Since most chemical and biological pro-
cesses take place in solution, a detailed understanding of molecular interactions involving the solute
and its immediate environment is required. It has long been known that physicochemical properties
of a molecule (solute) depend considerably on the nature of the interacting molecules in its imme-
diate environment. A solute molecule may be thought as a source of field that induces structural
changes in a solvent. Three distinct regions may be distinguished around a solute molecule, namely,
a primary region with completely oriented solvent molecules, a secondary region containing par-
tially oriented solvent molecules, and a bulk region where the solvent molecules are not under the
influence of the solute and normal distribution of solvent (as in pure solvent) prevails. The primary
and the secondary regions where the influence of the solute is felt are commonly referred to as the
cybotactic region [1]. The net molecular interaction taking place in this region is often referred to
as solvation interaction. In dilute solutions, the nature of these interactions is mainly solute–solvent
and solvent–solvent.
Both theoretical [2,3] and experimental [4] efforts have been made to understand such solvation
interactions. Central to the study of this problem of interaction of a solute and its local environment
is the question of how the free energy of a solute changes due to the presence of surrounding inter-
acting molecules of the medium. An approach to the problem at a molecular level is fraught with
461
462 Handbook of Surface and Colloid Chemistry
the inherent difficulty that only little information is available about the arrangement of molecules in
the cybotactic region. The vast literature of solvent effect on the kinetic, equilibrium, and spectro-
scopic properties of a solute indicate that one needs to consider only a few modes of solute–solvent
interaction [5,6]. One of them is a nonspecific long-range interaction due to the collective influence
of solvent as a dielectric medium. This involves electrostatic interactions arising due to Coulomb
and polarization forces. The continuum dielectric theory provides an isomorphic model for describ-
ing nonspecific interaction [5]. It has been found that this mode of interaction is determined by the
relative permittivity (ε) and the refractive index (n) or any function thereof. Another is the specific
interaction involving hydrogen bond donation (HBD) and the acceptance ability of solvents. Since
specific solvation interactions are considered to be chemical in nature, the use of a continuum model
for describing such interaction is severely limited [7]. Chemists have tried to understand the envi-
ronmental effect by introducing the concept of solvent polarity, which is supposed to represent the
overall solvation ability of the solvent and includes both nonspecific and specific modes of interac-
tion [4,8]. For this, a model process [1] (e.g., a chemical equilibrium or kinetic process or spectral
transition) is chosen, and changes in one of its parameters are recorded when the solvent is changed.
Of the various polarity scales, the Z [1,9] and ET (30) [10] have passed the test of time and are most
widely used. In view of the complexity of solute–solvent interaction, the representation of polarity
by a single parameter (e.g., Z or ET (30)) is not accurate and one has to use a multiparameter approach
[11,12]. In this approach, it is assumed that the change of a physicochemical parameter of a solute
(S) due to solvation, Δ(S, i), can be represented as
Δ ( S, i ) = P0 ( S ) + Σaα ( S ) Pα ( i ) (7.1)
where
i indicates a solvent
P0 and aα terms depend on the solute (S)
Pα’s are the solvent properties pertinent to the α mode of interaction of the solute with a solvent
Equation 7.1 is known as the linear solvation energy relationship (LSER) [4,5,11] and has
been found to be of great significance in the theory of solvent effect. Although the single or the
multiparameter approach discussed earlier has been applied widely and successfully for describing
solvation in pure solvents, their use in the case of heterogeneous media formed by the self-assembly
of molecules is rather limited. For example, enhanced solubility of organic molecules in the micel-
lar phase compared to the bulk aqueous medium points to the existence of significant interaction
of the molecules with the micellar environment [13]. Successful attempts, however, have been
made to find out the descriptors of specific and nonspecific modes of interaction of a solute with
the heterogeneous medium [14–18]. Knowledge of the parameters describing specific and non-
specific solvation may be helpful to understand the solubilization of organic molecules like drugs
in such media. The objective of this chapter is to discuss the possibility of applying the preceding
parametric approach to the case of heterogeneous media and explore the present status of the field.
The outline of the chapter is as follows. A brief discussion of the process of self-organization and
commonly used microheterogeneous media will be done in Section 7.1. This will be followed by
the common p rocedures for the description of solvation in a homogeneous medium in Section 7.2.
The description of solvation in a heterogeneous medium in terms of suitable parameters will be
discussed finally in Section 7.3.
in nature. In static self-assembly, the ordered state is formed as a system approaches equilibrium,
reducing its free energy. However in dynamic self-assembly, patterns of preexisting components
organized by specific local interactions are commonly described as self-organized assembles. Self-
assembly can be defined as the spontaneous and reversible organization of component molecular
units (building blocks) into ordered structures by noncovalent interactions. An important aspect of
self-assembly is the key role of weak molecular interactions (e.g., van der Waals, π → π, hydrogen
bonds) with respect to more traditional covalent, ionic, or metallic bonds. These weak molecular
interactions play an important role especially in biological systems. Examples of self-organization
in materials science include the formation of molecular crystals, colloids, lipid bilayers, phase-
separated polymers, and self-assembled monolayers [19]. A characteristic common to nearly all
self-assembled systems is their thermodynamic stability. For self-organization to take place without
the intervention of external forces, the process must lead to a lower Gibbs free energy; thus self-
assembled structures are thermodynamically more stable than the single, unassembled components.
A direct consequence is the general tendency of self-assembled structures to be relatively free of
defects. The existence of weak molecular interactions and the condition of thermodynamic stability
endows these systems with a special property, namely, the sensitivity to perturbations exerted by the
external environment, namely, change of salt concentration, temperature, pH of the medium, and
pressure. The weak nature of interactions is responsible for the flexibility of the architecture and
allows for rearrangements of the structure in the direction determined by thermodynamics. This
gives rise to an important property, namely, reversibility to such type of assemblies. The specific
molecular arrangements in self-assembled systems are crucial for a specific property (functionality)
of the system. Different possible organized assembles are reported in the literature, namely, normal
and reverse micelle formed by surfactants, cyclodextrins in aqueous and nonaqueous medium, and
vesicles [20–28]. Systems like normal aqueous micelle and vesicles have been dealt with in the
following section.
Gouy–
double layer
Hydrocarbon core
Stern layer
Electrical double layer
FIGURE 7.1 A model representing the cross section of a spherical ionic micelle.
many properties of the mixture. Adding salts to a colloid containing micelles can decrease the
strength of electrostatic interactions and lead to the formation of larger ionic micelles [30]. From
the thermodynamic point, the formation of micelles can be understood in terms of a balance
between entropy and enthalpy.
Among different possible micellar shapes, the most prevalent is the spherical one as pro-
posed by McBain [31] and supported by Hartley [32]; ellipsoidal and cylindrical shapes are also
possible [33–38]. A simplified model for ionic micelles as given by Hartley [32] is useful for
understanding features. The cross section of such a micelle is shown schematically in Figure 7.1.
The core of a normal micelle is liquid-like, consisting of the hydrocarbon chains of the surfac-
tants. A peripheral shell outside the core contains polar/ionic head groups. This shell, for ionic
surfactants, usually has a width up to a few angstroms and is called Stern layer; it consists of
surfactant head groups, bound counterions, and water molecules. Most of the counterions are
dissociated from the micelle and are located in the Gouy–Chapman double layer, which is out-
side the Stern layer and whose width is up to several hundred angstroms [39–41]. The surface
potential generated by the net charges of the Stern layer is usually in the range of 50–100 mV,
and this acts as an electrostatic barrier to the passage of charged ions to and across the micellar
surface. The presence of ionic groups at the micellar interface causes nearby water molecules to
solvate the micelles through ion–dipole interaction. For nonionic micelles the shell outside the
core is termed as the Palisade layer and consists of the polar head groups and hydrogen-bonded
water molecules. Recent solvation dynamics studies using time domain optical spectroscopy
indicate that the water molecules confined within a small volume of Stern layer/Palisade layer
of the micellar interface are fundamentally different from water molecules at the bulk [42,43].
Micelles have particular significance because of their ability to increase the solubility of spar-
ingly soluble substances in water. The spatial position of a solubilized substance in a micelle
depends on its polarity. Nonpolar molecules will be solubilized in the micellar core, and sub-
stances with intermediate polarity will be distributed along the surfactant molecules in certain
intermediate positions.
Apart from the formation of micellar aggregates in aqueous medium, certain naturally occurring
or synthetic phospholipids as well as completely synthetic surfactants can form organized assembles.
The assemblies formed from phospholipids are typically termed liposomes, and those formed from
synthetic surfactants are designated as vesicles [44]. Depending on the method of preparation, the
nature of surfactants, and the experimental conditions (pH, ionic strength, concentrations), different
vesicular structures can exist. These range from small (8200–500 Å diameter) or large (0.1–10 μ
diameter) single-walled bilayer structure to relatively larger (1000–8000 Å diameter) onion-like
multicomponent structure. Unlike micelles, vesicles once formed do not break down upon dilution.
Moreover, the hydrocarbon part of the vesicle is more ordered [44].
Solvation in Heterogeneous Media 465
where M1 and M2 are the two different electronic states of a molecule and the positive and nega-
tive sign refer to absorption and emission, respectively. The effect of solvation is reflected in
the spectroscopic behavior of the solute–solvent system. Interaction of different electronic states
with the surrounding solvent molecules is also different, and this phenomenon leads to significant
changes in the photophysical properties of the molecule. The observed spectral and photophysi-
cal parameters are thus characteristic of the molecule in its environment rather than the isolated
one. Thus, a study of a spectral parameter is supposed to provide information regarding solvation
interaction. It may be mentioned that while the initial state in a spectroscopic transition is an
equilibrium state, the final state is a nonequilibrium Franck–Condon state. A differential solva-
tion interaction between the two electronic states is responsible for the solvent sensitivity of a
spectral parameter. In an absorption process, the observed parameter represents the differential
solvation interaction of the solute molecule in the S0 and S1 state with an environment that is in
equilibrium with the S0 state. On the other hand, the observed emission parameter describes the
microenvironment that is in equilibrium with the S1 state. Moreover, the differential interaction
will be large when the e lectronic transition induces a substantial change in the electronic distri-
bution in the molecule. Thus, the optical response of molecules showing a charge transfer (CT)
has been found to depend significantly on solvation interaction. As such, these molecules act as
micropolarity reporters probing the local environment in a solution. Thus, a study of the param-
eters associated with the preceding process (e.g., band maximum and bandwidth in the case of
absorption; band maximum, bandwidth, quantum yield, lifetime, and anisotropy in the case of
fluorescence) provides information regarding solvation interaction. Spectroscopic parameters rep-
resenting solvent polarity have thus been derived from the solvent-sensitive property of standard
solutes (reporters/probes) absorbing light in the spectral region corresponding to UV–visible, IR,
ESR, and NMR spectra. Most commonly, the parameters are wavelength (frequency) maximum
466 Handbook of Surface and Colloid Chemistry
– R = –COOCH3 (Ia)
I
+
N R = –CN (Ib)
C2H5
R R
+ –
R N O R= (II)
R R
(III)
N N (IVa)
N (IVb)
N
CH3 H3C
of absorption/fluorescence band, intensity of band, etc. Sometimes the ratio of intensity of various
spectroscopic bands has been used (e.g., I1/I3 ratio for pyrene emission). In general, several modes
of solute–solvent interactions have been identified, namely, nonspecific (dipolarity and polariz-
ability) and specific (acidity, basicity) modes. If one carefully selects an appropriate, sufficiently
solvent-sensitive reference process, one may assume that this process would reflect faithfully all
possible solute–solvent interactions. It should, therefore, give an empirical measure of the solva-
tion capability of a particular solvent for the given reference process. Several compounds having
solvent-sensitive absorption/emission bands have been utilized as micropolarity reporters. Probes
may be classified as absorption probes or fluorescence probes depending on whether absorption
or fluorescence spectral characteristics are utilized. A list of some important reporter molecules
(probes) is given in Figures 7.2 and 7.3. Two main approaches for the description of micropolarity
have been discussed here.
the solvent increases. The Z-value was defined as the maximum transition energy in kilo calories
per mole corresponding to the longest wavelength CT band of this solute. Thus we have
28,590
Z kcal mol −1 =
( ) (7.3)
λ max ( nm )
1-Ethyl-4-cyanopyridinium iodide (Figure 7.2, Ib) has also been used to establish a parallel scale [46].
7.3.1.2 ET(30)-Scale
The indicator solute used by Kosower for the Z-scale is not quite soluble in certain nonpolar
solvents. To overcome this difficulty, Dimroth and Reichardt suggested the use of pyridinium-
N-phenol betaines as the indicator solute [10]. The ET (30) value of a solvent is the maximum tran-
sition energy in kilo calories per mole of the longest wavelength band in that solvent of the solute
shown in Figure 7.2, II. The ET (30) value is obtained from λmax values using an equation similar
to Equation 7.3. The ET (30) values of various neat solvents have been listed in Table 7.1. Owing
to the large displacement of the solvatochromic absorption band, the ET (30) values provide an
TABLE 7.1
List of Values of the Parameters ET(30), α, β, π*
Sl. No. Solvent ET(30)a (k cal mol−1) αb βb π*b
1 Water 63.1 1.17 0.47 1.09
2 Methanol 56.3 0.93 0.66 0.60
3 Ethanol 51.9 0.83 0.75 0.60
4 1-Propanol 50.7 0.78 0.80 0.52
5 2-Propanol 48.6 0.76 0.84 0.48
6 1-Butanol 50.2 0.79 0.82 0.47
7 1-Hexanol 48.8 — — —
8 1-Octanol 48.5 — — —
9 Cyclohexanol 46.9 — — —
10 Acetone 42.2 0.08 0.48 0.71
11 2-Pentanone 41.1 — 0.50 —
12 3-Pentanone 39.3 — 0.45 0.72
13 Acetonitrile 46.0 0.19 0.40 0.75
14 Butyronitrile 42.5 0.00 0.44 0.71
15 Chloroform 39.1 0.44 0.00 0.58
16 Dichloromethane 41.1 0.30 0.00 0.82
17 Formamide 55.8 0.71 0.70 0.97
18 Dimethylformamide 43.8 0.00 0.69 0.88
19 Dimethylsulfoxide 45.1 0.00 0.76 1.00
20 Dioxane 36.0 0.00 0.37 0.55
21 Tetrahydrofuran 37.4 0.00 0.55 0.58
22 Ethylacetate 38.1 0.00 0.45 0.55
23 Benzene 34.5 0.00 0.10 0.59
24 Toluene 33.9 0.00 0.11 0.54
25 Hexane 31.0 0.00 0.00 −0.08
26 Cyclohexane 31.2 0.00 0.00 0.00
27 Heptane 31.1 0.00 0.00 −0.08
a Refs. [8,10].
b Kamlet–Taft solvatochromic parameters, Refs. [45,67,70].
468 Handbook of Surface and Colloid Chemistry
excellent and very sensitive characterization of the polarity of the solvents [4,8]. The compound
can also act as polarity probe for the investigation of aqueous micellar solutions, microemulsions,
and phospholipid bilayers as well as for investigation of interfacial properties of lipid membranes.
The disadvantage of ET (30) probe is, however, that it has a low molar absorptivity and it is gener-
ally not fluorescent except under certain circumstances (dye embedded in polymer film or at a low
temperature [47]). However, intense fluorescence is essential when probes are used in biological
systems since they are mostly applied in highly diluted conditions. An important and widely used
scale of solvent polarity has been proposed using a fluorescence probe pyrene (Figure 7.3, III). It
has been known that the ratio of intensities of the first to third vibronic bands (I1/I3) of the S1 → S 0
emission of pyrene depends on the medium, and as such I1/I3 is often used as a measure of solvent
polarity [48–54]. The ratio I1/I3 ranges from ~0.6 in hydrocarbon media to ~1.67 in water. A number
of other fluorescence probes are known. Kessler and Wolfbeis introduced a group of ketocyanine
dyes (Figure 7.3, IVa and IVb) that act as good solvent-sensitive fluorophores [55,56]. Depending
on the solvent, these probes display a bright yellow to purple violet color and a strong green, yel-
low, and red fluorescence. Both the absorption and emission maxima are red shifted in going from
nonpolar to polar solvent.
Polarity probes in general should exhibit the general features mentioned in the following text [56,57]:
(1) High spectral sensitivity to polarity changes (a spectral shift of both excitation and emission
maxima is advantageous because of additional sensitivity and selectivity), (2) high fluorescence
quantum yield, (3) long wavelength absorption and fluorescence so that they can be conveniently
used in biological systems, and (4) photostability and thermal stability. Although a large number of
polarity probes have been reported, none of them appreciably satisfies all the four criteria. It may
be mentioned that the chemical structure of an ideal indicator compound should be such that all
important nonspecific (dipolar, dispersion) and specific (HBD, hydrogen bond acceptance [HBA],
and electron donor–acceptor) solute–solvent interactions are possible. Considering the wide varia-
tion in the chemical structure of the probe molecules used for characterization of solvent polarity,
the establishment of a universal, generally valid solvent polarity scale seems unattainable.
A ( S, i ) = A ( S ) + B ( S, i ) (7.4)
where A(S) is a solvent-independent part. The term B(S,i) is in general a complex function of both
solvent and solute comprising several modes of solute–solvent interaction. Experimental data indi-
cate that solvent effect on various kinetic, equilibrium, and spectroscopic parameters is essentially
similar in their very nature. The similarity is considered to show that there are comparatively few
mechanisms of physical interaction between solvent and solute [6,58–61]. In essence there are three
types of interactions: (1) nonspecific long-range solvent–solute interaction, (2) specific short-range
solute–solvent interaction, and (3) solvent–solvent interaction from the cavity effect. The most
effective nonspecific interactions are considered to be determined by macroscopic physical param-
eters of the solvent, that is, the relative permittivity or dielectric constant (ε) and refractive index (n).
The specific solvation is mainly determined by the acidity and basicity of the solvent, in terms of
the Lewis concept, which are measures of the solvent hydrogen bond ability to donate (HBD) and
to accept (HBA) a proton, respectively. Disruption and reorganization of solvent–solvent interaction
Solvation in Heterogeneous Media 469
are measured by the work necessary to separate solvent molecules to create a suitable cavity, large
enough to accommodate the solute.
It has been shown that under certain simplifying assumptions, the solute–solvent interaction
term B(S,i) may be factorized as [62]
B ( S, i ) = ∑a ( S ) P (i ) (7.5)
α α
where suffix α represents various modes of interaction, the parameters aα(S) and Pα(i) depending on
the solute and solvent, respectively. This type of expression, often called LSER, has been found to
be of great significance in the theory of solvent effect [11,62,63]. In general terms, LSER [8,64] is
expressed by the following equation:
XYZ = XYZ 0 + Cavity formation energy + ∑ Solute – solvent interaction energy (7.6)
In the preceding equation, XYZ is a property linearly related to the free energy of the system.
The term XYZ 0 denotes a constant and depends solely on the solute. Summation in the preceding
equation extends over all the possible modes of solute–solvent interaction [65,66]. Two approaches
in the use of the LSER equation may be distinguished, namely, the approach suggested by Koppel
and Palm (KP) [58] and that by Abraham, Kamlet, and Taft (AKT) [11]. In the KP approach, func-
tions of dielectric constant (ε) and refractive index (n) were used to describe the nonspecific interac-
tion. Thus Onsager reaction field parameter (ε − 1)/(2ε + 1) was used to describe the nonspecific
dipolar interaction, while (n2 – 1)/(n2 + 2) described the polarizability term. In the AKT approach,
the dipolarity and polarizability were described by the experimentally determined parameter π*
[62]. The specific interactions were described by the parameter “E” (electrophilic solvation ability)
and “B” (nucleophilic solvation ability) in the KP procedure [58]. But AKT have preferred the use of
HBD and HBA ability of solvent represented by the empirical parameter α and β, respectively [67].
The endothermic cavity formation term was taken in the AKT approach as equal to the solute molar
volume times the Hildebrand cohesive energy density (δ2H ) defined as the enthalpy of vaporization
per unit volume [68]. The original KP approach did not take this factor into account although this
term was introduced later by Makitra and Pirig [69]. Thus, the KP equation is
A = A 0 + yY + pP + eE + bB (7.7a)
2 ( ε − 1) n2 − 1
( )
Y= ; P= 2 (7.7b)
( 2ε + 1) n +2
( )
δ2H =
( ΔH – RT ) (7.8b)
V
where ΔH is the molar enthalpy of vaporization of the solvent whose molar volume is V. It has
been found that the various empirical polarity parameters, for example, ET (30) and Z, depend lin-
early on α-, β-, and π*-values [47,70]. The Kamlet–Taft solvatochromic parameters π*, α, and β
are usually determined by measuring the maximum transition energy of certain probes using the
470 Handbook of Surface and Colloid Chemistry
solvatochromic comparison method [71–73]. For the determination of π*, the commonly used probes
are 4-nitroanisole, 2-nitroanisole, 4-ethyl-nitrobenzene, and N,N-diethyl-4-nitroaniline. For these
probes, the energy of maximum absorption depends only on the dipolarity– polarizability modes of
interaction with solvent. For setting up the β scale, a spectroscopic or thermodynamic property of two
indicator solutes of differing H-bonding ability (e.g., 4-nitroaniline and N,N-diethyl-4-nitroaniline)
in a series of solvents was compared. The value of enhanced solvatochromic displacement normal-
ized with respect to a reference solvent hexamethylphosphoramide (HMPA) gave a dimensionless
parameter. To eliminate the specific role of solute, the procedure was repeated for multiple indica-
tors and the average of the dimensionless parameter gave β. A similar procedure using multiple
linear regression analysis (MLRA) was adopted for setting up the α-scale. Here due to various
complications, the requirements for the indicator solute were considered to be relatively stringent.
Sixteen diverse properties involving 13 indicators were used to calculate the final α-values [62].
Table 7.1 lists the relevant polarity parameters of some pure solvent. In fact, the LSER equations with
these parameters have been used in hundreds of studies involving the effect of solvent on kinetic
spectroscopic and equilibrium properties of a solute. Catalan has introduced the parameters SPP,
SA, and SB for representing the solvent dipolarity–polarizability (nonspecific mode), solvent acidity,
and solvent basicity (the specific modes) of a solvent. These parameters are obtained from spectro-
scopic measurements and their success in accounting for medium effects on reactivity and a host of
physiochemical properties is well documented and widely perceived [74–77].
For ionic micelles, the cation or anion of the amphiphilic molecule remains in equilibrium between
the two phases. An indicator solute distributes itself between the two phases. Any property (P) of
the solute that has dependence on its local environment can be used to reveal the equilibrium pro-
cess. Due to difference in polarity, the value of P is expected to be different in the two phases. The
value of P for such systems will be determined by the time average location of the probe in the two
phases [85]. The measured value of P in a micellar media can be assumed to be represented by a
mole fraction average of the aqueous phase property (Paq) and micellar phase property (Pm), as given
by the following equation:
P=
( naq Paq + nm Pm ) (7.10)
( naq + nm )
where
n refers to the number of moles of the solute
subscripts m and aq denote the micellar phase and aqueous phase, respectively
The calculation of the value of Pm and Paq requires knowledge of nm and naq, which in turn are
related to the distribution coefficient of the indicator solute between the two phases. It has been
shown that the parameters are related as follows [17,18]:
where
v is the molar volume of the surfactant
Cs is the concentration of the surfactant in the aggregated phase, as given by Cs = CT − cmc,
where CT is the total surfactant concentration
K is the molar-based distribution coefficient of the indicator between the two phases
472 Handbook of Surface and Colloid Chemistry
59.7
52.8
52.6 59.6
E(F ) (kcal. mol–1)
52.2 59.4
52.0
59.3
0.28
600,000
0.24
Fluorescence intensity
450,000
0.20
r
300,000
0.16
150,000
0.12
0
–2.2 –2.1 –2.0 –1.9 –1.8 –1.7 –1.6 –1.5 –3.2 –3.1 –3.0 –2.9
(c) log([DTAB]/M) (d) log([CTAB]/M)
FIGURE 7.4 Plot of the parameters energy of (a) maximum absorption/ (b) fluorescence, (c) fluorescence inten-
sity and (d) steady-state anisotropy (r) of (Va) in the aqueous solution of surfactants as a function of log CT. CT for
the surfactants have been denoted by surfactant. (From Kedia, N. et al., Spectrochim. Acta A, 131, 398, 2014.)
The variation of these parameters as a function of log CT can be studied. Figure 7.4 shows repre-
sentative plots describing such variation. Values of these parameters change abruptly after a certain
concentration. Data points can be fitted to Boltzmann sigmoid equation. It has been proposed that
the value of cmc can be obtained from the inflexion point of the Boltzmann sigmoid curve obtained
by plotting a solute property as a function of CT [86,87]. Truly speaking, P is related to log Cs by
Boltzmann equation as can be seen from Equation 7.11. Thus, the fitting of the experimental data
point to the Boltzmann sigmoid plot involving the parameter P and log CT is only approximate, and
hence the assignment of concentration at the inflexion point as cmc is likely to be in error. Equation
7.11 can be used for the determination of the cmc value [88]. An iterative procedure can be adopted
as follows. Equation 7.11 can be rearranged as
The value of concentration of surfactant at the inflexion point of P versus log CT plot is taken as
the guess value of cmc. The parameter Paq is then determined by linear regression analysis using
Equation 7.12. It follows from Equation 7.11 or 7.12 that P = Paq at CT = cmc. Thus, the value of the
concentration for which P = Pa was found out from the P versus log CT curve. This gives the output
value of cmc. The preceding procedure can be repeated by using the output cmc value as the input
Solvation in Heterogeneous Media 473
to get a new output cmc. The iteration can be continued until the input and output cmc values are
within a specified range (e.g., within ±2%). Once cmc is determined, values of Pm, Paq, and K can
be found out by a linear regression analysis using Equation 7.12. The procedure also requires the
value of molar volume of the surfactant. It is apparent from Equation 7.9 that values of parameters in
micellar solution, as found out from the iterative procedure, are close to the value of the parameter at
a concentration of surfactants much above the cmc value. But in view of the possibility of a change
in the shape of the micelle, the use of too high a concentration is not advisable.
E ( S, i ) = E0 ( S ) + p ( S ) π * ( i ) + a ( S ) α ( i ) + b ( S ) β ( i ) (7.13)
The Kamlet–Taft solvatochromic parameters π*, α, and β for a micellar medium are usually deter-
mined by measuring the spectroscopic parameters of certain probes using known correlation equa-
tions. The correlation equations are, however, obtained from studies in pure solvents. It is assumed
that the correlation equation obtained for pure solvents is also valid for micellar media. Both absorp-
tion [14–17] and fluorescence [18,90] probes have been used for estimating values of α, β, and π*
for the micellar media.
(34.12 − σ)
For 4-nitroanisole, π* = (7.14)
2.343
(32.56 − σ)
For 2-nitroanisole, π* = (7.15)
2.428
(37.67 − σ)
For 2-ethyl-nitrobenzene, π* = (7.16)
2.259
For determining β-values, the commonly used probes are 4-nitrophenol and 4-nitroaniline. The
transition energy of these probes depends on both β and π*. For example, for 4-nitroaniline the
wavenumber of transition is related to β as follows [17]:
(31.10 − 3.14π* − σ)
β= (7.18)
2.79
Similarly the α-values are determined by using ET (30), ET (33), and Fe(LL)2(CN)2 as α and
π*-probes (transition energy depends on α- and π*-values) [91–93]. The ET (33) probe is a bis
chlorosubstituted derivative of ET (30) probe. It has the advantage over the ET (30) probe because
of its water solubility.
Thus, relations (7.19) and (7.20) are valid for Dimroth–Reichardt ET (30) and ET (33)–betaine,
respectively:
In the preceding equations, δ is a correction factor to account for the polarizability of the medium
[4,67]. Its value is zero for nonaromatic and nonchlorinated pure organic solvents. For the Fe
complex, the transition energy (E) in kcal mol−1 is given by the following equation:
The value of α can be obtained using a known value of π* determined by using other probes. This
relies on the assumption that these α-probes reside in the same region in the micellar phase as do
the π*-probes.
⎛ I1 ⎞ *
⎜ I ⎟ = 1.304π + 0.46 (7.22)
⎝ 3⎠
This equation can be used to estimate π* for heterogeneous media and the values agree well with
those obtained by using other fluorescence probes [94]. The preceding empirical relation has often
been used to calculate π* employing the micellar phase value of I1/I3 for mixed micelle of varying
composition [90].
It is known that the position of the maximum fluorescence of ketocyanine dyes in a solution
is highly solvent sensitive and the energy of maximum fluorescence, E(F), is linearly dependent
on the solvatochromic parameters, α and π*. The dyes are solubilized in the micelle–water inter-
face [95]. It is believed that the polarity characteristics of the micelle–water interfacial region
resemble those of aqueous alkanols [49]. The values of E(F) of ketocyanine dyes have been
determined in various pure n-alkanols and mixed binary aqueous alkanol systems for which
the values of solvatochromic parameters (α and π*) are known [96]. An MLRA of E(F) with the
Solvation in Heterogeneous Media 475
O O
N (Va) N (Vb)
H3C
O O
(VI) (VII)
H3C N N CH3 N CH3
CH3 H3C H3C
O O
solvatochromic parameters has been performed. It has been found that the E(F) of the ketocyanine
dyes, (Va) and (Vb) (Figure 7.5), correlates well with the α- and π*-parameters for a series of pure
alkanols and water–alkanol mixtures. Thus, Equations 7.23 and 7.24 are found to be obeyed for
dyes (Va) and (Vb), respectively:
One can assume that these relationships are also valid for heterogeneous media. The value of α for
the micellar phase can be obtained using Equations 7.22 and 7.23 from the known values of E(F)
and π* for the micellar phase.
The coefficients p(s), a(s), and b(s) in Equation 7.13 depend on the solute and are related, respec-
tively, to dipolarity/polarizability and the HBA and HBD ability of the solute [12]. Equation 7.13 can
be used to treat the solvatochromic transition energy (E) for a fixed solute for a series of solvents
whose π*-, α-, and β-values are known. MLRA of E with π*, α, and β will then provide the p, a,
and b terms. This procedure is used to get the LSER equation for a particular solute [8]. Conversely,
the equation can be used to analyze the transition energy of a series of solutes of known p, a, and b
values in a fixed solvent (medium) to get the π*-, α-, and β-values of the medium. Ketocyanine dyes
provide an interesting series of solutes, which are characterized by solvatochromic fluorescence [55].
The spectroscopic transition leading to the fluorescence in these molecules has been established as
due to intramolecular charge transfer (ICT) involving the carbonyl oxygen and nitrogen atom of the
amino group [55,97–99]. Shannigrahi et al. studied the solvatochromic fluorescence energy of the
ketocyanine (Figure 7.3, IVa and IVb) and other structurally related dyes (Figure 7.5, Va, Vb, VI,
VII, VIII) extensively in various pure solvents [97–99]. MLRA of maximum fluorescence energy in
pure solvents has been done with solvent parameters π*, α, and β to get p, a, and b coefficients for
the solutes. MLRA of the energy of maximum fluorescence, E(F), of different solutes in the micel-
lar phase has been sought with the known solute parameters p, a, and b to get the values of π*, α,
and β for the micellar phase [18].
476 Handbook of Surface and Colloid Chemistry
For a particular micelle, (m), the energy of maximum transition for a particular solute, (s), is
given by the following equation:
⎡⎣ Em ( s ) − E0 ( s ) ⎤⎦ = a ( s ) α m + b ( s ) βm + p ( s ) π*m (7.25)
The values of α m, βm, and ≠*m that give the best correlation in the least-squares sense of [Em(s) − E 0(s)]
with a(s), b(s), and p(s) for the seven solutes were then determined. Alternatively, one has for the
micellar and aqueous phase:
A correlation analysis involving [Em(s) – Eaq(s)] in place of [Em(s) – E 0(s)] provides Δα = [α m − α aq],
Δβ = [βm − βaq], and Δπ* = [ π*m − π*aq ], that is, the change of the parameters as one goes from an
aqueous phase to a micellar phase.
TABLE 7.2
Solvatochromic Parameters for Various Micellar Phases
Micellar Phase α β π*
SDS 1.14a, 1.10b, 0.873c, 1.18f, 1.2g 0.70a, 0.42b, 0.401d 0.52a,f, 1.14b, 1.06d, 0.58g
CTAB 0.87a, 0.62c, 0.94g 0.53a, 0.47b 0.68a, 0.96c, 0.69g
DTAB 0.700e, 1.03g 0.486e 1.02a, 0.70g
TX100 1.06a, 0.92f 0.71a 0.56a, 0.72f
Water 1.17 0.47 1.09
a Ref. [18].
b Ref. [17].
c Ref. [16].
d Ref. [14].
e Ref. [15].
f Ref. [90].
g Ref. [88].
Solvation in Heterogeneous Media 477
respectively [14,15]. Similarly, Fuguet et al. reported α = 0.82 and 0.62 for SDS and CTAB, respec-
tively [16,17]. The values of β indicate that ionic micelles have relatively more HBA basic than the
aqueous phase. The neutral TX100 micelle on the other hand has HBA ability comparable to that
of the aqueous phase. π*-values for the micellar phase of the surfactant are less than that of the
aqueous phase for all the micelles. The micellar phase consists of a hydrocarbon core and a polar
region. The polar head groups, the bound counterions (in the case of ionic micelles), and water
molecules bound to the head groups mainly constitute the polar region. The polarity monitored by
the dye molecules used in the present study corresponds to this region of the micelles. The solvation
properties in the micellar phase arise partly due to the presence of bound water molecules and coun-
terions. The head group in an SDS micelle (–O–SO3−) can bind more water molecules than a head
group in CTAB (–N(CH3)3+). In the former case, the negative charge is distributed over the oxygen
atoms, while in the case of (–N(CH3)3+), the intervening methyl groups restrict the water molecules
from coming closer to the head group. Moreover, the counterion (Na+) in the case of SDS can bind
more water molecules by solvation than the corresponding counterion (Br−) in CTAB. The greater
HBD ability of the anionic SDS micelle relative to the cationic CTAB micelle can be rationalized
due to the presence of a larger number of bound water molecules in the Stern layer of the micelle.
On the other hand, the TX100 monomer surfactant molecule possesses some HBD strength due to
the presence of –OH groups. Thus, a higher value of R is expected for TX100 micelles. The higher
HBA ability of SDS relative to CTAB micelles can also be rationalized in terms of the presence
of water molecules as explained earlier. Moreover, in the case of SDS, the presence of (–O–SO3−)
head groups adds to the HBA basicity of the micelle formed. The head group of the neutral TX100
surfactant contains a phenyl group followed by a polyoxyethylene chain terminating in a hydroxyl
group. Thus, the polar region is expected to resemble an alcohol environment. Also, the determined
solvatochromic parameters are close to that of ethanol (α = 0.86, β = 0.75, π* = 0.54), which closely
resembles the chain end of the surfactant.
ACKNOWLEDGMENTS
The author thanks coworkers Dr. Mrinmoy Shannigrahi, Dr. Angahuman Maitra, Dr. Nipamanjari
Deb, Dr. Niraja Kedia, and Dr. Amrita Sarkar. The author is also grateful to Dr. Sanjib Sardar and
Dr. Amrita Sarkar for their sincere effort in preparing the manuscript.
478 Handbook of Surface and Colloid Chemistry
REFERENCES
1. E. M. Kosower, An Introduction to Physical Organic Chemistry, Wiley, New York (1968).
2. R. R. Doganadze, E. Kalman, A. A. Kornyshev, J. Ulstrup, eds., The Chemical Physics of Solvation,
Elsevier, Amsterdam, the Netherlands (1985).
3. A. Eilmes, Ab initio molecular dynamics simulations of ketocyanine dyes in organic solvents. In: Bubak,
M., Szepieniec, T., Wiatr, K. eds., Building a National Distributed e-Infrastructure—PL-Grid, Springer,
Berlin, Germany (2012), p. 276.
4. C. Reichardt, T. Welton, Solvents and Solvent Effect in Organic Chemistry, 4th edn., Wiley-VCH,
Weinheim, Germany (2011).
5. I. A. Koppel, V. A. Palm. In: N. B. Chapman, J. Shorter, eds., Advances in Linear Free Energy
Relationships, Chapter 5: Influence of Solvents on Chemical Reactivity, Plenum Press, London, U.K.
(1972).
6. R. M. C. Goncalves, A. M. N. Simoes, L. M. P. C. Albuquerque, J. Chem. Soc., Perkin Trans. 2 (1990)
1379.
7. G. Rauhut, T. Clark, T. Steinke, J. Am. Chem. Soc. 115 (1993) 9174.
8. C. Reichardt, Chem. Rev. 94 (1994) 2319.
9. E. M. Kosower, J. Am. Chem. Soc. 80 (1958) 3253.
10. K. Dimroth, C. Reichardt, T. Siepmann, F. Bohlmann, Liebigs Ann. Chem. 661 (1963) 1.
11. M. H. Abraham, P. L. Grellier, J. L. M. Abboud, R. M. Doherty, R. W. Taft, Can. J. Chem. 66 (1988)
2673.
12. R. W. Taft, J. L. M. Abboud, M. J. Kamlet, M. H. Abraham, J. Soln. Chem. 14 (1985) 153.
13. C. O. Rangel-Yagui, A. Pessoa, Jr., L. C. Tavares, J. Pharm. Pharm. Sci. 8(2) (2005) 147.
14. M. F. Vitha, J. D. Weckwerth, K. Odland, V. Dema, P. W. Carr, J. Phys. Chem. 100 (1996) 18823.
15. M. F. Vitha, P. W. Carr, J. Phys. Chem. B 102 (1998) 1888.
16. E. Fuguet, C. Rafols, M. Roses, Langmuir 19 (2003) 6685.
17. E. Fuguet, C. Rafols, E. Bosch, M. Roses, Langmuir 19 (2003) 55.
18. M. Shannigrahi, S. Bagchi, J. Phys. Chem. B 108 (2004) 17703.
19. D. Berillo et al., J. Colloid Interface Sci. 368 (2012) 226.
20. Y. Moroi, Micelles: Theoretical and Applied Aspects, Plenum, New York (1992).
21. M. M. Breuer, I. D. Robb, Chem. Ind. 13 (1972) 530.
22. M. J. Rosen, Surfactants and Interfacial Phenomena, Wiley Interscience, New York (1978).
23. D. Attwood, A. T. Florence, Surfactant Systems, Chapman & Hall, New York (1983).
24. E. G. Cockbain, Trans. Faraday Soc. 49 (1953) 104.
25. M. J. Schwuger, J. Colloid Interface Sci. 43 (1973) 491.
26. E. D. Goddard, K. P. Anantapadmanabhan, Application of Polymer Surfactant Systems, Marcel Dekker,
New York (1998).
27. E. D. Goddard, Colloids Surf. 19 (1986) 255.
28. M. N. Jones, Chem. Soc. Rev. 21 (1992) 127.
29. E. Dickinson. In: E. D. Goddard, K. P. Anantapadmanabhan, eds., Interactions of Surfactants with
Polymers and Proteins, CRC Press, Boca Raton, FL, Chapter 7: Proteins in solution and interfaces
(1993).
30. (a) N. J. Turro, A. Yekta, J. Am. Chem. Soc. 100 (1978) 5951; (b) D. Varadea, T. Joshi, V. K. Aswal, P. S.
Goyal, P. A. Hassand, P. Bahadur, Colloid Surf. A: Physicochem. Eng. Aspects 259 (2005) 95.
31. J. W. McBain, Colloid Science, D. C. Heath and Co., Boston, MA (1950).
32. G. S. Hartley, Aqueous Solution of Paraffin Chain Salts, Hermann and Cie, Paris, France (1936).
33. A. D. Abbot, H. V. Tartar, J. Phys. Chem. 59 (1955) 1193.
34. H. Schott, J. Pharm. Sci. 62 (1973) 162.
35. C. Tanford, J. Phys. Chem. 76 (1972) 3020.
36. R. G. Alargova, K. D. Danov, P. A. Kralchevsky, G. Broze, A. Mehreteab, Langmuir 14 (1998) 4036.
37. W. Philippoff, J. Colloid Sci. 5 (1950) 169.
38. W. D. Harkins, R. Mittelmann, J. Colloid Sci. 4 (1949) 367.
39. S. S. Berr, J. Phys. Chem. 91 (1987) 4760.
40. S. S. Berr, E. Caponetti, R. R. M. Jones, J. S. Johnson, L. J. Magid, J. Phys. Chem. 90 (1986) 5766.
41. G. D. J. Phillies, J. E. Yambert, Langmuir 12 (1996) 3431.
42. K. Bhattacharyya, in: V. Ramamurthy, K. S. Schanze, eds., Organic Molecular Photochemistry, Molecular
and Supramolecular Photochemistry Series, Vol. III, Marcel Dekker, New York (1999).
43. K. Bhattacharyya, Acc. Chem. Res. 36 (2003) 95.
Solvation in Heterogeneous Media 479
93. J. H. Park, A. J. Dallas, P. Chau, P. W. Carr, J. Phys. Org. Chem. 7 (1994) 757.
94. M. Shannigrahi, S. Bagchi, J. Phys. Chem. B 109 (2005) 14567.
95. D. Banerjee, P. K. Das, S. Mondal, S. Ghosh, S. Bagchi, J. Photochem. Photobiol. A 98 (1996) 183.
96. Y. Marcus, J. Chem. Soc., Perkin Trans. 2 (1994) 1751.
97. D. Banerjee, A. K. Laha, S. Bagchi, Indian J. Chem. 34A (1995) 94.
98. D. Banerjee, A. K. Laha, S. Bagchi, J. Photochem. Photobiol. A 85 (1995) 153.
99. P. K. Das, R. Pramanik, D. Banerjee, S. Bagchi, Spectrochim. Acta A 56 (2000) 2763.
100. A. Maitra, N. Deb, S. Bagchi, J. Mol. Liq. 139 (2008) 104.
8 State-of-the-Art Review
Water Purification Devices
CONTENTS
Nomenclature.................................................................................................................................. 482
Letters ................................................................................................................................. 482
Greeks ................................................................................................................................. 483
8.1 Bubble-Film Extraction as the Functional Constituent of Innovative Water Purifiers........ 483
8.2 Separation of Impurities in Known Water Purification Devices.........................................484
8.2.1 Functioning of a Sand Filter....................................................................................484
8.2.2 Filtration through Activated Carbon....................................................................... 488
8.2.3 Adsorption Kinetics on Activated Carbon.............................................................. 489
8.2.4 Adsorption Wave in Carbon Filter........................................................................... 490
8.2.5 Microbiological Activity in Filtration System......................................................... 492
8.2.6 Flotation Treatment of Polluted Water.................................................................... 494
8.3 Modern Technique of Water Treatment............................................................................... 495
8.4 Tap Water Quality Problems Still Remain.......................................................................... 499
8.5 Water Treatment by Sand Filtration in Recirculation Mode and Its Features..................... 499
8.6 Synergistic Effect of Filtration and Flotation Combination at Water Treating in
Recirculation Mode............................................................................................................. 502
8.7 Step-by-Step Description of Designed System.................................................................... 505
8.7.1 Constant Rate of Water Clarification at Direct-Flow Sand Filtration and
Calculation of Recirculation Filtration Unit Parameters......................................... 505
8.7.2 Constant Rate of Water Biofiltration at Direct-Flow Sand Filtration and
Calculation of Recirculation Biofiltration Unit Parameters....................................506
8.7.3 Generalized Constant Rate for Water Flotation and Calculation of
Recirculation Bubble-Film Extraction.....................................................................509
8.7.4 Water Degassing While Bubbling and Its Saturation by Atmospheric Oxygen...... 512
8.7.5 Water Desalination through the Bubble-Film Extraction Process.......................... 515
8.8 Schemes of Designed Devices and Their Operation Mode................................................. 520
8.A Appendix A.......................................................................................................................... 523
8.A.1 Equilibrium and Dynamic Adsorption of Surfactants at the Air–Water Interface............. 523
8.B Appendix B.......................................................................................................................... 530
8.B.1 Fast and Slow Surfactant Adsorption at the Air Bubble Surface and Transfer of
Adsorbed Matter Outside the Floated Water Bulk.............................................................. 530
8.B.2 Slow Surfactant Adsorption at the Air Bubbles and Transport of Adsorbed Molecules
out of the Floated Bulk........................................................................................................ 531
8.C Appendix C.......................................................................................................................... 536
8.C.1 Calculation of the Degree of Surface Filling by SAS Molecules Adsorbed from the
Side of Air............................................................................................................................ 536
481
482 Handbook of Surface and Colloid Chemistry
8.C.2 Calculation of the Degree of the Water–Air Interface Filling with SAS Molecules
Adsorbed from the Side of Water........................................................................................ 537
8.C.3 Ratio of Equilibrium Adsorption Constants as an Indicator of Surfactant Adsorption
Efficiency from Water Bulk and Air................................................................................... 538
8.C.4 Filling of Air–Water Interface by Surfactant Molecules Released from SAS Aerosol
Dispersed in Air.................................................................................................................. 539
8.D Appendix D..........................................................................................................................540
8.D.1 Property of the Double Electric Layer Formed at the Air–Water Interface by Dipoles
of Water Molecules, Hydroxyls, and Protons......................................................................540
References....................................................................................................................................... 542
NOMENCLATURE
Letters
a, b Filtering parameter, m−1
B′ Correlation coefficient, m3/kg
C Concentration, kg/m3
dG The quantity of sediment, formed in elementary filter’s cell, kg/m3
D Diffusion coefficient, cm2/s
F Volumetric flow rate, m3/s
F1 Airflow rate for bubbling, m3/s
G Surfactant mass flow, kg/s
i and j Anion/cation concentration ratio of fatty acid, m−3
j The flow of ions, mol/s
J Volatile substances flow, g/s
K Boltzmann constant, J/mol∙K
K
Integrated rate constant of the process, m3/s
K A and K AS Equilibrium dissociation constants of fatty acid, m3/mol
Kc The rate constant of the process, s−1
Keq (B) Adsorption equilibrium constant, kg/m3
Ki Rate constant of separate stages of water treatment, s−1
l The height of the filter, m
m Mass of the adsorbed molecules, kg
M Mass of bioconsumable impurities, kg
n The number of filters in the system
ns The adsorption capacity of i-type impurities
n∗s Number of adsorbed moles of i-type impurities
N Avogadro’s number, mol−1
Nα Total number of molecules adsorbed on the surface
NS Maximum sites that can be occupied by adsorbed molecules
P Partial pressure of gas, N/m2
Q Heat of adsorption, J/mol
r Air bubble’s radius, m
S Cross-sections area, m2
Sbio Total area of biofouling, m2
S i Specific area in monolayer space, m2
Si The surface area of one grain of sand, m2
Sp The surface area of porous media, m2
T Temperature, K
V Volume, m3
Water Purification Devices 483
V Sand volume, m3
Vi Volume of one grain of sand, m3
Vad The rate of adsorption, n∙ s−1
Vd The rate of desorption, n∙ s−1
w Superficial velocity of a fluid, defined as the volumetric flow rate divided by the
cross-sectional area of the filter, m/s
w1 Floating velocity of air bubbles, m/s
x Finite mass of bacteria, kg
x0 Initial mass of bacteria, kg
Greeks
α The portion of molecules that can be adsorbed
Γ Amount adsorbed surfactant, kg/m2
Γ∞ Maximum surfactant amount adsorbed at gas–liquid interface, kg/m2
ΔKi Efficiency increment of the separate stage at the expense of feedbacks, s−1
δ The thickness of adsorbed monolayer, m
δ1 The thickness of the diffusion layer, m
η, ϕ Empirical constants with dimensions, s∙ m−1 and s, accordingly
θ Part of the occupied sites
θp The degree of adsorbent surface filling by adsorbed molecules
μ Specific rate of biomass growth, s−1
ν0 The number of adsorption sites per unit surface area of adsorbent
ν1 Number of collisions between molecules of adsorbate with a unit of adsorbent surface
ν2 Number of molecules that are desorbed from adsorptive layer per unit time
σ The surface charge density, C/m2
τ Time, s
τ1 Duration of the time of one cycle, s
ϕ The electrostatic potential, V
Ω Total area of adsorption surface, m2
and thus, adsorbed substances evacuated more efficiently out of the treated water. The driving force
of this process is the compressed air, which is released from air bubbles inside the discharge chan-
nel. If necessary, one evacuation channel can be replaced by several channels.
Innovation was applied to prevent redispersion of floated concentrated solution from the sur-
face into the bulk water. As a result, the advanced flotation method was developed. The surfactant
concentration in water could be reduced to the level of hundredths of milligrams per liter. This
concentration is significantly less than that was achieved by other evacuation methods.
Bubble-film extractors were designed to remove dissolved endogenous and exogenous surfac-
tants from water. These devices have found practical application for quality improvement of tap
water at the points of its consumption. They were particularly useful when operating at peripheral
points of long pipelines.
The tests have shown that the use of bubble-film extractors is the best way to purify tap water from
pipelines’ corrosion products and other mineral admixtures, as well as from trihalomethanes (chlori-
nated derivatives) and different kinds of surface-active contaminants present in tap water. But water
purification degree from mineral impurities, especially from colloid rust, by means of bubble-film
extraction depends on concentration in the water of endogenous surfactants. These substances are
always present in tap water as a result of biofouling inside pipelines. When purifying water by bubble-
film extraction, the endogenous surfactants act as collectors and carriers of finely dispersed hydrox-
ides, oxides, and heavy metal ions. They also show the flocculation properties. Concentration of
endogenous surfactants in tap water varies with seasons: their maximum concentration was observed
in the summer months, and the minimum in winter [4]. To improve purification of tap water in the
case of lack of impurities complementary to endogenous surfactants (during the cold times of year), a
special synthetic surface-active cationic polyelectrolyte—polyhexamethylene guanidine hydrochlo-
ride (PHMGH)—can be added [3,5,6]. Its recommended dose is 2.5 mg/1 dm3 of water to be purified,
when the molecular mass of oligomers ranges from 1000 to 5000. PHMGH is the exogenous SAS.
Another approach consists in the compensation of endogenous surfactant deficiency in treated
water by means of biofouling incubation inside the filtration load. As was shown in [4], the surface-
active metabolic products of different bacteria cells act as cofactors of bubble-film water treatment
process. Rapid accumulation of endogenous surfactants in aqueous media occurs under favorable
ecological conditions. An appropriate niche for aerobic reproduction of heterotrophic bacteria is any
well-developed nontoxic granular material through which contaminated and oxygen-enriched water
passes. Here, the water contaminants are a source of feed for bacteria cells.
Hence, an implementation option of water treatment process on the principle of rotating wheel
has been substantiated. The main principle of such water treatment process is the recirculation of
treated water through cartridge packed with fine grain sand or charcoal and then through bubble-
film extraction module. Two units (sand or charcoal cartridge and bubble-film extractor), assembled
together and used for water treating in recirculation mode, form an emergent water purification
device. Functioning of this device as a whole has intensive influence upon the functionality of the
device’s parts, what makes the device as a whole more sustained.
S
w
dl l
C–dC
S ⋅ w ⋅ C ⋅ dτ − S ⋅ w ⋅ (C − dC ) ⋅ dτ = S ⋅ dl ⋅ dG, (8.1)
dl ⋅ dG
C − (C − dC ) = , (8.2)
w ⋅ dτ
or
dG dC
= w⋅ . (8.3)
dτ dl
As appears from (8.3), at a given filtration rate of turbid water through the sand filter, the effective-
ness of water clarification is determined by dC/dl. The high rate of water clarification is obtained at
high magnitudes of dC/dl, and at dC/dl → 0, the water clarification process has stopped.
486 Handbook of Surface and Colloid Chemistry
According to Mints’s theory, two different processes take place in each functioning layer of
the sand filter: (1) adhesion of sediment and colloidal particles onto filter grains and previously
entrapped impurities; (2) detachment of previously adhered particles by hydrodynamic forces
and their redispersion into filtered water. Mathematically, these phenomena are described as the
following:
where
dC1 is the decrease in dispersed impurity concentration in the filtering flow at the expense of
their adhesion to the grains of filtering material
dC2 is the increase in dispersed impurity concentration in the filtering flow at the expense of
particle detachment
The precipitation of dispersed particles onto grains of filter material is controlled by the gravi-
tational and hydrodynamic clamping forces, diffusion limitation, and Coulomb and dispersion
interactions [7,8]. Mechanisms of particle adhesion onto filter grains are described by theories of
Deryagin, Landau, Fairway, Overbek, and Dukhin [9]. To make the adherence of the dispersed
particles to the filter grains possible, the total sum of vector forces in regard to interacting objects
should result in the reduction of free energy of filtration system. Detachment of particles of dis-
persed phase from formed sediment increases the free energy of this system. The detachment
rate is increased as treated water is pumped faster through the pore space of filter grains. Thus,
the increase in the dispersed particle concentration, adherent to the filter grains in elementary
layer with thickness dl, is directly proportional to these particles’ concentration in filtered water
and inversely proportional to their filtration rate. It also depends upon the properties of the filter
packing grains and material impurities. To take these factors into account, the proportionality
coefficient (1/b) is added to the first component of the right-hand part of Equation 8.4, where b is
the filtering parameter. It is equal to the ratio of sedimentation rate constant and the surface water
flow w, which is defined as the volumetric flow rate divided by the cross-sectional area of filtration
bed. So, we shall obtain
And the amount of particles leaving the coagulated sediment in the filtering layer, whose thickness
is equals to dl, is proportional to the sediment mass G, the water flow rate, and the magnitude of
the adhesion forces between the particles in the sediment. All these factors are taken into account
by coefficient a, where a is the filtering parameter equal to the ratio of constant of liftoff speed of
sedimenting particles to the overall filtration rate, that is,
dC
= b ⋅ C − a ⋅ G. (8.7)
dl
∂ 2C ∂C ∂C
=b −a . (8.8)
∂l ⋅ ∂τ ∂τ ∂l
Water Purification Devices 487
And taking into account Equation 8.3, one can rewrite (8.8) as the following:
∂ 2C ∂C ∂C
+ a⋅w⋅ −b = 0. (8.9)
∂l ⋅ ∂τ ∂l ∂τ
Equation 8.9 is the main differential equation of filtration processes of dilute aqueous suspension
through granular packing. Integration of Equation 8.9 gives the solution as an infinite series [10],
and it is not useful for practical application. Therefore, it is expedient to use Equation 8.7 and get its
analytical solution for two extreme cases.
For the case of filter-fresh operating, we can consider that aG is negligible as compared with bC,
and when dead filter operating, the filtering conditions become such that aG is significantly greater
than bC. As a result, one can monitor the clarification efficiency of flow water, depending on the
filtering layer thickness at different hydrodynamic conditions.
Integrating Equation 8.7 between l = 0 and l, and between C = C0 to C for the case when bC is
significantly greater than aG, we shall get
C K
ln = −b ⋅ l = − c ⋅ l, (8.10)
C0 w
where Kc is the constant of settling process for filtered particles in the filter space.
Or
⎛ K ⎞
C = C0 exp(−b ⋅ l ) = C0 exp ⎜ − c ⋅ l ⎟ . (8.11)
⎝ w ⎠
And integrating Equation 8.7 in the same limits but for the case when bC is substantially less than
aG, we shall get
C
ln = a ⋅ G, (8.12)
C0
or
C = C0 exp(a ⋅ G ). (8.13)
Comparison of Equations 8.11 and 8.13 reveals that the concentration of dispersed phase in filtrate is
inversely proportional to the height of filtering layer in the case when adhesion of dispersed particles
to the filters grains is the dominated process. And when the settled particles pass into filtrate, then
the degree of water flow turbidity depends exponentially upon the amount of accumulated sediment.
This is the basic phenomenon, which occurs at filtration using sand and other granular packing.
Granular packing adsorbs dispersed impurities according to gravitational deposition mechanisms,
inertial and non-inertial heterocoagulation at low and medium filtration rates. And accumulated
impurities turn to the filtrate flow at very high filtration rate, or when intensive water backflow is
pushed up through the filter material forcing the sand particles into suspension.
According to most of the scientific studies and practical use of sand filters, the optimal height of
filtering layer ranges from 0.8 to 1.5 m. Particles of sand range in size from 0.3 to 1.4 mm in diam-
eter, and filtration rate should not exceed 10–20 m/h.
Backwashing of packing material is usually realized by increasing the velocity of reverse water
flow at which water passes back through the filter. This, in effect, blasts the clogged particles off of
488 Handbook of Surface and Colloid Chemistry
the filter. But in so doing, filtration velocity should not reach the free fall rate of sand particles in
filtrated flow in order to avoid the sand loss.
n*s K eq ⋅ C
= , (8.14)
ns 1 + K eq ⋅ C
where
n∗s is the number of moles of i-type impurities adsorbed per unit weight of activated carbon
ns is the adsorption capacity of activated carbon with respect to i-type impurities
Keq is the adsorption equilibrium constant
K eq ⋅ C
θp = . (8.15)
1 + K eq ⋅ C
The value of Keq, which appeared in Equations 8.14 and 8.15, is an exponential function of the ratio
of heat of adsorption Q of i-type impurities to the thermal energy RT, that is,
⎛ Q ⎞
K eq = Bʹ ⋅ exp ⎜ ⎟ , (8.16)
⎝ RT ⎠
The technological practice of the use of coal sorbents is based on these fundamentals. The
adsorption process is readily effected at room temperatures, and desorption is realized at high
temperatures. So, the adsorbent regeneration is carried out at high temperatures in vacuum or by
superheated steam treatment.
ν 0 ⋅ dθ
= Vad − Vd = α ⋅ ν1 ⋅ (1 − θ) − ν 2 ⋅ θ, (8.17)
dτ
where
τ is the adsorption process time
ν0 is the number of adsorption sites per unit surface area of adsorbent
or
dθ α ⋅ ν1 − α ⋅ ν1 ⋅ θ − ν 2 ⋅ θ
= = x. (8.18)
dτ ν0
dx α ⋅ ν1 + ν 2
=− = K c , (8.19)
dθ ν0
dx = − K c dθ. (8.20)
∫
0
∫
θp
x = − K c ⋅ (θ p − θ). (8.21)
Substituting the value x from (8.21) into (8.18), we shall get the required differential equation for
adsorption rate in the terms θ, Kc, τ:
dθ
= − K c ⋅ (θ p − θ), (8.22)
dτ
490 Handbook of Surface and Colloid Chemistry
or
dθ
= − K c ⋅ dτ. (8.23)
θp − θ
θp − θ
ln = − K c ⋅ τ, (8.24)
θp
or
In (8.25), the value θp is subject to (8.15), and therefore, θ depends on the product of two exponential
functions. Namely, it depends on the exponential member characterizing the gain of free energy in
the transition of adsorptive molecules from the environment to the adsorbent and from the expo-
nential function that consists of constant of transfer rate of adsorptive molecules to adsorbent. This
constant depends on the coefficients of surface and bulk diffusion of substances.
Thus, when in the medium with high concentration of adsorbate, which is characterized by high
adsorption heat, the adsorption process occurs quickly, and additional accumulation of adsorptive
molecules on the adsorbent surface is impossible. In practice, water is treated using adsorbents
in flowing column taking into account the adsorption rate and the stoichiometry of the saturation
adsorbent capacity.
a b c d
C0 C0 C0 C0
Ca Cb Cc Cd
C0
Concentration of adsorbate
Cd
tra on
n
cen luti
tio
in the output fluid
con let so
t
Ou
Cc
Ca Cb Breakthrough point
0 Volume of output fluid
FIGURE 8.2 The picture of MTZ motion across the adsorption bed. (a through d) The steps of the
MTZ motion across the adsorption bed (black—completely exhausted section, white—section with fresh
adsorbent).
a b c
t1 t1
t2
Height of the filter.
t2
t3
t1
t4
t3
t5
t6
FIGURE 8.3 The shapes of breakthrough curve on activated carbon for the substances with different adsorp-
tion heat. The values of adsorption heat: (a) high, (b) average, and (c) small.
(Figure 8.3a). Adsorption of proteins, lipids, polypeptides, and many other organic substances is
characterized by proportional tailing in time breakthrough curves (Figure 8.3b). Glucose, alco-
hol, and sucrose are practically unadsorbed on charcoal, and the adsorption wave is not observed
(Figure 8.3c).
The time required for the wave to reach the end of the filter bed characterizes the adsorption
capacity of the filter. This time depends on the height of packed bed, the quality of adsorption mate-
rial, concentration of adsorbate in the fluid phase, and the filtration rate. Sometimes the filtration
rate is substituted by total amount of solution that will pass through the filter with the requirement
492 Handbook of Surface and Colloid Chemistry
α
K = tg α
FIGURE 8.4 The filter protective effect depicted as a function of the height of packed bed.
that the concentration of adsorbate in the fluid will be decreased to the given level. Shilov equation
gives the relationship between adsorption capacity of the filter and height of packing:
τ = η⋅l − ϕ, (8.26)
where
τ is the exploitation time for filter (from new filtering material up to breakthrough point)
l is the height of packed bed
η, ϕ are the empirical constants depending upon the adsorbent type and its particle size, the type
and concentration of the adsorbate, and the filtration rate
Duration of filter protective effect as a function of the height of packed bed is shown in Figure 8.4.
When adsorption capacity of the filter reaches its limit (see Equations 8.14 and 8.25), no more
adsorption can take place in the bed.
At the presence of dissolved oxygen in filtered water, aerobic heterotrophic bacteria grow fast,
and as a result, the biofilm is rapidly formed on the grain surface of activated carbon. The phenom-
enological equations are as follows:
5
C6H12O6 + NH 3 → C5H 7NO2 + 3H 2O, (8.28)
6
⎛ K ⋅l ⋅ S ⎞
C = C0 ⋅ exp ⎜ − c , (8.30)
⎝ F ⎟⎠
where
Kc is the rate constant of biofiltration
F is the filtered water flow
l is the height of packed filter
S is the cross-sectional area of the biofilter
The first systematic tests of granular activated carbon as a biofilter were carried out in 1975 in
Germany. Mechanically pretreated water flow was directed through the activated carbon filter.
494 Handbook of Surface and Colloid Chemistry
Filtration velocity was 10 m/h, height of packed bed was 6 m, and filter diameter was 1 m. After a
while, a steady-state velocity of organic contaminant decomposition was reached. It was observed
that 1 m2 of surface area of activated carbon decomposes about 60–100 g of organic substances
with respect to dissolved organic carbon at the daily oxygen consumption 200–250 g. Comparison
of water quality before and after treatment has shown that the activated carbon not only is a vital
medium for microorganisms, but also acts as a buffer at sharply increasing concentration of organic
impurities in water.
⎛ K ⋅l ⋅ S ⎞
C = C0 ⋅ exp ⎜ − c , (8.31)
⎝ F ⎟⎠
where
C is the concentration of impurities in the outlet of the flotation column
C0 is the concentration of impurities in the inlet of the flotation column
Kc is the flotation constant
l is the height of flotation column
F is the flow rate of treated water
S is the cross-sectional area of the flotation unit
This equation is identical to Equation 8.11, which describes the work of a sand filter.
Water Purification Devices 495
FeCl3 NaOH
Filtration
Retainer
Borehole
Well pump
Diversion Pumping
station Fast
First filter Second Adsorption
ozonation ozonation filter
FIGURE 8.6 Technological scheme of drinking water preparation from Lake Zurich.
treatment technology described earlier is the approach that allows the production of drinking water
without the use of chlorine.
In modern water treatment technologies, the slow filters should preferably be located at the
final stages of the multibarrier process. This allows to essentially reduce the contaminant load
on biofilm thereby ensuring effective operation of the sand filters minimizing the numbers of
backwashings.
The slow filtration, if properly carried out, provides the high-quality drinking water. But the
use of slow filters is possible only in the presence of the required areas at the water treatment
plant.
In Switzerland, the water from Lake Zurich is treated by a multibarrier technology, which is
shown in Figure 8.6. According to this technology, the water is treated first with ozone, then filtered
through a fast filter, after which the water is reozonated, and then filtered through charcoal. After
these operations, the water flows through the final slow filter.
The first step of ozone treatment (1.1 mg O3/dm3 ± 25%) is used for disinfection and flocculation
of water contaminants before its fast filtering. Fast filters remove more than 95% suspended solids.
The next step of ozonation is carried out using 0.5 mg O3/dm3, and then water is treated by slow
filters with activated carbon.
The filtration zone consists of four layers of gravel and sand of different sizes, with the thickest
gravel forming the bottom layer and the finest gravel the top layer: a layer of fine sand (50–85 cm), a
layer of coarse sand (5 cm), a layer of fine gravel (5 cm), and a layer of coarse gravel (5 cm).
The main problem of sand filter operation is that these are prone to clogging. In the considered
technology, the filters were tested under two operating conditions: at flow rates 0.5 and 4–7 m/h. In
the first case, all tested filters have operated more than 9 years with periodic backwashing one time
every 2 or 3 years. In the second case, the filters have operated without the need for cleaning at a
period of 8 years. In operation, the hydraulic permeability of the filters was gradually reduced to
0.5 m/h, as a result of an increase in clogging. Therefore, the filters were washed, and the first 5 cm
of the sand layer were removed and replaced by new sand. This has made it possible to increase the
hydraulic permeability of filters by 1–2 m/h. However, the filtration capacity cannot be regenerated
completely by this method due to algae clogging of the sand filter bed. At least 10% of the pore
space in the top layer of filter bed was filled by biomass.
The total height of the sand bed reaches 100 cm. After 10–15 years of operation, the sand in the
filter must be regenerated. In Zurich, at the water treatment plant, there are 14 slow filters. The area
of the filter unit is 1.12 m2, the maximum filtering capacity of each unit is 16 m3/m2/day, and the
operating conditions are 4°C–8°C in the dark. Whole system provides about 250 m3 of clean water
per day.
An example of the drinking water production from highly polluted surface waters is the technol-
ogy patented in the United Kingdom. Its scheme is shown in Figure 8.7.
First, contaminated surface water is pretreated by coagulation and collected in a storage tank.
Then, water is filtered through the coarse filter. Next, the water passes through a two-stage filtration
Water Purification Devices 497
Coagulant batcher
UV-fabrication
Conditioning of water Reverse osmosis Adsorption filters
unit, using coagulation as a pretreatment process. The several steps of adsorption follow this on
activated carbon, and then the water is disinfected with ultraviolet rays. In the final stage, the reverse
osmosis system consisting of two units is used to purify water (one unit in service when one is
regenerating). Naturally, the end step of the process is the additional mineralization of treated water.
In Namibia, the surface water is mainly used for drinking water preparation. The technologi-
cal scheme is shown in Figure 8.8. It includes the following processes: ozonation at the first step,
then flocculation, flotation, and filtration through a double-layer filter. Next, the water is ozonated
again, and then it is filtered through the two-stage carbon filter and treated to safety standards with
ultraviolet.
Figure 8.9 describes the basic steps of water treatment process, which is briefly described in [16].
The water enters the plant and goes into accumulation tank. The treatment of water includes
mechanical filtration, coagulation and flocculation, flotation, and then water is transported to
biological treatment unit, where the organic impurities are decomposed due to heterotrophic
Coagulant batcher
Two-bed filter
Water source
Ozonation Ozonation
Coagulation
–flocculation Flotator
Coagulant batcher
Pump station
Water storage
Coarse filters Fine filters Biological treatment
Flotator
Drinking water
FIGURE 8.9 Scheme of drinking water preparation with the use of biofilters.
Coagulant batcher
FeCl3 FeCl3
Settler Coagulation
Ozonation Biolite filter
bacterial activity. Then, the treated water is filtered again to remove the trace amounts of sludge,
and after that, it is ozonated. Then, water flows through activated carbon filters (I and II steps). At
the first step, the organic compounds within a certain range of molecular weights and, at the second
step, the trace amounts of pesticides and other toxic substances are removed. The final step is water
processing by reverse osmosis.
At the drinking water treatment plant in the city of Ivry-sur-Seine (France), a multistage complex
technology is also used (Figure 8.10).
The pretreatment process comprises ozonation and then coagulation. Next treating processes
are primary filtration stage through biolith, and a secondary filtration stage through sand, with the
possibility of coagulation. The finishing treatment includes ozonation, biological filtration through
granular activated carbon, and disinfection with the use of chlorine.
As is seen from the earlier examples, the highest degree of purity and safety of drinking water
requires complex water treatment technologies.
Water Purification Devices 499
F
K1 K2 K3
FIGURE 8.11 Direct-flow scheme of water treatment. F: flow of treated water; K1, K2, K3: adsorption coef-
ficients of separate stages.
500 Handbook of Surface and Colloid Chemistry
The rational way out at water treatment is to apply water recirculation system instead of the
direct-flow system. This enables the use of compact sand filters and biofilters with the gravitational
or floating packing, small flotators, aerators, etc. At the same time, the efficiency of water treatment
is greatly increased due to arisen positive feedbacks between functional units in the created system.
The features and benefits of recirculation method at water processing can be demonstrated by
taking as an example the sand filter. Traditionally, to be effective, sand filter should have the height
of sand bed to be 0.7–2.0 m, flow rates in slow sand filters should be between 0.1 and 0.2 m/h, and
that in quick sand filters should be between 5 and 20 m/h.
According to [10], the output/input ratio of impurity concentrations C/C0 for the granular sand
filter is described by the following equation:
C ⎛ K ⋅l ⎞
= exp ⎜ − c ⎟ (8.32)
C0 ⎝ w ⎠
where
Kc is the rate constant of contaminant adsorption in the porous space of the filter bed
l is the height of the filter bed
w is the face filtration velocity
If we consider the filtration rate w as the water flow related to the cross-sectional area of the filter,
then we will obtain
C ⎛ K ⋅l ⋅ S ⎞
= exp ⎜ − c , (8.33)
C0 ⎝ F ⎟⎠
V = F⋅τ, (8.34)
or
C ⎛ n ⋅ Kc ⋅ l ⎞
= exp ⎜ − , (8.36)
C0 ⎝ w ⎟⎠
Equation 8.36 shows that the concentration of impurities at the outlet of the filtration system of
n filters connected in series is an inverse exponential function of the number of connected filters.
But the same equation describes the efficiency in a closed-loop filtration system with a single filter,
which has the following parameters: height of filter bed l, surface filtration rate w, and the ratio C/C0
determined by the nth filtration cycle. Therefore, for water volume V, filtered under recirculation
mode through the sand filter with bed height l and cross-section area S during cycle time τ1, the ratio
C/C0 is determined by the following equation [18,19]:
C ⎛ K ⋅ l ⋅ S ⋅ τ1 ⎞
= exp ⎜ − c ⎟ , (8.37)
C0 ⎝ V ⎠
where τ1 is the cycle time required for the realization of one filtration cycle.
And if the number of filtration cycles will be n, one shall obtain
C ⎛ n ⋅ K c ⋅ l ⋅ S ⋅ τ1 ⎞
= exp ⎜ − ⎟ . (8.38)
C0 ⎝ V ⎠
As the product of n and τ1 is the real duration of the recirculation filtration process, we’ll have
C ⎛ K ⋅l ⋅ S ⋅ τ ⎞
= exp ⎜ − c ⎟ , (8.39)
C0 ⎝ V ⎠
K c ⋅ l ⋅ S K c ⋅ l1 ⋅ S ⋅ τ
= . (8.40)
F V
C l l ⋅τ
≈ = 1 . (8.41)
C0 F V
Thus, the quality of water treatment at direct-flow filtration is directly proportional to the ratio of
height of the filter l to flow rate F. And the quality of water treatment under recirculation filtration is
directly proportional to the time of filtration τ, to the height of the filter l1, and inversely proportional
to the volume V of filtered water. The dependence of the water clarification degrees upon the filter-
ing mode and filter dimensions is shown in Figure 8.12.
Equation 8.41 indicates that the small sand filters can be used for effective clarification of cer-
tain volumes of contaminated water. And the decrease in the filter height is compensated by the
recirculation mode of its operation. This statement is true relative to the other units of multibarrier
502 Handbook of Surface and Colloid Chemistry
C0 , F F
S
C1, V1
C1, V1 l
l1= 0, 1l
C*1, V *1
V1 = F τ V *1 = V 1
C1 = C0 exp (– Kc l S/F) C*1 = C1 = C0 exp (– Kc 0.1l S 10τ/V1)
(a) (b)
FIGURE 8.12 The scheme of water treatment processes: (a) direct-flow filtration; (b) recirculation filtration.
technological scheme, that is, to floaters, bioreactors, and so on. The earlier described mechanism
shows that direct-flow multibarrier water treatment system, which has a large size, can be replaced
by the compact prototype, operating under recirculation condition. The recirculation technology
provides a new quality as compared with once-through filtration, biofiltration, flotation, etc., where
each unit in technological scheme performs only its own function.
ΔK1
ΔK2
F 1 2 3 F
K1 K2 K3
ΔK3
F1
FIGURE 8.13 The scheme of flotation–filtration water purification process under recirculation conditions.
1: bio sand filter; 2: aerator; 3: bubble-film extraction unit; F: flow of the treated water; F1: recirculation flow;
K1, K2, K3: rate constants of separate stages of water treatment; ΔK1, ΔK2, ΔK3: efficiency increments of the
stages at the expense of feedbacks in recirculation system.
the output of the filter is also increasing. But these products are natural flocculants and gatherers
(collectors) of ionic impurities of water. They provide additional aggregation of colloidal particles
and heavy ions in the apparatus of flotation and thereby ensure reduction of the required quantity of,
or even refuse from the use of, synthetic surface-active flocculating agents and gatherers. Therefore,
the index K3, which reflects the quality of work of the flotation apparatus, is increased by the value
ΔK1, and we come to the result (K3 + ΔK1).
But due to flotation (bubble-film extraction), the products of bacterial metabolism and the
products of bacterial degradation together with other water contaminants are removed continu-
ously from recirculation flow through the bubble-film extractor. As a result, another positive
feedback is realized in filtration–flotation system. The essence of this effect lies in the inhibition
of vital functions of bacteria with increasing microbial metabolite concentration according to the
law of chemical kinetics. And in accordance with the same law, bacterial activity is increased
as the products of bacterial metabolism are removed from treated water [20]. Thus, we are able
to add one more component, namely, ΔK3, to the magnitude (K1 + ΔK2). The component ΔK3
represents the increase in biofiltration efficiency due to bacterial inhibitors removed by means
of bubble-film extraction. In such a way, the resulting rate constant of the process takes the form
(K1 + ΔK2 + ΔK3).
So, there is a great synergy between the biofilter and the flotation unit (bubble-film extractor) in
closed-loop flotation and filtration system, and they provide each other with additional water treat-
ment possibilities when operated in the recirculation mode. In other words, the faster biofouling
growth of the granular filter leads to increasing natural flocculant concentration in treated water,
thereby improving the effectiveness of bubble-film extraction. (And the bubble-film extractions,
which saturate the treated water by atmospheric oxygen and remove degradation bacterial products,
were enhancing biofilter operation.)
Thus, the mathematical description of the direct-flow sand filtration is the following:
C ⎛ K ⋅l ⎞ ⎛ K ⋅l ⋅ S ⎞
= exp ⎜ − 1 ⎟ = exp ⎜ − 1 . (8.42)
C0 ⎝ w ⎠ ⎝ F ⎟⎠
C ⎛ K ⋅l ⋅ S ⎞
= exp ⎜ − 3 . (8.43)
C0 ⎝ F ⎟⎠
504 Handbook of Surface and Colloid Chemistry
⎛ K ⋅l ⋅ S ⎞ ⎛ K ⋅l ⋅ S ⎞ ⎛ (K ⋅ l ⋅ S + K3 ⋅ l ⋅ S ) ⎞
C = C0 ⋅ exp ⎜ − 1 ⋅ exp ⎜ − 3 = C0 ⋅ exp ⎜ − 1 ⎟ . (8.44)
⎝ F ⎟⎠ ⎝ F ⎟⎠ ⎝ F ⎠
C ⎛ ( K + ΔK 2 + ΔK 3 ) ⋅ l ⋅ S ⋅ τ ⎞ ⎛ ( K + ΔK 2 + ΔK 3 ) ⋅ l ⋅ S ⎞
= exp ⎜ − 1 ⎟ = exp ⎜ − 1 ⎟ , (8.45)
C0 ⎝ V ⎠ ⎝ F ⎠
and
⎛ ( K + ΔK1 ) ⋅ l ⋅ S ⎞
C = C0 exp ⎜ − 3 ⎟ . (8.46)
⎝ F ⎠
⎡ ( K + ΔK 2 + ΔK 3 ) ⋅ l ⋅ S + ( K 3 + ΔK1 ) ⋅ l ⋅ S ⎤
C = C0 exp ⎢ − 1 ⎥ . (8.47)
⎣ F ⎦
Comparing Equations 8.44 and 8.47, one can see the positive effect of combination of filter and
bubble-film extractor into recirculating system. The reason is the positive feedbacks arising in this
system due to recirculation mode of water treatment.
It is very important to note that in the recirculation filtration–flotation system, there is no need to
use more than one filter. One filtration module can operate as well as several filters in series. When
the total surface area of the grains in the filter bed is sufficient, the adsorption of water contaminants
takes place simultaneously through the following mechanisms: by gravitational sedimentation,
by pore-blocking, by heterocoagulation, and by bioprecipitation [7,8,12,21]. In this case, the total
adsorption efficiency is subjected to Equation 8.47, and the residual concentration of contaminants
at water treating depends upon the working capacities of the filtering and flotation modules.
Thus, there are three positive feedbacks between the filter and the bubble-film extractor in
the described water treatment system. Two of them are directed from the bubble-film extractor
to the filter and the third from the filter to the bubble-film extractor. Due to these feedbacks, the
adsorption rate of both filtration and bubble-film extraction stages is increased significantly. In
such a way, the water treatment under recirculation mode shows a new quality, which can be named
as emergence.
The magnitude of feedback contribution depends upon the amount of bacterial biomass in the
filter bed and its lifetime as well as the presence of impurities in water, which support bacterial
growth [11,20]. The organic compounds, easily digested by bacterial enzyme systems, are first
adsorbed from contaminated water. And then, bacteria digest the more complex organic substances.
The recirculation filtration–flotation system allows different impurities (with respect to dispersion
degree and nature) to be removed simultaneously. These impurities are the organic and inorganic
compounds, inorganic sediments, colloid and true solution of dissolved surfactants, dissolved gases,
salts, etc. [1,2].
The other important feature of the recirculation filtration–flotation water treatment system is
the possibility to minimize the overall dimensions of equipment maintaining the required water
quality.
So, if the direct-flow sand filter of height l enables 10-fold decrease in the concentration of impu-
rities in treated water, then the same result can be obtained after 10 cycles of recirculation through
Water Purification Devices 505
the filter with a height 0.1l, at the same filtering rate. And flotation device or any other water treat-
ment unit is functioned similarly under recirculation mode. In this case, the residual concentration
of any kind of water impurity can be reduced to acceptable level, considering that kinetics of with-
drawal of i-kind component obeys the inverse exponent law.
The graded quartz sand with a size between 0.3 and 2.0 mm is used in direct-flow sand filters.
The height of sand bed in these filters is between 0.5 and 2 m. The process of water clarification is
realized as quick sand filtration or slow sand filtration depending on the fraction of used sand. In
the first case, the filtration rate is about 5.5–14 m/h, and in the second case, it is about 0.1–0.2 m/h.
The parameters of these processes were taken from [22–24]. They are represented in Table 8.1.
The depicted data allow to determine the rate constants of water clarification at quick and slow
sand filtration using Equation 8.32. For slow sand filtration with the height of packed bed equal to
0.5 m, the calculation gives the value of rate constant of water clarification equal to 7.7⋅10−4 s−1,
if the initial turbidity level is 1500 mg/L, finish turbidity level is 1.5 mg/L, and filtration rate is
0.2 m/h. And for fast sand filtration, the value of rate constant of water clarification is equal to
5.26⋅10−4 s−1, if water turbidity is reduced by 50%, the height of filter bed is 2 m, and filtration rate
is 5.5 m/h.
Comparing the obtained results, one can see that the difference between the rate constants
of water clarification for both types of filtration is insignificant. These constants depend upon
the nature of filter material and contaminants of water but not upon the quick or slow filtering
conditions.
506 Handbook of Surface and Colloid Chemistry
TABLE 8.1
Water Clarification Using Standard Direct-Flow Sand Filters
Type of Height of Degree of Water Filtration Rate, Filter Backwash
Filter Packed Bed, mm Grain Size, mm Clarification m/h Period
Fast 700–800 0.5–1.2 The water quality 5.5–14 8–12 h
1300–1800 0.7–1.6
1800–2000 0.8–2.0 Initial turbidity 150 mg/L
Final turbidity 45–75 mg/L
Slow 500 0.3–1 Initial turbidity 1500 mg/L 0.1–0.2 10–30 days
50 1–2 Final turbidity 1.5 mg/L
TABLE 8.2
Calculated Parameters of Water Clarification in Recirculation Water Treatment System
Cross Real Integrated Water
Height Section Constant of Constant of Water Treatment
Volume of Degree of of Filter of the Water Clarification under Period under
Treated Clarification, Bed, l, Filter, S, Clarification, Recirculation, K, Recirculation
Water, V, dm3 % dm dm2 Кc, s−1 dm3/s Conditions, τ, h
1 2 3 4 5 6 7
20 90 0.5 6 6.48·10−4 1.94·10−3 6.5
200 90 5 30 6.48·10−4 9.72·10−2 1.3
2000 90 5 30 6.48·10−4 9.72·10−2 13
Known value of the rate constant of water clarification at the direct-flow water filtration allows
one to calculate the filter dimensions due to (8.39) and use the result for design recirculation filtra-
tion module. The results of these calculations are represented in Table 8.2.
It is seen that the use of recirculation filtration allows one to use the sand filter with smaller
dimensions and to obtain the required degree of water clarification. In particular, to obtain 90%
of water clarification (C/C0 = 0.9) when treating from 20 to 2000 L of water at the rate constant of
water clarification equal to 6.48·10−4 s−1, the calculated dimensions of the sand filter and required
times of water treating are shown in the third, fourth, and seventh columns of Table 8.2. So, if one
has enough time to implement the water treatment process in recirculation mode, then one can
greatly reduce the dimensions of required devices, material consumption, and correspondingly, the
cost of the designed equipment.
Adsorption of bacterial cells from the filtered water to the grain surface in the filter material leads to
bacterial growth under favorable ecological conditions. As a result, the biofilms appear at each grain
surface. During the time, the biofilms gradually fill the entire surface of porosity space within the
filter body. Thickness of biofilms on the surface of filter grains ranges from 50 to 10,000 nm [20].
The vital activity of bacterial cells inside the space of biofouling (see Equations 8.27 through
8.29) leads to removal of impurities from filtered water, which is the substrate of microbial metabo-
lism. The products of bacterial metabolism and dead bacterial cells diffuse back into the filtered
water bulk.
Water Purification Devices 507
The process of biofilm growth obeys the law of chemical auto-synthesis [20]. Its basic equation is
where
x is the finite mass of bacteria
x0 is the initial mass of bacteria
μ is the specific rate of biomass growth
τ is the current time
The magnitude of μ is a constant value under favorable conditions for microbial vital activity, and
thus, the biofilm can growth exponentially.
But the sign and the magnitude of μ are changed at the deterioration of the factors influencing the
activity of bacteria, such as decreasing concentration of nutrient substrates, fluctuation in tempera-
ture, pH, and redox potential of the filtered water. In this case, the magnitude of μ may decrease to
zero. It is due to the lack of food for bacterial cells and their autolysis. So, there is a quick response
of bacterial cells to the change of external conditions. Those bacteria and their colonies are able to
survive, which have the largest amount of vital nutrients in the filtered water. Therefore, biofiltra-
tion takes an important part in modern water treatment technologies that is equivalent to water
self-purification.
The technical characteristics of some selected examples of industrial biofilters are shown in
Table 8.3. Using the data represented in this table, one can determine the rate constant of b iofiltration
and the productivity of biofiltration devices. For example, the rate constant of direct-flow biofiltra-
tion through sand filter is equal to 1.91⋅10−3 s−1 (see the first row of Table 8.3). This result was
obtained from Equation 8.30 at l = 2.5 m, V = 10 m/h, and C/C0 = 25/140.
TABLE 8.3
Dimensions and Technical Characteristics of some Direct-Flow Biofilters
Name of the Productivity of
Device Dimensions Filtration Rate Biofiltration Notice
Direct-flow Height of the 10 m/h—through The nitrate ion
biofilter with filter about coarse-grained packing, concentration decreases—
microorganisms 2.5 m 20–30 m/h—through from 140 to 25 mg/L.
in the sand bed fine-grained packing
Gravity biofilter Height of the Surface filtration 60–100 g of organic
with activated filter bed—6 m rate—10 m/h substances (with respect
coal to dissolved organic
carbon) are decomposed
on 1 m2 of filtering
surface at oxygen daily
consumption 200–250 g.
Membrane Denitrification rate: 5.8 g Initial NO3−
bioreactor NO3−—N/m²/day. concentration was
4 g NO3−—N/m²/day. 200 mg/dm³; the
6.1 g NO3−—N/m²/day. degree of removal
of NO3− was 90%.
If the average diameter of the biofilter grains is ~1 mm, and its cross-sectional area is equal to
1 m2, then the total area of biofouling in the filter bed reaches 11,250 m2 or about 4,500 m2/1 m3 of
sand. The result was obtained using the following formula:
where
Sp is the surface area of porous media of the filter bed
0.75 is the void ratio of packing material
V is the sand volume
Vi is the volume of one grain of sand
Si is the surface area of one grain of sand
When the surface rate of biofiltration is equal to 10 m/h and conversion biofiltration efficiency
is equal to 0.178 (C/C0 = 25/140 = 0.178), 1 m2 of biofouling surface will convert daily 2.45 g of
nitrates. This result was obtained by using the equation
w ⋅ S ⋅ ΔC ⋅ τ
M= , (8.50)
Sbio
where
M is the mass of bioconsumable impurities accounting to 1 m2 of biofouling
w is the surface rate of biofiltration
ΔC is the difference of input–output concentrations of bioconsumable impurities
S is the cross-sectional area of the biofilter
τ is the process duration
Sbio is the total area of biofouling in the filter bed
In small-size sand filter, which is characterized in the first row of Table 8.2, the surface area of
porous space of the filter bed is equal to 18 m2. On this surface, covered by biofilm, 11 g of nitrates
can be destructed during 6 h filtration time (or 0.1 g of nitrates by 1 m2 of biofouling surface per
1 h) provided having the same hydrodynamic conditions and nitrate concentration as indicated in
Table 8.3.
The activated carbon filters are more effective for biofiltration. As is seen from the second row
of Table 8.3, when a filter with a volume of filter bed equal to 6 m3 is used, 60–100 g of organic
substances (with respect to dissolved organic carbon) can be decomposed at oxygen daily consump-
tion equal to 200–250 g. This is due to the fact that the porous surface of grained activated carbon
available for biofouling is 20–70 m2/g (see Section 8.2.2). Therefore, total biofilm area in such filter
could reach 4⋅108 m2. And using only 150 g of charcoal grained packing, the surface area reaches a
value from 3,000 to 10,500 m2. It is about 166–583 times as higher as the surface available for bio-
fouling in the sand filter, for which packing material mass is 6 kg and which consists of grains with
a radius 0.5 mm (see the first row of Table 8.2). The foregoing information shows that combining
sand and charcoal packing for water filtration at recirculation mode, one can achieve a significant
improvement in the water quality due to the microfloral activity in the filter bed.
The biofiltration process under recirculation conditions does not require additional equipment.
The only need is to get a developed nontoxic porous surface in the filter bed and to saturate treated
water with the oxygen of air to create required living conditions for aerobic microorganisms. But in
any case, to maintain this fouling at acceptable levels, one need to effectively remove the products
of microbial metabolism from the treated water.
Water Purification Devices 509
⎛ C ⎞
Γ = Γ∞ ⎜ ⎟ ⋅ {1 − exp[ −( K ↑ + K ↓) ⋅ τ]}, (8.51)
⎝ K eq + C ⎠
where
Γ is the amount of surfactant adsorbed at time τ (its surface concentration)
Γ∞ is the maximum surfactant amount adsorbed at the interface
C is the surfactant concentration in bulk solution
Keq is the adsorption equilibrium constant of surfactant
K↑ is the constant rate of surfactant transfer to the interface from the subphase
K↓ is the constant rate of the reverse surfactant transition into the subphase
τ is the duration of adsorption process
The sum of the quantities K↑ and K↓ in Equation 8.51 is nothing else than the total adsorption rate of
surfactant at the interface. It is measured in s−1 and is noted by Kc. This, considering Equation 8.51,
is transformed into the following form:
⎛ C ⎞
Γ = Γ∞ ⎜ ⎟ ⋅ {1 − exp(− K c ⋅ τ)}, (8.52)
⎝ K eq + C ⎠
In (8.51) and (8.52), the value of (Γ∞⋅(C/Keq + C)) represents the equilibrium value of adsorption at
the air–water interface according to Langmuir model. And the component represented in the curved
brackets is the correction factor, which takes into account the surfactant adsorption reducing in time
τ, as compared to its equilibrium value. The phenomenon is arisen from the real time of surfactant
molecule diffusion from subphase to the interface under the influence of the concentration gradi-
ent. The difference between these values is the greater, the smaller the diffusion coefficient of the
surfactant molecules in the aqueous medium. The mechanism of this phenomenon is described in
Appendices 8.A and 8.B.
If at a small value of τ, the magnitude (−Kcτ) of Equation 8.52 is much greater than 1, the com-
ponent represented in the curved brackets can be assumed to be 1. This means that the surfactant
adsorption occurs without any potential barriers overcoming and diffusion restrictions. And the
time factor does not have noticeable influence on the adsorption process, that is, adsorption of sub-
stances immediately reaches its equilibrium value at any volume concentration of surfactant. As a
result, Equation 8.52 transforms to the classical equation, which describes the equilibrium adsorp-
tion following the Langmuir model:
⎛ C ⎞
Γ = Γ eq = Γ ∞ ⎜ ⎟ , (8.53)
⎝ K eq + C ⎠
510 Handbook of Surface and Colloid Chemistry
But if the exponent in Equation 8.52 is much less than 1, that is, (−Kc⋅τ) → 0, the analysis of adsorp-
tion process in this case must be carried out by Equation 8.52 or by the following equation, which
is obtained from (8.52)
⎛ C ⎞
Γτ = Γ∞ ⎜ ⎟ ⋅ K c ⋅ τ, (8.54)
⎝ K eq + C ⎠
And if the value of C is much less than Keq, that is, conformed by real conditions of contaminated
water treatment when C ≪ Keq, then Equation 8.54 is transformed into the equation known as
Henry’s adsorption isotherm:
C
Γ = Γ∞ ⋅ ⋅ K c ⋅ τ, (8.55)
K eq
where
Γ∞ is the maximum surfactant concentration at the interface
Keq is the adsorption equilibrium constant for the Langmuir model
Kc is the adsorption rate constant
τ is the duration of the adsorption process (in the case of flotation, it is equal to the time of air
bubbles floating through the water volume)
Thus, taking into account that the magnitude of τ is the ratio of time to air bubbles passing through
water bulk to their floating velocity, then Equations 8.52 and 8.55 will be written as follows:
⎛ C ⎞ ⎧ ⎡ l1 ⎤ ⎫
Γ = Γ∞ ⎜ ⎟ ⋅ ⎨1 − exp ⎢ − K c ⋅ ⎥ ⎬ , (8.56)
K
⎝ eq + C ⎠ ⎩ ⎣ w 1 ⎦⎭
and
C l
Γ = Γ∞ ⋅ ⋅ K c ⋅ 1 , (8.57)
K eq w1
where
l1 is the distance of air bubbles passing through water bulk
w1 is the floating velocity of air bubbles
At given values of Γ, Γ∞, C, Keq, l1, and w1, we have no problem at calculating the magnitude of Kc
following Equations 8.56 and 8.57 (see Appendices 8.A and 8.B).
In order to check the influence of Kc, the adsorption of sodium dodecyl sulfate (SDS), which is a
typical detergent, was investigated in detail [3]. The diameter of air bubbles was 0.004 m, the float-
ing velocity of the bubbles was about 0.20 m/s, and the distance of air bubbles passing through water
bulk was 0.2 m. The results are shown in Table 8.4.
If one knows the adsorption rate and degree of surfactant adsorption at the surface of air bubbles
at their floating, then one can calculate the surfactant mass flow from treated water depending upon
the intensity of bubbling, air bubble’s radius, and time of bubble floating. The surfactant mass flow
change can be expressed by the following equation [3]:
dG F C l F
= −Γ ⋅ 3 1 = −Γ ∞ ⋅ ⋅ K c ⋅ 1 ⋅ 3 1 , (8.58)
dτ r K eq w1 r
Water Purification Devices 511
TABLE 8.4
Adsorption of SDS at the Surface of Air Bubbles of Radius 2⋅10−3 m Rising from a
Depth of 0.2 m with a Velocity of 0.2 m/s
Surfactant Adsorption of Degree of Surfactant Adsorption on
Concentration in Surfactant on Air Adsorption of Air Bubble Surface during Their
Water Bulk, C, mg/L Bubble, Γ, mol/cm2 Surfactant, Γ, mg/cm² Floating, θ = S0/S1, at S0 = 30 A²
0.1 6.66·10−13 1.92·10−7 0.0012
0.2 1.33·10−12 3.84·10−7 0.0024
0.3 2·10−12 5.76·10−7 0.0036
and
dG F ⎛ C ⎞ ⎧ ⎛ l1 ⎞ ⎫ F1
= −Γ ⋅ 3 1 = Γ ∞ ⋅ ⎜ ⎟ ⋅ ⎨1 − exp ⎜ − K c ⋅ ⎟ ⎬ ⋅ 3 , (8.59)
dτ r ⎝ K eq + C ⎠ ⎩ ⎝ w1 ⎠ ⎭ r
where
F1 is the air flow rate for bubbling
r is the air bubble’s radius
However, in the case when constant water volume is treated under recirculation conditions, the
surfactant mass flow is expressed as
dG dC
=V , (8.60)
dτ dτ
dG dC F C l F
=V = −Γ ⋅ 3 1 = −Γ ∞ ⋅ ⋅ K c ⋅ 1 ⋅ 3 1 , (8.61)
dτ dτ r K eq w1 r
dG dC F ⎛ C ⎞ ⎧ ⎛ l1 ⎞ ⎫ F1
=V = −Γ ⋅ 3 1 = −Γ ∞ ⋅ ⎜ ⎟ ⋅ ⎨1 − exp ⎜ − K c ⋅ ⎟ ⎬ ⋅3 , (8.62)
dτ dτ r ⎝ K eq + C ⎠ ⎩ ⎝ w1 ⎠ ⎭ r
dG dC F C ⎧ ⎛ l ⎞⎫ F
=V = −Γ ⋅ 3 1 = −Γ ∞ ⋅ ⋅ ⎨1 − exp ⎜ − K c ⋅ 1 ⎟ ⎬ ⋅ 3 1 . (8.63)
dτ dτ r K eq ⎩ ⎝ w 1 ⎠⎭ r
Equation 8.61 is valid in the case when the surfactant is characterized by very low rate constant of
its adsorption at the interface and the concentration of flotation solution is very low. Equation 8.62
describes the situation when the rate constant of surfactant adsorption and the concentration of flota-
tion solution are comparable. And Equation 8.63 corresponds to the condition when the surfactant
concentration in water is low, but the magnitudes of rate constant of surfactant adsorption are quite
different.
Selecting (8.63), one simulates the natural process of tap water purification from its surface-
active contaminants with different molecular mass, ranging from simple detergents to finishing
proteins.
512 Handbook of Surface and Colloid Chemistry
TABLE 8.5
Parameters of the Flotation (Bubble-Film Treatment) under Recirculation Conditions
Adsorption Distance
Velocity of Air Integrated
Adsorption Constant of Bubbles Velocity
Treated Cycle Time Extraction Equilibrium Surfactant Passing Air Flow Constant
Water of Water Degree of Constant, at Air through for of
Volume, Treatment, Surfactant, Keq, Bubble, Kc, Water Bubbling, Flotation,
No V, L τ, h C/C0 mol/cm3 s−1 Bulk, l, cm F1, L/min K, cm³/s
1 20 5.07 0.1 0.269·10−6 4.46 30 4.4 2.52
2 200 12.63 0.1 0.269·10−6 4.46 30 17.6 10.095
3 2000 12.5 0.1 0.269·10−6 4.46 150 176 101.4
By separating variables in (8.63) and then integrating this equation, one shall get
C τ
dC 1 ⎧ ⎛ l ⎞⎫ F
⋅ ⎨1 − exp ⎜ − K c ⋅ 1 ⎟ ⎬ ⋅ 3 1 dτ, (8.64)
∫
C0
C
= −Γ ∞ ⋅
V ⋅ K eq ⎩ ⎝ w1 ⎠ ⎭ r ∫ 0
C 1 ⎧ ⎛ l ⎞⎫ F
ln = −Γ ∞ ⋅ ⋅ ⎨1 − exp ⎜ − K c ⋅ 1 ⎟ ⎬ ⋅ 3 1 ⋅ τ. (8.65)
C0 V ⋅ K eq ⎩ ⎝ w 1 ⎠⎭ r
Using Equation 8.65, one can calculate the basic parameters of the flotation process, that is, the
magnitude of C/C0 depending upon Q, Kc, Keq, r, l1, τ, and V, for the case of ideal (100%) removal
of surfactants from the treated water under recirculation conditions. The data were obtained as is
shown in Appendix 8.A. These devices were designed to transform the flow of air bubbles with
adsorbed surfactant on their surfaces into the flow of surfactant-contained thin liquid films.
At constant intensity of air bubbling Q and known values of Γ∞, Kc, Keq, r, l1, and w1, the right-
hand side of Equation 8.65 can be written as
C ⎛ K ⋅ τ ⎞ (8.66)
= exp ⎜⎜ − ⎟⎟ ,
C0 ⎝ V ⎠
8.7.4 Water Degassing While Bubbling and Its Saturation by Atmospheric Oxygen
At the course of bubble-film extraction alongside with adsorption of surfactants at the air bubble
surface, the volatile substances dissolved in water penetrate to the bubbles’ internal surface. Due
to this process, the rapid exchange absorption takes place in the bubbling water. As a result,
the concentration of volatile substances in water is decreased. And the concentration of air is
increased [28].
Water Purification Devices 513
The difference in the concentration of volatile substances at the air–water interface is the driving
force of their diffusion. According to Fick’s law,
dC
J = −D , (8.67)
dx
where
J is the volatile substance flow through the interface
D is the diffusion coefficient of volatile substances
dC/dx is the concentration gradient
Most commonly, the mass transfer process in the absorption system is described by two-film model
[29]. In this case, the diffusion of volatile substances through gaseous and liquid films is considered
individually. But the diffusion rate through water is decreased by about four orders of magnitude
as in atmosphere. Therefore, the main diffusion resistance exists in the aqueous phase, exactly in
the thin layer around the floating bubbles. So, the concentration gradient in the studied system is
represented by the following equation:
dC C − Cg
=± , (8.68)
dx δ1
where
C is the volatile substance concentration in the aqueous phase
Cg is the volatile substance concentration in the bubbles
δ1 is the thickness of the diffusion layer around the surface of floating air bubbles
In Equation 8.68, the plus sign concerns the case when volatile substance is transferred from aqueous
phase into the bubble volume and the minus sign concerns the case when the diffusion of volatile
substance occurs in opposite direction. And the concentration gradient has the plus sign when the
process of water degassing is analyzed (i.e., the evolution of chlorine, chloroform, ammonia, meth-
ane, hydrogen sulfide, and other gases analyzed, if their concentration in air is considerably lower
than that in aqueous phase).
Equations 8.67 and 8.68 describe the degassing process of water by means of air bubbling. The
gas flow J through the air–water interface is described by the following equation:
V dC
J= ⋅ , (8.69)
S dτ
where
V is the degassing water volume
S is the total area of the air–water interface in two-phase system
C is the concentration of dissolved gas in water
τ is the time
dC ⎛ C − Cg ⎞
V = ± ⎜ −D ⋅ S ⋅ , (8.70)
dτ ⎝ δ1 ⎟⎠
514 Handbook of Surface and Colloid Chemistry
or
dC ⎛ D S ⎞
= ± ⎜ − 1 ⋅ dτ ⎟ . (8.71)
C − Cg ⎝ δ V ⎠
C − Cg ⎛ D S ⎞
ln = ± ⎜ − 1 ⋅ ⋅ τ ⎟ , (8.72)
C0 − C g ⎝ δ V ⎠
⎛ D S ⎞
C = ±(C0 − Cg ) ⋅ exp ⎜ − 1 ⋅ ⋅ τ ⎟ + Cg , (8.73)
⎝ δ V ⎠
where
C0 is the initial gas concentration in aqueous phase
C is the final gas concentration in aqueous phase
Cg is the initial gas concentration in the space of bubble
⎛ D S ⎞
P = ±( P0 − Pg ) exp ⎜ − 1 ⋅ ⋅ τ ⎟ + Pg , (8.74)
⎝ δ V ⎠
where
P is the current partial pressure of gas in aqueous phase
P0 is the initial partial pressure of gas aqueous phase
Pg is the partial pressure of gas in air bubbles
Equations 8.73 and 8.74 show that the concentration of gas in aqueous phase at bubbling is the expo-
nential function of τ, area of air–water interface, and airflow. Since the total area of the air–water
interface in the system is
l1 F
S= ⋅ 3 1 , (8.75)
w1 r
⎛ D l F τ ⎞
C = ±(C0 − Cg ) exp ⎜ − 1 ⋅ 1 ⋅ 3 1 ⋅ ⎟ + Cg , (8.76)
⎝ δ V r w1 ⎠
where
l1 is the distance of air bubbles passing through water bulk
w1 is the floating velocity of air bubbles
F1 is the air flow
r is the mean radius of air bubbles
Equation 8.76 allows us to estimate the effectiveness of bubble aeration (i.e., oxygen saturation
in water) and degassing of water depending upon the airflow, the distance of air bubbles passing
Water Purification Devices 515
TABLE 8.6
Parameters of Absorption Component in the Circulation System
Integrated
Time of Specific Rate Distance of Air Velocity
Treated Water Degree of Constant of Bubbles Passing Air Flow for Constant of
Water Treating, Degasation, Mass Transfer, through Water Bulk, Bubbling, F1, Absorption,
Volume, V, l τ, h C/C0 D/δ, m/s l, m L/min K, cm3/s
20 1.93 0.1 2⋅10−5 0.3 4.4 3.3⋅10−6
200 5.75 0.1 2⋅10−5 0.3 14.8 1.1⋅10−5
2000 5.8 0.1 2⋅10−5 1.5 29.58 1.1⋅10−4
through water bulk, their radius, and the process duration. The values of the diffusion coefficient D
and boundary layer thickness δ1 can be assumed as 10−9 m²/s and 0.5⋅10−4 m, respectively [30,31].
The results of calculations for the bubbling systems with different capacities are shown in Table 8.6.
It is shown that the surface-active impurities are removed from treated water alongside with water
degassing, and water also is saturated with oxygen of air.
TABLE 8.7
Magnitudes of Potential Surface of Water
χ H2O, V References
+0.50 Eley D.D., Evans M.G.//Trans. Faraday Soc. 1938. V.34. P.1093.
+0.29 Passoth G.//Z. Phys. Chem. 1954. Bd.203. S.275.
+0.20 Stokes R.H.//J. Amer. Chem. Soc. 1964. V.86. P.979.
+0.06 Krishtalik L.I., Alpatova N.M., Ovsyannikova E.V.//J. Electroanal. Chem. 1992. V.329. P.1.
−0.08 Noyes R.M.//J.Amer. Chem. Soc. 1964. V.86. P.971.
−0.17 Измайлов Н.А.//Докл. AH CCCP. 1963. T.149. P.1364.
−0.30 De Lidny C.L., Alfenaar M., Veen V.D. 1968. V.87. P.585.
−0.30 Hash N.S.//Austr. J. Sci. Res. 1948. V.A1. P.480.
−0.30 Мищенко К.П., Квят Э.И.//Журн. физ. хим. 1954. P.28. P.1451.
−0.34 Latimer W.M., Pitzer K.S., Slansky C.M./J. Chem. Phys. 1939. V.7. P.108.
−0.36 Strelov H.//Z. Electrochem. 1952. Bd.56. P.119.
−0.39 Halliwell H.F., Nyburg S.C.//Trans. Faraday Soc. 1963. V.59. P.1126.
−0.48 Vervey E.//Rec. trav. Chim. 1942. V.61. P.127.
−1.10 Solomon M.//J. Phys. Chem. 1970. V.74. P.2519.
But the situation will change if the air bubble surface will be filled with densely packed
adsorption monolayer of ionized surfactants. In this case, the charge density of monolayer
can reach 1 C/m 2 . And the equal quantity of charges of opposite sign forms the diffuse part
of the double electric layer. Due to this phenomenon, the flotation device allows to remove
some ionic impurities from treated water, which contain sufficient amount of surfactants.
And the exhaustion of surfactant at air bubble surface leads to the end of ion evacuation with
bubble-film flow. The dissolved ions are removed due to the action of surfactants as cofactors
of flotation.
Thus, the formation of powerful electric double layer on the surfaces of floating air bubbles,
covered by adsorbed surfactants, makes the removal of ions released from dissociated mineral
impurities possible. The selected surfactants should be strong fatty acids, strong fatty bases, or other
ionizable suitable compounds with high affinity to the air–water interface.
We have examined the behavior of the adsorption fatty acid m onolayer at the air bubble surface
in respect to coexistence of ionized and neutral molecules and surface-inactive ions. We can write
the dissociation equation for fatty acid, which is the object of the study, as follows:
where
ROH is the fatty acid nonionized molecule
RO− is the surface-active fatty acid anion
H+ is the surface-inactive cation
To evaluate the ion-transporting ability of air bubble coated with a monolayer of fatty acid, we shall
write the expressions for the dissociation constant of fatty acid in the bulk of water and in the state
of adsorbed monolayer. They are as follows:
⎡ ROν− ⎤ ⋅ ⎡ H ν+ ⎤
KA = ⎣ ⎦ ⎣ ⎦ , (8.78)
[ ROH ν ]
Water Purification Devices 517
⎡ ROS− ⎤ ⋅ ⎡ H S+ ⎤
K AS =⎣ ⎦ ⎣ ⎦ , (8.79)
[ ROH S ]
where K A and K AS are the equilibrium dissociation constants of fatty acid in aqueous bulk and in
adsorption monolayer, respectively.
The indexes S and v represent the surface and the bulk concentration of fatty acids and their
dissociation products. The relationships between these concentrations are as follows:
⎡ H S+ ⎤ = ⎡ H ν+ ⎤ ⋅ δ ⋅ j. (8.82)
⎣ ⎦ ⎣ ⎦
where
i and j are the anion/cation concentration ratio of fatty acids in the space of adsorbed monolayer
δ is the thickness of adsorbed monolayer
And then,
⎡ ROν− ⎤ ⋅ δ ⋅ i ⋅ ⎡ H ν+ ⎤ ⋅ δ ⋅ j
K AS =⎣ ⎦ ⎣ ⎦ . (8.83)
[ ROH ν ] ⋅ δ
At the air–water interface, the molecules of fatty acid and its anions form the negatively charged
plate of the electric double layer (for the case when water does not contain any additional electro-
lytes). And hydrated protons (i.e., H3O+) form the diffuse part of the electrical double layer, that is,
these ions forming a second plate of the electrical condenser. And the local proton density depends
exponentially on the electrostatic potential according to the following equation:
e ⋅ ϕZ ⎞
⎡ H Z+ ⎤ = ⎡ H ν+ ⎤ ⋅ exp ⎛⎜ − ⎟ , (8.84)
⎣ ⎦ ⎣ ⎦ ⎝ KT ⎠
where
[ H Z+ ] is the local concentration of protons in the diffuse part of the electrical double layer
φz is the electrostatic potential at the point of determination of the local proton concentration
So
e ⋅ ϕS
⎡ H S+ ⎤ = ⎡ H ν+ ⎤ ⋅ j ⋅ δ = ⎡ H ν+ ⎤ exp ⎜⎛ − ⎞
⎟ ⋅ δ, (8.85)
⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎝ KT ⎠
where n is the total concentration of fatty acid and surfactant anions at the air–water interface,
expressed in moles per unit area of the monolayer.
Regarding the surface concentration of fatty acid anions, solution of (8.86) yields
n
⎡ ROS− ⎤ = . (8.87)
⎣ ⎦ ⎡ H ν+ ⎤ exp(−e ⋅ ϕS /KT ) ⋅ δ
1+ ⎣ ⎦
K AS
But when [ H ν+ ] = 10 − pH, n = 1/Si ⋅ N, where Si is the area per molecule of fatty acid in adsorbed mono-
layer, and N is Avogadro’s number, then (8.87) is transformed to the following form:
−1
1 ⎡ 10 − pH ⋅ exp(−e ⋅ ϕS /KT ) ⋅ δ ⎤
⎡ ROS− ⎤ =
⎦ Si ⋅ N ⎢1 + ⎥ . (8.88)
⎣
⎣ K AS ⎦
Multiplication of both sides of Equation 8.88 by Avogadro’s number N and electron charge e
describes the relationship between the surface charge density σ and the boundary potential jump φS
in the monolayer of fatty acid:
−1
1 ⎡ 10 − pH ⋅ exp(−e ⋅ ϕS /KT ) ⋅ δ ⎤
⎡ ROS− ⎤ ⋅ N ⋅ e = σ = − ⋅ e ⋅ ⎢1 + ⎥ . (8.89)
⎣ ⎦ Si ⎣ K AS ⎦
And substitution of (8.89) into the Gouy–Chapman equation gives the function of the surface
potential φS upon the surface charge density σ in fatty acid monolayer and the packing density of
molecules at different ionic concentration in subphase:
−1
2 KT (1/Si ) ⋅ e ⋅ ⎡⎣1 + 10 − pH
⋅ exp(−e ⋅ ϕS /KT ) ⋅ (δ /K AS ) ⎤⎦
ϕS = Arsh , (8.90)
e A⋅ C
where A = 8 N ⋅ K ⋅ T ⋅ ε0 ⋅ ε 1 .
The transcendental equation (8.90) allows to calculate the surface charge density σ and surface
potential jump φS at the air bubble surface covered with the monolayer of fatty acid or any other
ionizable surfactant, depending upon the values of pH, Si, K AS, δ, and C.
The results of the calculation of the charge density σ at bubble surface, obtained by Equation
8.90 at C = 10−6 –10−1 M, Si = 26⋅10−20 m², K AS = 8⋅10−14 M/dm2, and δ = 3⋅10−9 dm, have showed
calculated magnitudes in the range 1⋅10−2–6⋅10−1 K/m2.
The surface charge density of adsorbed fatty acid monolayer increases proportionally with
the increase in their ionization constants, ionic strength of the subphase, pH, and packing density
of the surface-active ions in the space of monolayer. Maximum value of the surface charge density
corresponds to the case when K AS /δ = 1⋅10−1 M and the minimum when K AS /δ = 1⋅10−5 M.
The strong fatty acid leads to the formation of monolayer with strong ionization negative charge
whose value under investigated conditions is virtually independent of ionic strength and pH of
subphase. The weak fatty acid leads to the formation of monolayer with negative charge of smaller
magnitude and strongly dependent upon ionic strength and pH of subphase. The same results, but
with charges of opposite sign, could be observed in the case of study of monolayer consisting of
surface-active bases.
The results obtained by Equation 8.90 allow one to conclude that close-packed adsorption mono-
layer of strong fatty acid or strong fatty base, localized at the air bubble surface, generates powerful
Water Purification Devices 519
surface charges. These surface charges are compensated by counterions of the diffuse part of the
double electric layer in the subphase on the mechanisms of ion exchange and trivial shielding.
In the flotation device, the flow of charged air bubbles acts as collector and transporter of
ionic impurities of treated water. Through the bubble-film extractor, this flow is removed from
treated water. So, the mineral composition of treated water can be corrected, and the target mineral
impurities can be extracted. The efficiency of demineralization of treated water can be evaluated if
the relationship of bubble charge density upon pH and ionic concentration of water is known. An
appropriate empiric equation is the following:
σ = σ0 + B ⋅ C, (8.91)
where
σ is the surface charge density in the ionized monolayer at a given ionic strength of subphase
σ0 is the hypothetical surface charge density in the monolayer at zero ionic concentration of the
subphase
B is the correction factor
σ 3F1 3 F1
j= ⋅ = (σ 0 + B ⋅ C ) ⋅ ⋅ , (8.92)
e⋅ N r e⋅ N r
where
j is the flow of ions collected by air bubbles
F1 is the air volumetric flow rate
r is the radius of air bubbles
e is the elementary charge
N is Avogadro’s number
On the other hand, the flow of ions at air bubble surface is the reason of decrease in their concentra-
tion in treated water volume:
V ⋅ dC
j= . (8.93)
dτ
V ⋅ dC 3 F1
= (σ 0 + B ⋅ C ) ⋅ ⋅ . (8.94)
dτ e⋅ N r
dC 3F1 ⋅ dr
∫ X = ∫ e ⋅ N ⋅ r ⋅ V . (8.95)
520 Handbook of Surface and Colloid Chemistry
1 1 3F1 ⋅ τ
⋅ ln(σ0 + B ⋅ C1 ) − ⋅ ln(σ0 + B ⋅ C2 ) = , (8.96)
B B e ⋅ N ⋅ r ⋅V
or
⎡⎛ σ + B ⋅ C1 ⎞⎤ 3F1 ⋅ τ
ln ⎢⎜ 0 ⎟ ⎥ = B ⋅ e ⋅ N ⋅ r ⋅ V . (8.97)
⎣ ⎝ σ 0 + B ⋅ C2 ⎠⎦
Equation 8.97 allows one to determine the time required to reduce the ion concentration from C1 to C2
in water volume V (without return of extracted substances to the treated water through the bubble-
film extractor), depending on the air volumetric flow rate F1 and the radius of floating air bubbles r, if
these bubbles are filled with monolayers of fatty acid or fatty base with known parameters σ0 and B.
Maximum ion-transporting efficiency of bubble flow depending upon the volumetric airflow rate
and the radius of floating air bubbles can be found from the following equation:
V ⋅ dC σ 3F1
j= = ⋅ , (8.98)
dτ e⋅ N r
or
e ⋅ N ⋅ r ⋅V
Δτ = ⋅ ΔC. (8.99)
3F1 ⋅ σ
Here σ is the ultimately realized surface charge density in a monolayer of strong fatty acid or a
strong fat base. It will be about l.0 K/m2, if the specific area of the monolayer will be equal to
16·10−20 m2. In this case, 1 m3 of the air sprayed to form bubbles with radius 1 mm allows us to
transfer about 8⋅10−2 g of salt impurities. And when the radius of air bubbles will be 0.1 mm, then
1 m3 of air will be able to remove 0.8 g of salt impurities. It will be cations or anions depending
on the nature of surfactants adsorbed at the bubble surfaces. Monolayers of ionized fatty acids will
entrain the cations, and monolayers of ionized fatty bases will entrain the anions.
In conclusion, it is reasonable to note that when the aerosol consisting of concentrated solution
of suitable fatty acid (or of suitable fatty base) is injected into the ejector’s channel (which creates a
stream of air bubbles in the space of the flotation unit), the ionized close-packed adsorption mono-
layer of fatty acid or fatty base at the surface of air bubbles is generated automatically.
10 11
9
2 5
FIGURE 8.14 The scheme of filtration–flotation device. 1: unit case; 2: centrifugal pump; 3: drainage device;
4: sand; 5: pump inlet line; 6: ejector; 7: bubbling compartment; 8: bubble-film extractor; 9: water inlet;
10: UV sterilizer; 11: water outlet.
water. The pump 2 is connected with a drainage unit 3 and generates the water flow through the
space of granular filter material 4, which consists of quartz sand or mixture of sand and granular
mesoporous activated carbon.
At first, the packing material functions as the adsorption filter and hetero-coagulator. But as the
biofouling grows on the filter grains, the sand filter functions as biofilter that combines the functions
of bio-precipitator and bio-catalytic reactor.
Discharge line 5 directs the water flow after the pump through ejector 6 into the bubble-film
extraction module. This module consists of bubbling compartment 7 and the real bubble-film extrac-
tor 8. The water flow is directed through Laval nozzle to form a high velocity flow, and in ejector
6, the air is injected to form the water–air mixture, which is supplied into bubbling compartment 7.
During floating time, the air bubbles adsorb the surfactants and inorganic impurities, which are
complementary to adsorbed surfactants, and absorb dissolved gases. The treated water is also satu-
rated with the oxygen of air bubbles.
The impurities collected at air bubble surface are removed from the flotation space through
the bubble-film extractor. And treated water enriched with oxygen in the bubbling compartment 7
then pumped through the filter. The processes of filtration and flotation are repeated as long as the
concentration of impurities in treated water is reduced to the required level. The time needed to
purify the given water volumes are determined by the integral rate constant of withdrawal of i-kind
of impurities, and by their initial and residual concentration. Time is calculated according to (8.47).
At the functioning of designed water purifier, the residual concentration of i-type contaminants
in the treated water can be reduced to the required level, taking into account that the kinetics of
impurities removal obeys the inverse exponential function. Thus, the basic mechanisms of impurities
522 Handbook of Surface and Colloid Chemistry
removal from water volume through the sand are the same as in the case of any other granular filter.
In particular, the coarse impurities are adsorbed by filter’s surface to form loose sediment and block
partially the pores of filter media [21]. Colloidal particles form coagulum inside the filter media
according to the hetero-coagulation mechanisms [8,12,35].
But the distinguished feature of the use of granular filter in conjunction with flotator (bubble-
film extractor) under recirculation conditions is the fact that the water is enriched with oxygen of air
and then is fed continuously through the porous space of the filter material. And at the same time,
the surface-active products of microbial metabolism are removed from that water flow due to the
bubble-film extraction.
Saturation of filtered water with oxygen of air and withdrawal of surfactant inhibitors of micro-
bial metabolism lead to growth of biomass within the filter sand, and thus the sand filter is trans-
formed into the biofilter. Thus, the positive feedback is realized between the functional units of the
complex filtration–flotation device. The contribution of these components to the efficiency of water
purification under recirculation conditions depends upon the age and form of the species of biofoul-
ing and also upon the presence of impurities in the filtered water flow, which are the substrate supply
for bacteria [19].
First of all the biofouling absorbs from the water flow the organic substances that are easily
decomposed by enzymes of bacterial cells. As the concentration of these substances is reduced, bio-
fouling digests the more complex organic compounds. Due to this process, the water becomes much
more transparent at the filter outlet. And the waste products of biofouling, such as carbon dioxide,
endogenous surfactants, and the hardly decomposed fragments of dead bacterial cells, are entered
into the filter output. These substances are removed from treated water volume through the bubble-
film extractor. At the same time, the mineralization level of the treated water is also decreased.
Thus, the created devices allow us to remove impurities from treated water, which belong to dif-
ferent classes in accordance with their dispersion degree, origin, and nature. Finite sterilization of
water can be carried out by ultraviolet using bactericidal bulbs of suitable capacity. The different
processes of impurities removal through the created devices are illustrated in Figure 8.15.
UV-sterilizer
Transformation of heavy metals salts Adsorption removal of dissolved gases
into insoluble form of hydroxides at and volatile organic substances by air
the expense of pH shifting to alkaline bubbles flow
values in treated water.
Bubble-film extractor
FIGURE 8.15 Topography of the processes of impurities removal through the created devices (1, 2, 3: posi-
tive feedbacks in the system).
Water Purification Devices 523
8.A APPENDIX A
8.A.1 EQUILIBRIUM AND DYNAMIC ADSORPTION OF
SURFACTANTS AT THE AIR–WATER INTERFACE
This section concerns the equilibrium adsorption of surfactants and kinetics of this process in aque-
ous systems. Flotation of SASs is based on this phenomenon. Equilibrium adsorption of surfactants
at the air–water interface is described by Gibbs’s theory, which shows the relationship between the
surface excess of surfactants (SAS), surface tension of SAS solution, and its concentration. Many
authors studied the adsorption phenomenon that occurs at the interfaces. Numerous mathematical
models have been proposed.
In the following, we consider the equilibrium and kinetics of adsorption of surfactants at the
air–water interface on the basis of Langmuir’s theory of adsorption of gases on solids. According
to Langmuir’s theory, it is assumed that the adsorption surface consists of NS sites, which can be
occupied by adsorbed molecules. These sites correspond to the minimum of surface free energy. At
achieving the balance between the adsorbed molecules and molecules of gas, only some parts of the
potentially available adsorption sites are occupied by gas molecules. This part is equal to θ, and the
total number of molecules Nα adsorbed on the surface obeys the ratio
N a = N S ⋅ θ. (8.A.1)
The surface sites occupied by adsorbed molecules are not available for other molecules. In this
case, only free sites of the surface are available for further adsorption. This part is equal to (1 − θ).
And if we assume that the adsorption sites are equivalent and have unit fixed occupancy, then in
accordance with the Hertz–Knudsen equation [9], the adsorption rate of molecules from the gaseous
phase is determined by the following equation:
1/ 2
⎛ dN a ⎞ ⎛ KT ⎞
⎜ dτ ⎟ = Ω ⋅ n ⋅ (1 − θ) ⋅ ⎜ 2π ⋅ m ⎟ , (8.A.2)
⎝ ⎠ads ⎝ ⎠
where
Ω is the total area of adsorption surface
n is the number of adsorbed molecules per volume unit of the subphase (air space)
K is the Boltzmann constant
T is the temperature
m is the mass of the adsorbed molecules
And if no other molecules can be adsorbed, then those which velocity components V exceed thresh-
old value V0, then Equation 8.A.2 is transformed into the following form:
1/ 2
⎛ dN a ⎞ ⎛ kT ⎞ ⎛ 1 m ⋅ V02 ⎞
⎜ dτ ⎟ = Ω ⋅ n ⋅ (1 − θ) ⋅ ⎜ 2π ⋅ m ⎟ ⋅ exp ⎜ − ⋅ ⎟. (8.A.3)
⎝ ⎠ads ⎝ ⎠ ⎝ 2 kT ⎠
According to Langmuir, the rate of desorption, which is proportional to value NS, can be written as
⎛ dN a ⎞ ⎛ Ψ ⎞
⎜ dτ ⎟ = θ ⋅ N S ⋅ ν ⋅ exp ⎜ − kT ⎟, (8.A.4)
⎝ ⎠des ⎝ ⎠
where ν exp(−Ψ/kT) is the rate constant of desorption per one molecule, or the average probability
of the desorption of molecule into subphase bulk per 1 s.
524 Handbook of Surface and Colloid Chemistry
The sum of Equations 8.A.3 and 8.A.4 describes the resulting adsorption rate with respect to the
total number of molecules adsorbed on the surface and molecules desorbed from the surface, that is,
1/ 2
dN a ⎛ kT ⎞ ⎛ m ⋅ V02 ⎞ ⎛ Ψ ⎞
= Ω ⋅ n ⋅ (1 − θ) ⋅ ⎜ ⎟ ⋅ exp ⎜ ⎟ − θ ⋅ N S ⋅ ν ⋅ exp ⎜ − ⎟. (8.A.5)
dτ ⎝ 2π ⋅ m ⎠ ⎝ 2kT ⎠ ⎝ kT ⎠
In accordance with (8.A.5), this velocity has the dimension of molecules per second, and the first
component of the right-hand side of Equation 8.A.5 describes the flow of molecules from the sub-
phase to the adsorption surface, and the second component of this equation describes the reverse
flow of molecules (i.e., from the surface into gaseous space). Taking into account that the pressure P
and the concentration n of adsorbed molecules in the gaseous phase are related by ratio,
P = n⋅k⋅T, (8.A.6)
the following general expression of adsorption rate for the adsorbed gas in the terms of pressure was
obtained by Langmuir:
dN a 1 ⎛ m ⋅ V02 ⎞ ⎛ Ψ ⎞
= Ω ⋅ P ⋅ (1 − θ) ⋅ ⋅ exp ⎜ − ⎟ − θ ⋅ N S ⋅ ν ⋅ exp ⎜ − ⎟. (8.A.7)
dτ 2π ⋅ m ⋅ k ⋅ T ⎝ 2kT ⎠ ⎝ kT ⎠
And for the state when the flow of molecules is equal in both directions (the right-hand side of (8.A.7)
is equal to zero), Langmuir has derived the equation of equilibrium adsorption. It relates the degree of
adsorption surface filling θ and pressure P of the adsorbed gas, and it is derived from (8.A.7):
or
P
θ= , (8.A.9)
N S ⋅ ν ⋅ exp(−Ψ /kT )
P+
Ω ⋅ (1/ 2π ⋅ m ⋅ k ⋅ T ) ⋅ exp −m ⋅ V02 / 2kT
( )
or
1
θ= . (8.A.10)
N S ⋅ ν ⋅ exp(−Ψ /kT )
1+
Ω ⋅ n kT / 2π ⋅ m ⋅ exp −m ⋅ V02 / 2kT
( )
1
θ= . (8.A.11)
N S ⋅ N ⋅ ν ⋅ exp(−Ψ /kT )
1+
Ω ⋅ h ⋅ N ⋅ n ⋅ (1/h) kT / 2π ⋅ m ⋅ exp −m ⋅ V02 / 2kT
( )
Here
N S /Ω ⋅ N ⋅ h = CS∞ is the maximal adsorption capacity, mol/m3
n/N = C is the adsorptive concentration, mol/m3
Water Purification Devices 525
Thus,
C
θ= , (8.A.12)
ν ⋅ exp(−Ψ/kT )
C +C ∞
S
kT / 2π ⋅ m ⋅ h2 ⋅ exp −m ⋅ V02 / 2kT
( )
or
C
θ= . (8.A.13)
C + CS∞ ( K ↓ / K ↑)
Here
K↓ = ν ⋅ exp(−Ψ/kT), s−1
K ↑= kT / 2π ⋅ m ⋅ h2 ⋅ exp(−m ⋅ V02 / 2kT ), s−1
The symbol K↑ in Equation 8.A.13 corresponds to the rate constant of adsorptive molecules trans-
ferred from the subphase to the adsorption surface, and the symbol K↓ corresponds to rate constant
for the desorption process. The ratio K↓/K↑ is nothing else than the dimensionless equilibrium con-
stant of the adsorption, and the value CS∞ ⋅ ( K↓ /K ↑) is the equilibrium adsorption constant that has
the dimension of concentration.
Similarly, the derivation of the Langmuir equation of gas adsorption allows us to consider the
kinetics and the equilibrium adsorption of surfactants in the water–air system. In this system,
the diffusion resistance layer controls the transfer of surfactant molecules from aqueous phase to its
surface. The diffusion flux density J obeys Fick’s equation, that is,
dC
J = −D , (8.A.14)
dx
where
D is the diffusion coefficient
dC/dx is the adsorptive concentration gradient in the diffusion layer
At the linear profile of surfactant concentration in the diffusion layer of thickness δ, the total number
of surfactant molecules with sufficient energy to overcome this layer through area Ω per time unit
is subjected to the following equation:
⎛ dN a ⎞ C
⎜ dτ ⎟ = D ⋅ δ ⋅ Ω ⋅ N ⋅ (1 − θ), (8.A.15)
⎝ ⎠ads
where
C is the surfactant concentration in subphase
(dN a /dτ)ads is the number of surfactant molecules that overcome the diffusion layer
If the additional energy for adsorption of surfactant molecules is required due to the potential
barriers in the adsorption system, then, similarly with (8.A.3), the exponential factor is introduced
into Equation 8.A.10, that is,
⎛ dN a ⎞ C ⎛ Ua ⎞
⎜ dτ ⎟ = D ⋅ δ ⋅ Ω ⋅ N ⋅ (1 − θ) ⋅ exp ⎜ − kT ⎟, (8.A.16)
⎝ ⎠ads ⎝ ⎠
And the number of surfactant molecules in desorption flow (dN a /dτ)des is still described
by Equation 8.A.4. Thus, the equilibrium state in the system is described by the following
equation:
dN a ⎛ dN a ⎞ ⎛ dN ⎞ C ⎛ U ⎞ ⎛ Ψ⎞
=⎜ ⎟ − ⎜ a ⎟ = D ⋅ ⋅ Ω ⋅ N ⋅ (1 − θ) ⋅ exp ⎜ − a ⎟ − N S ⋅ ν ⋅ θ ⋅ exp ⎜ − ⎟ = 0.
dτ ⎝ dτ ⎠ads ⎝ dτ ⎠des δ ⎝ kT ⎠ ⎝ kT ⎠
(8.A.17)
D ⋅ (C / δ) ⋅ Ω ⋅ N ⋅ exp(−U a /kT )
θ= , (8.A.18)
D ⋅ (C /δ) ⋅ Ω ⋅ N ⋅ exp(−U a /kT ) + N S ⋅ ν ⋅ exp(−Ψ/kT )
or
C
θ= . (8.A.19)
N S ⋅ ν ⋅ δ ⋅ exp ( −Ψ /kT )
C+
D ⋅ Ω ⋅ N ⋅ exp ( −U a /kT )
Then, if the thickness of the adsorbed surfactant layer is designated by h, and when the numera-
tor and the denominator of the second term of divider of Equation 8.A.19 are multiplied by h, we
obtain
C
θ= , (8.A.20)
C + ( δ ⋅ h / D ) ⋅ C ⋅ ν ⋅ exp − (U a − Ψ )/kT
∞
S ( )
K↓
K p = CS∞ , (8.A.21)
K↑
where
K↓ = ν ⋅ exp(−Ψ/kT)
K↑ = (D/δ⋅h)exp(−Ua /kT)
When dividing the numerator and the denominator of (8.A.20) by CS° , then the expression for
equilibrium constant of the adsorption process becomes dimensionless:
K↓
Kp = . (8.A.22)
K↑
Water Purification Devices 527
C /CS∞
θ= . (8.A.23)
C /C + ( K ↓ / K ↑)
∞
S
C /CS∞
Γ = Γ∞ ⋅ , (8.A.24)
C /C + ( K ↓ / K ↑ )
∞
S
where
NS 1
Γ∞ = = . (8.A.25)
Ω ⋅ N Si ⋅ N
In (8.A.25), Γ∞ value corresponds to the maximum adsorption capacity and has dimensionality
of moles per unit area. And Si is the area per molecule of surfactant in the adsorption layer when
the maximum adsorption capacity is reached. The solution of (8.A.24) in relation to K↓/K↑ is the
following:
K ↓ C ⎛ Γ∞ ⎞
= − 1⎟ . (8.A.26)
K ↑ CS∞ ⎜⎝ Γ ⎠
Equation 8.A.26 yields the value K↓/K↑, which is based on the values Γ, Γ∞, and CS° . These param-
eters are easily determined from the equilibrium adsorption isotherm by Gibbs equation:
C dγ 1 dγ
Γ= ⋅ = ⋅ , (8.A.27)
n ⋅ R ⋅ T dC n ⋅ R ⋅ T d ⋅ ln C
where
γ is the surface tension of the surfactant solution with concentration C
R is the universal gas constant
T is the temperature
n is the factor depending on the type of the adsorbed substance
d[ A]
= − K ↑ ⋅[ A] + K ↓ ⋅[ B], (8.A.28)
dτ
528 Handbook of Surface and Colloid Chemistry
where [A] and [B] correspond to the concentrations of the adsorbed molecules in subphase [A] and
in the adsorption layer [B].
At the beginning of the adsorption process, all surfactant molecules are located in the s ubphase,
and their initial concentration is A0. In the course of adsorption, the surfactant molecules are
accumulated at the interface, and their concentration in subphase is decreased:
[ B] = [ A0 ] − [ A]. (8.A.29)
d[ A]
= − K ↑ ⋅[ A] + K ↓ ([ A0 ] − [ A])
dτ
= K↓ ⋅[ A0 ] − ( K ↑ + K ↓) ⋅ [ A]
⎛ K↓ ⎞
= −( K ↑ + K ↓) ⋅ ⎜ [ A] − ⋅ [ A0 ] ⎟
⎝ K ↑ + K ↓ ⎠
= −( K ↑ + K ↓) ⋅ ([ A] − [ Aeq ]). (8.A.30)
[ Beq ] [ A0 ] − [ Aeq ] K ↑
= = , (8.A.31)
[ Aeq ] [ Aeq ] K↓
and:
K↓
[ Aeq ] = ⋅ [ A0 ]. (8.A.32)
K ↑ + K↓
[ A] τ
d[ A]
= ( K ↑ + K ↓) ⋅ dτ,
−
∫
[ A0 ]
[ A] − [ Aeq ] ∫
0
(8.A.33)
and
[ A0 ] − [ Aeq ]
ln = ( K ↑ + K ↓) ⋅ τ. (8.A.34)
[ A] − [ Aeq ]
The value [A0]−[Aeq] in Equation 8.A.34 is none other than the Γeq, that is, it is the adsorption
of the surfactant at the interface, and the value [A]−[Aeq] is the difference between the values
of the equilibrium adsorption and the current value of adsorption. Therefore, (8.A.34) can be
written as
Γ eq
ln = ( K ↑ + K ↓) ⋅ τ, (8.A.35)
Γ eq − Γ
Water Purification Devices 529
or
Γ eq − Γ
= exp[ −( K ↑ + K ↓) ⋅ τ], (8.A.36)
Γ eq
and
This equation shows the kinetics of surfactant accumulation at the interface. Taking into account
(8.A.37), we shall finally obtain
1 ⎛ C ⎞
Γτ = ⋅⎜ ∞ ⎟ ⋅ {1 − exp[ −( K ↑ + K ↓) ⋅ τ]}. (8.A.38)
Si ⋅ N C
⎝ S ( K ↓ / K ↑ ) + C ⎠
As is followed from (8.A.38), if CS∞ ⋅ ( K ↓ /K ↑) C , and (K↑ + K↓)⋅τ ≫ 1, the following expression
is valid:
K↑
Γτ ≈ C ⋅ ⋅ h. (8.A.39)
K↓
This means that the adsorption process is completed almost immediately under indicated
conditions. It does not require any length of time to achieve the equilibrium state in the adsorption
system. This dynamic is typical for the adsorption of small surfactant molecules and inorganic
ions or gases.
As is also followed from (8.A.38), when CS∞ ⋅ ( K ↓ /K ↑) C and (K↑ + K↓)⋅τ is small, but not so
small, that the inverse exponent turns into a real unit, then
K↑
Γτ ≈ C ⋅ ⋅ h ⋅ ( K ↑ + K ↓) ⋅ τ = C ⋅ K τ ⋅ τ, (8.A.40)
K↓
C
Γ p = Γ∞ ⋅ . (8.A.41)
C ( K ↓ /K ↑ ) + C
∞
S
In this case, the kinetics of equilibrium achieving is predetermined only by the exponential term
in Equation 8.A.38. So, the sets of Gibbs and Langmuir adsorption equations are the physical–
chemical basis for the calculation of any flotation process.
530 Handbook of Surface and Colloid Chemistry
8.B APPENDIX B
8.B.1 FAST AND SLOW SURFACTANT ADSORPTION AT THE
AIR BUBBLE SURFACE AND TRANSFER OF ADSORBED
MATTER OUTSIDE THE FLOATED WATER BULK
The different surfactants can exist at air–water interface in a form of ions and neutral molecules. When
the SASs are adsorbed in the ionized state with the charge of their surface-active ions +1 or −1, then the
adsorption Γi is calculated vs. concentration C using Equation 8.A.27 from Appendix 8.A at n = 2, that is,
C dγ
Γi = ⋅ . (8.B.1)
2 RT dC
And if the SAS is adsorbed in unionized state, Equation 8.A.27 is used at n = 1, that is,
C dγ
Γi = ⋅ . (8.B.2)
RT dC
In the case, when the maximum (Γ∞) and equilibrium (Γi) adsorption at certain surfactant concentra-
tion C is known, the value of adsorption equilibrium constant is determined by following the equation:
⎛Γ ⎞
K eq = C ⋅ ⎜ ∞ − 1 ⎟ . (8.B.3)
⎝ Γ ⎠
C
Γ = Γ∞ ⋅ . (8.B.4)
K eq + C
The transport of surfactant molecules with the bubble flows from the water bulk to air–water inter-
face is the dynamic process. Its rate depends upon the SAS adsorption velocity at each air bubble
and upon the total area of all floating bubbles, that is,
3F1
dG = Γ ⋅ dτ. (8.B.5)
r
Here
G is the mass transfer of surfactant with air bubble flow
Γ is the adsorption of surfactant at the bubble surface during their floating
F1 is the volumetric air flow rate
r is the air bubble radius
τ is the time
In accordance with (8.B.4), Equation 8.B.5 may be transformed to the following form:
1 ⎛ ⎡ l ⎤ ⎞ 3F1
dG = Γ ∞ ⋅ C ⋅ ⎜ 1 − exp ⎢ − K ↑ ⎥ ⎟ ⋅ dτ. (8.B.6)
K eq ⎝ ⎣ ω ⎦⎠ r
The expression contained in the parentheses in (8.B.6) makes correction for the process rate because of
the restricted diffusion and potential barriers to the penetration of surfactant molecules to the interface.
Water Purification Devices 531
For the case of low surfactant concentration in aqueous solution, the large diffusivity of sur-
factant molecules, and the absence of potential barriers to these molecules to be adsorbed at the
air–water interface, the expression can be simplified to the following form:
1 3F
dG = Γ ∞ ⋅ C 1 dτ. (8.B.7)
K eq r
And taking into account that
dG = V ⋅ dC, (8.B.8)
Γ ∞ 3F1
⋅ =K (8.B.10)
K eq r
and
C ⎛ K ⋅ τ ⎞
= exp ⎜ − ⎟. (8.B.11)
C0 ⎝ V ⎠
Equation 8.B.11 shows the reduction of surfactant concentration in the given volume V of floated solu-
tion under experimental conditions when the extraction efficiency is equal to 100%. Equation 8.B.11
is identical to 8.A.41. Equation 8.B.7, which leads to (8.B.10) and (8.B.11), describes the adsorption
process for small molecules with large diffusivity when the essential potential barriers are absent.
dG 1 ⎧ ⎡ l ⎤ ⎫ 3F
= Γ∞ ⋅ ⋅ C ⋅ ⎨1 − exp ⎢ −( K ↑ + K ↓) ⋅ ⎥ ⎬ 1 , (8.B.12)
dτ K eq ⎩ ⎣ ω ⎦⎭ r
where
Γ∞ is the maximum surfactant adsorption at the interface
Keq is the adsorption equilibrium constant of SAS
C is the surfactant concentration in the water bulk
K↑, K↓ are the adsorption/desorption velocity constants for surfactant
l is the path length of floating air bubbles
ω is the velocity of floating air bubbles
F1 is the volume air flow for bubbling
r is the radius of air bubbles
532 Handbook of Surface and Colloid Chemistry
The values of Γ∞ and Keq are calculated from experimental data of equilibrium surface tension of
surfactant solution upon its concentration according to (8.A.27), of Appendix 8.A. The values of r,
F1, and C are obtained from experimental conditions. And the values of K↑ and K↓ are calculated
using (8.B.12), if it is written as the following equation:
1 ⎧ ⎡ l ⎤ ⎫ 3F dA m
Γ∞ ⋅ ⋅ C ⋅ ⎨1 − exp ⎢ −( K ↑ + K ↓) ⋅ ⎥ ⎬ 1 = ⋅ .
K eq ⎩ ⎣ ω ⎦⎭ r dτ Si ⋅ N
(8.B.13)
Here
dA/dτ is the rate of formation of adsorption monolayer at the interface
Si is the specific area in the adsorption monolayer at a given surface pressure π
m is the molecular mass of adsorbed surfactant
N is Avogadro’s number
The left-hand side of Equation 8.B.13 is used to characterize the surfactant flow arisen as a result of
their adsorption from the water bulk to the bubble surface floating up to the interface of the aque-
ous solution. The right-hand side of this equation is used to characterize the increase in the mass of
adsorption monolayer arisen at the water surface.
The air bubbles with adsorbed surfactants are destroyed when reaching the air–water interface,
and the surfactants from air bubble surfaces are accumulated at the interface. Figure 8.16 shows
the experimental assembly used for the study of the SAS adsorption process described by (8.B.13).
The process of SAS adsorption at the air bubble surfaces occurs in flotation cylinder, which is
connected with Langmuir trough, as it is shown in Figure 8.16a and b. And after that, the floated
SASs are accumulated as a monomolecular layer at the air–water interface in Langmuir trough.
The experimental assembly allows one to realize the flotation process using both recirculating and
continuous-flow conditions. The increase in SAS concentration in adsorption monolayer arisen at
the air–water interface in Langmuir trough is controlled by movable barrier and Wilhelmy balance.
At low magnitudes of surface pressure in arisen monolayers, the specific area is proportional to the
surface pressure [3].
Figure 8.17 shows the dA/dτ plots concentration of different surfactant solutions, studied in the
experimental assembly. The charts represent the rate of subphase’s surface filling with adsorption
SAS monolayers, compresses to surface pressures of 2.5–5.0 mN/m.
In the case of the adsorption of SDS at π = 4 mN/m, the specific area is about 100⋅10−16 cm2.
The following values of parameters were determined by experiment and calculation to obtain the
(a) (b)
FIGURE 8.16 The schema of experimental assembly used for the realization of the adsorption process.
(a) Recirculation operating conditions; (b) continuous flow operating conditions.
Water Purification Devices 533
4 3
dA/dτ, cm2 s
1
2
1
0
0 100 200 300
C, m kg/dm3
FIGURE 8.17 The dependence of dA/dt upon C for the following solutions: 1: sodium decyl sulfate; 2:
sodium dodecyl sulfate; 3: alkyl sulfonate; 4: chemical DC–10.
adsorption rate constant of SDS at the surface of air bubbles in flotation system (Figure 8.16): Г∞ =
3.1⋅10−10 mol/cm2; Keq = 0.269⋅10−6 mol/cm3; C = 0.3 mg/dm3 = 3⋅10−4 mg/cm3; dA/dτ = 0.75 cm2/s;
π = 4 mN/m; si = 100 A2 = 100⋅10−16 cm2; m = 288⋅103 mg/mol; l = 10 cm; ω = 25 cm/s; Wв =
250 cm3/min; r = 0.1 cm; N = 6.02⋅1023 mol−1.
To a first approximation, the sum of the direct and reverse velocity constants of adsorbed
molecules transferred from (8.B.13) was expressed as the resultant adsorption rate constant Kτ,
that is,
K ↑ + K ↓= K τ . (8.B.14)
Then, transforming Equation 8.B.13, one can obtain
⎛ l ⎞ dA m K r
1 − exp ⎜ − K τ ⎟ = ⋅ ⋅ eq ⋅ , (8.B.15)
⎝ ω ⎠ dτ Si ⋅ N Γ ∞ ⋅ C 3F1
l ⎡ dA m K eq ⋅ r ⎤
−Kτ = ln ⎢1 − ⋅ ⋅ , (8.B.16)
ω ⎣ d τ Si ⋅ N Γ ⋅ C ⋅ 3F1 ⎥⎦
∞
⎛ 0.348 ⎞
= ln ⎜ 1 − ⎟
⎝ 0.419 ⎠
= ln(1 − 0.832)
= ln 0.168.
On the other hand, the resultant rate adsorption constant Kτ can be expressed as
D ⎛ U ⎞
K ↑ + K ↓= K τ = exp ⎜ − a ⎟. (8.B.17)
δ⋅h ⎝ KT ⎠
Then,
⎛ U ⎞ K ⋅ δ ⋅ h 4.46 ⋅ 0.01 ⋅ 18 ⋅ 10 −8
exp ⎜ − a ⎟ = τ = = 8.028 ⋅ 10 −3 ,
⎝ KT ⎠ D 1 ⋅ 10 −6
K↓
K eq = Cs∞ ,
K↑
where
Keq = 0.269·10−6 mol/cm3
CS∞ = N S /Ω ⋅ N ⋅ h = 1.722 ⋅ 10 −3 mol /cm 3
K↑ = 4.46
K eq 0.269 ⋅ 10 −6 ⋅ 4.46
K ↓= ⋅ K ↑= = 6.96 ⋅ 10 −4.
CS∞ 1.722 ⋅ 10 −3
So,
K↓ = 6.96·10−4 s−1,
and
K↑ = 4.44 s−1.
K↓ ⎛ ΔF ⎞ K ↓ 6.96 ⋅ 10 −4
But = exp ⎜ − ⎟; = = 1.56 ⋅ 10 −4.
K↑ ⎝ RT ⎠ K↑ 4.46
Then,
ΔF ΔF
− = ln 1.56 ⋅ 10 −4 ; = 8.76; ΔF = 8.76 ⋅ 8.31 ⋅ 293 = 21.32 kJ/mol.
RT RT
Water Purification Devices 535
So, we have got the following characteristics of the adsorption system: Keq = 0.269⋅10−6 mol/cm3;
Г∞ = 3.1⋅10−10 mol/cm2; CS∞ = 1.722 ⋅ 10 −3 mol /cm 3; Kτ = 4.46 s−1; K↓ = 6.96⋅10−4 s−1; K↑ = 4.46 s−1;
ΔF = 21.32 kJ/mol; Ua = 11.73 kJ/mol.
In such a way, the kinetics of SDS adsorption in the water–air system is fast. The degree of
interface filling with detergent molecules, calculated according to (8.B.12), reaches 67% of its
equilibrium value at τ = 0.4 s, l = 10 cm, and V = 25 cm/s. And it reaches 97% and 99% of its equi-
librium value at l = 20 cm, l = 40 cm, and V = 25 cm/s (adsorption time 0.8 and 1.6 s, respectively).
Therefore, if the flotation device is used to extract the defined surfactant from its aqueous solution,
the value of K (8.A.41), (8.B.1) through (8.B.11) at l ≥ 40 cm, r ≥ 0.1 cm, and V = 25 cm/s can be
calculated according to
Γ 3F
K = ∞ ⋅ 1 .
K eq r
(8.B.18)
Here the value {1−exp[−(K↑ + K↓)(l/ω)]}, which takes into account the contribution of the track
length of floating bubbles, is little affected upon the adsorption, and it may be taken as 1. And if the
flotation device is used for the treatment of solutions of high-molecular-weight surfactants that have
a small diffusion coefficient, the generalized rate constant of flotation process is calculated accord-
ing to the following formula:
1 ⎧ ⎡ l ⎤ ⎫ 3F1
K = Γ ∞ ⋅ ⎨1 − exp ⎢ −( K ↑ + K ↓) ⋅ ⎥ ⎬ ⋅ .
K eq ⎩ ⎣ ω ⎦ ⎭ r (8.B.19)
The expression represented in square brackets of (8.B.19) corresponds to the product of rate constant
of SAS adsorption and duration of air bubbles floating through the water bulk. When the absolute
value of this product is small, Equation 8.B.19 is transformed to the following form:
K ⋅ l 3F
K = Γ ∞ ⋅ τ ⋅ 1 ,
K eq ⋅ ω r
(8.B.20)
or
K ↑ ⋅ l 3F1
K = Γ ∞ ⋅ ⋅ .
K eq ⋅ ωV r
(8.B.21)
So, to calculate the value of generalized rate constants K (8.A.41), (8.B.1) through (8.B.11) of the
removal of the defined surfactants from water bulk, it is essential to know the value of the following
parameters: constants Keq and Kτ, the track length l of floating bubbles, their floating velocity ω, the
value of maximum adsorption Г∞, air flow F1, and the radius of the bubbles r. If it is necessary to
remove SASs of the known molecular weight, the concentration of n-type impurities can be calcu-
lated according to (8.A.41), (8.B.1) through (8.B.11), using the values of K,
Keq, Kτ, and Г∞ for n-type
impurities. And as a whole, the value of K must be calculated as in the case of difficult-to-remove
contaminants.
536 Handbook of Surface and Colloid Chemistry
8.C APPENDIX C
8.C.1 CALCULATION OF THE DEGREE OF SURFACE FILLING BY
SAS MOLECULES ADSORBED FROM THE SIDE OF AIR
If one is able to get very fine surfactant dispersion in the space of air, then the degree of air–water
interface filling with surfactant molecules from air side is described by the following equation:
C
θ= . (8.C.1)
v ⋅ exp(−ψ /KT )
C +C ⋅ ∞
s
KT /2π ⋅ m ⋅ h2 ⋅ exp −m ⋅ V02 / 2 KT
( )
Here,
Cs∞ = 1/Si ⋅ N ⋅ h
Si is the specific equilibrium area in adsorbed monolayer
h is the thickness of adsorbed monolayer
N is Avogadro’s number
m is the mass of adsorbed molecules
At the adsorption of long-chain surfactants, the value of Si is varied from 20·10−20 to 60·10−20 m2.
For SDS, octyl sulfate, and tetradecyl benzenesulfonate, the value of Si must start with 53.6·10−20 m2
per molecule [36].
The magnitude of h depends upon the molecular structure of adsorbed surfactant: for palmitic
acid (C15H31COOH), h = 23 Å; and for stearic acid (C17H35COOH), h = 26 Å. Therefore, in the case
of study of the surfactants with different lengths of its hydrocarbon chains, one can use the follow-
ing formula:
26
hi = i,
17
where i is the number of carbon atoms in the hydrocarbon chain of adsorbed SAS molecule.
The value of m is calculated from the relationship m = M/N, where M is the molecular mass of
adsorbed surfactant. The value of the second component of denominator in (8.C.1) is the equilibrium
adsorption constant Keq, which has dimensions mol/cm3, mol/dm3, and mol/m3. So, we can write
ν exp(−ψ /KT )
K eq = Cs∞ ⋅ , (8.C.2)
KT / 2π ⋅ m ⋅ h2 ⋅ exp −m ⋅ V02 / 2 KT
( )
or
or
1
K eq = Cs∞ ⋅ ⋅ X, (8.C.4)
KT / 2π ⋅ m ⋅ h2
Water Purification Devices 537
where
⎛ m ⋅ V02 − 2ψ ⎞
X = ν exp ⎜ ⎟.
⎝ 2 KT ⎠ (8.C.5)
C
θ= , (8.C.6)
1 ν exp(−ψ /KT )
C + Cs∞ ⋅ ⋅
D /δh exp(−U a /KT )
or
C
θ= 1
, (8.C.7)
C + K eq
where
1 ν exp ( −ψ / KT )
1
K eq = Cs∞ ⋅ ⋅ , (8.C.8)
( D / δh ) exp ( −U a / KT )
or
1 ⎛U −ψ ⎞
1
K eq = Cs∞ ⋅ ⋅ ν exp ⎜ a ⎟. (8.C.9)
( D / δh ) ⎝ KT ⎠
And
1
1
K eq = Cs∞ ⋅ ⋅ X. (8.C.10)
( / δh )
D
⎛U −ψ ⎞
Here X = ν exp ⎜ a ⎟. (8.C.11)
⎝ KT ⎠
Analogous to (8.C.5), the expression (8.C.11) describes the energy–frequency behavior of desorption
of adsorbed molecules from adsorption layer into the water bulk.
538 Handbook of Surface and Colloid Chemistry
K p Cs∞
=
( D / δh )( X ) .
(8.C.12)
K 1p Cs∞ KT /2πmh2 ⋅ ( X )
Here
D is the diffusion coefficient of surfactant
δ is the thickness of diffusion layer at the interface
h is the thickness of adsorption layer at Г → ∞
⎛ mV02 − 2ψ ⎞ ⎛U −ψ ⎞
X ≈ ν exp ⎜ ⎟ and X ≈ ν exp ⎜ a ⎟. (8.C.13)
⎝ 2 KT ⎠ ⎝ KT ⎠
Equation 8.C.12 allows one to evaluate the ratio of adsorption constants for any surfactant. For
example, for the adsorption of SDS at the following experimental conditions: D = 1·10−6 cm2/s;
δ = 0.01 cm (at radius of air bubble ~1 mm and its floating velocity 0.25 m/s); h = 18·10−8 cm; K =
1.38·10−23 J/K; T = 293 K; m = 0.288 kg/mol; N = 6.02·10−23 mol−1, one can obtain
K eq 1 ⋅ 10 −6 / 0.01 ⋅ 18 ⋅ 10 −8 555.5
= = = 2.73 ⋅ 10 −8 , (8.C.14)
1
K eq (1.38 ⋅ 293 ⋅ 6.02 / 2 ⋅ 3.14 ⋅ 0.288) ⋅ (1/18 ⋅ 10 )
−10
2
2 . 03 ⋅ 1010
that is,
1
K eq K eq
1
= 2.73 ⋅ 10 −8 and = 3.6 ⋅ 107.
K eq K eq
It means that in the case when the equilibrium adsorption constant of SDS is equal to 1.16·10 −3
mol/dm3 (at adsorption of SDS from bulk water), the value of equilibrium adsorption constant
of SDS at adsorption from air is 3.6·107 as smaller as at adsorption from water bulk, and it is
equal to
1
K eq
K eq = = 3.22 ⋅ 10 −11 †mol/dm 3 .
3.6 ⋅ 107
Water Purification Devices 539
So, in the case of essential SAS concentration in air, the adsorption of SDS at the air–water interface
will result in fast formation of close-packed adsorption monolayer (8.B.13).
KT 1, 38 ⋅ 10 −23 ⋅ 293
2. K a ↑1mcm = =
2πmh2 2 ⋅ 3,14 ⋅ 4 / 3 ⋅ 3,14 ⋅ (1 ⋅ 10 −6 )31000 ⋅ (18 ⋅ 10 −10 )2
KT 1, 38 ⋅ 10 −23 ⋅ 293
3. K a ↑0,1mcm = =
2πmh2 2 ⋅ 3,14 ⋅ 4 / 3 ⋅ 3,14 ⋅ (1 ⋅ 10 −7 )31000 ⋅ (18 ⋅ 10 −10 )2
Here, Ka↑i is the rate transfer constant of aerosol particles with the mentioned sizes from air to the
interface.
And,
K↑ 555, 5 K↑ a
1. = = 0, 0806. At a = 1 ⋅ 10 −5 m, = 12, 4.
K a ↑ 6, 889 ⋅ 10 3
K↑
K↑ 555, 5 K↑ a
2. = = 2, 55 ⋅ 10 −3. At a = 1 ⋅ 10 −6 m, = 390, 63.
K ↑ a 2,17 ⋅ 10 5
K↑
K↑ 555, 5 K↑ a
3. = = 8, 06 ⋅ 10 −5. At a = 1 ⋅ 10 −7 m, = 12401.
K ↑ a 6, 889 ⋅ 10 6
K↑
where K↑ = (D/δh) = 555,5 s−1. (This value characterizes the diffusion component of rate constant
of SDS adsorption from water bulk at the surface of floating air bubble with radius 1 mm, when
δ = 0,01 cm.)
The comparison of obtained data shows that at a given radius of aerosol particles of SAS s olution,
the rate adsorption constants of this aerosol at its adsorption from air are essentially higher than
the rate adsorption constants from water bulk. This allows one to realize technically the flotation
process using the close-pack ionized adsorption monolayer of surfactants, specially obtained at air
bubble surface by aerosol dispersion and, thus, to control this process efficiently.
540 Handbook of Surface and Colloid Chemistry
8.D APPENDIX D
8.D.1 PROPERTY OF THE DOUBLE ELECTRIC LAYER FORMED
AT THE AIR–WATER INTERFACE BY DIPOLES OF WATER
MOLECULES, HYDROXYLS, AND PROTONS
In the system of desalted water–air, when water doesn’t contain any surfactants, the double electric
layer at the air bubble surface is formed due to the structure and charge distribution in water mol-
ecules and due to the presence of charged hydroxyls and protons arisen as a result of dissociation
of water molecules. In order to estimate the maximum charge density at the bubbles owing to water
molecule dissociation, let’s use the expression for the dissociation constant of water in its bulk and
at the interface, respectively:
⎡OH v− ⎤ ⋅ ⎡ H v+ ⎤
KA = ⎣ ⎦ ⎣ ⎦, (8.D.1)
⎡⎣ HOH v ⎤⎦
⎡OH s− ⎤ ⋅ ⎡ H s+ ⎤
K AS =⎣ ⎦ ⎣ ⎦, (8.D.2)
⎡⎣ HOH s⎤
⎦
where K A and K AS are the dissociation constants of water in the water bulk and at the interface,
respectively. The indexes s and v designate the surface and volume concentrations:
⎡ H s+ ⎤ = ⎡ H v+ ⎤ δ ⋅ j. (8.D.5)
⎣ ⎦ ⎣ ⎦
In this case,
⎡OH v− ⎤ δ ⋅ i ⎡ H v+ ⎤ δ ⋅ j
K AS = ⎣ ⎦ ⎣ ⎦ , (8.D.6)
⎡⎣ HOH v ⎤⎦ δ
where
i and j are coefficients of hydroxyl and proton concentration at the interface
δ is the thickness of the surface layer
The outside plate of the double electric layer is formed from water molecules and hydroxyl ions,
which are located at the interface. At that, the hydrated protons (H3O+) form the diffusion part of the
double electric layer in the subphase. Their local concentration in this part of double electric layer
depends exponentially upon electrostatic potential:
eϕ
⎡ H z+ ⎤ = ⎡ H v+ ⎤ ⋅ exp ⎛⎜ − z ⎞
⎟, (8.D.7)
⎣ ⎦ ⎣ ⎦ ⎝ KT ⎠
where
[ H z+ ] is the local concentration of hydrated protons (H3O+)
φz is the electrostatic potential in the point of local concentration determination
Water Purification Devices 541
eϕ
⎡ H s+ ⎤ = ⎡ H v+ ⎤ ⋅ j ⋅ δ = ⎡ H v+ ⎤ exp ⎛⎜ − s ⎞
⎟ δ, (8.D.8)
⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎝ KT ⎠
Here, n is the total concentration of water molecules and hydroxyl ion concentration at the interface.
Solution of (8.C.9) with respect to the hydroxyl ion surface concentration gives
n
⎡OH s− ⎤ = . (8.D.10)
⎣ ⎦ ⎡ H ⎤ exp(−eϕs /KT )δ
+
v
1+ ⎣ ⎦
K AS
But when [ H v+ ] = 10 − pH, and n = 1/(Si⋅N), where Si is the area per water molecule at the interface, and
N is Avogadro’s number, Equation 8.C.10 transforms to the following form:
−1
1 ⎡ 10 − pH exp(−eϕs /KT )δ ⎤
⎡OH s− ⎤ =
⎦ Si N ⎢1 + ⎥ . (8.D.11)
⎣
⎣ K AS ⎦
Multiplication of both the parts of (8.C.11) by Avogadro’s number N and electron charge N shows
the relationship between the surface charge density σ, the boundary potential jump φs at air bubble,
and pH of desalted water:
−1
1 ⎡ 10 − pH exp(−eϕs /KT )δ ⎤
⎡OH s− ⎤ N ⋅ e = σ = − e ⎢1 + ⎥ . (8.D.12)
⎣ ⎦ Si ⎣ K AS ⎦
And the substitution of (8.D.12) into Gui–Chapman equation gives the analytical dependence of sur-
face potential φs and surface charge density σ upon packed density of water molecules in the surface
layer 1/Si, dissociation constant of water K AS, and concentration of dissolved salt C:
−1
2 KT (1/Si )e ⎡⎣1 + 10 − pH ⋅ exp(−eϕs /KT )(δ /K AS ) ⎤⎦
ϕs = Arsh , (8.D.13)
e A C
where A = 8 NKTε0ε1 .
Equation 8.D.13 allows one to calculate the boundary potential jump φS and surface charge
density σ at the air bubble depending upon pH, Si, KAS, δ, and C. The results of the calculation of
σ at pH 1–14, C = 10−6–10−1 M, Si = 6⋅10−20 m², KA = 1,8⋅10−16 mol/dm3, KAS = 5,4.10−25 mol/dm2,
and δ = 3⋅10−9 dm show that the surface charge density at the air bubble in this case will be vanish-
ingly small.
542 Handbook of Surface and Colloid Chemistry
REFERENCES
1. Birdi K.S. 2009. Handbook of Surface and Colloid Chemistry. CRC Press, Taylor & Francis Group, Boca
Raton, FL.
2. Gevod V.S. 2003. NATO Science Series. Modern tools and methods of water treatment for improving
standards. IV. Earth Environ. Sci., 48, 137.
3. Гевод В.С., Решетняк И.Л., Шклярова И.Г., Хохлов А.С., Гевод С.В. 2002. Поверхностно-
активные и другие загрязнения в водопроводной питьевой воде: свойства, мониторинг, причины
накоплений и экономичное удаление. Изд-во УГХТУ, Днепропетровск.
4. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г., Нефедов В.Г. 2004. Вопросы химии и хим.
технологии, 4, 213.
5. Гевод В.С., Решетняк И.Л., Гевод С.В., Хохлов А.С., Шклярова И.Г. 2001. Вопросы химии и хим.
технологии, 6, 152.
6. Гевод В.С., Решетняк И.Л., Гевод С.В., Хохлов А.С., Шклярова И.Г. 2005. Вопросы химии и хим.
технологии, 1, 209.
7. Русанов А.И. 1981. Поверхностное разделение веществ: Теория и методы. Химия, Ленинград.
8. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г. 2007. Вопросы химии и хим. технологии, 1, 178.
9. Мелвин-Хьюз Э.А. 1962. Физическая химия. Иностранная литература, Москва.
10. Касаткин А.Г. 1971. Основные процессы и аппараты химической технологии. Химия, Москва.
11. Голубовская Э.К. 1978. Биологические основы очистки воды. Высшая школа, Москва.
12. Гончарук В.В., Дешко И.И., Герасименок Н.Г. и др. 1998. Химия и технология воды, 1, 19.
13. Гевод В.С., Решетняк И.Л., Хохлов А.С., Гевод С.В., К.С. Бирди. 1998. Вопросы химии и хим.
технологии, 3, 55.
14. Гевод В.С., Решетняк И.Л., Хохлов А.С., Гевод С.В., К.С. Бирди. 1998. Вопросы химии и хим.
Технологии, 3, 58.
15. Craun G.F. 2001. Microbial Pathogens and Disinfection by Products in Drinking Water: Health Effects
and Management of Risks. ILSI Press, Washington, DC.
16. Graham H.J., Colloins R.D. 1996. Advances in Slow Sand Alternative Biological Filtration. John Wiley
& Sons Inc., Chichester, U.K.
17. Woolschlager J., Rittmann B., Piriou P., Kiene L., Schwartz B. 2000. A comprehensive disinfection and
water quality model for drinking water distribution systems. First World Water Congress of the IWA,
Paris, France. ISBN: 2-9515416-0-0. EAN: 9782951541603. NP-119.
18. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г. 2007. Вопросы химии и хим. технологии, 5, 222.
19. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г., Нефедов В.Г. 2009. Вопросы химии и хим.
технологии, 3, 162.
20. Перт С.Дж. 1978. Основы культивирования микроорганизмов и клеток. Мир, Москва.
21. Жужиков В.А. 1980. Фильтрование: теория и практика разделения суспензий. Химия, Москва.
22. Запольский А.К. 2005. Водопостачання, водовідведення та якість води. Вища школа, Київ.
23. Кастальский А.К., Минц Д.М. 1962. Подготовка воды для питьевого и промышленного
водоснабжения. Высшая школа, Москва.
24. Кульский Л.А. 1986. Технология очистки природных вод. Вища школа, Киев.
25. Кинеле, Хартмут. 1984. Активные угли и их промышленное применение. Химия, Ленинград.
26. Слипченко В.А., Малицкая Т.Н. 1992. Химия и технология воды. 1, 35.
27. Гончарук В.В., Клименко Н.А., Савчина Л.А. и др. 2006. Химия и технология воды. 1, 44.
28. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г., Хохлов А.С., Губенко И.И. 2002. Вопросы
химии и хим. технологии, 1, 121.
29. Perry R.H., Green D.W. 2008. Perry’s Chemical Engineers’ Handbook. McGraw-Hill Book Company,
New York.
30. Гороновский И.Т. 1987. Краткий справочник по химии: справочник. Наукова думка, Киев.
31. Кошкин Н.И. 1966. Справочник по элементарной физике. Наука, Москва.
32. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г., Хохлов А.С. 2002. Вопросы химии и хим.
технологии, 6, 161.
33. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г., Гринев В.М., Нефедов В.Г. 2004. Вопросы
химии и хим. технологии, 3, 215.
34. Birdi K.S. 2010. Introduction to Electrical Interfacial Phenomena. CRC Press, Taylor & Francis Group,
Boca Raton, FL.
35. Гевод В.С., Решетняк И.Л., Гевод С.В., Шклярова И.Г. 2007. Вопросы химии и хим. технологии, 2, 227.
36. Gaines G.L., Jr. 1966. Insoluble Monolayers at Liquid–Gas Interface. Interscience Publishers, New York.
9 Preparation, Characterization, and
Thermally Sensitive Particles
CONTENTS
9.1 Introduction...........................................................................................................................544
9.2 Synthesis of Reactive Thermally Sensitive Latex Particles...................................................544
9.2.1 Rapid State of the Art................................................................................................546
9.2.2 Kinetic Study............................................................................................................. 547
9.2.2.1 Effect of Initiator........................................................................................ 549
9.2.2.2 Effect of Temperature................................................................................. 550
9.2.2.3 Effect of the Cross-Linker Agent................................................................ 551
9.2.2.4 Functionalization Studies and Operating Methods.................................... 552
9.2.2.5 Polymerization Mechanism........................................................................ 553
9.2.2.6 Thermally Sensitive Core–Shell Particles.................................................. 556
9.3 Colloidal Characterization..................................................................................................... 557
9.3.1 Morphology............................................................................................................... 558
9.3.2 Effect of Temperature on Hydrodynamic Particle Size............................................. 559
9.3.3 Electrokinetic Study.................................................................................................. 560
9.3.4 Volume Phase Transition Temperature (TVPT)........................................................... 562
9.3.5 Colloidal Stability...................................................................................................... 562
9.4 Immobilization of Biomolecules...........................................................................................564
9.4.1 Protein Adsorption..................................................................................................... 565
9.4.1.1 Effect of Temperature on the Adsorption of Protein onto
Poly[NIPAM] Particles............................................................................... 565
9.4.1.2 Effect of Ionic Strength on the Adsorption of Protein onto
Poly[NIPAM].......................................................................................... 566
9.4.1.3 Desorption Study........................................................................................ 566
9.4.2 Adsorption of Nucleic Acids...................................................................................... 567
9.4.2.1 Adsorption Kinetic...................................................................................... 567
9.4.2.2 Influence of pH and Ionic Strength............................................................. 567
9.4.2.3 Desorption Study of Pre-adsorbed Nucleic Acids...................................... 569
9.4.2.4 Specific Extraction of Nucleic Acids.......................................................... 570
9.4.3 Amplification of Nucleic Acids................................................................................. 571
9.5 Thermally Sensitive Particles or Nanogels for In Vivo Drug Release.................................. 572
9.6 Conclusion............................................................................................................................. 579
Acknowledgments........................................................................................................................... 580
References....................................................................................................................................... 580
543
544 Handbook of Surface and Colloid Chemistry
9.1 INTRODUCTION
Polymer latex particles are widely used as a solid support in numerous applications and especially
in the biomedical field, due to the existence of various polymerization processes (emulsion, disper-
sion, microemulsion, etc.) for preparing latex that is well-defined in terms of particle size, reactive
groups, surface charge density, colloidal stability, etc. Since 1986, precipitation polymerization of
alkylacrylamide and alkylmethacrylamide derivatives (water-soluble monomers) has been found to
be a convenient method for producing submicronic, functionalized thermally sensitive latex hydro-
gel particles, as reported by Pelton and Chibante [1]. Since then, thermally responsive microgel latex
particles have played a particular and considerable role in academic research and industrial appli-
cations. In academic research, studies are mainly focused on the polymerization mechanism and
colloidal characterization in dispersed media. From an application point of view, stimuli-responsive
microgels have been principally explored in drug delivery as a carrier in therapy.
The implication of such stimuli-responsive particles as a solid polymer support of biomolecules
in the biomedical field is probably due to various factors: (1) easiest to prepare via precipitation
polymerization (hydrogel particles) or a combination of emulsion and precipitation polymerizations
(core–shell particles); (2) the colloidal properties are related to the temperature and to the medium
composition (i.e., pH, salinity, surfactant, etc.); (3) the adsorption and the desorption of antibodies
and proteins are principally related to the incubation temperature; (4) the covalent binding of pro-
teins onto such hydrophilic and stimuli-responsive particles can be controlled easily by temperature;
and finally (5) the hydrophilic character of the microgel particles is an undeniably suitable environ-
ment for immobilized biomolecules.
The main objective of this chapter is to report on the preparation, characterization of thermally
sensitive particles, and the pertinent aspects that should be considered before their utilization as a
polymer support in the biomedical field. This will be followed by an examination of the prepara-
tion of such hydrophilic thermally sensitive latex particles bearing reactive groups. Subsequently,
the colloidal characterizations that are to be taken into consideration will be presented. Finally, the
chapter will be closed by presenting and illustrating recent applications of thermally sensitive poly-
mer colloids as solid supports in the biomedical field.
FIGURE 9.1 Scheme of thermally sensitive linear polymer as a function of temperature. (From Pelton, R.H.
and Chibante, P., Colloid Surf., 20, 247, 1986.)
FIGURE 9.2 Illustration of particles morphology: (a) hard sphere (i.e., polystyrene), (b) microgel struc-
ture (polyNIPAM), (c) core–shell like (i.e., polystyrene core-polyNIPAM shell), and (d) composite particles
(i.e., hard magnetic polymer core and water soluble polymer shell).
In addition, the LCST of thermosensitive polymer has been widely studied using different
physical methods—including turbidity, fluorescence, dynamic light scattering, and calorimetric
measurements—in order to demonstrate the relationship between the polymer properties and sol-
vent conditions, as reported by Schild [3] in a very thorough review concentrated on poly(NIPAM).
The elaboration of thermally sensitive colloidal particles has been largely studied as can be
evidenced by the numerous reported publications. In fact, three kinds of thermally sensitive colloi-
dal particles have been reported: hydrogel and core–shell particles and some composite thermally
structured core–shell particles (Figure 9.2).
The most examined microgels are poly(N-isopropylacrylamide) [1,4], poly(N-isopropylmeth-
acrylamide) [5], poly(N-ethylmethacrylamide) [6], and more recently, poly(N-vinyl caprolactam)-
[7–10] based hydrogel particles.
This section is principally devoted to the preparation of thermally sensitive hydrogel particles
(microgels and core–shells) using batch polymerization process. The effect of each reactant and
parameter (initiator, temperature, cross-linker agent) on the polymerization process (polymeriza-
tion kinetic, conversion, final particle size, morphology, water-soluble polymer, etc.) is presented
and discussed. Special attention will be focused on the functionality of the elaborated thermally
sensitive particles.
546 Handbook of Surface and Colloid Chemistry
0.5 μm
FIGURE 9.3 Transmission electron microscopy of cationic poly(NIPAM) microgel particles deposited from
water at a temperature below the TVPT. Due to the hydrophilic character of such particles, hexagonal arrange-
ment is generally observed when the particles are dried at a temperature below the TVPT. (From López-León,
T. et al., J. Phys. Chem. B, 110, 4629, 2006.)
Thermally Sensitive Particles 547
+ Cross-linker
T < LCST
Cross-linked gel
FIGURE 9.4 Polymerization in the aqueous phase of amphiphilic-like monomers, leading to thermally
sensitive material preparation (T is the polymerization temperature).
explored [6]. Due to the high reactivity of MBA compared to NEMAM, high water soluble
polymer formation has been observed and with almost no detectable particles in medium.
Consequently, to prepare well-defined microgel particles, it is of paramount importance to take
into account not only the reactivity of the principal monomer but also the reactivity of the used
cross-linker agent.
More recently, special attention has been focused on the elaboration of vinyl caprolactam (VCL)–
based microgels. Poly(N-vinylcaprolactam) is thermally sensitive in nature (in the same LCST range
as for polyNIPAM) and biocompatible as reported [14]. To prepare VCL-based particles, cross-
linker agents such as MBA and charged initiator are used. To some extent, all the study reported
on the preparation of polyNIPAM-based microgels (the most studied system) has been extended to
VCL monomer. The use of classical water-soluble cross-linker agent (i.e., MBA) to elaborate poly-
VCL microgel may lead to non-biocompatible microgels for in vivo applications.
9.2.2 Kinetic Study
To prepare stable microgel particles, the polymerization should be absolutely higher than the LCST
of the corresponding homopolymer (or copolymer). In fact, without the precipitation of the formed
chains (or lowly cross-linked chains) during the polymerization process, no real particles can be
obtained. For instance, the lower critical polymerization temperatures (LCPT) in the case of NIPAM,
NIPMAM, and NEMAM monomers are 65°C, 70°C, and 80°C, respectively. The observed LCPT
is related to the properties of the formed oligomers (chain light, molecular weight, composition, and
charge distribution). This point is clearly illustrated in the polymerization mechanism described in
the following text, but not really established experimentally.
Water-soluble alkylacrylamide or alkylmethacrylamide monomers conversion during the
polymerization process can be easily determined or followed using both gas chromatography
and 1H-NMR techniques. According to the high reactivity of the generally used water-soluble
cross-linkers (i.e., MBA) and the low concentration used in the polymerization recipe, only the
main monomer conversion can be easily followed as a function of polymerization time. For most
alkyl(metha)acrylamide monomers studied, the polymerization kinetic was generally found to be
rapid. In fact, only 30 min were needed to reach total conversion in the case of NIPAM/MBA/V50
or KPS (polymerization temperature from 60°C to 80°C) [7] and 120 min in the case of NIPMAM/
MBA/KPS (polymerization temperature from 60°C to 80°C) [5]. The difference in polymerization
rate between NIPAM and NIPMAM is related to the high propagation rate constant (kp) for NIPAM
compared to NIPMAM [4,15], as well as for the known reactivity of alkylacrylamides compared
548 Handbook of Surface and Colloid Chemistry
Cross-linker MBA
K+–3OS–O–O–SO3– K+
Initiator KPS
O NH NH O
O NH
NIPAM PNIPAM
to alkylmethacrylamides derivatives (kp ~ 18,000 L mol−1 s−1 for acrylamide [16,17] and kp ~ 800 L
mol−1 s−1 for methacrylamide [18] at 25°C) (Figure 9.5).
The detailed analysis of particle size versus polymerization time and conversion is one of the
best methodologies generally used in polymerization in dispersed media to point out the particle
evolution during the polymerization process. The diameter of hydrogel particles (NIPMAM/MBA/
KPS at 70°C) [5] versus (conversion)1/3 reported in Figure 9.6 was linear relationship, reflecting
the fact that the number of particles remained constant and no new particles were formed during
the polymerization process. In fact, the particle size (R) is related to the particle number (Np) and the
polymer mass (m, i.e., polymer conversion) by assuming the constancy of the particle density during
the polymerization process:
The observed behavior confirmed by scanning electron microscopy (SEM) analysis of the particle
size distribution versus polymerization time is illustrated in Figure 9.7. The utilization of the previ-
ously described investigations of both particle size versus conversion1/3 and size distribution versus
time using TEM or SEM are related to the difficulty of particle number concentration determination
(or estimation) induced by the swollen property of the microgels.
1000
QELS at T < TVPT
800
Diameter (nm)
200
0
0 1 2 3 4 5
Conversion1/3
FIGURE 9.6 Illustration of particle size of hydrogel versus conversion in the case of classical alkylacryl-
amide monomers. See Duracher’s work for experimental data with NIPMAM monomer. (From Duracher, D.
et al., J. Polym. Sci. A: Polym. Chem., 37, 1823, 1999.)
Thermally Sensitive Particles 549
FIGURE 9.7 Scanning electron microscopy (SEM) images of poly(NIPMAM) particles at different polym-
erization times. (From Duracher, D. et al., J. Polym. Sci. A: Polym. Chem., 37, 1823, 1999.)
R p ⊕[ I ]a (9.2)
100
Monomer conversion (%)
80
60
40
Low concentration
20 High concentration
0
0 5 10 15 20 25 30
Times (min)
FIGURE 9.8 Monomer conversion versus polymerization time for two initiator concentrations. Example for
poly(NIPAM) particles: polymerization temperature T = 70°C, [NIPAM] = 48.51 mmol, [MBA] = 3 mmol,
[V50] = 0.1 (—) to 1 mmol (- - - -) for total volume = 250 mL. (From Meunier, F., Synthèse et caractérisation
de support polymères particulaires hydrophiles à base de N-isopropylacrylamide, Elaboration de conjugués
particules/ODN et leur utilisation dans le diagnostic médical, Thèse, University of Lyon-1, France, 1996.)
550 Handbook of Surface and Colloid Chemistry
10
0.1
10–4 10–3 10–2
–3
Initiator concentration (mol dm )
FIGURE 9.9 Dependence of polymerization rate on the initiator concentration in log–log scale. Temperature =
70°C, [NIPAM] = 48.51 mmol, [MBA] = 3 mmol, and total volume = 250 mL (Rp ≈ [I]0.18, I: V50 initiator
concentration). (From Meunier, F., Synthèse et caractérisation de support polymères particulaires hydrophiles
à base de N-isopropylacrylamide, Elaboration de conjugués particules/ODN et leur utilisation dans le diag-
nostic médical, Thèse, 1996.)
For instance a ≈ 0.18 for NIPAM/MBA/KPS at 70°C (Figure 9.9), and a ≈ 0.4 in the case of emul-
sion polymerization. With such a complex system, it is more appropriate to consider the relationship
between the number of particles (Np) and the reactant concentration, then:
N p ⊕[ M ]a [ I ] b [CL ]c (9.3)
where
[M] is the monomer
[CL] is the cross-linker agent
a, b, and c are the scaling exponents
N dec = I (1 − e − kd t ) (9.4)
where
kd is the decomposition rate constant of the initiator at a given temperature
I is the initial concentration of initiator
Ndec is the amount of decomposed initiator after a given time and at a given temperature
Thermally Sensitive Particles 551
10
0.1
50 60 70 80 90 100
T (°C)
FIGURE 9.10 Polymerization rate versus polymerization temperature in semilog scale. [NIPAM] = 48.51
mmol, [I] = 0.3 mmol, [MBA] = 3 mmol, and total volume = 250 mL. (From Meunier, F., Synthèse et carac-
térisation de support polymères particulaires hydrophiles à base de N-isopropylacrylamide, Elaboration de
conjugués particules/ODN et leur utilisation dans le diagnostic médical, Thèse, 1996.)
R p = k p [ M ] ⋅ nN
p (9.5)
50
30
20
10
0
0 5 10 15 20 25
Cross-linker (%)
FIGURE 9.11 Effect of MBA (2%–20% w/w) on water-soluble polymer formation using NIPMAM (1 g)
monomer and KPS (2% w/w) initiator, polymerization temperature 70°C. (From Duracher, D. et al., J. Polym.
Sci. A: Polym. Chem., 37, 1823, 1999.)
1 2 3 3 1
Cross-linker Cross-linker
distribution distribution
• Batch polymerization: Polymerizations performed in a closed reactor, with all the ingre-
dients being introduced at the beginning of a single step. This method, apart from certain
exceptions, is of little interest, since a large part of the functional monomer is consumed,
providing substantial quantities of water-soluble polymers that disturb nucleation and the
final stabilization of the particles.
Thermally Sensitive Particles 553
CH3
HC CH2 H2C CH
C O
N H
CH2 COOH
CH2 N CH
CH2 COOH H3C CH3
(a) (b)
FIGURE 9.13 Structure of N-(vinylbenzylimino)-diacetic acid (IDA) (a) and N-isopropylmethacrylamide (b).
• Semicontinuous addition methods: These are very useful for copolymerizations requir-
ing the control of homogeneous compositions at the level of the chain and the particle.
Variable composition gradient processes permit the introduction of the functional mono-
mer to a suitable conversion, favoring incorporation on the surface or inside the particles.
• Multistage polymerizations, including the deferred addition of an ionic comonomer
(constituting the basic latex), favoring a highly efficient surface functionalization.
• Seed functionalization: These methods consist of the functionalization of a latex seed by
a monomer or monomer mixture. This often permits surface incorporation to be increased
and is well adapted for formulating controlled charge density model colloids.
• Postreaction on reactive latexes: This process is very useful for modifying the functional-
ity of a given latex if it cannot be obtained directly.
The preparation of functionalized thermally sensitive microgel particles has not been sufficiently
investigated, since only a few works have been reported. The first systematic work on the preparation
of functionalized poly(NIPAM) hydrogel particles was studied using NIPAM/MBA/AEMH (ami-
noethylmethacrylate hydrochloride) in a batch polymerization process, and first reported by Meunier
et al. [4] and then by Duracher et al. [21] during an investigation of the effect of N-(vinylbenzylimino)-
diacetic acid (IDA) on the polymerization of NIPMAM/MBA/KPS (Figure 9.13). The effect of the
charged functional monomer on the precipitation polymerization of such alkylacrylamide and alkyl-
methacrylamide monomers (i.e., NIPAM and NIPMAM) was found to resemble the effect of the
ionic (or water-soluble) monomer on the emulsion polymerization of styrene, for instance. In fact, the
increase in the functional monomer concentration leads to rapid polymerization, high polymeriza-
tion conversion (>95%), low particle size (Figure 9.14), and high water-soluble polymer formation.
In addition to batch functionalization, the shot-grow process was also performed in order to pre-
pare amino-containing thermally sensitive poly(NIPAM) hydrogel [4] and core–shell (polystyrene
core and poly(NIPAM) shell) particles [22]. The obtained results revealed a good functionalization
yield with a non-negligible amount of water-soluble polymer formation.
As in the case of emulsion polymerization, an increase in the functional monomer concentration
leads to a reduction in the final hydrodynamic particle size and enhanced water-soluble polymer
formation, as illustrated in Figures 9.15 and 9.16 for the batch polymerization of NIPMAM/MBA/
IDA/KPS. The reduction in particle size versus functional monomer has been attributed to the
enhancement of precursor formation and the number of stable particles, which rapidly become the
polymerization loci.
(a) (b)
(c) (d)
FIGURE 9.14 The influence of IDA (charger monomer) on final particle size and size distribution as exam-
ined by scanning electron microscopy (SEM) [21] (the photos are in the same scale): (a) 0%, (b) 0.6%, (c) 1.2%,
and (d) 2%.
400
350 10°C
60°C
300
Diameter (nm)
250
200
150
100
50
0
0 0.005 0.01 0.015 0.02 0.025 0.03
Functional monomer (IDA)
FIGURE 9.15 Effect of functional monomer (IDA in g) on hydrodynamic particle diameter measured by
quasi elastic light scattering (QELS) at 10°C and 60°C. (From Duracher, D. et al., Macromol. Symp., 150,
297, 2000.)
Thermally Sensitive Particles 555
40
35
25
20
15
10
0
0 0.005 0.01 0.015 0.02 0.025 0.03
Functional monomer (IDA)
FIGURE 9.16 The influence of IDA (in g) on water-soluble polymer (WSP) formation. (From Duracher, D.
et al., Macromol. Symp., 150, 297, 2000.)
• Water phase polymerization: Before adding the water-soluble initiator, the medium was
totally homogeneous and relatively limpid, even at the polymerization temperature. After
adding the initiator, the polymerization requires a low induction period (t*) (generally less
than 3 min in the case of NIPAM/MBA/KPS or V50 system using recrystallized reac-
tants), probably due to the presence of oxygen traces in the aqueous phase. Above the t*
period, the decomposed initiator leads to rapid oligomer formation. The oligomer concen-
tration is principally governed by the initiator decomposition rate and efficiency.
l2 2l* Oligomers
T > LCST formation
Precursors Primary
formation particles
Growth of particles
Oligomer
End of polymerization
Monomer
WSP
T < LCST
• Nucleation step: The nucleation period is mainly due to the precipitation of the oligomers
formed when the critical chain lengths (or molecular weight and composition) are reached
or when the cross-linked chains start to precipitate. In fact, the scanning electron micros-
copy (SEM) (Figure 9.7) analysis of the particle size and size distribution revealed rapid
and narrow particle distribution formation (i.e., generally in less than 5 min in the case of
NIPMAM/MBA/KPS and NIPAM/MBA/V50 systems). In accordance with highly reac-
tive water-soluble cross-linkers such as MBA, it may be assumed that the first oligomers
contain a significant amount of the cross-linker agent, and of primary particles.
• Growth of particles: After the rapid nucleation period, the polymerization of residual
monomer may take place at various domains (i.e., water phase, on the surface of the formed
particle, and in the swelled parts of the particles). In fact, the partition of the monomer
between the water phase and polymer particles should be considered: (1) polymerization
in the water phase leads to the formation of small oligomers that can be cross-linked on
the formed particles when there is a sufficient residual cross-linker agent concentration,
(2) and (3) the possible polymerization of the monomer in the swollen state. When the
cross-linker agent is totally consumed, the oligomers formed contribute to water-soluble
polymer formation. The water-soluble polymer formed may originate from the desorption
process of adsorbed chains onto the particles when the dispersion is cooled (temperature <
TVPT or < LCST), or from low molecular weight chains that are highly water-soluble even
above the LCST of the considered homopolymer.
(a) (b)
(c) (d)
FIGURE 9.18 TEM micrographs of the final PNEMAM microgel particles as a function of the cross-linker
concentration (wt% EGDMA) [25]. For low cross-linked microgels, the hexagonal arrangement was clearly
observed as for polyNIPAM microgel: (a) 8% EGDMA (D TEM = 20 nm), (b) 15% EGDMA (D TEM = 180 nm),
(c) 30% EGDMA (D TEM = 193 nm), and (d) 45% EGDMA (D TEM = 211 nm).
reactivity of the used hydrophobic cross-linker and then constitute the core, whereas, the shell is
formed composed of poly(NEMAM) network [6,25,26]. As unexpected, spherical and monodis-
perse particles are obtained irrespective of the amount of hydrophobic cross-linker as illustrated in
Figure 9.18 via the TEM analysis of crude particles.
9.3.1 Morphology
The morphological characterization of structured latexes is a fundamental aspect of their study, since
(1) it provides very useful information on the nature of the mechanisms that regulate the particle’s
constitution, and (2) knowledge of the organization of the polymer within the particle is the essential
foundation for the theoretical interpretation of the behavior of the resulting latex films (mechanical
properties, permeability, etc.). From this perspective, there are a great many techniques that require
examination in order to eliminate artifacts and wrong conclusions deduced from their use.
Various methods have been reported and published. Of these methods, electronic microscopy is
still preferred for studying structured latexes, since improvements in observation techniques (sam-
ple preparation, introduction of selective marking of one of the components) permit directly an
increasingly refined analysis of polymers inside particles.
Scanning electron microscopy (SEM) can also be used for both surface morphology and particle
size and size distribution, as illustrated in Figure 9.19, for thermally sensitive polystyrene core-
cross-linked poly(NIPAM) shells. In addition, atomic force microscopy can be used for investigat-
ing colloidal particle morphology, as reported by Duracher et al. [22].
In the case of batch radical polymerization of NEMAM/EGDMA/KPS, the thickness layer is
mainly governed by the amount of EGDMA in the polymerization recipe as examined by atomic
0.50
0.25
μm
0.00 0.8
0.8 0.4
μm
μm 0.4
0
0
(a) (b)
FIGURE 9.19 (a) Scanning electron microscopy (SEM) and (b) atomic force microscopy (AFM) of poly-
styrene (core)-cross-linked poly(NIPAM) shell microspheres. (From Duracher, D. et al., Colloid Polym. Sci.,
276, 219, 1998.)
Thermally Sensitive Particles 559
0.9
0.8
0.7
0.6
Peak/valley
0.5
0.4
0.3
0.2
0.1
0
0 5 10 15 20 25 30 35 40 45 50
Cross-linker amount (%)
FIGURE 9.20 Peak to valley ratio versus EGDMA concentration in NEMAM/EGDMA/KPS system
(1% KPS and 15% wt/wt EGDMA). (From Hazot, P. et al., J. Polym. Sci. A: Polym. Chem., 40, 1808,
2002.)
force microscopy (AFM) via the determination of pick to valley ratio (Figure 9.20). The measured
ratio was found to increase (from 0.4 to 0.9) with increasing the cross-linker amount. Pick to valley
ratio is close to 1 for 45% cross-linker amount. Consequently, the increase in the EGDMA concen-
tration leads to more rigid particles.
kT
D= (9.6)
3πηDh
where
D is the diffusion coefficient
k is the Boltzmann constant
T is the absolute temperature
η is the viscosity of the medium
Light scattering technology (QELS) is generally suitable for low particle sizes (i.e., diameter <
1 µm). For polystyrene latex particles bearing low charges, density, and hydrophobic surface, the
particle size does not generally depend on temperature and salinity. However, in the case of hydro-
philic and poly(N-alkylacrylamide) derivatives, particle size and swelling ability are affected by the
solvent quality of the dispersion medium, temperature, pH, and salinity. By way of illustration, the
particle size for various types of particle poly(NIPAM) as a function of temperature is reported in
Figure 9.21.
This decrease (in particle size vs. T) is more marked in the case of large particles (with low
cross-linking properties), due to the pronounced size difference below and above the volume
phase transition temperature (TVPT) between the shrunken and the expanded state. In addition,
560 Handbook of Surface and Colloid Chemistry
1200
1000
Dh (nm) 800
600
400
200
20 25 30 35 40 45
T (°C)
FIGURE 9.21 Hydrodynamic size of negative (•) and positive (◾) poly(NIPAM)-based microgels as a func-
tion of temperature in a 10 −3 M NaCl solution. Vertical lines have been drawn to guide the eye toward the TVPT
values. Below the TVPT χ12 = 0 in good solvent (extended microgel) and above the TVPT χ12 > 0.5 in poor solvent
(Shrunken microgel) [9.3].
the effect of temperature is more marked in the case of total hydrogel microspheres compared
to the core–shell particles, as illustrated earlier (Figure 9.21). The behavior observed has already
been reported by many authors with regard to poly(NIPAM) microgel particles, and attributed to
the breaking of hydrogen bonds between polymer network and water molecules during the heating
process, which induces a decrease in hydrodynamic particle size. Such light scattering experi-
ments can provide two major pieces of information concerning the dispersion: (1) the effect of
temperature on particle size (i.e., volume phase transition domain) and (2) the swelling ability of
the considered colloidal particles. The swelling ratio calculated from the hydrodynamic particle
size above and below the volume phase transition temperature reflects the cross-linking density
of the particle but not the cross-linker distribution, as recently reported by Guillermo et al. [20]
using the NMR technique as described in Section 9.2.2.3 (Figure 9.12). To compare the swelling
capacity of the microgel particles (or core–shell latexes) with different functional monomer con-
centrations, charge densities, and internal structures, the swelling ratio (Sw) was introduced and
defined by the following equation:
V
Sw = (9.7)
Vc
where
V and Vc represent the particle volume calculated from the hydrodynamic radius as determined
by QELS
Vc represents the collapsed volume
9.3.3 Electrokinetic Study
The investigation of electrophoretic mobility versus pH and temperature can also be considered as
a key point for the analysis of surface charge density variation. In fact, the electrophoretic mobil-
ity of thermally sensitive microgel particles is drastically affected by the medium temperature.
Thermally Sensitive Particles 561
The decrease in particle size, and subsequent increase in surface charge density, results in an
increase in electrophoretic mobility, as expressed by the following relationship between (µe) and
particle size (R):
Ne
μe ≈ (9.8)
4πηR 2 κ
where
N is the number of charged groups per particle of hydrodynamic radius R
e is the electron charge
κ is the reciprocal of Debye length thickness
η is the viscosity of the medium.
The effect of temperature on the electrophoretic mobility of polyNIPAM microgel particles has
been investigated as first reported by Pelton and Chibante [1], Kawaguchi et al. [12], and then by
Meunier et al. [4].
As illustrated in Figure 9.22, the electrophoretic mobility increased (in the absolute value)
together with the temperature, irrespective of the nature of the surface charge. Such behavior is
related to the surface charge density versus temperature. The amplitude of the measured transition
in the electrokinetic property was more marked for low cross-linked thermally sensitive particles
(i.e., high swelling ability). The electrokinetic study of such thermally sensitive particles needs
particular attention in order to demonstrate the location of charges implicated in electrophoretic
mobility and what is the relationship between the volume phase transition temperature and the elec-
trokinetic transition temperature.
To point out the relationship between the electrophoretic mobility (i.e., charge density) and
the reduction of the particle size, the electrophoretic mobility has been examined as a function of
(rh−2). When the linearity (μ e vs. rh−2) is observed, it reflects the homogeneous charge distribution
on the microgel “surface” and constant reduction of the particles size. When the nonlinearity is
observed, it may reflect the complexity of the microgel structure and also the charge distribution
in the vicinity of the immobile water molecules [13].
5
4
3
2
μe (10−8 m2 V−1 s−1)
1
0
–1
–2
–3
–4
–5
20 25 30 35 40 45
T (°C)
FIGURE 9.22 Electrophoretic mobility of the negative (λ) and positive (ν) poly(NIPAM) microgels as a
function of temperature (10 −3 M NaCl). (From López-León, T. et al., J. Phys. Chem. B, 110, 4629, 2006.)
562 Handbook of Surface and Colloid Chemistry
1 Linear
0.6
0.4
0.2
0
20 25 30 35 40 45 50 55 60
Temperature (°C)
FIGURE 9.23 Normalized optical density versus temperature for both linear and microgel poly(NIPAM)
polymer (for highly diluted polymer and low salinity concentration).
9.3.5 Colloidal Stability
The colloidal stability has been generally investigated via the stability factor (i.e., Fush factor, W)
determination as a function of salinity and temperature using the turbidity-based method or any clas-
sical spectrophotometer equipment. The dispersion was highly diluted in water at given pH and tem-
perature. The aggregation rate constants (dOD/dt) were determined by measuring the optical density
(OD) variation at 600 nm wavelength as a function of time, after adding NaCl solution. The critical
coagulation concentration (CCC) of the dispersion was deduced from the stability factor (W) varia-
tion as a function of ionic strength (below and above the LCST) plotted in log–log scale. The stability
factor was determined using the following classical equation relating W to aggregation rate constants:
(dOD /dt ) f
W= (9.9)
(dOD /dt )s
Thermally Sensitive Particles 563
where (dOD/dt)f and (dOD/dt)s are the initial slopes of the optical density (OD) variation as a func-
tion of time (t), for fast and slow aggregation processes, respectively. The same methodology can be
also used in order to investigate the effect of pH or any additive on the colloidal stability.
The colloidal stability of thermally sensitive colloidal microgels has been largely examined in
terms of experimental studies. In this direction, interesting and complete work has been recently
published by López-León et al. [13]. Interestingly, the salinity affects not only the colloidal stability
but also the swelling ability of the microgel. In fact, the increase in the salinity reduced the solubil-
ity of thermally sensitive polymer (i.e., polyNIPAM, polyNIPMAM, etc.) and consequently induced
reduction of the LCST and then the colloidal stability of the particles. Since the salinity affects the
volume phase transition temperature of microgels, it is interesting to take into account not only the
effect of salinity on the colloidal stability but the effect of indiscernible temperature and salinity
couple on the colloidal stability as illustrated by the stability diagram (Figure 9.24a).
As a general tendency, the colloidal stability of thermally sensitive latex particles is related to both
temperature and salinity. In fact, in the case of linear N-alkylacrylamide or methacrylamide-based
40
35 Unstable
T (°C)
30
25
Stable
20
0 200 400 600 800 1000
(a) NaCl (mM)
High
temperature Stable
Collapsed state
TVPT Unstable
Extended state
Low Stable
temperature
Low salinity Salinity High salinity
(b)
FIGURE 9.24 (a) Stability diagram for the negative (λ) and positive (ν) microgels as a function of salt con-
centration and temperature. (From López-León, T. et al., J. Phys. Chem. B, 110, 4629, 2006.) (b) Schematic
illustration of diagram stability of thermally sensitive particles as a function of temperature and salinity.
564 Handbook of Surface and Colloid Chemistry
polymers, the increase in temperature leads to the precipitation polymer chains as discussed earlier.
In addition, the increase in salinity reduces the lower critical solution temperature (LCST) of con-
sidered polymer in water. The earlier presented effects (temperature and salinity) for linear polymer
are also available for microgel and core–shell thermally sensitive particles. The increase in salinity
reduces the volume phase temperature (TVPT) and for high salinity medium, the colloidal stability
may be affected and the aggregation process may take place as illustrated in Figure 9.24b, in which
the colloidal stability domains are depicted versus salinity and temperature. The critical coagula-
tion concentration (CCC) of such a system is high below the volume phase transition temperature in
which the particles are in extended state and sterically stabilized. Whereas above the volume phase
transition temperature, the CCC is low revealing that the particles are electrostatically stabilized.
In addition, to some extent such behavior is reversible by cooling the dispersion and reducing the
salt concentration.
9.4.1 Protein Adsorption
Various kinds of colloidal particles have been used in biomedical domains. In analytical chemistry,
they are used as solid supports for sample preparation. In the drug delivery field, nanocolloids, and
particularly stimuli responsive polymer-based nanogels, have been intensively explored as protein car-
riers for in vivo applications. The reported studies in this direction are mainly focused on the release
efficiency rather than on the driven forces involved in the loading and the release of the loaded proteins.
Due to its low critical solubility temperature (LCST) around 32°C, poly(N-isopropylacrylamide)-
based material has long been studied in regard to life science such as versatile tools for the sepa-
ration and purification of proteins [30], drug delivery [31,32], control of enzyme activity [33–36],
therapeutics, and diagnostics [29,37,38]. Pioneering work of Kawaguchi et al. [30] pointed out the
temperature dependency of protein loading onto anionic polyNIPAM microgel particles, the results
being interpreted in terms of hydrophobic interactions caused by the dehydration of the colloidal
microgels. However, Elaïssari et al. [39] reported that the loading of proteins onto charged ther-
mally sensitive similar microgels was mainly governed by electrostatic interactions.
0.8
(Ns/Ns,max)
0.6
0.4
0.2
0
10 20 30 40 50 60
Temperature (°C)
FIGURE 9.25 Reduced amount of HIV-1 capsid P24 protein adsorbed onto thermally sensitive polystyrene
core-cationic cross-linked poly(NIPAM) shell as a function of temperature (at pH 6.1, 10 mM phosphate buffer).
566 Handbook of Surface and Colloid Chemistry
100
60
40
20 At 40°C
At 20°C
0
10–6 10–5 10–4 10–3 10–2 10–1 100
NaCl concentration (mol/L)
FIGURE 9.26 Effects of electrolyte concentration [NaCl] on modified HIV-1 protein (P24) adsorp-
tion onto cationic thermally sensitive core–shell microspheres at 40°C, pH 6.1 at 20°C (⦁) and 40°C (◽).
(From Duracher, D. et al., Langmuir, 13(23), 9002, 2000.)
temperature is well controlled in the case of attractive electrostatic interactions and only the plateau
is drastically affected by the pH, salinity, and surface charge density.
Protein
T > TVPT
T < TVPT
Highly hydrated surface Latex particle Low hydrated surface
Low surface charge density bearing thermally High surface charge density
No or low adsorption sensitive shell High adsorption
tangible interpretation of the protein adsorption and desorption processes should take account
the colloidal particle properties (i.e., cross-link density, charge distribution, and hydrophilic–
hydrophobic balance) and the protein characteristics (flexibility, charge density, distribution of
hydrophobic domains, etc.).
0.8
High charged particles
Ns/Ns,max
0.6
0.4
0.2
Low charged particles
0
3 4 5 6 7 8 9 10
pH
FIGURE 9.28 Reduction in adsorption of nucleic acids onto high cationic polystyrene latexes and low
cationic poly(NIPAM) microgel particles as a function of pH at 20°C and 10 −3 M ionic strength. (From
Elaissari, A. et al., J. Biomater. Sci. Polym. Edn., 10, 403, 1999.)
adsorption of polyelectrolytes onto oppositely charged solid supports. The attractive electrostatic
interactions are the driving forces in the adsorption process of DNA, RNA, and ssDNA [42,43] onto
oppositely charged polymer supports. The variation of the quantity of nucleic acids adsorbed onto
cationic thermally sensitive poly(NIPAM) latex particles as a function of both pH and ionic strength
are shown in Figures 9.28 and 9.29, respectively.
As for classical polyelectrolytes, the adsorption of oligodeoxyribonucleotides (ssDNA) is
basically related to the ssDNA and the adsorption energy as described by the following equation:
where
k is a constant characterizing the studied system
n is the polymerization degree (i.e., number of bases, for instance, dT[35], n = 35)
ΔG is the adsorption energy per monomer (per base)
0.8
0.6
Ns/Ns,max
0.4
0.2
FIGURE 9.29 Reduction in amount of nucleic acids adsorbed onto (amidine groups 5 µmol/g, ⦁) and amine
(amine and amidine groups, 14 µmol/g, ◾) poly(NIPAM) microgel particles, and (amine and amidine groups 5)
thermally sensitive magnetic bearing poly(NIPAM) shells [28] as a function of NaCl concentration (at pH 4.5
and at 20°C). (From Ganachaud, F. et al., Langmuir, 13, 701, 1997.)
Thermally Sensitive Particles 569
The adsorption energy in the case of nucleic acids/polymer particles is the sum of hydrophobic
adsorption energy (Ψhydrophobic) attributed to the staking adsorption process (ssDNA/negatively
charged polystyrene latex) and electrostatic adsorption energy (Ψelectrostatic) related to the charge–
charge interaction:
In the case of such hydrophilic thermally sensitive microgel particles, the hydrophobic adsorption
can be totally neglected and the electrostatic term can be described by the following relationship:
where σssDNA and σlatex are the charge density of ssDNA and latex particles, respectively. The charge
density of ssDNA fragment and the surface charge density of latex particles can be approximately
expressed as follows:
n|e|
σssDNA ≈ (9.13)
L
εκσ
σlatex ≈ (9.14)
4π
where
L is the chain length of a given ssDNA
ζ is the zeta potential of latex particles
κ is the inverse of double layer thickness
ε is the dielectric constant
In the pH range generally investigated in the adsorption study, the σssDNA is negatively charged,
whereas, the σlatex is pH dependent as evidenced from any electrokinetic study (zeta potential versus
pH). The electrostatic adsorption energy is thus expected to vary linearly with respect to the latex
surface charge density leading to linear variation of log(Ns) versus latex surface charge density or
zeta potential as well evidenced by Elaissari et al. [29].
Ds(mg/g) 3
0
4 5 6 7 8 9 10
pH
FIGURE 9.30 The effect of pH on the desorption of preadsorbed nucleic acid molecules onto cationic ther-
mally sensitive magnetic latex particles. Adsorption was performed using 10 mM phosphate buffer, pH ~ 5.2,
10 −3 M NaCl at 20°C, with an incubation time of 180 min. (From Elaissari, A. et al., J. Magn. Magn. Mater.,
225, 127, 2001.)
The extraction of the RNA is a real stake in the biomedical diagnosis where the sensitivity,
specificity, and rapidity (fast analysis) are required. There are techniques allowing the specific
capture of nucleic acid molecules, but in this case, it is necessary the use of solid phase on which
well-defined oligonucleotides sequence are fixed, for instance, the use of reactive colloidal parti-
cles [40] bearing oligonucleotides (polyT) for the capture of mRNA (eukaryotes having polyadenyl
[polyA] tail). The extraction of the nucleic acids (DNA or RNA) is also performed by using the
differential precipitation based on centrifugation on cesium chloride. In this case, only the RNA is
recovered. Lastly, the most widespread technique called mini-chromatography based the use of a
mini-column material, which fixes RNA. These extracted nucleic acids thereafter are purified and
eluted in a low volume, which led to concentrated nucleic acid molecules.
The principal problem generally resides in the specific extraction of RNA from any biological
sample containing DNA. The single exit in this case is the selective extraction of ARN that requires
examining the chemical difference between DNA and RNA molecules. The ribose in 3′ position of
RNA presents a cis-diol function, which led to a complex formation with boronic acid, compound,
this in a specific way, even in the presence of DNA.
The specific recognition of cis-diol function by the boronic acid derivatives was the subject of
many researches. In fact, certain authors used this affinity for controlling the reversible immobiliza-
tion of proteins, enzymes, or all biomolecules bearing glucose site. This interaction is sensitive to
the pH of the medium due to the pKa of the boronic acid involved in such affinity. This is due to the
effect of the pH on the boronic stereochemistry of the acid. Indeed, it is trigonal in form at low pH
and tetragonal at basic pH (i.e., pH > pKa = 8.8). Thus, the complexation reaction is favored at basic
pH because of the stability of boronat form. The shift to acid medium lead to the decomplexation
and thereafter the release of the cis-diol-containing molecules. It should be noted that the yield of
the complexation reaction between the boronic acid and the polyols compound depends on two
major factors: (1) the availability of the cis-diol functions and (2) the nature of the charged site close
to the boronic acid.
The boronic acid (or phenyl boronic acids) derivatives can be introduced onto particles surface
via various processes starting from simple chemical grafting onto preformed reactive magnetic
particles to direct incorporation into the used recipes for particle elaboration.
Tuncel et al. [46–48] have reported more recent work. The polymeric microspheres prepared by
this group were used for the immobilization of the RNA. The principle of such extraction is illus-
trated in Figures 9.31 and 9.32.
The specificity and efficiency have been examined by Tuncel et al. [46–48] as a function of
various parameters such as temperature, salinity, pH, and also the degree of particle functionaliza-
tion (Figure 9.33). The immobilization of RNA onto phenyl boronic acid containing polyNIPAM
microgel was found to be high compared to nonreactive polyNIPAM microgels. The desorption of
pre-adsorbed RNA was examined as a function of pH, temperature, and also the salinity and nature
of the salt used. As a general tendency, the desorption of RNA increases with increasing the pH and
the salinity of the incubation medium [49].
RNA
RNA
DNA DNA
RNA
RNA RNA RNA in the supernatant
RNA adsorbed particles redispersed RNA at 37°C
in fresh HEPES buffer at +4°C (pH:8.5)
RNA
Temperature increase
Thermoflocculation and desorption of RNA
RNA
RNA
+4°C 37°C
RNA
Temperature decrease
Shrunken poly (NIPA-co-VPBA)
particles in precipitate at 37°C
FIGURE 9.31 Illustration of specific extraction of RNA using phenyl boronic acid containing thermally
sensitive microgel particles. (From Elmas, B. et al., Colloid Surf. A: Physicochem. Eng. Aspects, 232, 253,
2004.)
The inhibition of direct amplification of adsorbed nucleic acids on the colloidal particles could
be due to the factors mentioned earlier, and also to (1) the high affinity between the enzymes and
the particles, and (2) the nature of the support (high hydrophobicity, non-coated iron oxide, or
denaturizing domains), as well as the high concentration of colloidal particles in the PCR medium.
The potential application of hydrophilic, cationic thermally sensitive particles is schematically
summarized in Figure 9.34.
OH OH
+OH–
N B N B
OH –OH– OH
H H OH
Trigonal form Tetrahedral form
C–
–O
NH2
+
N
OCH2
O
NH2
N N
OH OH
N N
O O
–O
C– P O P OCH2
O
NH2 O– OH
+
N
OCH2
O OH OH
NH2
N N
OH OH
N N
O O
P O P OCH2
–
O
O OH
O O
B OH
N H
FIGURE 9.32 Schematic illustration of specific immobilization of RNA onto phenyl boronic acid contain-
ing polyNIPAM microgels. (From Uguzdogan, E. et al., Macromol. Biosci., 2(5), 214, 2002.)
Nanogels are the nanosized particles produced via physically or chemically cross-linked polymer
networks, which are swell in a good solvent. The materials that have the base of Nanogels possess
elevated drug loading capacity, biocompatibility, and biodegradability, which are the main issues
to design a drug delivery system effectively. It should keep in mind that they can be administered
through oral, pulmonary, nasal, parenteral, intra-ocular, and topical routes [55].
Due to following reasons, nanogels have been known as more advanced drug delivery system
than others:
1. To avoid fast clearance by phagocytic cells, allowing both passive and active drug targeting
their size and surface properties can be manipulated.
2. Drug release control at the target site, therapeutic efficacy enhancement, and decreasing
side effects. Drug loading is relatively high and might be achieved exclusive of chemical
reactions, which is an important factor for maintaining the drug activity.
574 Handbook of Surface and Colloid Chemistry
16
14
10
0 5 10 15 20 25 30 35
Temperature (°C)
FIGURE 9.33 The influence of VPBA amount (in polyNIPAM microgels) on the fixation of RNA as a func-
tion of temperature.
DNA or RNA
Protein
Biological
sample
Anionic – – – + + + Cationic
+ latex
latex –– –– +
+++ pH < 5
T > LCST –
T < LCST
FIGURE 9.34 Schematic illustration of nucleic acids and proteins extraction, purification, and concentration
using thermally sensitive microgels.
Thermally Sensitive Particles 575
Stimuli
Drug Change in Pendant Drug
release Temperature pH acid/basic
temperature release
group
FIGURE 9.35 Various stimuli responsible for controlling drug release from smart polymeric drug delivery
systems. (From Honey Priya, J. et al., Acta Pharm. Sin. B, 2, 120, 2014.)
or
FIGURE 9.36 Drug release model from nanogel. (From Sultana, F. et al., Appl. Pharm. Sci., 3, S95, 2013.)
Ability of these nanogels to reach the smallest capillary vessels and to penetrate the tissues either through
the paracellular or the transcellular pathways [56] is mainly due to their tiny volume. In addition, these
particles are also biodegradable. A model of drug discharge from nanogel is shown in Figure 9.36.
In general, the size of nanogels are ranged between 20 and 200 nm in diameter and therefore are
useful in avoiding the quick renal exclusion but are small enough to prevent the uptake by the reticu-
loendothelial system. The have good penetration capabilities because of their extremely small size.
More particularly, it can cross the blood brain barrier (BBB). Furthermore, nanogels are capable to
solubilize hydrophobic molecules and diagnostic agents in their core or gel networks. Hydrophilic
and hydrophobic drugs and charged solutes might be given via nanogel. These properties of nanogel
are considerably influenced by temperature, existence of hydrophilic/hydrophobic groups in the
polymeric networks, the cross-linking density of the gels, surfactant concentration, and type of
cross-links present in the polymer networks.
Frequently nanogels are classified into two major ways. The primary classification is according
to their reactive manners that might be either stimuli-responsive or nonresponsive:
ξ
Mesh size
Mc
Covalent link
(chemical)
Junction
(physical)
Entanglement
(physical)
Hydrogel
FIGURE 9.37 A cross-linked hydrogel structure with the mesh size ξ and the average molecular weight
between the cross-linking points MC, respectively. (From Buenger, D. et al., Prog. Polym. Sci., 37, 1678,
2012.)
The mechanisms of drug release from nanogels are encompassed diffusion, nanogel degradation,
and displacement through ions present in the environment [58]. Nanogels, sub-micron hydrogel
particles with colloidal properties, are motivating candidates as drug delivery carrier because
the hydrogel nanostructure of nanogels can be designed to achieve controlled opening structures,
chemical topologies, and swelling responses to environmental incitements, whereas the colloidal
nanoparticle macrostructure of nanogels suggests the advantages of high specific surface areas and
ready injectability.
Besides photo cross-linking hydrogels, externally “smart” sensitive systems of drug delivery
have been studied as a new controlled delivery method to control the release of drugs in answer to
alters in the environment [59,60].
Nanogels, which formed exceptional nanovehicles in pharmaceutics are mainly used to encap-
sulate water soluble active molecules. Discharged nanogels in a swollen state enclose significant
amount of water. Loading of biological agents is often attained by mechanisms of self-assembly
including electrostatic, van der Waals, and/or hydrophobic interactions among the agent and the
polymer matrix. Different nanogels have been demonstrated to deliver their payload inside cells and
cross biological barriers. By reason of rise up stability inside cells, nanogels shown good potential
for improving of drug bioavailability in digestive tube and brain of low molecular drugs and bio-
macromolecules [61].
Nowadays, research is focused on preparing smart biocompatible and biodegradable poly-
mers possessing certain desirable characteristics like stimuli-sensitive polymers suffering phase
transitions in response to changes in pH, ionic strength, light, electric field, irradiation, or tem-
perature. Particularly, among the temperature-sensitive hydrogels reported to date, poly(N-
isopropylacrylamide) (PNIPAM) and its copolymers have been commonly used for cell separation
also for pharmaceutical and tissue engineering applications because of their distinctive thermal
properties [59,62].
Short half-lives, physical and chemical instability, and poor bioavailability are frequently the
reasons that can limit the pharmaceutical and biological therapeutics.
Usually it is useful to employ the polymers, which might react with the stimuli that exist intrinsi-
cally in the natural systems. In Table 9.1 diverse smart polymeric drug delivery systems have been
summarized [52].
Thermally Sensitive Particles 577
TABLE 9.1
Advantages and Disadvantages of the Thermally Sensitive Polymers
Advantages Disadvantages
Toxic organic solvents avoidance High burst release of drug
Delivery of hydrophilic and lipophilic drugs Low mechanical strength of the gel leading to potential dose-dumping
Reduced systemic side effects Lack of biocompatibility of the polymeric system and gradual lowering of
pH of the system due to acidic degradation
Site-specific delivery of the drugs
Sustained release properties
Sources: Schmaljohann, D., Adv. Drug Deliv. Rev., 58, 1655, 2006; Ruel-Gariépy, E. and Leroux, J.-C., Eur. J. Pharm.
Biopharm., 58, 409, 2004.
The requirement for the new approaches to control the delivery of the compounds such as pep-
tides, proteins, plasmid DNA, antisense oligodeoxynucleotides, and immunotoxins has been created
through the development of different advanced medications over the previous decade. Their abil-
ity to get the targeted regions is important for the activity of these molecules, nevertheless, once
they enter into the body system proteases or DNA-degrading enzymes in vivo simply degraded the
mentioned polymers [65].
Because of their ability to employ as controlled drug delivery systems, poly(N-isopropylacrylamide)
(PNIPAM) is the most popular temperature-sensitive polymer among microgels. Additionally, these
polymers shows a sharp lower critical solution temperature at 32°C, which might be shifted to body
temperature via formulating with surface active agents or additives. Being cytotoxic because of
quaternary ammonium existence in its structure, non-biodegradability and its ability of activation
platelets upon contact with body fluids have limited the utilization of the polyNIPAAM [52]. Newly,
another thermo-sensible polymer, poly(N-vinylcaprolactam) (PVCL), which demonstrates supe-
rior biocompatibility in comparison with PNIPAM, is being widely studied, additionally poly(N-
vinylcaprolactam) (PVCL) is particularly interesting because of the reality that is too stable against
hydrolysis [58,66]. Thermally responsive polymers are large molecules that dissolve in cold water
but collapse and precipitate upon heating of the aqueous solution over its lower critical solution tem-
perature (LCST) [58]. Prevention of organic toxic solvents, delivery ability of both hydrophilic and
lipophilic drugs, decreased systemic side effects, targeting delivery, and sustained release properties
are the main advantages of the thermosensitive polymeric systems. Despite these advantages, cer-
tain disadvantages such as high-burst release of drug, small mechanical strength of the gel leading
to potential dose-dumping, and lack of biocompatibility of the acidic degradation are accompanied
with these systems [63,64].
Hydrogels are achieved through cross-linking of these polymers that after heating above the
critical temperature reversibly shrink. In a dispersed medium, responsive nanoparticles could be
prepared through conducting the polymerization of the suitable monomer and a cross-linker. Latex
is the product, in this instance nanosized gel particles dispersed in water [58]. The several responsive
characteristics of nanoparticles, like those composed of poly(N-isopropylacrylamide) (PNIPAM)
have led to them becoming the source of significant interest in the area of drug delivery. They are
considered as “smart” polymers because of the conformational changes they suffer in response to
changes in environmental conditions such as temperature and pH [67].
Thermosensitive nanogels based on poly(N-isopropylacrylamide) (PNIPAM) offer the more
potential advantage of undergoing a volume phase transition as they are heated above a critical tem-
perature, resulting in significant gel de-swelling and (in some cases) thermally triggered aggregation
of the nanogel particles to form aggregates and/or physically cross-linked hydrogels. Since thermo-
sensitive nanogels can exhibit triggerable changes in pore size and colloidal stability, they have been
578 Handbook of Surface and Colloid Chemistry
widely investigated for the “on-demand” delivery of drugs or model drugs, including acetylsalicylic
acid, fluorescein-labeled dextran, insulin, and bovine serum albumin, among others [68].
Nanogels of N-isopropylacrylamide (NIPAM) are extremely responsive to temperature nearby
the physiological temperature. These nanogels not only mimic biological systems but they respond
to small environmental variations of body pathological state as well, because of their unique lower
critical solution temperature (LCST) in the vicinity of human body temperature. Both proper-
ties of liquids and solids describe through the superb swelling and de-swelling comportment of
N-isopropylacrylamide (NIPAM)-based materials.
In an entirely swollen state, these gels act as soft materials and behave like liquids as the poly-
mer network is solvated by a large amount of trapped solvent and the polymer chains exhibit great
mobility. Additionally, the nature of the cross-linked structure contributes to keeping their solid
form. The significant interest of N-isopropylacrylamide (NIPAM)-based gels in controlled drug
release uses has been created by their swelling and de-swelling comportment [61].
The inclusion of the thermoresponsive poly(N-isopropylacrylamide) (PNIPAM) constituent
suggests the advantages of: (1) degradation mechanism modulation of biodegradable polymers,
(2) cytotoxicity reduction of polycationic polymers, (3) thermally localizing medication to tar-
get sites after systemic administrations while their lower critical solution temperature (LCST) is
tailored to temperatures between 37°C (body temperature) and 42°C (used habitually in clinical
hyperthermia), and (4) drugs released at diverse profiles in answer to stimuli such as temperature,
pH, etc. [69].
In Figure 9.38, an example of a drug delivery process is illustrated, collapsed microgel particles
are positioned in the aqueous drug solution at temperature beneath the volume phase transition
temperature (VPTT), and the microgel particles will swell and/or, due to the polymer–drug interac-
tion, small drug molecules will freely penetrate the pores of the polymer network. Subsequently, the
drug-loaded microgel particles can be easily removed by centrifugation leaving a less concentrated
drug solution behind. The microgel particle will squeeze out the encapsulated drug if the solution is
heated up to a temperature above volume phase transition temperature (VPTT).
One research that has been done by Abu Samah and Heard in 2013 showed that in vitro perme-
ation data of caffeine-loaded poly(N-isopropylacrylamide) copolymerized with 5% (w/v) of acrylic
Swollen microgel
Drug-loaded
microgel
removed
Collapsed
microgel
FIGURE 9.38 Illustration of a temperature-sensitive microgel particle-based drug release process. (From
Imaz, A. and Forcada, J., N-Vinylcaprolactam-Based Microgels for Biomedical Applications, Vol. 48,
J. Polym. Sci., Part A: Polym. Chem., 2010, pp. 1173–1181.)
Thermally Sensitive Particles 579
acid (AAc) at 2°C–4°C enhanced the delivery of the loaded caffeine across epidermis in comparison
to the saturated solution of caffeine, through 3.5 orders of magnitude. In a study, it has been found
that “In solid tumors with hyperthermia therapy designed thermo-responsive and biodegradable
linear-dendritic nanoparticles have a big potential for thermally targeted and sustained release of
Ceramide C6” [69].
According to another in vitro and in vivo study, the ocular controlled release in a desired quantity
was gained by the thermosensible hydrogel. Moreover, the release was without adverse effects and
completely tolerated [70,71].
9.6 CONCLUSION
This chapter covers the preparation, characterization, and biomedical application of thermally
sensitive particles. Thermosensitive hydrogel is prepared by precipitation polymerization of
N-alkylacrylamide or N-alkylmethacrylamide as a principal water-soluble monomer, a water-
soluble cross-linker (for instance, N-methylenebisacrylamide), and an initiator (such as Azobis-
amidinopropane derivatives, potassium persulfate, or basically any charged initiator). The core–shell
latexes are produced by a combination of emulsion and precipitation polymerization, such as the
preparation of polystyrene core and poly(N-isopropylacrylamide) shell or encapsulation of colloidal
seed using alkylacrylamide derivatives. During the elaboration of such stimuli-responsive particles,
various aspects should be considered: (1) a water-soluble cross-linker is needed, (2) the polymeriza-
tion temperature should be higher than the LCST of the corresponding linear polymer, and (3) the
production of water-soluble polymer (which can be controlled by monitoring the polymerization
conditions). The polymerization mechanism has been clearly discussed and well illustrated, but the
nucleation step remains questionable and requires further work.
The colloidal characteristics of N-alkylacrylamide or N-alkylmethacrylamide-based p articles are
temperature related. In fact, the swelling ability, charge density and charge distribution, hydrophilic–
hydrophobic balance, hydration and dehydration property, particle size, surface polarity, colloidal
stability, water content, turbidity, electrokinetic and rheological properties are indiscernibly tem-
perature dependent. Such polymer particles can be used as a stimuli-responsive model for the inves-
tigation of colloidal properties and for theoretical studies.
As can be seen from this chapter, the adsorption and desorption of proteins and nucleic acids can
be monitored by controlling the key point governing the driving forces involved in the adsorption
process.
The adsorption of proteins onto charged thermally sensitive particles is greatly affected by the
incubation temperature. Protein adsorption onto highly hydrated thermosensitive particles below
the volume phase transition temperature is negligible. However, the affinity and the amount of
protein increase together with the temperature reveal the complex adsorption process. Desorption
is easily favored by cooling the temperature and controlling the adsorption time. In addition, the
adsorption and desorption processes are pH, time, and ionic strength dependent.
The adsorption process of nucleic acids onto such cationic hydrophilic thermally sensitive col-
loids is governed by the attractive electrostatic interaction as the driving forces. The adsorption of
DNA, RNA, and ssDNA are related to the surface charge density and accessible adsorption sites.
The desorption of pre-adsorbed nucleic acid molecules onto polymer particles is favored by reduc-
ing the attractive electrostatic interactions by altering the pH and the salinity of the medium. Such
hydrophilic particles can be used for specific adsorption and the concentration of nucleic acids from
any biological sample containing a complex mixture (proteins, lipids, membrane fragments, etc.).
In addition, colloidal particles bearing hydrophilic and cross-linked thermally sensitive shells can
be used directly in nucleic amplification processes (i.e., PCR) without any inhibition phenomena.
Thermally sensitive polymers under colloidal nano and microgels are also started to be used
in drug delivery and the most examined polymers are poly(N-alkylacrylamide) derivatives.
Unfortunately, such polymers are biocompatible but not biodegradable, and they are mainly used
580 Handbook of Surface and Colloid Chemistry
as model rather than for real in vivo applications. Thermally sensitive polymers have lower criti-
cal transition temperature close to the physiological value that suggests several possibilities in the
biomedical field. Biocompatibility, biodegradability, permeability, and physical characteristics of
hydrogels are the properties that made them good candidates as biomaterials for use in many medi-
cal applications, including drug delivery.
ACKNOWLEDGMENTS
I would like to gratefully acknowledge all the students (F. Meunier, L. Holt, V. Bourrel, F. Sauzedde,
D. Duracher, L. Housni, P. Hazot, G. Zhou, Y. Cuie, T. Taniguchi, T. Leon, and G. Levourch)
involved in the elaboration, the characterization, and the use of thermally sensitive colloidal par-
ticles for biomedical applications.
REFERENCES
1. Pelton RH, Chibante P. Preparation of aqueous latices with N-isopropylacrylamide. Colloid Surf.
1986;20:247–256.
2. Heskins M, Guillet JE. Solution properties of poly(N-isopropylacrylamide). J. Macromol. Sci. Chem.
1968;A2(8):1441–1455.
3. Schild HG. Poly(N-isopropylacrylamide): Experiment, theory and application. Prog. Polym. Sci.
1992;17:163–249.
4. Meunier F, Elaissari A, Pichot C. Preparation and characterization of cationic poly(n-isopropylacryl-
amide) copolymer latexes. Polym. Adv. Technol. 1995;6:489–496.
5. Duracher D, Elaissari A, Pichot C. Preparation of poly(N-isopropylmethacrylamide) latexes kinetic
studies and characterization. J. Polym. Sci. A: Polym. Chem. 1999;37:1823–1837.
6. Hazot P, Delair T, Elaissari A, Pichot C, Chapel JP, Davenas J. Synthesis and characterization of
functionalized poly(N-ethylmethacrylamide) thermosensitive latex particles. Macromol. Symp.
2000;150:291–296.
7. Loos W, Verbrugghe S, Goethals EJ, Du Prez FE, Bakeeva IV, Zubov VP. Thermo-Responsive
Organic=Inorganic Hybrid Hydrogels based on poly(N-vinylcaprolactam). Macromol. Chem. Phys.
2003;204:98–103.
8. Bronstein L, Kostylev M, Tsvetkova I, Tomaszewski J, Stein B, Makhaeva EE, Khokhlov AR, Okhapkin
AR Core–shell nanostructures from single poly(N-vinylcaprolactam), Macromolecules: Stabilization
and Visualization. Langmuir 2005;21:2652–2655.
9. Van Durme K, Verbrugghe S, Du Prez FE, Van Mele B. Light scattering and micro calorimetry studies on
aqueous solutions of thermo-responsive PVCL-g-PEO copolymers, Polymer 2003;44:6807–6814.
10. Van Durme K, Verbrugghe S, Du Prez FE, Van Mele B. Influence of poly(ethylene oxide)-grafts on
kinetics of LCST-behavior in aqueous poly(N-vinyl caprolactam) solutions and networks studied by
Modulated Temperature DSC. Macromolecules 2004;37:1054–1061.
11. Mc Phee W, Tam KC, Pelton R. Poly(N-isopropylacrylamide)latices prepared with sodium dodecyl
sulfate. J. Colloid Interf. Sci. 1993;156:24–30.
12. Kawaguchi H, Kawahara M, Yaguchi N, Hoshino F, Ohtsuka Y. Hydrogel microspheres I. Preparation of
monodisperse hydrogel microspheres of submicron or micron size. Polym. J. 1988;20:903–909.
13. López-León T, Ortega-Vinuesa JL, Bastos-Gonzales D, Elaissari A. Cationic and anionic poly(N-
isopropylacrylamide) based submicron gel particles: Electrokinetic properties and colloidal stability.
J. Phys. Chem. B 2006;110:4629–4636.
14. Vihola H, Laukkanen A, Valtola L, Tenhi H, Hirvonen J. Biomaterials 2005;26:3055–3064.
15. Wu X, Pelton RH, Hamielec AE, Woods DR, McPhee W. The kinetics of poly(N-isopropylacrylamide)
microgel latex formation. Colloid Polym. Sci. University of Lyon-1, France, 1994;272:467–477.
16. Dainton FS, Tordoff M. The polymerization of acrylamide in aqueous solution part 3. The hydrogen
peroxide photosensitized reaction at 258°C. Trans. Faraday Soc. 1957;53:499.
17. Currie DJ, Dainton FS, Watt WS. The effect of pH on the polymerization of acrylamide in water. Polymer
1965;6:451.
18. Dainton FS, Sisley WD. Polymerization of methacrylamide in aqueous solution part-1-hydrogen-
peroxide photosensitized reaction. Trans. Faraday Soc. 1963;59:1369.
Thermally Sensitive Particles 581
45. Charles MH, Charreyre MT, Delair T, Elaissari A, Pichot C. Oligonucleotide-polymer nanoparticle
conjugates: Diagnostic applications. STP Pharm. Sci. 2001;11(4):251–263.
46. Uguzdogan E, Denkbas EB, Tuncel A. RNA-sensitive N-isopropylacrylamide/vinylphenylboronoic acid
random copolymer. Macromol. Biosci. 2002;2(5):214–222.
47. Çamli ST, Senel S, Tuncel A. Nucleotide isolation by boronic acid functionalized hydrophilic supports.
Colloid Surf. A: Physicochem. Eng. Aspects 2002;207:127–137.
48. Elmas B, Onur MA, Senel S, Tuncel A. Thermosensitive N-isopropylcarylamide-vinylphenyl boronic
acid copolymer latex particles for nucleotide isolation. Colloid Surf. A: Physicochem. Eng. Aspects
2004;232:253–259.
49. Elmas B, Onur MA, Senel S, Tuncel A. Temperature controlled RNA isolation by N-isopropylacrylamide-
vinylphenyl boronic acid copolymer latex. Colloid and Polym. Sci. 2002;280(12):1137–1146.
50. Aguilar MR, Elvira C, Gallardo A, Vázquez B, Román JS. Smart polymers and their applications as
biomaterials. Top. Tissue Eng. 2007;3:2.
51. Medeiros SF, Santos AM, Fessi H, Elaissari A. Thermally-sensitive and magnetic poly(N-
vinylcaprolactam)-based nanogels by inverse miniemulsion polymerization. Colloid Sci. 2012;1: 2.
52. Honey Priya J, Rijo J, Anju A, Anoop KR. Smart polymers for the controlled delivery of drugs—A
concise overview. Acta Pharm. Sin. B 2014;2:120–127.
53. Yong Q, Kinam P. Environment-sensitive hydrogels for drug delivery. Adv. Drug Deliv. Rev.
2012;64:49–60.
54. Nguyen KT, West JL. Photopolymerizable hydrogels for tissue engineering applications. Biomaterials
2002;23:4307–4314.
55. Sultana F, Manirujjaman M, Imran-Ul-Haque MA, Sharmin S. An overview of nanogel drug delivery
system. Appl. Pharm. Sci. 2013;3:S95–S105.
56. Gonçalves C, Pereira P, Gama M. Self-assembled hydrogel nanoparticles for drug delivery applications.
Materials 2010;3:1420–1460.
57. Buenger D, Topuz F, Groll J. Hydrogels in sensing applications. Prog. Polym. Sci. 2012;37:1678–1719.
58. Viholaa H, Laukkanenb A, Hirvonena J, Tenhub H. Binding and release of drugs into and from thermo-
sensitive poly(N-vinyl caprolactam) nanoparticles. Eur. J. Pharm. Sci. 2002;16:69–74.
59. Kavanagha CA, Rochevb YA, Gallaghera WM, Dawsonc KA, Keenan AK. Local drug delivery in
restenosis injury: Thermoresponsive co-polymers as potential drug delivery systems. Pharmacol. Ther.
2004;102:1–15.
60. Kim MR, Jeong JH, Park TG. Swelling induced detachment of chondrocytes using RGD-modified
poly(N-isopropylacrylamide) hydrogel beads. Biotechnology 2002;18:495–500.
61. Kabanov AV, Vinogradov SV. Nanogels as Pharmaceutical Carriers, Vol. 4. Springer Link, Springer,
New York, 2008, pp. 67–80.
62. Alam MA, Rabbi MA, Miah MAJ, Rahman MM, Rahman MA, Ahmad H. A versatile approach on
the preparation of dye-labeled stimuli-responsive composite polymer particles by surface modification.
Colloid Sci. Biotechnol. 2012;1:225–234.
63. Schmaljohann D. Thermo- and pH-responsive polymers in drug delivery. Adv. Drug Deliv. Rev.
2006;58:1655–1670.
64. Ruel-Gariépy E, Leroux J-C. In situ-forming hydrogels—Review of temperature-sensitive systems.
Eur. J. Pharm. Biopharm. 2004;58:409–426.
65. Ramanan RMK, Chellamuthu P, Tang L, Nguyen KT. Development of a temperature-sensitive compos-
ite hydrogel for drug delivery applications. Biotechnol. Prog. 2006;22:118–125.
66. Imaz A, Forcada J. N-Vinylcaprolactam-Based Microgels for Biomedical Applications, Vol. 48. 2010,
pp. 1173–1181.
67. Abu Samah NH, Heard CM. Enhanced in vitro transdermal delivery of caffeine using a temperature and
pH-sensitive nanogel, poly(NIPAM-co-AAc). Int. J. Pharm. 2013;453:630–640.
68. Hoare T, Young S, Lawlor MW, Kohane DS. Thermoresponsive nanogels for prolonged duration local
anesthesia. Acta Biomater. 2012;8:3596–3605.
69. Stover TC, Kim YS, Lowe TL, Kester M. Thermoresponsive and biodegradable linear-dendritic nanopar-
ticles for targeted and sustained release of a pro-apoptotic drug. Biomaterials 2008;29:359–369.
70. Hamcerencu M, Popa M, Costin D, Bucatariu P, Desebrieres J, Riess G. New ophthalmic insert.
Oftalmologia 2009;53:83–90.
71. Ichikawa H, Fukumori Y. New applications of acrylic polymers for thermosensitive drug release.
STP Pharm. Sci. 1997;7:529–545.
10 Microemulsions and Their
Applications in Drug Delivery
Ziheng Wang and Rajinder Pal
CONTENTS
10.1 Introduction........................................................................................................................... 583
10.1.1 Continuous and Dispersed Phases............................................................................. 585
10.1.2 Surfactants................................................................................................................. 585
10.1.3 Co-Surfactant and Linker.......................................................................................... 587
10.1.4 Microemulsion and Nanoemulsion: Their Similarities and Differences................... 588
10.2 Characterization of Microemulsions..................................................................................... 589
10.2.1 Types of Microemulsions........................................................................................... 589
10.2.2 Hydrophilic Lipophilic Difference Method to Prepare Microemulsions.................. 590
10.2.3 Self-Microemulsifying Drug Delivery System.......................................................... 591
10.3 Formulation Developments.................................................................................................... 592
10.3.1 Interaction between the Two Phases.......................................................................... 592
10.3.2 Phase Scan................................................................................................................. 593
10.3.3 Pseudo-Ternary Phase Diagram................................................................................ 594
10.4 Challenges of Microemulsion-Based Drug Delivery System................................................ 597
10.4.1 Toxicity and Safety of SMEDDS............................................................................... 597
10.4.2 Scale-Up and Manufacturing..................................................................................... 597
References....................................................................................................................................... 598
10.1 INTRODUCTION
Microemulsions are being increasingly used as vehicles for lipophilic drugs. They can be used for
parenteral drug delivery systems (Date and Nagarsenker 2010) or could be encapsulated into softgel
capsules as a convenient solid dosage form (Gullapalli 2010). In contrast to conventional emulsions,
microemulsions are small droplet size (typically between 20 and 200 nm) (Talegaonkar et al. 2008)
and exhibit long-term stability (McClements 2012). They have been proven to promote the gastric-
intestinal (GI) absorption of lipophilic drugs and consequently enhance the bioavailability of some
active pharmaceutical ingredients (API) (Lawrence and Rees 2000). Food and Drug Administration
in the United States (U.S. FDA) has classified APIs into four groups based on the Biopharmaceutical
Classification System (BCS). APIs in class II (high permeability, low solubility) and class IV (low
permeability, low solubility) are ideal candidates for microemulsion-based drug delivery systems
due to their poor solubility in the aqueous phase. Table 10.1 gives examples of drugs that have been
successfully delivered in microemulsion form.
Microemulsions are isotropic, thermodynamically stable systems. They appear to be trans-
parent (or translucent) (Lawrence and Rees 2000). They can be easily prepared and identified in
different ways. For example, when 7 mL of a household liquid detergent, 14 mL of white spirit,
583
584 Handbook of Surface and Colloid Chemistry
TABLE 10.1
Examples of Drugs Delivered in Microemulsion Form
Drug Name Usage References
Paclitaxel Anticancer Gao et al. (2003)
Fenofibrate Antihyperlipidemic Liang et al. (2006)
Cholesterol ester transfer protein (CETP) inhibitors Gumkowski et al. (2005)
Atorvastatin Shen and Zhong (2006)
Fluvastatin Benameur et al. (2003)
Rapamycin Immunosuppressive drug Fricker et al. (2006)
Cyclosporine Ward and Cotter (1987)
Felodipine Antihypertensive drug Von Corswant (2003)
Nifedipine Rudnic et al. (1999)
Indomethacin Analgesic drug Farah and Denis (2000)
Ibuprofen Bauer et al. (2002)
Naproxen Mulye (2002)
Tipranavir Anti-HIV drug Chen and Gunn (2003)
Progesterone Hormone Gao and Morozowich (2003)
Testosterone Gao and Morozowich (2003)
Fish oil Nutrition supplement Mishra et al. (2001)
Acyclovir Antiviral drug Burnside et al. (1999)
Melatonin Immunomodulator Eugster et al. (1996)
and 4 mL of n-pentanol (amyl alcohol) or n-butanol are gently mixed, a two-phase system is pro-
duced at equilibrium (Makoto 1998) where the upper oil phase of the system exhibiting a strong
Tyndall effect is identified as the microemulsion phase (Makoto 1998). Microemulsion-based drugs
normally exhibit long shelf-life due to the high stability of microemulsions. Other advantages of
using microemulsion as a lipophilic drug carrier include the enhanced bioavailability (due to small
droplet size) and the ease of formation (due to low interfacial tension) (Lawrence and Rees 2000;
Talegaonkar et al. 2008). The small droplet size of microemulsion also facilitates the permeability
of the drug passing through the mucous membrane. Microemulsion-based drugs can be delivered
through different routes such as oral delivery as softgel capsules (Kovarik et al. 1994), topical or
transdermal delivery as lotions (Gupta et al. 2005), and parenteral delivery as intramuscular and
intravenous injections (Von Corswant et al. 1998).
Microemulsions are ternary systems containing oil, water, and surfactant. The terms “oil” and
“water” in a microemulsion system normally refer to “oil phase (oil and oil soluble components such
as cyclosporine)” and “aqueous phase (water and water soluble components such as sodium chlo-
ride),” respectively. The phase behavior of water–oil–surfactant mixtures was extensively studied
by Winsor (1948). Based on his experimental observations, Winsor classified equilibrium mixtures
of water–oil–surfactant into four systems: (1) type I (Winsor I) system where water continuous or
oil-in-water (O/W) type microemulsion coexists with the oil phase. In these systems, the aqueous
phase is surfactant-rich; (2) type II (Winsor II) system where oil continuous or water-in-oil (W/O)
type microemulsion coexists with the aqueous phase. In these systems, the oil phase is surfactant-
rich; (3) type III (Winsor III) system where bicontinuous type microemulsion (also referred to as
surfactant-rich middle-phase) coexists with excess oil at the top and excess water at the bottom; and
(4) type IV (Winsor IV) system where only a single-phase (microemulsion) exists. The surfactant
concentration in type IV microemulsion is generally greater than 30 wt%. Type IV microemulsion
could be water continuous, bicontinuous, or oil continuous depending on the chemical composition.
The phase behavior of microemulsions is often described as a fish diagram shown in Figure 10.1
(Komesvarakul et al. 2006).
Microemulsions and Their Applications in Drug Delivery 585
Type II (W/O)
Type III X
+ (bicontinuous) Type IV
3 3 1
Aqueous phase
Oil phase
Microemulsion
2
– Type I (O/W) phase
Excess phase
+
Surfactant concentration
FIGURE 10.1 Fish diagram: 1—single-phase region; 2—two-phase region (upper bar: microemulsion phase
at the top; lower bar: microemulsion phase at the bottom); and 3—three-phase region. X: tri-critical point. This
diagram assumes that the density of the aqueous phase is greater than that of the oil phase. (Adapted from
Komesvarakul, N. et al., J. Cosmet. Sci., 55, 309, 2006.)
10.1.2 Surfactants
Surfactants are amphiphilic molecules composed of both hydrophilic part (e.g., ethylene oxide
group) and hydrophobic part (e.g., alkyl chain). There are three types of surfactants: ionic (anionic
and cationic), nonionic, and zwitterionic. The hydrophilic part of surfactant bears either negative
charge (anionic) or positive charge (cationic) in the case of ionic surfactant, no charge in the case
of nonionic surfactant, and both negative and positive charges in the case of zwitterionic surfac-
tant. At a very low concentration of surfactant, the surfactant molecules exist as monomers in
the base liquid. The surfactant molecules also adsorb onto the interface between two immiscible
phases (oil–water or gas–liquid interfaces). With the increase in surfactant concentration, the
surfactant molecules begin to aggregate. As an example, consider the addition of surfactant to
water. The surfactant molecules initially disperse in water as monomers and also adsorb on to the
air–water interface till they reach surface saturation. Then monomers in water begin to aggregate
as clusters with their hydrophobic groups toward the interior of clusters and their hydrophilic
groups toward water. These clusters are called micelles and the concentration where micelles
586 Handbook of Surface and Colloid Chemistry
begin to form is called the critical micelle concentration (cmc) of the surfactant (Rosen 2004).
The cmc of a surfactant makes it unique from other amphiphilic molecules. For example, the
molecules of a co-surfactant are also amphiphilic but co-surfactants generally have no cmc. With
continuous addition of surfactant to water above cmc, changes in the microstructure of surfac-
tant/water system occur as shown in Figure 10.2a from spherical micelle solution to cylindrical
micelle solution to hexagonal liquid crystal (LC) to cubic liquid crystal and to lamellar liquid
crystal. Note that liquid crystal (LC) is a material in semisolid state that has both organized
structure (like solid) and disordered structure (like liquid) at a molecular level. The changes in
the microstructure summarized in Figure 10.2a are normally accompanied by changes in the
viscosity of the system. For example, the viscosity of LC is much higher than that of micellar
solutions; hexagonal LC has the highest viscosity among the three types of liquid crystals (Rosen
2004). It should be noted that the lamellar structure could be either flat or curved depending on
the surfactant structure (Kik et al. 2005).
When surfactant is added to oil, the surfactant molecules form micelles as usual at surfactant
concentrations above cmc but the micelles formed in oil are called reverse micelles as they have
a reverse arrangement of surfactant molecules as compared with the arrangement of surfactant
molecules in micelles formed in aqueous systems. In reverse micelles, the hydrophilic heads of the
Micelle Micelle
Monomer (spherical) (cylindrical)
LC-L
LS
LC-H LC-C
0 cmc 100
(a) Weight% of surfactant
0 cmc 100
(b) Weight% of surfactant
FIGURE 10.2 (a) Phase diagram of surfactant–water system. LC-H, liquid crystal-hexagonal region; LC-C,
liquid crystal-cubic region; LC-L, liquid crystal-lamellar region; LS, liquid surfactant region (water in surfac-
tant); and S, solid region. (b) Microstructural changes in surfactant–oil system with the increase in surfactant
concentration. (Adapted from Rosen, M.J., Surfactants and Interfacial Phenomena, 3rd edn., John Wiley &
Sons, Inc., New York, 2004, pp. 1–33, 110–113, 208–234.)
Microemulsions and Their Applications in Drug Delivery 587
surfactant molecules are inside the micelles and the hydrophobic tails of the surfactant molecules
extend away from the core of the micelles to the oil phase. Figure 10.2b shows the microstructural
changes that occur upon continuous addition of surfactant to oil (Rosen 2004).
The adsorption of surfactant at the interface alters the interfacial free energy. The interfacial free
energy is defined as the minimum work required to create the interface (Rosen 2004). Interfacial
tension (γ), which is the interfacial free energy normalized by area, is the minimum work required
to create the interface of unit area. Physically, interfacial tension is a measure of the strength of
interaction between two phases. The higher the interfacial tension, the weaker is the interaction
between the two phases. For example, pure trichloroethylene (TCE) and water are immiscible
phases with γ = 39.6 mJ/m2 at 25°C (Ma et al. 2008). To create interfacial contact area between
the two phases, that is, to formulate TCE–water emulsions, energy input exceeding 39.6 mJ/m2 is
required. With the addition of certain amounts of anionic surfactant sodium oleate (SO) and cat-
ionic surfactant benzethonium chloride (BC), γ can be lowered to as low as 3.7 ± 0.4 mJ/m2 (Wang
and Acosta 2013). This reduction in interfacial tension occurs because the surfactant partitions
between TCE and water phases, and as a result, an increase in interaction between TCE and water
phases is observed. Due to a decrease in the interfacial tension, less energy input is required to cre-
ate new interfacial contact area. It should be noted that γ should be very low (as low as 1 mJ/m2 or
lower) in order to formulate microemulsions (Upadhyaya et al. 2006). Due to a very low interfacial
tension, microemulsions are normally formed under gentle agitation with ultralow energy input
(e.g., stomach agitation in vivo is sufficient). To achieve ultralow values of interfacial tension, it is
often necessary to add a co-surfactant to the system. Thus, microemulsions are usually formulated
using a combination of surfactant and co-surfactant.
In the pharmaceutical field, nonionic surfactants are widely used as they are less irritative than
ionic surfactants (Mason et al. 2006). Before the design and formulation of any microemulsion-
based drug delivery system, it is important to consult the inactive ingredient guide (IIG) published
by FDA on their website to check the limits for different surfactants.
butanol (1:1 in weight ratio) is often selected as a co-surfactant in order to disrupt the order of the
LC structure while maintaining the original phase behavior (Salager et al. 2005).
With the increase in alkyl chain length of a co-surfactant, the mixture of surfactant and
co-surfactant becomes more lipophilic. However, the adsorption of co-surfactant at the interface
becomes less significant. This is because a long-chain alcohol has more affinity for the oil phase.
Therefore, instead of replacing surfactant at the oil–water interface by pushing the head groups of
surfactant molecules apart at oil–water interface, a long chain co-surfactant preferentially stays
with the hydrophobic part of the surfactant (e.g., tail group). This type of co-surfactant with long
alkyl chain length (>C10) is called a lipophilic linker (Salager et al. 2005). A linker is different from
co-surfactant in that the linker is either hydrophilic or lipophilic enough to co-adsorb at only one
side of the oil–water interface whereas co-surfactant interacts with both oil and water phases and
replaces the surfactant at the oil–water interface. Due to the co-adsorption effect, microemulsions
formulated using linkers may exhibit larger solubilization capacity than the ones without a linker
(Acosta et al. 2005). Typical linkers include hexyl glucoside (hydrophilic linker), sorbitan mono-
oleate (lipophilic linker), and long chain (>C10) alcohols (lipophilic linker) (Acosta et al. 2005).
where
ΔGI (J/mol) is the change in interfacial free energy
−TΔS (J/mol) is the entropic contribution to free energy of formation
The entropic contribution is due to the change in configuration from two stratified phases to an
emulsion with a large number of droplets. The change in the interfacial free energy ΔGI is given as
follows:
where
γ is the interfacial tension
ΔA is the increase in the interfacial contact area between the phases due to the formation of a
microemulsion or a nanoemulsion
The change in interfacial free energy ΔGI is always positive in the formation of a microemulsion or
a nanoemulsion as both γ and ΔA are positive; positive ΔA is due to an increase in the interfacial
area between oil and water when droplets are formed. The entropy contribution −TΔS is always
negative as both T and ΔS are positive; positive ΔS is due to an increase in the disorder of the
Microemulsions and Their Applications in Drug Delivery 589
system when droplets are formed. Therefore, ΔG formation can be either positive or negative depending
on the balance between ΔGI and −TΔS. In the case of microemulsions, ΔG formation is negative and
therefore microemulsions are formed spontaneously and are thermodynamically stable systems.
The ΔG formation is negative in microemulsions due to an ultralow value of interfacial tension between
oil and water. Unlike microemulsions, nanoemulsions are thermodynamically unstable systems as
ΔG formation is positive due to high interfacial tension between the oil and water phases. Although
nanoemulsions are thermodynamically unstable systems, they can be made kinetically stable due to
steric stabilization of the droplets.
Kinetically, the rate of separation of nanoemulsion into two separate phases can be described by
the Arrhenius equation (Missen et al. 1999):
k = Ae − Ea /RT (10.3)
where
k is the rate constant
A is the pre-exponential factor
Ea is the activation energy
R is the universal gas constant
T is the absolute temperature (in K)
At constant temperature, the rate of separation of nanoemulsion into separate phases depends on
the activation energy.
Laser pointer
Laser pointer
Excess phase
Laser pointer
Laser pointer
Light pathway
FIGURE 10.3 Tyndall effect in type I, II, III, and IV microemulsions. These observations assume that the
density of the oil phase is smaller than that of the water phase.
590 Handbook of Surface and Colloid Chemistry
microemulsions have a high conductivity (say greater than 10 μS/cm) due to water as the continuous
phase (Krauel et al. 2005).
The phase transition from type IV oil continuous microemulsion to type IV water continuous
microemulsion could be captured by observing changes in the viscosity (Watanabe et al. 2004). For
example, the viscosity of type IV oil continuous (W/O) microemulsion rises slowly initially with
the increase in water concentration. With further addition of water the viscosity begins to increase
sharply. The increase in viscosity is mainly due to the transition from type IV oil continuous (W/O)
microemulsion to type IV bicontinuous microemulsion. The viscosity reaches a maximum value at
some water concentration. Upon further increase in water concentration, the transition of type IV
bicontinuous microemulsion to type IV water continuous (O/W) microemulsion occurs resulting
in a sharp decrease in viscosity. The viscosity of type IV water continuous (O/W) microemulsion
continues to decrease with further increase in aqueous fraction (Watanabe et al. 2004).
Other methods such as dying (Ho et al. 1996) and cryo-field emission scanning electron micros-
copy (Cryo-FESEM) (Krauel et al. 2005) could also be used to identify the types of microemulsions.
The droplet size distribution of microemulsions can be determined using the dynamic light
scattering (DLS) technique. The average droplet size of about 80–120 nm is an acceptable level for
microemulsion-based drug delivery systems (Wang and Pal 2014).
Mh
HLB = 20 × (10.4)
M
where
Mh is the molecular weight of the hydrophilic part of the surfactant
M is the molecular weight of the whole surfactant molecule
According to Equation 10.4, the HLB scale ranges from 0 to 20. The surfactant with HLB value
between 12 and 16 is considered to be suitable for the formulation of O/W (oil in water) microemul-
sions. Surfactants with HLB values between 7 and 11 are more suitable for the preparation of W/O
(water in oil) microemulsions. The HLB concept is simple to use and the database is available for a
large number of surfactants. However, the main disadvantage of the HLB concept is that it does not
consider the impact of other factors on microemulsion formulation, such as temperature, aqueous
phase salinity, oil-chain length, and co-surfactant. For example, water continuous microemulsion
could be formulated using Cremophor EL (HLB = 13.5) at room temperature (20°C –25°C) (Wang
and Pal 2014). However, one may end up with bicontinuous type microemulsion using the same
formulation at 37°C in vivo due to the change in temperature. In summary, the impact of factors
such as co-surfactant, chemical nature of oil, temperature, salinity of aqueous phase, etc., on micro-
emulsion type are not captured in the HLB concept.
To overcome the limitations of the HLB method to formulate microemulsions, a different
approach called hydrophilic–lipophilic difference (HLD) was developed by Salager et al. (1983,
2000). The HLD approach captures the impact of various factors on microemulsion type. The HLD
value for nonionic and ionic surfactants can be calculated as follows (Acosta and Bhakta 2009):
Nonionic surfactants:
Ionic surfactants:
where
S is the salinity (g/100 mL) of aqueous phase
b is an empirical constant equal to 0.13 for NaCl and 0.1 for CaCl2 (Acosta and Bhakta 2009)
K is a constant that ranges from 0.1 to 0.2 (normally is around 0.17)
NCO is the alkane carbon number (ACN) of oil, which is a measure of hydrophobicity of the oil
phase (Ontiveros et al. 2013)
The more hydrophobic (e.g., long chain length) the oil phase, the higher is the NCO value. In the case
of non-alkane oils, an equivalent alkane carbon number (EACN) is used (Acosta and Bhakta 2009);
ϕ(A) or f(A) is a factor that takes into account the influence of co-surfactant. It is related to the par-
titioning of co-surfactant in the two phases. If there is no partitioning of co-surfactant observed in
the formulation (e.g., using sec-butanol as noted earlier) or no co-surfactant is present in the system,
ϕ(A) or f(A) is zero (Salager et al. 2005; Acosta and Bhakta 2009); cT and αT are temperature factors,
equal to 0.06 and 0.01 K−1, respectively; ΔT is temperature difference from 25°C; and CC is called
characteristic curvature of the surfactant. Like HLB, CC is also a measure of the hydrophilicity of
surfactant. However, the CC scale is different from that of HLB (Acosta et al. 2008). The smaller the
CC value, the more hydrophilic is the surfactant (e.g., the CC value of sodium oleate is −1.7 whereas
oleic acid has a CC value of 0) (Acosta et al. 2008).
The HLD criteria for the formulation of different types of microemulsions are as follows: for type
I microemulsion, HLD < 0; for type III microemulsion, HLD ≈ 0; and for type II microemulsion,
HLD > 0 (Salager et al. 2000; Mason et al. 2006; Ontiveros et al. 2013). According to Equations
10.5 and 10.6, oil with a long-chain structure (high alkane carbon number) is p referred for the
formulation of type I microemulsion, if other factors such as temperature, surfactant, co-surfactant,
and electrolyte concentration are kept the same. In the case of nonionic surfactant, the tempera-
ture effect is more pronounced than the electrolyte concentration effect. The hydrophilic parts of
nonionic surfactants are more sensitive to temperature changes than to electrolyte concentration
changes. The breakage of hydrogen bonds at high temperature makes nonionic surfactant more lipo-
philic (Nilsson and Lindman 1983). On the contrary, ionic surfactant is more sensitive to changes
in electrolyte concentration than to temperature changes. A high electrolyte concentration can
compress the electrical double layer of the hydrophilic part and make the ionic surfactant more
lipophilic (Srinivasan and Blankschtein 2003).
Although the HLD method has a limited database (CC value is known only for a limited number
of surfactants; likewise the NCO value is known for a limited number of oils) as compared with the
HLB method, this concept can play an important role in the formulation design.
Drugs are known to solubilize at the interface of microemulsion droplets or micelles. Research
work by Spernath et al. (2003) has demonstrated that reverse micelles have a higher drug solubi-
lization capacity than that of the individual components. They entrapped lycopene (model drug)
in the reverse micelles of R-(+)-limonene (model oil) and surfactant polysorbate 60 (Tween 60).
The drug capacity reached 1500 ppm as compared with 700 ppm in R-(+)-limonene alone (Spernath
et al. 2002). The reason for the increased capacity of drug solubilization is that the drug can now
distribute at the surface of the reverse micelles rather than occupy the core. However, drug solubili-
zation at the surface of reverse micelles is highly dependent on the physical properties of surfactant,
co-surfactant, and drug. The interactions between the drug and surfactant, and the hydrophobicity
of surfactant are also important. Different components can result in different solubilization capac-
ity of drugs (Narang et al. 2007). It should also be noted that drug solubilization may reduce after
oral administration of SMEDDS due to aqueous phase dilution and structural changes in vivo from
reverse micelles to O/W microemulsion.
γ I = ( γ aa − γ ab ) + ( γ bb − γ ab ) = γ aa + γ bb − 2 γ ab (10.7)
It should be noted that the interaction energy between the like molecules (i.e., γaa or γbb) is always
greater than the interaction energy between the unlike molecules (γab). When the phase consist-
ing of “a” molecules is a gas phase, γaa ≈ 0 and γab ≈ 0, and therefore γI is the surface tension of
phase “b” given as γbb. Likewise, γaa is the surface tension of phase “a” (Rosen 2004). According to
Equation 10.7, the interfacial tension γI increases with the decrease in γab as the surface tensions γaa
γaa
γab
Molecules “a”
Molecules “b”
γab
γbb
FIGURE 10.4 Interaction between two immiscible liquids. (Adapted from Rosen, M.J., Surfactants and
Interfacial Phenomena, 3rd edn., John Wiley & Sons, Inc., New York, 2004, pp. 1–33, 110–113, 208–234.)
Microemulsions and Their Applications in Drug Delivery 593
and γbb are fixed values. Therefore, interfacial tension (γI) is a measure of interaction between the
two phases. The higher the γI, the weaker is the interaction between the two phases.
To formulate a microemulsion, the interfacial tension (γI) should be ultralow (<1 mJ/m2)
(Upadhyaya et al. 2006) and the interaction between the two phases (γab) must be high. Therefore, a
certain amount of surfactant is needed to increase the interaction between the two phases to a level
where a microemulsion is formed spontaneously. This concentration is called critical microemul-
sion concentration (cμc), which is the minimum surfactant concentration required to formulate a
microemulsion (Aveyard et al. 1989).
10.3.2 Phase Scan
The purpose of a phase scan is to determine the temperature (in case of nonionic surfactants) or
salinity (in case of ionic surfactants) that can produce a microemulsion of desired type (O/W, W/O,
or bicontinuous) and properties (such as solubilization). Figure 10.5 shows a typical phase scan
(Rosen 2004). Phase scan is normally run from type I (Winsor I) system to type II (Winsor II)
system by increasing the temperature (in case of nonionic surfactant) or salinity (in case of ionic
surfactant). As shown in Figure 10.5, oil solubilization increases from sample 1 to 2 with the increase
in temperature or salinity. With the increase in temperature/salinity, the surfactant becomes more
lipophilic due to the increased dehydration. Consequently, the interaction between oil and water
phase (γab) increases and the interfacial tension (γOW) decreases. As surfactant becomes continu-
ously more lipophilic with the increase in temperature/salinity, a BM phase (bicontinuous or middle
phase) begins to separate from WM phase (water continuous O/W microemulsion phase). At the
start of the separation (sample 3), the interfacial tension between oil and BM phase (γOB) is still high
and the interfacial tension between BM and water phase (γBW) is zero (BM and water are m iscible).
3-phase region
FIGURE 10.5 Phase scan with temperature (for nonionic surfactant) or salinity (for ionic surfactant). OE ,
excess oil phase; OM, microemulsion (W/O) phase; WE , excess water phase; WM, microemulsion (O/W) phase;
BM, microemulsion (bicontinuous) phase; γOW, interfacial tension between oil and water phase; γBW, interfacial
tension between bicontinuous and water phase; and γOB, interfacial tension between oil and bicontinuous
phase. (Adapted from Rosen, M.J., Surfactants and Interfacial Phenomena, 3rd edn., John Wiley & Sons, Inc.,
New York, 2004, pp. 1–33, 110–113, 208–234.)
594 Handbook of Surface and Colloid Chemistry
The apparent interfacial tension between oil and water (γOW) in the three-phase region, with a mid-
dle phase present between oil and water phases, can be calculated as follows (Rosen 2004):
γ OW = γ OB + γ BW (10.8)
The apparent interfacial tension in the three-phase region is shown as a dashed line in Figure 10.5.
When the temperature/salinity is increased to a certain level (sample 4 in Figure 10.5), most of the
surfactants from WM migrate to the middle phase resulting in two excess phases (OE and WE) and
minimum apparent γOW. This Winsor III formulation shown as sample 4 in Figure 10.5 is the opti-
mal formulation. The temperature corresponding to the optimal formulation is the phase-inversion
temperature (PIT) and the corresponding salinity is the optimum salinity (Rosen 2004).
With further increase in temperature/salinity from sample 4 to 5, the surfactant starts migrating
from the middle phase (BM phase) to the oil phase (OE phase) and type II microemulsion (OM phase)
begins to form. As surfactant becomes more lipophilic, the middle phase (BM phase) disappears.
Water solubilization decreases and γOW increases with further increase in temperature/salinity from
sample 6 to 7. These changes are due to the transfer of surfactant to the oil phase and the reduction
in the interaction between oil and water phases.
As already pointed out, the purpose of the phase scan is to determine the optimal temperature/
salinity that can produce microemulsion of desired type and properties. Once the optimal tempera-
ture/salinity is established, one can further investigate the effects of composition on phase behavior
using the ternary phase diagram.
(a) (b)
FIGURE 10.6 Liquid crystal structure under crossed polarized microscope. (a) Oil streaks-lamellar LC and
(b) angular and striated textures-hexagonal LC. (Adapted from Cistola, D.P. et al., Biochemistry, 25(10), 2804,
1986.)
Surfactant (wt%)
0 100
10 90
20 80
30 70
A
40 60
50 50
Titration path 1
2
60 40
3
4
70 30
5
6
80 20
7
8
90 9 10
10
100 0
0 10 20 30 40 50 60 70 80 90 100
Aqueous (wt%) Oil (wt%)
FIGURE 10.7 Ternary diagram. Point A represents 30 wt% of aqueous phase, 10 wt% of oil phase, and 60
wt% of surfactant phase.
diagram is referred to as a pseudo-ternary diagram. The ternary diagram can be read following
the solid lines shown in the figure. For example, point A corresponds to a composition of 30%
water phase, 60% surfactant phase, and 10% oil phase. The phase region to which point A belongs
depends on the particle size and the appearance of the sample. In order to mark different phase
regions and boundaries on the ternary diagram, a titration technique is employed (Wang and Pal
2014). The titration begins by fixing two components and varying the third component. For e xample,
the dashed line shown in Figure 10.7 is followed with the addition of water. The titration procedure
begins with zero loading of aqueous phase and ends up at a point of 100% aqueous phase loading.
596 Handbook of Surface and Colloid Chemistry
TABLE 10.2
An Example of Titrating Water to Surfactant + Oil System
(Points Are Shown in Figure 10.7 as Black Dots)
# Surfactant (wt%) Oil (wt%) Water (wt%)
1 50 50 0
2 45 45 10
3 40 40 20
4 35 35 30
5 30 30 40
6 25 25 50
7 20 20 60
8 15 15 70
9 10 10 80
10 5 5 90
The titration procedure is repeated for different ratios of surfactant phase to the oil phase. As an
example, Table 10.2 gives the compositions of mixtures for one set of titration experiments repre-
sented by black dots shown in Figure 10.7.
Figure 10.8 shows the typical ternary phase diagram with phase regions and boundaries. The
phase diagram is developed using the titration method (Prajapati et al. 2012). Three different regions
(ME, LC, and EM) are identified on the phase diagram. Titration begins at different surfactant/
oil ratios with zero aqueous phase loading. At the start of the titration, type IV oil-continuous
microemulsion (W/O) is formed. With continuous addition of water, phase transitions from W/O
Surfactant (wt%)
0 100
10 90
20 80
30 70
40 60
50 50
60 ME
LC (W/O) 40
70 30
ME
80 (O/W) 20
90 10
EM
100 0
0 10 20 30 40 50 60 70 80 90 100
Aqueous (wt%) Oil (wt%)
FIGURE 10.8 Ternary phase diagram of water (aqueous phase), PEG-35 castor oil (Cremophor EL, surfac-
tant phase), and mixture of glycerol monocaprylocaprate and caprylic/capric triglycerides (1:3) (oil phase) at
37°C. (Adapted from Prajapati, H.N. et al., Pharm. Res., 29(1), 285, 2012.)
Microemulsions and Their Applications in Drug Delivery 597
microemulsion to LC and from LC to O/W microemulsion are observed provided that the surfactant/
oil ratio is high at the start of the titration (e.g., 80 wt% of surfactant/20 wt% of oil). The continuous
addition of water changes the spontaneous curvature of surfactant in oil phase and induces phase
inversion from W/O to O/W microemulsion (Fernandez et al. 2004). Note that at a certain water
wt%, water and oil phases begin to merge and intertwine together to form LC. This point is called
the emulsion inversion point (EIP) and the interfacial tension is minimum at this point (Forgiarini
et al. 2001). The formulation ends up with type IV water-continuous microemulsion (O/W) with fur-
ther addition of water. It should be noted that at low to medium surfactant concentrations (<40 wt%
of surfactant at the start of the titration), the formulation may end up in the coarse emulsion (EM)
region. Thus, as noted earlier, a certain amount of surfactant (critical microemulsion concentration)
is always required to formulate a microemulsion.
REFERENCES
Acosta, E.J., A.S. Bhakta. 2009. The HLD-NAC model for mixtures of ionic and non-ionic surfactants.
J Surfactants Deterg 12(1): 7–19.
Acosta, E.J., T. Nguyen, A. Witthayapanyanon, J.H. Harwell, D.A. Sabatini. 2005. Linker-based bio-compatible
microemulsions. Environ Sci Technol 39(5): 1275–1282.
Acosta, E.J., J.S. Yuan, A.S. Bhakta. 2008. The characteristic curvature of ionic surfactants. J Surfactants
Deterg 11: 145–158.
Aveyard, R., B.P. Binks, P.D.I. Fletcher. 1989. Interfacial tensions and aggregate structure in pentaethylene
glycol monododecyl ether/oil/water microemulsion systems. Langmuir 5: 1210–1217.
Bauer, K., C. Neuber, A. Schmid, K.M. Volker. 2002. Oil in water microemulsion, US Patent 6426078-B1.
Bavière, M., R.S. Schechter, W.H. Wade. 1981. The influence of alcohol on microemulsion composition.
J Colloid Interface Sci. 81: 266–279.
Benameur, H., V. Jannin, D. Roulot. 2003. Method and formulation for decreasing statin metabolism, US Patent
6652865-B2.
Burnside, B.A., C.E. Mattes, C.M. McGuinness, E.M. Rudnic, G.W. Belendiuk. 1999. Oral acyclovir delivery,
US Patent 5883103-A.
Burskirk, G.A., V.P. Shah, D. Adair. 1994. Workshop III report: Scale-up of liquid and semisolid disperse
systems. Eur J Pharm Biopharm 40: 251–254.
Chen, S., J.A. Gunn. 2003. Oral dosage self-emulsifying formulations of pyranone protease inhibitors, US
Patent 6555558-B2.
Cistola, D.P., D. Atkinson, J.A. Hamilton, D.M. Small. 1986. Phase behavior and bilayer properties of fatty
acids: Hydrated 1:1 acid-soaps. Biochemistry 25(10): 2804–2812.
Constantinides, P.P. 1995. Lipid microemulsions for improving drug dissolution and oral absorption: Physical
and biopharmaceutical aspects. Pharm Res 12: 1561–1572.
Date, A.A., M.S. Nagarsenker. 2010. Design and evaluation of microemulsions for improved parenteral delivery
of propofol. AAPS PharmSciTech 9(1): 138–145.
Eugster, C., C.H. Eugster, W. Haldemann, G. Rivara. 1996. Spontaneously dispersible concentrates and aque-
ous microemulsions with steryl retinates having anti-tumor activity, US Patent 5496813-A.
Farah, N., J. Denis. 2000. Orally administrable composition capable of providing enhanced bioavailability
when ingested, US Patent 6054136-A.
Fernandez, P., V. André, J. Rieger, A. Kühnle. 2004. Nano-emulsion formation by emulsion phase inversion.
Colloids Surf A 251(1): 53–58.
Forgiarini, A., J. Esquena, C. Gonzalez, C. Solans. 2001. Formation of nano-emulsions by low-energy emulsi-
fication methods at constant temperature. Langmuir 17: 2076–2083.
Fricker, G., B. Haeberlin, A. Meinzer, J. Vonderscher. 2006. Galenical formulations, US Patent 7025975-B2.
Gao, P., W. Morozowich. 2003. Self-emulsifying formulation for lipophilic compounds, US Patent 6531139-B1.
Gao, P., B.D. Rush, W.P. Pfund et al. 2003. Development of a supersaturable SEDDS (S-SEDDS) formulation
of paclitaxel with improved oral bioavailability. J Pharm Sci 92(12): 2386–2398.
Gullapalli, R.P. 2010. Review-soft gelatin capsules (softgels). J Pharm Sci 99: 4107–4148.
Gumkowski, M.J., L. Franco, S.B. Murdande, M.E. Perlman. 2005. Self-emulsifying formulations of choles-
teryl ester transfer protein inhibitors, US Patent 6962931-B2.
Gupta, R.R., S.K. Jain, M. Varshney. 2005. AOT water-in-oil microemulsions as a penetration enhancer in
transdermal drug delivery of 5-fluorouracil. Colloids Surf B 41: 25–32.
Ho, H.O., C.C. Hsiao, M.T. Sheu. 1996. Preparation of microemulsions using polyglycerol fatty acid esters as
surfactant for the delivery of protein drugs. J Pharm Sci 85(2): 138–143.
Jones, S.C., K.D. Dreher. 1976. Co-surfactant in micellar systems used for tertiary oil recovery. Soc Petrol Eng
J 16: 161–167.
Kang, B.K., J.S. Lee, S.K. Chon et al. 2004. Development of self-microemulsifying drug delivery systems
(SMEDDS) for oral bioavailability enhancement of simvastatin in beagle dogs. Int J Pharm 274:
65–73.
Kik, R.A., J.M. Kleijn, F.A.M. Leermakers. 2005. Bending moduli and spontaneous curvature of the monolayer
in a surfactant bilayer. J Phys Chem B 109: 14251–14256.
Komesvarakul, N., M.D. Sanders, E. Szekeres et al. 2006. Microemulsions of triglyceride-based oils: The effect
of co-oil and salinity on phase diagrams. J Cosmet Sci 55: 309–325.
Kovarik, J.M., E.A. Mueller, J.B. Van Bree, W. Tetzloff, K. Kutz. 1994. Reduced inter and intra intraindi-
vidual variability in cyclosporine pharmacokinetics from a microemulsion formulation. J Pharm Sci 83:
444–446.
Microemulsions and Their Applications in Drug Delivery 599
Krauel, K., N.M. Davies, S. Hook, T. Rades. 2005. Using different structure types of microemulsions for the
preparation of poly(alkylcyanoacrylate) nanoparticles by interfacial polymerization. J Control Release
106(1–2): 76–87.
Lawrence, M.J., G.D. Rees. 2000. Microemulsion-based media as novel drug delivery systems. Adv Drug Deliv
Rev 45: 89–121.
Li, P., A. Ghosh, R.F. Wagner, S. Krill, Y.M. Joshi, A.T.M. Serajuddin. 2005. Effect of combined use of nonionic
surfactant on formation of oil-in-water microemulsions. Int J Pharm 288: 27–34.
Liang, L., A.H. Shojaei, S.A. Ibrahim, B.A. Burnside. 2006. Self-emulsifying formulations of fenofibrate
and/or fenofibrate derivatives with improved oral bioavailability and/or reduced food effect, US Patent
7022337-B2.
Ma, H., M. Luo, L.L. Dai. 2008. Influences of surfactant and nanoparticle assembly on effective interfacial
tensions. Phys Chem Chem Phys 10: 2207–2213.
Makoto, T. (ed.). 1998. Disperse Systems. Wiley-VCH, Weinheim, Germany, pp. 244–245, ISBN: 3-527-29458-9.
Mason, T.G., J.N. Wilking, K. Meleson, C.B. Chang, S.M. Graves. 2006. Nanoemulsions: Formation, structure,
and physical properties. J Phys Condens Matter 18: 635–666.
McClements, D.J. 2011. Edible nanoemulsions: Fabrication, properties, and functional performance.
Soft Matter 7(6): 2297–2316.
McClements, D.J. 2012. Nanoemulsions versus microemulsions: Terminology, differences, and similarities.
Soft Matter 8: 1719–1729.
Mishra, A., I. Moussa, Z. Ramtoola, N. Clarke. 2001. Pharmaceutical compositions containing an omega-3
fatty acid oil, US Patent 6284268-B1.
Missen, R.W., C.A. Mims, B.A. Saville. 1999. Introduction to Chemical Reaction Engineering and Kinetics.
John Wiley & Sons, Inc., New York, p. 44.
Mulye, N. 2002. Self-emulsifying compositions for drugs poorly soluble in water, US Patent 6436430-B1.
Narang, A.S., D. Delmarre, D. Gao. 2007. Stable drug encapsulation in micelles and microemulsions. Int J
Pharm 345: 9–25.
Nilsson, P.G., B. Lindman. 1983. Water self-diffusion in nonionic surfactant solutions. Hydration and obstruc-
tion effects. J Phys Chem 87(23): 4756–4761.
Ontiveros, J.F., C. Pierlot, M. Catté et al. 2013. Classification of ester oils according to their equivalent alkane
carbon number (EACN) and asymmetry of fish diagrams of C10E4/ester oil/water systems. J Colloid
Interface Sci 403: 67–76.
Pasquali, R.C., M.P. Taurozzi, C. Bregni. 2008. Some considerations about the hydrophilic-lipophilic balance
system. Int J Pharm 356(1): 44–51.
Prajapati, H.N., D.M. Dalrymple, A.T. Serajuddin. 2012. A comparative evaluation of mono-, di-and triglyc-
eride of medium chain fatty acids by lipid/surfactant/water phase diagram, solubility determination
and dispersion testing for application in pharmaceutical dosage form development. Pharm Res 29(1):
285–305.
Rosen, M.J. 2004. Surfactants and Interfacial Phenomena, 3rd edn. John Wiley & Sons, Inc., New York,
pp. 1–33, 110–113, 208–234.
Rudnic, E., J. McCarty, B. Burnside, C. McGuinness, G. Belenduik. 1999. Emulsified drug delivery systems,
US Patent 5952004-A.
Salager, J.L., R.E. Antón, D.A. Sabatini, J.H. Harwell, E.J. Acosta, L.I. Tolosa. 2005. Enhancing solubilization
in microemulsions-state of the art and current trends. J Surfactants Deterg 8(1): 3–21.
Salager, J.L., N. Marquez, A. Graciaa, J. Lachaise. 2000. Partitioning of ethoxylated octylphenol surfactants in
microemulsion-oil-water systems: Influence of temperature and relation between partitioning coefficient
and physicochemical formulation. Langmuir 16(13): 5534–5539.
Salager, J.L., M. Minana-Perez, M. Perez-Sanchez, C.I. Rojas. 1983. Surfactant-oil-water systems near the
affinity inversion part III: The two kinds of emulsion inversion. J Disp Sci Technol 4(3): 313–329.
Shen, H., M. Zhong. 2006. Preparation and evaluation of self-microemulsifying drug delivery systems
(SMEDDS) containing atorvastatin. J Pharm Pharmacol 58(9): 1183–1191.
Spernath, A., A. Yaghmur, A. Aserin, R.E. Hoffman, N. Garti. 2002. Food-grade microemulsions based on
nonionic emulsifiers: Media to enhance lycopene solubilization. J Agric Food Chem 50: 6917–6922.
Spernath, A., A. Yaghmur, A. Aserin, R.E. Hoffman, N. Garti. 2003. Self-diffusion nuclear magnetic resonance,
microstructure transitions, and solubilization capacity of phytosterols and cholesterol in Winsor IV food-
grade microemulsions. J Agric Food Chem 51: 2359–2364.
Srinivasan, V., D. Blankschtein. 2003. Effect of counterion binding on micellar solution behavior:
2. Prediction of micellar solution properties of ionic surfactant-electrolyte systems. Langmuir 19(23):
9946–9961.
600 Handbook of Surface and Colloid Chemistry
Swenson, E.S., W.J. Curatolo. 1992. Means to enhance penetration: (2) Intestinal permeability enhancement
for proteins, peptides and other polar drugs: Mechanisms and potential toxicity. Adv Drug Deliv Rev
8: 39–92.
Swenson, E.S., W.B. Milisen, W. Curatolo. 1994. Intestinal permeability enhancement: Efficacy, acute local
toxicity and reversibility. Pharm Res 11: 1132–1142.
Talegaonkar, S., A. Azeem, F.J. Ahmad et al. 2008. Microemulsions: A novel approach to enhanced drug deliv-
ery. Recent Pat Drug Deliv Formul 2: 238–257.
Upadhyaya, A., E.J. Acosta, J.F. Scamehorn, D.A. Sabatini. 2006. Microemulsion phase behavior of anionic-
cationic surfactant mixtures: Effect of tail branching. J Surfactants Deterg 9(2): 169–179.
Von Corswant, C. 2003. Microemulsions for use as vehicles for administration of active compounds, US Patent
6602511-B2.
Von Corswant, C., P. Thoren, S. Engstrom. 1998. Triglyceride based microemulsion for intravenous administra-
tion of sparingly soluble substances. J Pharm Sci 87: 200–208.
Wang, Z., E. Acosta. 2013. Formulation design for target delivery of iron nanoparticles to TCE zones. J Contam
Hydrol 155: 9–19.
Wang, Z., R. Pal. 2014. Enlargement of nanoemulsion region in pseudo-ternary mixing diagrams for a drug
delivery system. J Surfactants Deterg 17: 49–58.
Ward, M.V., R. Cotter. 1987. Rapid acting intravenous emulsions of omega-3 fatty acid esters, US Patent
4678808-A.
Watanabe, K., A. Noda, M. Masuda, T. Kimura, K. Komatsu, K. Nakamura. 2004. Bicontinuous microemulsion
type cleansing containing silicone oil. II. Characterization of the solution and its application to cleansing
agent. J Oleo Sci 53(11): 547–555.
Winsor, P.A. 1948. Hydrotropy, solubilisation and related emulsification processes. Trans Faraday Soc
44: 376–398.
11 A Quantum Mechanics–Based
Adhesion and Wetting
Approach
Costas G. Panayiotou
CONTENTS
11.1 Introduction........................................................................................................................... 601
11.2 Molecular Surface Tension Components...............................................................................604
11.2.1 Derivation of the PSP Equation for the Work of Adhesion....................................... 612
11.3 Applications........................................................................................................................... 615
11.4 Discussion and Conclusions................................................................................................... 619
Acknowledgments........................................................................................................................... 620
References....................................................................................................................................... 620
11.1 INTRODUCTION
Adhesion and wetting of solid surfaces are major fields in interface science with a remarkable
range of technological applications. In spite their importance, their precise determination still
remains a challenge due, primarily, to the challenges associated with the practical implementation
of an operational definition of the surface tension, γS, of solids. The latter is often determined from
extrapolations from the melt state and, most commonly, indirectly from the contact angle of liquid
drops deposited on the surface and the application of a model/equation for the interfacial tension,
γSL , with the liquid. In general, the work of adhesion of two unit surfaces or the negative of the sur-
face free energy change upon the formation of a unit interface ij from the component unit surfaces
i and j is obtained from the Dupre equation [1,2]:
The work of adhesion of a solid with a liquid, forming a contact angle θ with the solid surface, is
given by the classical Young equation [3]:
As seen, a combination of Equations 11.1 and 11.2 does not give the interfacial tension unless we
know the surface tension of the solid surface, and vice versa. The use of additional liquids will
provide for more equations, but one unknown will always be left undetermined. Thus, it is essential
to have an additional model/equation for the interfacial tension or the work of adhesion.
601
602 Handbook of Surface and Colloid Chemistry
One of the most, if not the most, widely used equations for the interfacial tension is the van
Oss–Chaudhury–Good (vOCG) equation [4–6]:
γ SL = ( γ SLW − γ LW
L ) + 2( γ SA − γ LA )( )
γ SB − γ LB (11.3)
or
WSLadh = 2 ⎡⎢ γ SLW γ LW
L + γ SA γ LB + γ SB γ LA ⎥⎤ (11.4)
⎣ ⎦
Superscript LW in these equations stands for Lifshitz–van der Waals and indicates that in the
vOCG model this term encompasses the contributions from the weak London dispersion inter-
actions, as well as the Debye dipole–induced dipole and the Keesom dipole–dipole orientation
interactions. Superscripts A and B stand for acid and base, respectively, and indicate that Lewis
acid–base interactions are also considered in the vOCG model but via the asymmetric product of
the second term on the rhs of Equation 11.3 and not by a quadratic always positive term like the
first term. This model gives the working Equations 11.3 and 11.4 but does not give equations for the
separate LW and A or B surface tension components of the compounds in contact. The latter are, in
practice, determined by a simultaneous multiparameter fit to experimental contact angle data for
various liquid–solid pairs or to experimental interfacial tension data for liquid–liquid interfaces.
Through this multifitting process it is attempted to obtain unique values for the surface tension
components valid for all kinds of interfaces, including pure liquid–vapor or pure solid–vacuum
(or air) interfaces. This results, very often, to very high values of base surface tension components,
which are difficult to reconcile with our current understanding of intermolecular interactions.
Numerous corrective attempts have been made in the literature [7–9] for reducing the values of
these base components but the problem still remains. An often common element to these corrective
attempts is the drastic increase of the acid surface tension component of water and the concomitant
reduction of the base component, which were equal in the original vOCG model [4–6]. In addi-
tion to the preceding problems, there is experimental evidence (and, thus, criticism of the vOCG
model for not accounting) for polarization of the solid surface by the contacting liquid, that is, for
the change of the surface tension components of solids in response to the interaction with (some)
contacting liquids [10,11].
This chapter presents a quantum mechanics–based alternative to the vOCG approach whose pre-
dictive capacity offers a new way of addressing the previously mentioned controversial interfacial
issues. The new approach is built upon the recently introduced molecular descriptors called partial
solvation parameters (PSP) [12–16]. The rationale and the working equations of partial solvation
parameters are described in this series of recent papers [12–16] where the reader is referred to for
the details. A short review is presented in Chapter 3.
Briefly, the PSP approach heavily resides on the quantum mechanics–based COSMO-RS the-
ory of solutions [17–22]. The COSMO model belongs to the class of continuum solvation models
(CSM) of quantum mechanics. For the solvation picture, it considers the molecule embedded in a
conductor of infinite permittivity that screens perfectly the molecular charges on the surface of its
molecular cavity. This molecular cavity is characterized by a volume, Vcosm, and a molecular surface
area, Acosm. The crucial information is contained in the so-called COSMO file of each compound
obtained from quantum chemical calculations at various levels of theory. COSMO files give the
detailed surface charge distribution or the σ-profile of each molecule. The σ-profile may be ana-
lyzed into its moments of various orders, known as COSMOments, out of which a large number of
properties may be calculated, among them the molecular descriptors of Abraham’s QSPR/LSER
model [23,24].
Adhesion and Wetting 603
There are two PSP schemes, the s and the σ-scheme. The two schemes have identical hydrogen
bonding parameters but they differ in the way they partition the non-hydrogen-bonding interactions
of the molecule. In the s-scheme this partitioning leads to the dispersion, sd, and polar, sp PSP, which
are equivalent to the more familiar Hansen’s dispersion solubility parameter, δd, and polar solubility
parameter, δp, respectively [25]. In the σ-scheme the partitioning leads to the van der Waals, σW, and
polarity/polarizability, σpz PSP. The van der Waals PSP is, simply, the weak van der Waals energy
density:
EvdW
σW = (11.5)
Vmol
where
EvdW is the weak van der Waals molar energy
Vmol is the molar volume of the compound
Although Equation 11.6 entails a slight recalculation of partial solubility parameters, it is useful
in orienting the reader on the physical content of each quantity and in obtaining σpz from a simple
subtraction. The thus calculated σpz PSP is close to the dielectric PSP obtained from the equation:
Ediel
σdiel = (11.7)
Vmol
where Ediel is the dielectric molar energy of the compound. Both, EvdW and Ediel, are directly available
from COSMO-RS theory [19], which is the basic source of information for getting PSPs. The name,
van der Waals, has been retained for the corresponding, σW, PSP although it clearly accounts for
the atom-specific weak dispersive or London interactions. Similarly, the name, dispersion PSP, has
been retained for sd in order to specify the equivalence with the more familiar partial solubility
parameter, δd. This is not a very accurate terminology, however, since sd (and δd) accounts for the
London as well as the Debye induction forces. On the other hand, sp (and δp) account for perma-
nent dipole–dipole (Keesom orientation) interactions and is equal to zero for compounds with zero
dipole moment. Thus, σpz accounts for, both, Keesom and Debye interactions.
In the PSP approach the compounds are divided into two major classes, the homosolvated and
the heterosolvated. Heterosolvated are the compounds that can hydrogen bond only with another
(heteron, in Greek) different compound, that is, they cannot self-associate. All other compounds are
homosolvated. The hydrogen bonding PSPs, σGa for the proton donor (electron acceptor or acidic)
and σGb for the proton acceptor (electron donor or basic) capacity of the compound, are obtained
from the corresponding Abraham’s LSER descriptors, A and B, which are in turn obtained from the
COSMOments of the molecule [19]. For water, A = 0.676 and B = 0.663 or A = 1.0198B. The PSP
approach considers perfect neutrality for water and this requires a slight shift of the B scale by ca.
2%. By setting m-SUM = A + 1.0198B, the following defining equations for our hydrogen bonding
PSPs are obtained:
2 2 2
σGhb = σGa + σGb (11.8)
604 Handbook of Surface and Colloid Chemistry
2 2
σGa σGb A
= 1 − = (11.9)
2
σGhb m-SUM
2
σGhb
m-SUM
σGhb = 5.5 (11.10)
Vcosm
It is clear from the above description that PSPs may be obtained in a straightforward manner as long
as the COSMO file of the molecule is available. COSMO files are already available for thousands
of compounds in free or commercial databases. They may also be obtained directly from quan-
tum chemical calculation suites such as the TURBOMOLE [26] or the DMol3 incorporated in the
MaterialStudio suite of Accelrys® and other suites. The PSPs for a number of common solvents are
reported in Table 11.1. The PSPs for high polymers cannot be obtained at present from quantum
chemical calculations due to their enormous number of conformers and to cpu time limitations.
They may be obtained, however, from inverse-gas chromatography (IGC) and other experimental
methods [15,16]. It should be made clear that PSPs heavily depend on COSMO-RS model. Any
changes in COSMO-RS parameterization or any miscalculations of LSER descriptors in the multi-
linear regressions with COSMOments will have an impact on PSPs. However, in Reference 14, the
PSP approach has been developed into a full activity-coefficient model for concentrated as well as
infinitely dilute systems. In this development, the free energy change upon formation of a hydrogen
bond between acidic molecule i and basic molecule j is given by the universal equation
in direct analogy with the QSPR/LSER model [23]. Thus, PSPs may be checked now against experi-
mental data on activity coefficients and derived thermodynamic quantities.
With this introduction and the tables with PSPs we may now proceed to the presentation of the
new approach to interfacial energies, wetting, and adhesion. This is done in the next section while
examples of calculations are shown in Section 11.3. A preliminary version of the new approach is
presented in the Appendix of Reference 15.
TABLE 11.1
Partial Solvation Parameters (PSP) of Pure Compounds
Part A: The PSPs for Non-hydrogen-Bonded Compounds
Compound σW (MPa1/2) σpz (MPa1/2) Compound σW (MPa1/2) σpz (MPa1/2)
n-Pentane 13.65 5.00 Cyclohexane 13.85 8.35
n-Hexane 13.75 6.50 Cycloheptane 13.85 8.55
n-Heptane 13.75 7.00 1-Pentene 13.80 4.35
n-Octane 13.75 7.00 1-Hexene 13.85 5.75
n-Nonane 13.75 7.50 Benzene 14.84 11.43
n-Decane 13.75 7.50 Toluene 14.59 11.09
n-Undecane 13.75 7.50 Ethylbenzene 14.41 10.77
n-Dodecane 13.75 8.00 n-Propylbenzene 14.30 10.36
n-Hexadecane 13.75 8.00 n-Butylbenzene 14.24 10.17
n-Octadecane 13.75 8.00 n-Pentylbenzene 14.17 10.22
n-Eicosane 13.75 8.50 n-Hexylbenzene 14.13 10.20
n-Tetracosane 13.75 8.50 m-Xylene 14.40 10.89
n-Octacosane 13.75 8.50 o-Xylene 14.40 11.45
n-Dotriacontane 13.75 8.50 p-Xylene 14.38 10.67
Isobutane 13.17 2.90 Styrene 14.70 12.06
Isopentane 13.37 3.65 Naphthalene 14.43 13.04
2-Methylpentane 13.48 5.10 Carbon tetrachloride 16.97 6.06
3-Methylpentane 13.48 5.80 Methyl chloride 16.92 10.14
2.3-Dimethylpentane 13.37 6.40 Monochlorobenzene 15.51 11.57
2.3.4-Trimethylpentane 13.22 6.95 Benzyl chloride 15.45 12.89
Cyclopentane 14.10 8.05 Carbon disulfide 18.42 8.77
Part B: The PSPs for Hydrogen-Bonded Compounds
Compound σW σpz σGb σGa
Homosolvated
Methanol 15.18 10.62 18.34 15.01
Ethanol 14.65 10.92 15.12 11.28
1-Propanol 14.45 10.96 13.17 9.87
1-Butanol 14.30 11.33 11.91 8.74
1-Pentanol 14.23 10.46 11.19 8.59
1-Hexanol 14.16 9.79 10.51 7.96
1-Heptanol 14.10 10.42 9.95 7.41
1-Octanol 14.07 10.23 9.49 6.96
1-Nonanol 14.03 10.96 9.10 6.58
1-Decanol 13.98 10.82 8.83 6.27
Isopropanol 14.23 10.03 13.39 9.57
2-Butanol 14.23 9.24 11.39 8.43
2-Methyl-1-propanol 14.11 7.76 11.21 9.24
2-Methyl-2-propanol 13.87 8.81 11.82 7.75
2-Pentanol 14.07 10.29 10.61 7.73
3-Pentanol 14.07 10.30 10.27 6.83
Cyclohexanol 14.10 14.04 10.91 7.42
Phenol 14.95 13.93 8.35 11.89
Benzyl alcohol 14.84 15.44 10.63 8.07
m-Cresol 14.73 13.63 8.07 10.73
o-Cresol 14.76 12.21 7.71 10.62
(Continued)
606 Handbook of Surface and Colloid Chemistry
stable and preferred ones in the liquid state, not only for entropic reasons but, mainly, because
they allow for a more efficient and extensive intermolecular hydrogen bonding network. Across the
interface, the molecules will assume various conformations and orientations between those shown
in Figure 11.1 that will ultimately lead to surface free energy minimization.
Molecules like ethanol or acetone and the higher member of their homologous series may pref-
erably orient themselves and keep away from the liquid–vapor interface their polar sites. In this
way, these molecules (and the majority of polar molecules) will exhibit a polar character on the
FIGURE 11.1 Conformers of ethylene glycol and glycerol with and without intramolecular hydrogen bonds
(shown by dashed lines).
608 Handbook of Surface and Colloid Chemistry
(a) (b)
FIGURE 11.2 The COSMO surfaces of (a) water and (b) diiodomethane (DIM). The charge densities vary
from highly negative (electron donor or proton acceptor) sites typically marked in deep red over oxygen
atom to highly positive (electron acceptor or proton donor) sites marked in deep blue over hydrogen atoms.
Intermediate densities are marked with intermediate colors of the visible color spectrum.
interface lower than its corresponding character in the bulk liquid. Such a possibility, however, is
not very much available for molecules like water whose COSMO surface is shown in Figure 11.2
with its surface charge density marked in colors. No matter how oriented at the interface, the water
molecule will always expose its polar/hydrogen-bonding sites away from the bulk liquid phase. As
a consequence, the water molecule is expected to exhibit nearly its full hydrogen bonding capacity
at the interface. On the other hand, apolar or less polar molecules, like diiodomethane shown also
in Figure 11.2, are relatively more indifferent for specific orientations at interfaces other than those
dictated by stereochemical constraints or space filling requirements. Since these two molecules are
extensively used in the interface/wetting/adhesion literature, it is worth commenting briefly on their
σ-profiles shown in Figure 11.3.
According to COSMO-RS convention, only charge densities in excess of 0.01 e/A2 in absolute
values are capable of participating in hydrogen bonds. As observed in Figure 11.3, the σ-profile of
water molecule is nearly symmetric and extended beyond the hydrogen bonding cutoffs on both
sides of the profile. This is in contrast to arguments often appearing in the literature for a much more
acidic than basic character of water. The nearly symmetric σ-profile of water leads to nearly equal
A and B LSER molecular descriptors, as we have seen in the previous section and corroborates our
16
15 diiodomethane_c0.cosmo
14 h2o_c0.cosmo
13
12
11
10
9
p(σ)
8
7
6
5
4
3
2
1
0
–0.03 –0.02 –0.01 0.00 0.01 0.02 0.03
2
σ (e/Å )
FIGURE 11.3 The sigma profiles of water and diiodomethane. The charge densities refer to the screening
charges, that is, to the counter-charges of the real molecular charges.
Adhesion and Wetting 609
selection of acidity/basicity scale based on the water neutrality. On the other hand, the σ-profile
of the much used diiodomethane is somewhat asymmetric with a minor protrusion into the acidic
domain, although its main polarity sites are populating the opposite domain of the profile. In other
words, diiodomethane is not an apolar molecule and in no way we may consider it interacting with
dispersion forces only.
From the preceding exposition, it becomes clear that preferential orientations and molecular
conformations are of central importance at interfaces and ultimately dictate interfacial free ener-
gies. This free energy character of interfacial tension makes difficult the mapping of PSPs onto
the corresponding surface tension components, γW, γpz, γa, and γb. We may proceed, however, by
confining ourselves, first, to the weakest van der Waals interactions, namely, the London dis-
persive interactions and the mapping of σW into the dispersion or van der Waals surface tension
component, γW.
From its very definition, σW is an energy density (energy per unit volume), while the correspond-
ing surface tension component, γW, will be energy per unit area. In a real pure liquid, we could
allocate to each molecule an average solvation area and an average solvation volume. The COSMO
area, Acosm, and COSMO volume, Vcosm, could be considered good measures for this solvation area
and the solvation volume, respectively. We could, then, relate the preceding components through
the basic defining equations:
or
Vcosm 2
γW = σW (11.13)
Acosm
Equation 11.13 is an approximate equation, but it may satisfactorily serve our purpose. Of course,
the other components of surface tension cannot be obtained by such a simple equation and, in order
to proceed, we should turn to their free energy character. The appropriate equations in this case
should resemble Equation 11.11 giving the hydrogen bonding free energy of interacting acidic and
basic molecules. Quite generally, this free energy character may simply be expressed as follows:
γ = γ W + γ pz + γ hb = Γ h − TΓ s (11.14)
The components with superscript h will be referred to as enthalpic or energetic components and
those with superscript s as entropic ones. γpz and γhb are the surface tension components arising from
polarity/refractivity and hydrogen bonding interactions, respectively. Regarding hydrogen bonding
quantities of interacting molecules i and j, their relation to PSPs could then be obtained from defin-
ing equations of the form (cf. Equations 11.11 and 11.13):
Vi V j
Γijh = hσa,i σb, j = 2 Γ ha,i Γ hb, j (11.15)
Ai A j
and
Vi V j
γ hb,ij = fσGa,i σGb, j + g = 2 γ a,i γ b, j (11.16)
Ai A j
610 Handbook of Surface and Colloid Chemistry
The ratios under the square roots are the solvation volume-to-surface area ratios for each type of
interaction. It is not obvious how these ratios could be obtained but we will not need them here.
Subscripts hb, a, and b in these equations are referring to the overall hydrogen bonding, to acidic,
and to basic components, respectively. For simplicity, in this work, we will consider only one type
of acidic and/or basic group per molecule. For pure hydrogen-bonded compounds, then, the defining
Equation 11.16 becomes
γ hb = 2 γ a γ b (11.17)
For a molecule with similar acidic and basic character, there is no reason to expect that it will lose
this similarity at the interface. For the purposes of this work then and on account of Equation 11.16,
we will adopt the following equations:
If γhb were known, Equations 11.17 and 11.18 would form a system of two equations with two
unknowns that could be solved for γa and γb. Even then, however, we would need additional infor-
mation in order to calculate γpz. From the preceding discussion, we may expect that, as the mol-
ecule at the pure liquid–vapor interface is adopting orientations that will minimize its surface
energy, its polar or hydrogen bonding groups are more likely to be oriented toward the bulk liquid
phase rather than the interface. As a consequence, this tendency will affect, both, hydrogen bond-
ing and polarity/polarizability components of its surface tension, at least to the extent that high
surface charge densities may be counted either as hydrogen bonding sites when interacting with
complementary h ydrogen bonding sites or as polarity/polarizability sites when interacting with non-
hydrogen b onding counterparts [17–21]. In view of this COSMO picture of real solvents and in order
to p roceed, we will also adopt the following equations:
γ pz σ2pz γ pz σ2pz
= , = (11.19)
2 γ a uσGa σGhb 2 γ b uσGbσGhb
or
γ pz σ2pz
= (11.20)
2 γ a γ b uσGhb σGa σGb
The factor u should be, in general, a characteristic property of the molecule and should reflect the
peculiarities of the molecular structure and the surface charge distribution that are not accounted
for by the PSP terms in the rhs of the equations. This assumption facilitates calculations very much.
If the total surface tension, γ, is known, then, Equation 11.14 along with Equations 11.17 through
11.19 form a system of four equations with four unknowns, namely, γpz, γhb, γa, and γb. The solution
of this system is facilitated by noting from Equations 11.17 and 11.20 that
⎛ σ σ σ ⎞
γ pz + γ hb = γ pz ⎜ 1 + u Ghb 2 Ga Gb ⎟⎟ = γ − γ W or
⎜ σ pz
⎝ ⎠
γ − γW
γ pz = (11.22)
(1 + u ( σ Ghb σGa σGb /σ2pz ))
Then, from Equations 11.19 and 11.22 we may obtain γa, γb, and γhb. It is worth mentioning that the
preceding equations hold for, both, homosolvated and heterosolvated compounds. For the latter,
Equation 11.21 gives γhb = 0 while Equation 11.22 gives γpz = γ − γW. The important point is that the
method (Equation 11.19) provides also with γa or γb components of heterosolvated compounds, for
which γhb = 0.
The experimental surface tensions of many common liquids are available in the open literature
or in critical compilations such as the DIPPR database [27]. From these total surface tensions,
one may then obtain the partial surface tensions by using the preceding set of simple equations.
Equation 11.13 should be used first in order to obtain γW. The value of factor u, however, has not
been specified yet.
It is important to point out that for a compound with equal acidic and basic PSPs (σGa = σGb),
the preceding method implies that its corresponding acidic and basic surface tension components
will also be equal. This is the case for water in the PSP approach. The COSMO volume and area
for water are equal to 0.0256 and 0.4306 nm2, respectively [19]. Substituting in Equation 11.13,
we obtain γW = 15.71 mN/m. The DIPPR value for the total surface tension of water at 20°C is
72.8 mN/m. By replacing in Equations 11.19 through 11.22, we obtain the other three surface ten-
sion components as functions of factor u. The results are reported in Table 11.2 while in Figure 11.4
is shown the variation of the hydrogen bonding component, γα or γb, with factor u. In the same figure
is also shown that the van Oss [4–6,28] value of 25.5 mN/m for these components corresponds to
a value near 2 for the factor u. In order, then, to further simplify our method and have a nearly
common reference point with the vOCG approach [4–6,28], we set u = 2 and consider it a universal
constant applicable to all compounds.
In Table 11.3 are reported the COSMO volumes and areas and the surface tension components
for a number of common solvents as calculated by the preceding method. In the last column are
reported the total surface tensions as obtained from DIPPR [27]. It should be stressed that the values
of the surface tension components in Table 11.3 are predicted ones and not fitted to experimental
data other than the overall surface tension of the compound.
TABLE 11.2
The Polar and Hydrogen-Bonding Surface Tension Components of Water for
Various Values of the Factor u (Equations 11.19 through 11.22)
u γpz γa, γb γhb/γtotal
0.75 12.47 22.31 0.61
1 9.90 23.60 0.65
2 5.42 25.84 0.71
3 3.73 26.68 0.73
4 2.84 27.12 0.75
612 Handbook of Surface and Colloid Chemistry
28
26 van Oss
γa (mN/m)
24
22
0 2 4
u
FIGURE 11.4 The HB component of surface tension, γa or γb, of water as a function of the factor u in the
defining Equations 11.19 through 11.22. As shown, the adopted value of 25.5 mN/m by the vOCG approach
[4–6,28] corresponds to a value of u near 2.
The major part of the sum of γW and γpz, presumably, corresponds to the widely used Lifshitz–van
der Waals surface tension [4–6,28], or
γ W + γ pz ≅ γ LW (11.23)
The calculations with Equation 11.23 and Table 11.3 are in reasonably good agreement with the
reported values for the Lifshitz–van der Waals surface tension component in the literature [4–6,28].
As an example, the γLW calculated from Equation 11.23 for methanol, ethanol, glycerol, and water
are 17.23, 18.03, 31.97, and 21.13 mN/m, respectively, while the corresponding reported values
[28] are 18.5, 20.1, 34.0, and 21.8 mN/m, respectively. Having the partial surface tensions of pure
solvents, we may now proceed to the next step and propose a method for the surface-tension char-
acterization of polymers and solid surfaces.
The preceding procedure may be applied to polymers and smooth solid surfaces although there
are some noticeable difficulties. The problems in this case start with the estimation of γW. As already
mentioned, it is not possible at present to run quantum chemical calculations for high polymers. The
required ratio of COSMO volume-to-surface area of a polymer could be obtained, as an example,
from its monomer or its oligomer analogs through the following equation:
⎛ VS ⎞ ⎛V ⎞ ⎛r⎞ ⎛q⎞
⎜A ⎟ = ⎜ cosm ⎟ ⎜ ⎟ ⎜ ⎟ (11.24)
⎝ S ⎠polymer ⎝ Acosm ⎠monomer ⎝ q ⎠polymer ⎝ r ⎠monomer
q, r being the UNIQUAC/UNIFAC-type parameters [29,30]. The higher the oligomer analog the
better is expected to be the approximation with Equation 11.24. When the molecular structures of
the solid surface are more complex (branched, cross-linked, dendritic, etc.), Equation 11.24 will be a
rather poor approximation. The simplest solution in these cases is the use of one or two experimen-
tal data of contact angles with non-hydrogen bonding solvents in order to obtain γW and substitute
in Equations 11.19 through 11.22 in order to obtain all other surface tension components. However,
this requires first an equation for the contact angle or for the work of adhesion or for the interfacial
tension. This is done in the next section.
TABLE 11.3
The Partial Surface Tensions of Common Pure Compounds
Vcosm Acosm γW γpz γb γa γ (mN/m)
Compound (nm3) [19] (nm2) [19] (mN/m) (mN/m) (mN/m) (mN/m) [27]
n-Hexane 0.1457 1.569 17.56 0.35 0 0 17.91
n-Heptane 0.1676 1.7689 17.91 1.90 0 0 19.83
n-Octane 0.1895 1.9684 18.21 2.90 0 0 21.13
n-Nonane 0.2113 2.1679 18.46 3.95 0 0 22.40
n-Decane 0.2333 2.3678 18.68 4.70 0 0 23.39
n-Dodecane 0.2768 2.7664 18.97 5.95 0 0 24.93
n-Hexadecane 0.3664 3.5686 19.49 7.65 0 0 27.15
n-Octadecane 0.4089 3.9657 19.49 8.55 0 0 28.01
n-Eicosane 0.4527 4.3648 19.46 9.10 0 0 28.56
n-Docosane 0.4965 4.7636 19.45 9.50 0 0 28.97
n-Tetracosane 0.5403 5.1626 19.42 9.80 0 0 29.29
n-Octacosane 0.6274 5.9555 19.36 10.35 0 0 29.71
Cyclohexane 0.1263 1.3149 18.20 6.45 0 0 24.64
Cycloheptane 0.1498 1.4863 19.25 7.75 0 0 26.99
Benzene 0.1100 1.2137 19.95 8.25 0 0 28.21
Toluene 0.1318 1.4055 19.95 8.00 0 0 27.92
Ethylbenzene 0.1539 1.5860 20.14 8.40 0 0 28.59
n-Propylbenzene 0.1762 1.7865 20.18 8.40 0 0 28.50
n-Butylbenzene 0.1980 1.9860 20.20 8.45 0 0 28.64
Acetaldehyde 0.0643 0.8288 15.93 4.83 6.62 0 20.76
1-Propanal 0.0853 1.0150 16.73 5.23 8.01 0 21.96
1-Butanal 0.1074 1.2153 17.32 7.63 10.02 0 24.95
1-Pentanal 0.1294 1.4194 17.69 7.64 6.50 0 25.33
Acetone 0.0863 1.0268 16.72 6.32 6.57 0 23.04
Methyl ethyl ketone 0.1070 1.2065 17.34 6.62 6.62 0 23.96
Methanol 0.0484 0.6756 16.51 0.72 2.76 2.26 22.22
Ethanol 0.0700 0.8811 17.04 0.99 2.36 1.76 22.10
1-Propanol 0.0918 1.0805 17.04 0.99 2.36 1.76 23.39
1-Butanol 0.1139 1.2814 17.75 1.37 2.47 1.85 24.37
1-Pentanol 0.1355 1.4820 18.18 1.85 2.53 1.86 25.30
1-Hexanol 0.1578 1.6831 18.51 1.92 2.78 2.13 25.90
1-Heptanol 0.1794 1.8818 18.79 2.02 2.92 2.21 26.71
1-Octanol 0.2014 2.0812 18.96 2.62 2.97 2.22 27.10
1-Decanol 0.2449 2.4766 19.15 2.81 3.00 2.20 28.38
2-Pentanol 0.1354 1.4548 18.42 1.55 2.04 1.49 23.45
Cyclohexanol 0.1391 1.4226 19.43 6.35 4.63 3.15 33.42
Ethylene glycol 0.0805 0.9784 18.92 10.82 10.23 8.14 47.99
Glycerol 0.1118 1.2151 19.51 22.46 12.88 8.83 63.30
Tetraethylene glycol 0.2468 2.5551 19.94 8.05 14.77 5.87 46.61
Ethyl acetate 0.1173 1.3316 16.9 6.34 7.33 0.00 23.24
n-Propyl acetate 0.1393 1.5329 17.31 6.55 6.52 0.00 23.86
n-Butyl acetate 0.1611 1.7323 17.65 7.1 5.55 0.00 24.75
Carbon tetrachloride 0.1284 1.3421 26.00 1.00 0 0 27.00
Diiodomethane 0.1164 1.2537 32.71 17.21 0 0 49.92
n-Propylamine 0.0968 1.1245 16.16 1.29 3.67 1.05 21.38
n-Butylamine 0.1186 1.3228 17.2 1.93 4.17 1.21 23.62
Formamide 0.0575 0.7675 17.56 12.74 14.79 13.34 58.40
Dimethyl sulfoxide 0.0988 1.1167 21.14 8.90 40.34 1.03 42.95
Water 0.0256 0.4306 15.71 5.42 25.84 25.84 72.80
614 Handbook of Surface and Colloid Chemistry
polarity/polarizability interactions are contributing with quadratic terms of PSP differences while
the hydrogen bonding interactions with logarithmic terms of equilibrium constants, Kij, for the
hydrogen bonding between a proton donor (acidic) molecule i and a proton acceptor (basic) molecule
j. Assuming analogous contributions at interfaces, Equation 69 of Reference 14 for the residual free
energy at infinite dilution per unit volume or the compatibility of components i and j, rewritten here
for convenience,
ΔGijr,∞ ⎪⎧ RT ⎛ a ⎞ ⎪⎫
= ⎨(σWi − σWj )2 + (σ pzi − σ pzj )2 − ln ⎜ 1 + K ij ⎟ ⎬ (69, Reference 14)
Vmol, i ⎪⎩ Vmol, i ⎝ rj ⎠ ⎪⎭
could be translated into the following equation for the interfacial tension (or compatibility of sur-
faces i and j):
⎧⎪ 2 2 RT ⎛ a ⎞ ⎫⎪
γ ij = ⎨ ( γ W ,i − γ W , j ) +( γ pz, i − γ pz, j ) − ln ⎜ 1 + K ij ⎟ ⎬ (11.25)
Vmol, i ⎝ rj
⎩⎪ ⎠ ⎪⎭
Assuming one interacting site per segment (or, a/rj = 1) and equilibrium constants significantly larger
than 1, Equation 11.25 reduces to the following simple equation for cross-associating compounds:
⎧ 2 2 RT ⎫
γ ij = ⎨
⎩
( γW,i − γW, j ) +( γ pz, i − γ pz, j ) −
Vmol, i
ln( K ij ) ⎬ (11.25a)
⎭
The corresponding equation for molecules able to self-associate and cross-associate becomes:
⎧⎪ 2 2 RT ⎛K ⎞ RT ⎛ K ⎞ ⎫⎪
γ ij = ⎨ ( γ W, i − γ W, j ) +( γ pz, i − γ pz, j ) − ln ⎜ ij
Vmol, i ⎝ K ii ⎟−V ln ⎜ ji ⎟ ⎬ (11.26)
⎠ mol, j ⎝ K jj ⎠ ⎭⎪
⎩⎪
The equilibrium constants Kij are related to hydrogen bonding free energies through the classical
equation:
− RT ln K ij = GijH (11.27)
Substituting from Equation 11.27, Equations 11.25a and 11.26 further simplify to the following
forms:
⎧⎪ 2 2 GijH ⎫⎪
γ ij = ⎨
⎪⎩
( γ W, i − γ W, j ) +( γ pz, i − γ pz, j ) + ⎬ (11.25b)
Vmol, i ⎪⎭
and
The last terms, however, in these equations are energy densities and are more appropriate to PSPs
rather than to surface tension components that are energies per unit area. Thus, they also need to
Adhesion and Wetting 615
be translated into γ terms. This translation is facilitated by the inspection of Equations 11.11 and
11.16 or 11.17, which are the defining equations for the bulk and interface free energy changes upon
hydrogen bond formation. Substituting the analogous terms, Equations 11.25b and 11.26a become
γ ij = {( γW,i − γW, j
2
) +( γ pz, i − γ pz, j )
2
}
+ 2 γ a, i γ b, j (11.25c)
and
γ ij = {( γ W, i − γ W, j
2
) +( γ pz, i − γ pz, j )
2
+2 ( ) (
γ a, i γ b, j − γ a, i γ b, i + 2 γ a, j γ b, i − γ a, j γ b, j )}
(11.26b)
or
γ ij = {( γ W, i − γ W, j
2
) +( γ pz, i − γ pz, j )
2
+2 ( γ a, i − γ a, j )( γ b, i − γ b, j )} (11.26c)
Equation 11.25c is a special case of Equation 11.26c and holds for molecules that cross-associate but
do not self-associate (heterosolvated; we avoid the terms monopolar and bipolar used with vOCG
model because polar interactions are distinct from hydrogen bonding interactions here). Our discus-
sion will focus then on the general Equation 11.26c. As observed, the last term in this equation is
identical to the hydrogen bonding term of the vOCG equation [4–6,28]. The preceding derivation
was useful in delineating the assumptions behind this term. Expanding terms in Equation 11.26c
and replacing for pure components, we obtain
γ ij = γ i + γ j − 2 { γ W, i γ W, j + γ pz, i γ pz, j + γ a, i γ b, j + γ a, j γ b, i } (11.27)
which, combined with Dupre equation (11.1), gives for the work of adhesion:
Wijadh = 2 { }
γ W, i γ W, j + γ pz, i γ pz, j + γ a, i γ b, j + γ a, j γ b, i (11.28)
From Young’s Equation 11.2, then, we obtain for the contact angle of a drop of liquid, L, on a solid
surface, S, the following working equation:
γ L (1 + cos θ) = 2 { γ W, L γ W, S + γ pz, L γ pz, S + γ a, L γ b, S + γ a,S γ b, L (11.29) }
This is the desired equation, which may be now applied for the characterization of polymer and
solid surfaces.
11.3 APPLICATIONS
In this section, we apply Equation 11.29 for the characterization of polymer surfaces on the basis
of their PSPs. PSPs for some common polymers [15,16] are reported in Table 11.4. In order to use
Equation 11.29 for the prediction of surface tension components of polymers, as we did for liquids
in the previous section, we must have first γW, which in turn requires Acosmo and Vcosmo values. Since
616 Handbook of Surface and Colloid Chemistry
TABLE 11.4
Partial Solvation Parameters of Common Polymers
Polymer σW σpz σGb σGa
Polyethylene, HDPE 13.75 9.00 0.0 0.0
Polyethylene, LDPE 13.75 7.75 0.0 0.0
Polypropylene, PP 14.05 5.05 0.0 0.0
Poly(ethyl ethylene), PEE 14.60 5.00 0.0 0.0
Polystyrene, atactic, PS 15.20 7.00 0.0 0.1
Poly(methyl acrylate), PMA 18.00 15.80 9.85 ± 0.40 0.0
Poly(ethyl acrylate), PEA 18.30 11.50 9.0 ± 0.15 0.0
Poly(methyl methacrylate), PMMA 18.50 7.80 7.55 ± 0.70 0.0
Poly(butyl methacrylate), PBMA 17.80 7.50 7.6 ± 0.25 0.0
Poly(hydroxy-ethylacrylate), PHEA 20.80 19.55 4.0 ± 0.20 7.5
Poly(vinyl chloride), PVC 18.50 13.95 0.0 4.2 ± 1.2
Polyepichlorohydrin, PECH 17.30 15.45 7.5 3.0 ± 1.1
Poly(ε-caprolactone), PCL 18.55 9.80 8.25 ± 0.25 0.0
Hyperbranched, Boltorn H2O 19.95 5.65 10.45 ± 0.15 0.0
Poly(dimethyl siloxane), PDMS 14.50 3.70 5.15 ± 0.15 0.0
Poly(vinyl methyl ether), PVME 16.55 11.25 6.9 ± 0.20 0.0
Polytetrahydrofuran, PTHF 14.85 11.20 9.2 0.0
Poly(ethylene oxide), PEO 16.40 15.70 10.6 0.0
these values cannot be obtained directly for high polymers, we should resort to approximate values
from analog oligomers. In Table 11.5 are reported such values for a number of analogs. As observed,
there is a non-negligible scatter of values around the average ratio of 0.096. Even for polyethylene,
the uncertainty from using data from its analogs may be as high as 5%. From Table 11.3, we observe
that γW tends to a limiting value of 19.4 mN/m as the size of n-alcane increases. On the other hand,
using the two ratios of Table 11.5 and the value of 13.75 MPa1/2 for the σW PSP of polyethylene
TABLE 11.5
The Vcosmo/Acosmo Ratio for Analog Oligomers
Polymer Analog Acosmo (nm2) Vcosmo (nm3) Vcosmo/Acosmo γW, Pred (mN/m)
PE Tetradecane 3.1652 0.3211 0.101 19.5 ± 0.5
PE Hexatriacontane 7.5481 0.8034 0.106
PP 2,4-Dimethylpentane 1.847 0.190 0.101 20.0 ± 1.0
PP 2,4,6-Trimethylheptane 2.2058 0.2355 0.107
PVC Vinyl chloride 0.934 0.076 0.082 28.0 ± 4.0
PVC 2,4-Dichloropentane 1.6962 0.1707 0.101
PEO Triethyleneglycol n-propylether 2.6554 0.2561 0.096 26.5 ± 0.5
PEO Tetraethyleneglycol monomethylether 2.7662 0.2687 0.097
PEO Tetraethyleneglycol monobutylether 3.4051 0.3352 0.098
PS Styrene 1.5445 0.1476 0.096 22.5 ± 1.0
PS Ethylbenzene 1.5902 0.1544 0.097
PVAc Vinylacetate 1.296 0.1127 0.087 23.0 ± 4.0
PVAc Ethylacetate 1.3345 0.1175 0.088
PMMA Methylmethacrylate 1.4487 0.1342 0.093 33.0 ± 1.0
Average 0.096
Adhesion and Wetting 617
[15,16], we obtain for γW the values of 19.1 and 20.0 mN/m. These results lead to an estimated or pre-
dicted value of 19.5 ± 0.5 mN/m for the γW of polyethylene. Following this procedure, predicted val-
ues for γW have been obtained for other polymers and are reported in the last column of Table 11.5.
The observed large uncertainties for poly(vinyl acetate) (PVAc) and poly(vinyl chloride) (PVC)
are due to the large discrepancies between the σW of high polymers and of their analog oligomers.
Such discrepancies are sometimes unavoidable in view of the fact that the available inverse-gas
chromatography (IGC) data are for relatively high temperatures. In these cases of large uncertain-
ties, one may resort to the use of Equation 11.29 with experimental contact angle data for a couple
of solvents. Having γW, in the one or in the other way, one may follow the procedure of the previous
section and predict all other surface tension components of the polymers. Surface tensions of pure
polymer surfaces are, however, needed in this procedure.
In their seminal paper, Owens and Wendt [31] have reported surface tension and contact angle
data for a number of polymers, which have been widely used as reference data in the literature.
These data are reproduced in Table 11.6 for the polymers with available PSPs. In each row of the
table there are two sets of data. The upper set uses the experimental contact angle data and calcu-
lates the surface tension components of the polymers on the basis of these experimental data. The
lower set (in italics) gives the predicted surface tension components by the present approach and
uses these components in order to predict the contact angles as well. As observed, the calculated
surface tension components are in rather satisfactory agreement with the predicted ones. On the
other hand, the predicted contact angles with diiodomethane (DIM) are in very good agreement
with the experimental ones. As regards contact angles with water, this approach overpredicts them
for polystyrene and to a lesser extent for poly(vinyl chloride). The failure for polystyrene may be due
to the fact that COSMO-RS and, thus, PSP does not consider the interactions with the pi-electrons
of the aromatic ring as hydrogen bonding. This holds true, unless this particular polystyrene surface
had acquired extraneous polar groups, which is often the case with such data in the literature. The
failure for PVC originates from a rather too high value for σW from IGC data, which leads to a high
value for γW, which, in turn, leads to a low value for γpz and, thus, a low value for γa. The overall
picture, however, is rather satisfactory. The important message of the calculations reported in Table
11.6 is that Equation 11.29 along with the surface tension components of solvents from Table 11.3
may be used for a rather reliable estimation of the surface tension components of polymers. In fact,
they may be used for the characterization of any smooth solid surface regardless of whether PSP
data are available or not.
Before proceeding further or before embarking into calculations of surface tension components
for surfaces with unknown PSPs, we should clarify the methodology and set guidelines for the
procedure. The first thing we should keep in mind is the potential existence of numerous surfaces
TABLE 11.6
The Surface Tension Components of Polymers
Polymer γtotal (mN/m) θ, Water θ, DIM γW γpz γa γb
Polyethylene 33.2 104 53 19.7 13.5 0 0
106 51 19.5 13.7 0 0
Poly(vinyl chloride) 41.5 87 36 24.4 17.1 3.2 0
92 35 28.0 13.5 1.2 0
Poly(methyl methacrylate) 40.2 80 41 32.7 7.5 0 7.0
81 40 33.0 7.2 0 6.8
Polystyrene 42.0 91 35 23.5 18.2 0 1.5
100 35 22.5 19.5 0 0
Source: Data by Owens, D.K. and Wendt, R.C., J. Appl. Polym. Sci., 13, 1741, 1969.
618 Handbook of Surface and Colloid Chemistry
for the same polymer, depending on intrinsic and extrinsic factors ranging from molecular size
distribution, microstructure or end groups, up to specific surface treatment and conditioning. The
second thing to keep in mind is that a theoretical approach like PSP can predict unique properties of
surfaces if there are such unique properties. If the surface is not clean and not at equilibrium or the
probe affects the surface properties, it is difficult to extract unique surface properties. An inspec-
tion in the compilation of experimental contact angles on polymers by Lyklema [32] reveals this
non-unique property problem. As an example, even for polyethylene, the reported advancing contact
angle with water ranges from 28 to 116. Obviously, these polyethylene surfaces are not identical
(if the measurements have been conducted properly).
The probe effect on the surface properties is a crucial issue in the study of these interfacial phe-
nomena. It is then worth examining if the current approach can predict contact angles with other
solvents on the polymers of Table 11.6. In Table 11.7 are reported such predicted contact angles with
four additional solvents.
The results in Table 11.7 deserve several comments. First of all, there are two sets of predic-
tions. Those marked in italics are absolute predictions directly from the corresponding PSPs.
The other set was obtained from the calculated (modified) surface tension components reported
in Table 11.6 in order to reproduce the experimental contact angle data with water and/or DIM.
As observed, the predictions with both sets fall, essentially, within the range of measured values.
With the expected exception of PVC, the predictions by the two sets are very similar. A simi-
lar to PVC discrepancy was expected for polystyrene, on the basis of the data in Table 11.6. As
observed in Table 11.7, the absolute predictions for polystyrene are much better than the predic-
tions with the modified surface tension components. This probably indicates that the surface of
the polystyrene sample in Table 11.6 may have not been very clean. It is essential to point out that
with both sets of predictions, the defining Equations 11.18 through 11.20, which interrelate the
polarity/polarizability, acidity, and basicity components, were respected. This is important when
a hydrogen bonding solvent is used as a probe of the corresponding character of the solid surface
as it cannot attribute arbitrary values in one surface tension component disregarding the others.
This in turn guarantees that the hydrogen bonding character of all surfaces, liquid or solid, will
be compatible with their bulk-phase hydrogen bonding character and, thus, the information may
be interchanged.
TABLE 11.7
Experimental and Predicted Contact Angles on the Polymers of Table 11.6
θ, Ethylene Glycol θ, Formamide θ, Glycerol θ, DMSO
Polymer Predicted Experim. Predicted Experim. Predicted Experim. Predicted Experim.
Polyethylene 72 69 a 85 a77–81 80 a79–94 63
72 85 80 r53–56 63
Polystyrene 52 60a 67 a69–88 64 a71–80 44
62 77 r35–70 71 49
PVC 45 45.7b 63 a55–66 59 a67–69 <0 Swollenb
52 69 56.9a 65 61.8a 14
PMMA 43 36.8b 61 a51–64, 41a 63 a60–69 43 12b
44 61 63 58.1b 45
Experimental data from the compilation of Lyklema [32] unless otherwise specified. Advancing and receding contact angle
data are marked with a and r in front, respectively. Calculations in italics are absolute predictions.
a Dann [33].
TABLE 11.8
Comparison of Hydrogen Bonding Character in Bulk and in Liquid–Vapor Interface
σ2hb γ hb
B= S=
Alkanol σ + σ 2pz
2
W γW + γpz B/S
Methanol 1.636 0.290 5.645
Ethanol 1.066 0.226 4.716
Ethanol 1.066 0.226 4.716
1-Propanol 0.823 0.223 3.685
1-Butanol 0.656 0.217 3.025
1-Pentanol 0.638 0.238 2.680
1-Hexanol 0.587 0.244 2.399
1-Heptanol 0.501 0.238 2.104
1-Octanol 0.458 0.234 1.956
1-Decanol 0.375 0.227 1.652
620 Handbook of Surface and Colloid Chemistry
of the pure polymer surface while the former its “polarized” state. Contact angles of absolute pre-
dictions are, thus, expected to be closer to advancing contact angles while the ones predicted with
the modified surface tension components could be considered to be closer to the receding ones if
one could exclude solvent molecules adsorbed on the surface in the receding front.
There are two important points to be made in relation with the vOCG model [4–6,28]. As we
have seen, in the present approach there is a direct correspondence of the acidity/basicity surface
character of a compound to its corresponding bulk character. In contrast, in the vOCG approach,
this correspondence is lost and the basicity character appears disproportionately prevailing. In par-
allel, the water neutrality hypothesis (equal acidity and basicity) of the original vOCG model is
sound and fully supported by quantum chemical calculations and QSPR/LSER data. Any attempts
in the literature to drastically alter this neutrality may have improved the fitting performance but
they may have obscured the physical basis and the insight in the studied interfacial phenomena.
It should be stressed again that the polymer PSPs used in this work are those obtained from IGC
(cf. Table 11.4) at relatively high temperatures. No attempt was made to correct for any temperature
dependence. In a refinement of the model, one could certainly account for it. Such an account of
temperature dependence could bring σW from IGC for high polymers closer to the corresponding
one from oligomers.
In conclusion, the present approach to surface tension components has some apparent similarities
with the vOCG model [4–6,28] but, in essence, it is a drastically different approach. The key feature
is that it provides with a sound basis for the interactions dictating interfacial phenomena, which is
fully compatible with the picture we have about intermolecular interactions from quantum chemi-
cal calculations. The new approach has not been tested extensively yet, but the tests of this work
are indicating a rather satisfactory agreement with experiment. Work is underway in our laboratory
toward a more extensive testing of the new approach against experimental data.
ACKNOWLEDGMENTS
The author is grateful to Otto Monsteds Fond for financial support of this work and to Professor
G. Kontogeorgis/DTU for valuable discussions.
REFERENCES
1. Dupre, A., Theorie Mechanique de la Chaleur, Gauthier, Villars, Paris, France, 1869.
2. Hiemenz, P.C., Rajagopalan, R., Principles of Colloid and Surface Science, 3rd edn., Marcel Dekker,
New York, 1997.
3. Young, T., Philos. Trans. R. Soc. Lond., 95, 65–87, 1805.
4. van Oss, C.J., Chaudhury, M.K., Good, R.J., Adv. Colloid Interface Sci., 28, 35, 1987.
5. Good, R.J., J. Adhesion Sci. Technol., 6, 1269, 1992.
6. van Oss, C.J., Interfacial Forces in Aqueous Media, 2nd edn., CRC Press, Boca Raton, FL, 2006.
7. Della Volpe, C., Siboni, S., J. Colloid Interface Sci., 195, 121, 1997.
8. Della Volpe, C., Siboni, S., J. Adhesion Sci. Technol., 14, 235, 2000.
9. Lee, L.H., J. Adhesion, 67, 1, 1998.
10. Yasuda, T., Okuno, T., Yasuda, H., Langmuir, 10, 2435, 1994.
11. Carre, A., J. Adhesion Sci. Technol., 21(10), 961, 2007.
12. Panayiotou, C., Phys. Chem. Chem. Phys., 14, 3882, 2012.
13. Panayiotou, C., J. Chem. Thermodynamics, 51, 172, 2012.
14. Panayiotou, C., J. Phys. Chem. B, 116, 7302, 2012.
15. Panayiotou, C.G., J. Chromatogr. A, 1251, 194, 2012.
16. Panayiotou, C., Polymer, 54, 1621, 2013.
17. Klamt, A., Jonas, V., Buerger, T., Lohrenz, J.C.W., J. Phys. Chem., 102, 5074, 1998.
18. Klamt, A., COSMO-RS from Quantum Chemistry to Fluid Phase Thermodynamics and Drug Design,
Elsevier, Amsterdam, the Netherlands, 2005.
19. COSMObase Ver. 1401, COSMOlogic GmbH & Co., K.G., Leverkusen, Germany, 2013.
Adhesion and Wetting 621
20. Lin, S.T., Sandler, S.I., Ind. Eng. Chem. Res., 41, 899, 2002.
21. Grensemann, H., Gmehling, J., Ind. Eng. Chem. Res., 44, 1610, 2005.
22. Pye, C.C., Ziegler, T., van Lenthe, E., Louwen, J.N., Can. J. Chem., 87, 790, 2009.
23. Abraham, M.H., Chem. Soc. Rev., 22, 73, 1993.
24. Zissimos, A.M., Abraham, M.H., Klamt, A., Eckert, F., Wood, J., J. Chem. Inf. Comput. Sci., 42, 1320,
2002.
25. Hansen, C.M., Hansen Solubility Parameters: A User’s Handbook, CRC Press, Boca Raton, FL, 2007.
26. Ahlrichs, R., Baer, M., Haeser, M., Horn, H., Koelmel, C., Chem. Phys. Lett., 162, 165, 1989.
27. Daubert, T.E., Danner, R.P., Physical and Thermodynamic Properties of Pure Compounds: Data
Compilation, Hemisphere, New York, 2001.
28. van Oss, C.S., Chaudhury, M.K., Good, R.J., Chem. Rev., 88, 927, 1988.
29. Abrams, D.S., Prausnitz, J.M., AIChE J., 21, 116, 1975.
30. Sorensen, J.M., Arlt, W., Liquid–Liquid Equilibrium Data Collection, DECHEMA, Chemistry Data
Series, Vol. 5, Frankfurt, Germany, 1979.
31. Owens, D.K., Wendt, R.C., J. Appl. Polym. Sci., 13, 1741, 1969.
32. Lyklema, J., Fundamentals of Interface and Colloid Science, Vol. III, Academic Press, London, U.K.,
2000.
33. Dann, J.R., J. Colloid Interface Sci., 32, 302, 1970.
34. Della Volpe, C., Siboni, S., Maniglio, D., Morra, M., Cassinelli, C., Anderle, M., Speranza, G. et al.,
J. Adhesion Sci. Technol., 17(11), 1425, 2003.
12 Surface Chemistry of Oil
Recovery (Enhanced Oil
Recovery)
K.S. Birdi
CONTENTS
12.1 Introduction........................................................................................................................... 623
12.1.1 Surface Chemistry Aspects of Enhanced Oil Recovery........................................... 627
12.1.1.1 Interfacial Tension in Oil Reservoir........................................................... 630
12.1.1.2 Viscosity of Reservoir Oil.......................................................................... 634
12.1.2 Surface-Active Substances (Surfactants) Applied in EOR........................................ 634
12.1.3 Polymers and Surfactants in EOR............................................................................. 636
References....................................................................................................................................... 637
12.1 INTRODUCTION
The demand for energy by mankind is the most important issue of this age. As the world’s popu-
lation increases (ca. threefold every 50 years), the demand for energy increases drastically. The
energy demand is increasing at about 2% per year. Among energy resources (e.g., oil, gas, coal,
atomic energy, hydropower, wind energy, and solar energy), oil is the most important at the pres-
ent stage. Oil supplies about 45% of the world’s energy (petroleum) while natural gas (mostly
as methane, CH4) supplies about 20%. At present, worldwide, about 80 million barrels of oil is
consumed every day. Most of the oil produced today comes from mature reservoirs. It is esti-
mated that some 1012 barrels of oil remain in the reservoirs worldwide after the primary production
(1 barrel = 0.159 m3). Hence, it is a great challenge to scientists to invent methods and procedures to
recover these vast amounts of residual oil reserves. There are chemical processes called enhanced
oil recovery (EOR) that have been used to recover the residual oil in the past decades (Ahmed,
2001; Bavière, 2007; Birdi, 1999; Carcoana, 1992; Green, 1998; Kenneth, 2012; Lake, 1996; Lake
and Walsh, 2008; Lichtner et al., 1996; Mark and Lake, 2003; Rao, 2012; Scheng, 2013; Schramm,
2010; Smith, 1966; Taber, 1968; Taber et al., 1997; Tunio et al., 2011) (Figure 12.1). Further, the
EOR methods have developed in various ways as the price of oil has increased over the past decades
(from few dollars to around 100 USD/barrel). The oil reservoir production history comprises vari-
ous phases (Figure 12.1):
• Primary phase
• Secondary phase
• EOR
623
624 Handbook of Surface and Colloid Chemistry
The conventional oil reservoir is found to be a complicated system (under high pressures and temper-
atures [which depends on the depth and other parameters]) (Figure 12.2) (Alvarado and Manrique,
2010; Bavière, 1991; Birdi, 1999, 2010; Douglas and Tiratsoo, 1975; Green, 1998; Kalfayan, 2008;
McCain, 1990; Moritis, 2008; Rao, 2012; Scheng, 2010, 2013; Taber et al., 1997; van Poolen, 1980;
Wilhite, 1986):
FIGURE 12.2 Oil reservoir rock: oil (circles) in large pores/narrow pores (schematic).
Surface Chemistry of Oil Recovery (Enhanced Oil Recovery) 625
The recovery of oil from reservoir involves the flow of fluid through pores under a given temperature
and pressure. However, it is known that high pressure alone sometimes does not help in the flow
of oil (oil recovery) due to various chemicophysical forces. In these reservoirs, there are various
specific surface forces (besides other forces), which are important for the oil recovery. In order to
recover the vast amounts of residual oil in reservoirs, over the past decade enhanced oil recovery
(EOR) methods are being applied. The conventional oil reservoirs during the past decades are
now being subjected to EOR and with very positive results. The application of surface and colloid
chemistry principles has been applied in EOR. In very simple terms, one can further describe the
essential surface forces involved in EOR:
Of course, the reservoir characteristics such as pressure and temperature will also determine the
EOR procedure to be applied. It is thus seen that the degree of oil recovery will depend on these
factors or a combination of these different parameters. The EOR process has thus to resolve these
forces in the reservoirs in order to increase the oil recovery.
What Is Oil?
It is of interest to mention briefly about the origin and composition of oil as found in the earth reser-
voirs (Rao, 2012). This is important since the EOR method will be dependent on the composition and
chemical characteristics of the oil and the reservoir rock. The composition of oil as found in different
parts of the world has been analyzed. Many different trace amounts of molecules have been identified
as markers for the source of oil. It is believed that oil (fossil fuel) is formed by natural processes as
decomposition of organic matter (such as wood, grass, leaves, etc.). These processes are also known to
have taken place over hundreds of millions of years. The composition of fossil fuels is high in carbon.
It is thus evident that some millions of years ago, CO2 from air (at present, CO2 concentration in air
is about 400 ppm [0.04%]) was used to produce plants through photosynthesis (CO2–water–sunlight:
plants, wood, etc.), which after decomposition created coal (solid), oil (liquid petroleum), and natural
gas (mostly methane as gas). This also shows that the state (solid–liquid–gas) at which the decomposi-
tion ended depends on the physicochemical conditions (i.e., pressure and temperature) in the reservoir.
In the past decades, natural gas and oil have become the most important energy sources. In the United
States, gas recovery has increased dramatically in the past decade. This increase arose from the new
unconventional oil and gas recovery technology applied to tight gas as found in shale gas. Natural
gas is recovered by hydraulic fracturing (fracked) technology, which allows natural gas to escape
through fractures (Rao, 2012). The International Energy Agency (IEA) estimates show that more than
hundreds of trillion cubic meters (tcm) of technically recoverable unconventional gas reserves are
available worldwide. The United States already produces shale gas, which amounts to about 30% of
the domestic demand. Other countries, such as China, Australia, Canada, and the United Kingdom are
currently developing similar gas recovery projects. Further, conventional reservoirs are much smaller
in volume than the source rocks (such as shale oil and gas). This means that as regards potential oil and
gas supply, mankind can recover much larger amounts of oil and gas from shale deposits in the future
(estimates indicate that these deposits may be of the order of many decades larger than the conven-
tional reservoirs). Shale oil technology has also developed during the past decades, which now adds a
considerable contribution to the world supply (Rao, 2012; van Poolen, 1980).
Due to various physicochemical forces, only about 30% of the total oil in a reservoir (worldwide)
has been recovered with the conventional techniques. The primary recovery process is based on the
626 Handbook of Surface and Colloid Chemistry
reservoir pressure as present initially, and the recovery decreases (and finally stops) as the pressure
diminishes. At this stage, mankind has an urgent need to sustain energy supply and therefore the
old (i.e., mature) reservoirs with large residual oil reserves left unrecovered require enhanced oil
recovery (EOR) methods. In the past decades, a great amount of research has been carried out on
EOR processes and the literature shows that this is an important ongoing subject worldwide. This
chapter describes the most essential surface chemistry aspects about EOR technology. The impor-
tant surface chemistry parameters are delineated.
General recovery history of oil reservoirs worldwide:
Following are the distinct stages of oil recovery in the history of most of the reservoirs:
I. In this stage the oil is flowing freely under pressure as found at the well site.
II. As the pressure drops and the flow slows down, additional pressure (by using a gas or water
injection) is applied.
As the oil recovery stops after stage II, the reservoir needs to be applied to enhance oil recovery (EOR),
if the residual oil is to be recovered. A typical EOR oil recovery in a conventional reservoir shows an
increase in oil flow, as shown in Figure 12.3. The data from many different oil reservoirs have been ana-
lyzed by using various mathematical models. These models fit the EOR flow pattern very satisfactorily.
Different EOR Processes
• Gas injection
• Inert gas (nitrogen)
• Hydrocarbon gas injection
• Miscible liquid propane
• CO2 flooding
• Improved water flooding processes
• Surfactant/polymer flooding
• Low IFT flooding
• Alkaline solutions
• Polymer flooding
• Gels
• Microbial injection
• Foams
Water
flooding
Oil recovery
EOR
Time
FIGURE 12.3 A typical oil flow pattern after EOR application to a reservoir (schematic).
Surface Chemistry of Oil Recovery (Enhanced Oil Recovery) 627
Gas
Steam
Polymer
Reservoir depth (ft)
CO2
Surfactant injection
Nitrogen
15,000
FIGURE 12.4 Correlation between EOR methods and oil reservoir characteristics (schematic).
• Thermal methods
• In situ combustion
• Steam and hot water
These different enhanced oil recovery methods become even more involved when the combination
of two or more different techniques is used, for example, when CO2 injection is combined with sur-
factant–polymer flooding. A simple schematic correlation has been found between the EOR method
and the depth of the reservoir and oil viscosity (Figure 12.4). Obviously, in such a complex system
no simple correlation can be absolutely valid, but it can be a useful guideline. Field operations have
provided evidence for a good relation as depicted in Figure 12.4.
(a) (b)
For instance, the pressure inside a soap bubble is higher than outside. The difference in pressure,
ΔPbubble, is given as:
2γ liquid
ΔPbubble = (12.1)
R radius
where
γliquid is the surface tension of the liquid
R radius is the radius of the bubble
It is thus seen that the magnitude of ΔPbubble will be larger in a smaller bubble than inside a larger
bubble. Furthermore, the magnitude of ΔP will be zero at a flat surface (since R radius is very large).
This also indicates that when two bubbles collide, the smaller bubble will merge into the larger
bubble. For example, it is known that the height of meniscus of water in a capillary of smaller radius
is higher than in a larger size capillary (Figure 12.6). The magnitude of ΔPcapillary at the surface of
fluid is related to the height, hcapillary, a liquid will rise with a given surface tension, γliquid:
γ liquid
h capillary = (12.2)
(2R radiusρ liquid ggravity )
where
ρliquid is the density of the fluid
ggravity is the acceleration of gravity (Birdi, 1999, 2010)
Oil
FIGURE 12.6 Rise of a fluid (oil) in large and small capillary (schematic).
Surface Chemistry of Oil Recovery (Enhanced Oil Recovery) 629
These results indicate that oil will be easily produced (which corresponds to the primary oil
production) from large size pores than in narrow pores (residual oil will be expected to be present in
narrow pores). These data thus indicate that the residual oil is present in very fine pores of the rock.
In simple terms, one may describe this process as follows:
• Fluid as found under the surface of the earth is under high pressure and temperature, and
will move upward.
• Fluid has to move through fissures or fractures (capillary forces).
• The viscosity of fluid will give resistance to flow.
• Interaction between the fluid and the surface of the rock will determine the process (wet-
ting or nonwetting characteristics).
In the reservoir, besides oil there is also some amount of brine (water with a high content of salt).
It is thus obvious that in the system
Oil–Rock–Brine
there are interfacial forces involved between the surface tensions of γrock, γoil, and γbrine. These surface
forces can be analyzed by contact angle measurements (Birdi, 1999, 2002, 2009, 2010). The fact that
most of the oil (50%–70%) in a petroleum reservoir remains un-recovered in the ground is due to
some main chemicophysical phenomena:
EOR History
The various EOR methods, which have been applied over the past decades, have been regulated
by the economics of the oil prices. Mostly, these developments have been regulated by the price
increase through time (from ca. a dollar per barrel to over 100 USD/barrel at present stage). The
present high price of oil has been the driving factor for the application of more advanced EOR tech-
niques. Furthermore, it must be stressed that these EOR processes are also necessary, considering
that about 70% of oil is left behind in the mature oil wells. The EOR methods are dependent on
various factors, which are as follows:
In order to apply the most suitable EOR procedure, one thus needs the knowledge of the various
physicochemical data of the reservoir. Steam injection reduces the viscosity of oil (due to higher
temperature), thus helping the increased production. The addition of polymers gives rise to higher
viscosity of the water phase (used for pushing the oil). More recently, there have been some recovery
methods where microbial applications have been applied in EOR (Banat, 1995; Bryant and Lockhart,
2002; Donaldsen, 1991; Jimoh, 2012). The microbial EOR is complicated and the exact mechanism is
630 Handbook of Surface and Colloid Chemistry
also difficult to analyze. However, it is known that microbes produce various chemicals and some of
these molecules may exhibit surface activity, which will reduce the oil IFT, and thus help oil recovery.
Formation of oil from organic matter (Source rock, such as shale oil)–movement of oil toward
the surface of the earth–trapped in pores and as found in conventional reservoirs
The production of petroleum from a reservoir may be divided into different phases. The first stage
of oil recovery is where the flow is under the reservoir pressure. Very early in the life of a reservoir,
energy must usually be supplied to the porous medium, which bears the crude oil so that it contin-
ues to flow to the producing wells. This energy is brought into the reservoir by injection of water
or gas (nitrogen, methane, CO2) (Tunio et al., 2011). With these secondary recovery methods, about
30%–40% of the original oil in place may be recovered, while the rest may be left in the earth.
Water Oil
(a)
Water
(b) Oil
FIGURE 12.7 Water-flooding of oil in a circular shape (a) or square type (end view) pore (b).
These data show that the magnitude of IFT is about 50 mN/m for alkanes and changes very little
with chain length. On the other hand, the magnitude of IFT changes to much lower values in the
case of alcohol/water systems. The reduction of IFT gives rise to oil recovery from oil as present
in the very small pores (with high capillary pressure). The magnitude of IFT of alkane/surfactant
solution can be much lower than 1 mN/m (Birdi, 1999, 2002, 2009, 2010), depending on the
system.
The high degree of water-by-pass is also indicative of that the pores are more nonspherical type.
The degree of by-pass is larger in square-type pores than in the circular ones. Since the pores in any
reservoir are expected to be nonspherical, one should also expect a large degree of water-by-pass.
Actually, this is also observed in real reservoirs and laboratory experiments.
The capillary forces in square tubing have been investigated as a function of size of tubing and
surface tension of fluids (Bera et al., 2011; Birdi et al., 1988; Cayias et al., 1976). Of course, one
may expect that the pores in reservoirs are a mixture of spherical and nonspherical (square, triangle,
pentagon, etc.) shapes. The water-by-pass in rectangular capillary is largely present at the corners
(which are absent in circular capillaries) (Figure 12.7). However, this is expected to decrease as
the magnitude of IFT decreases (due to the application of suitable surfactant). As can be seen in
Figure 12.2, the nonwetting phase (oil) has to pass small throats before it can leave the pore space.
It is the diameter of these pore throats that determines the capillary pressure that has to be overcome
to displace the nonwetting phase by the wetting phase.
where
δliquid is the density
ggravity is the force of gravity
Slength is the side length of the square tubing
Hrise is the rise of liquid in the tubing
Cconstant (=1.089) is a capillary constant for square shapes
632 Handbook of Surface and Colloid Chemistry
γoil
Oil
CA
γrock γoil/rock
Rock
FIGURE 12.8 Wetting characteristics of oil on rock surface as depicted by contact angle (CA), Θ, of surface
tension of oil (γoil), of solid (rock) (γrock), and of oil/solid (rock) (γoil/rock).
The different surface forces, which are present when a drop of oil is placed on a solid (such as
reservoir rock) are as follows:
γ oil
γ rock = γ oil cos Θ + (12.4)
γ rock
where
γrock is the surface tension of rock
γoil is the surface tension of oil
γoil/rock is the surface tension of rock–oil interface
contact angle, Θ, is where the three different surface forces reach an equilibrium (Figure 12.8)
(Birdi, 2010; Chattoraj and Birdi, 1984)
The contact angle corresponds to the state where the different surface forces are at equilibrium. The mag-
nitude of γrock can be estimated if the value of contact angle, Θ, and the γoil is known. The d etermination
of surface tension of solids (such as γrock) has been extensively studied and its application in real-world
systems has been extensively delineated in the current literature (Birdi, 1999, 2010; Chattoraj and Birdi,
1984). If the contact angle is less than 90°, then the liquid is known to wet the solid surface. However,
if the contact angle is larger than 90°, then the surface is nonwetting. For example:
Water drop placed on the following solid surfaces exhibits different contact angles:
This clearly shows that the surface of TEFLON is nonwetting as regards water. Similar consider-
ations will apply when considering the wetting properties of oil–rock systems in any EOR system.
The capillary pressure (i.e., a fluid inside a capillary tubing) is given as:
(2γ cos Θ)
Pcapillary = (12.5)
R pore
where
γoil is the interfacial tension
Θ is the contact angle
Rpore is the pore radius
Surface Chemistry of Oil Recovery (Enhanced Oil Recovery) 633
(a)
(b)
FIGURE 12.9 State of capillary pressure of oil and water in a cylindrical (a: symmetrical and b: asymmetric
tubing) (schematic).
If the diameter of the pores were uniform (such as a cylinder), capillary pressure differences would
be zero, as shown in Figure 12.9a. However, in an oil drop that is driven by water injection through
a pore (with radii R1 and R2), the difference in capillary pressure is given as
⎛ 1 1 ⎞
ΔPcapillary = 2γ oil cos Θ ⎜ − ⎟ (12.6)
R
⎝ 1 R 2 ⎠
In a cylindrical pore, R1 = R2, which means ΔPcapillary = 0. The capillary pressure gives rise to water-
by-pass, as also observed in real reservoir flooding. It is thus seen that (Figure 12.9b) the quantity
capillary pressure counteracts oil recovery. This means that the amount of oil that may be recovered
is proportional to the pressure drop applied to the porous medium and inversely proportional to
capillary pressure. This can be expressed as follows:
⎛ ΔP ⎞ ⎛ R ⎞
1 – Scapillary = ⎜ ⎟ ⎜ ⎟ (12.7)
⎝ 1 ⎠⎝ γ ⎠
where
Scapillary is the remaining oil or residual oil saturation
R is the radius of the pore in the rock (Birdi, 1999)
where
k is the permeability
ε is the porosity
634 Handbook of Surface and Colloid Chemistry
(kDarcy ΔP )
Vflow = (12.9)
(μ fluid dx solid )
Gravel: 100,000 D
Sand: 1 D
Granite: less than 0.01 micro-Darcy (μD = 10−6 D)
Laboratory experiments have been carried out to (Birdi, 1991; Taber, 1968) investigate the influence
of pressure drop and interfacial tension on residual oil saturation. It was found that the oil recovery
from a core of reservoir rock is proportional to the pressure drop across the core and capillary pres-
sure. These experiments were performed on 420 mD Berea-sandstone and 2000 mD sand pack.
Though both series of experiments were performed on different materials and in different ways,
they showed similar results.
100 1
0 10
Degree of flooded pore volumes
FIGURE 12.10 Variation of oil recovery (% of oil in place [OIP]) versus flooded pore volumes.
10,000
Viscosity of oil (mPa s)
Reservoir
1
2 4
Reciprocal temperature (1/T 103 K–1)
FIGURE 12.11 Variation of viscosity (mPa s) of a typical heavy oil versus temperature (1/T 103 K−1).
thus reduce the interfacial tension by a few decades. Oil–water interfacial tension may be reduced
by these surfactants from 40 to 10−4 –10−5 mN/m, depending on the system. The application of sur-
factants (which gives rise to lower IFT) has been found to be a very important procedure to recover
more oil from petroleum reservoirs. There are many other chemicals that exhibit properties similar
to those of surfactants (Alvarado and Manrique, 2010; Birdi, 1999; Taber et al., 1997). Surfactants,
which have been used in EOR, have to be stable at the oil reservoir conditions (especially high
temperature and pressure) (Table 12.1). For instance, arylalkyl sulfonic acids (formed from olefin
and SO3) are found to be stable at high temperatures.
These molecules are also called as amphiphiles, due to their dual nature of properties. The
hydrophobic part, which is the alkyl chain, dislikes water (hydrophobic) or is insoluble in water,
which makes it useful for oil recovery, as this part induces lower IFT. The hydrophilic part (like
water) allows the surfactant to disperse in water to form large aggregates, micelles (Austad et al.,
1994; Birdi, 1999, 2010).
The magnitude of temperatures in oil reservoirs is found to be relatively high (70°C–130°C). The
water in these reservoirs is found to contain very high concentrations of electrolytes (mostly NaCl).
Further, in some reservoirs one finds besides quartz sand, significant amounts of clay minerals.
636 Handbook of Surface and Colloid Chemistry
TABLE 12.1
Typical Surfactants Used in EOR
Hydrophobic Part Intermediate Part Hydrophilic Part Counter Ions
Iso-octylphenolic Ethyleneoxide Carboxylate Alkaline
Dodecyl-phenolic Propyleneoxide Sulfate Amine
Di-nonylphenolic Ethylene-propyleneoxide Sulfonate Quartz
The surfactant molecule is characterized by its two parts: hydrophobic part and hydrophilic part.
Surfactants have a strong tendency to adsorb rocks in the reservoirs (due to the characteristics of
the molecule) (Chattoraj and Birdi, 1984). This is not desirable and therefore one needs to apply
surfactants that exhibit low adsorption characteristics. In order to reduce the magnitude of interfa-
cial tension, γoil/water, between water and oil, surfactants of various compositions have been tested,
and additional surfactants have been developed (Table 12.1). Depending on the oil properties, the
salinity of the reservoir brine, and the temperature, the surfactants are tailor-made based on the
composition of reservoir fluids.
( k water /ηwater )
Mmobility = (12.10)
( k oil /ηoil )
(where kwater and koil are the effective permeabilities to water and oil, respectively, and ηwater and ηoil
are the viscosities of water and oil, respectively) influences the areal sweep efficiency. Apart from
these two effects, the vertical sweep efficiency determines the performance of a flood. Before the
mechanism that improves the vertical sweep efficiency during polymer flooding is discussed, some
introductory remarks on the rheology of pseudo-plastic liquids are necessary (Birdi, 1999; Rao,
2012). For water and oil, the viscosity is, in most cases, a constant value. For polymer solutions, this
is more or less not the case; the viscosity is a function of the rate of shear strain. In regard to the
viscous flow of a liquid, the rate of shear strain is a function of both flow geometry and flow velocity.
For flow in porous media, this means that in narrow pores the rate of shear strain is higher than in
Surface Chemistry of Oil Recovery (Enhanced Oil Recovery) 637
larger pores, or, in terms of petroleum reservoir engineering, that at the same Darcy velocities the
shear rate in less permeable zones is higher than in zones having good permeabilities. Hence, for
example, for a Darcy velocity of 0.2 m/day in a sandstone with a porosity of 25% and a permeability
of 2000 mD, the rate of shear strain is 8.6 s−1 for a polymer solution with a viscosity of 40 mPan
(n = 0.6). For the same polymer solution at the same Darcy velocity in sandstone with a porosity of
20% and a permeability of 200 mD, the rate of shear strain is 30 s−1. This means that the viscosity
of the flowing polymer solution is higher in the sandstone with the high permeability (19.9 mPa·s)
than in the sandstone with the low permeability (10.2 mPa·s).
The reservoir oil recovery by various EOR methods has been described for the water (or polymer
or surfactant) flooding techniques in EOR (Ahmed, 2011; Alvarado and Manrique, 2010; Birdi, 1999;
Taber et al., 1997). Most commonly, the highly permeable zones are invaded by the flood water dur-
ing secondary operations or natural water drive, and low permeable zones are not flooded so that oil
is left in these parts of a reservoir. This means that since the IFT of oil–water (ca. 50 mN/m) is high,
the capillary pressure inhibits the water from entering reservoir regions with smaller pores. During
polymer flooding (high viscosity), a poor vertical sweep efficiency may be improved because the
polymer solution, of course, first follows the paths prepared by water and then, because of its high
viscosity, tends to block these parts of the reservoir, so that oil that was previously immobile starts
flowing. In the next phase of oil recovery, the surfactant-polymer solution flooding improves the oil
recovery due to lower IFT values (less than 1 mN/m) (Ahmed, 2011).
Different oil recovery stages:
• Water flooding:
• Oil recovery from high permeable zones
• Polymer solution flooding:
• Oil recovery from low permeable zones
• Surfactant–polymer solution flooding:
• Oil recovery from very low permeable zones
The pressure gradient in the reservoir and especially in those zones where oil was immobile becomes
higher in a polymer flood than it was during the water drive. It has been reported (Sandiford, 1977)
that in a system of two parallel flooded cores of different permeabilities, the oil recovery due to
polymer flooding is significantly higher than during a flood in one core of uniform permeability.
REFERENCES
Ahmed, T., Reservoir Engineering Handbook, 2nd edn., Gulf Professional Publishing, Boston, MA, 2001.
Alvarado, V., Manrique, E., Energies, 1529, 3, 2010.
Austad, T., Fjelde, I., Veggeland, K., Taugbol, K., J. Petrol. Sci., 255, 10, 1994.
Banat, I.M., Biores. Technol., 1, 51, 1995.
Baviere, M., Basic Concepts in Enhanced Oil Recovery Processes, Society of Chemical Industry, London,
U.K., 1991.
Baviere, M., Basic Concepts in Enhanced Oil Recovery Processes, Elsevier Applied Science, London, U.K.,
2007.
Bera A., Ojha, K., Mandal, A., Kumar, T., J. Colloid Surf. A, 114, 383, 2011.
Birdi, K.S., ed., Handbook of Surface and Colloid Chemistry, CRC Press, Boca Raton, FL, 1999.
Birdi, K.S., ed., Handbook of Surface and Colloid Chemistry, 2nd edn., CRC Press, Boca Raton, FL, 2002.
Birdi, K.S., ed., Handbook of Surface and Colloid Chemistry, 3rd edn., CRC Press, Boca Raton, FL, 2009.
Birdi, K.S., Surface and Colloid Chemistry, Principles and Applications, CRC Press, Boca Raton, FL, 2010.
Birdi, K.S., Vu, D.T., Noerregaard, A., Colloid Polym. Sci., 470, 266, 1988.
Bryant, S.L., Lockhart, T.P., Soc. Pet. Eng. Reserv. Eval., 365, 5, 2002.
Carcoana, A., Applied Enhanced Oil Recovery, Prentice Hall, New York, 1992.
Cayias, J.L., Schechter, R.S., Wade, W.H., Soc. Petrol. Eng. J., 351, 16, 1976.
Chattoraj, D.K., Birdi, K.S., Adsorption and the Gibbs Surface Excess, Plenum Press, New York, 1984.
638 Handbook of Surface and Colloid Chemistry
Donaldsen, E.C., ed., Microbial Enhancement of oil Recovery—Recent Advances, Development in Petroleum
Science, New York, 1991.
Douglas, H., Tiratsoo, E.N., Introduction to Petroleum Geology, Scientific Press, Beaconsfield, U.K., 1975.
Dyes, A.B., Caudle, B.H., Ericson, R.A., Trans. AIME, 201, 81, 1954.
Green, D.W., Enhanced Oil Recovery (SPE Textbook Series), Vol. 6, Society of Petroleum Engineers,
Richardson, TX, 1998.
Jimoh, I.A., Microbial enhanced oil recovery, PhD thesis, Luma Print, Aalborg University, Esbjerg, Denmark,
2012.
Kalfayan, L., Production Enhancement with Acid Stimulation, 2nd edn., Pennwell Corp., 2008.
Kenneth, S.D., Hubberts Peak: The Impending World Oil Shortage, Princeton University Press, Princeton, NJ,
2012.
Lake, L.W., Enhanced Oil Recovery, Prentice Hall, New York, 1996.
Lake, L.W., Walsh, M.P., Enhanced oil recovery (EOR), Tech. report, Society of Petroleum Engineers,
Richardson, TX, 2008.
Lichtner, P.C., Steefel, C.I., Oelekers, E.H., Reactive Transport in Porous Media, Mineralogical Society of
America, Washington, DC, 1996.
Mark, W., Lake, L.W., A Generalized Approach to Primary Hydrocarbon Recovery, Elsevier, Amsterdam, the
Netherlands, 2003.
McCain, W.D., The Properties of Petroleum Fluids, 2nd edn., Pennwell Publishers, Tulsa, OK, 1990.
Moritis, G., Oil Gas J., 60, 96, 2008.
Rao, V., Shale Gas, RTI Press, New York, 2012.
Sandiford, B.B., Enhanced oil and surface chemistry, in: Shah, D.O., Schechter, R.S., eds., Improved Oil
Recovery by Surfactant and Polymer Flooding, Academic Press, New York, 1977, p. 487.
Scheng, J., Modern Chemical Enhanced Oil Recovery: Theory and Practice, Gulf Professional Publishing,
New York, 2010.
Scheng, J., Enhanced Oil Recovery Field Case Studies, Gulf Professional Publishing, New York, 2013.
Schramm, L.L., Surfactants: Fundaments and Applications in the Petroleum Industry, Cambridge University
Press, Cambridge, UK, 2010.
Smith, C., Mechanics of Secondary Oil Recovery, Reinhold Pub. Corp., New York, 1966.
Taber, J.J., SPE 2098, SPE Reprint Ser. No. 24 (Surfactant/Polymer Flooding I), Society of Petroleum
Engineers, Richardson, TX, 1968.
Taber, J.J., Martin, F.D., Seright, R.S., SPE Res. Eng., 12, 189, August 1997.
Tunio, S.Q., Tunio, A.H., Ghirano, N.A., El Adawy, Z.M., Int. J. Appl. Sci. Technol., 143, 5, 2011.
van Poolen, H., Fundamentals of Enhanced Oil Recovery, Pennwell Corp., Tulsa, OK, 1980.
Wilhite, G.P., Waterflooding, Textbook Series, Society of Petroleum Engineers, Richardson, TX, 1986.
13 A Review of Polymer–
Surfactant Interactions
Ali A. Mohsenipour and Rajinder Pal
CONTENTS
13.1 Introduction........................................................................................................................... 639
13.2 Basic Aspects of Polymers and Surfactants..........................................................................640
13.2.1 Polymers....................................................................................................................640
13.2.2 Surfactants.................................................................................................................640
13.3 Interaction of Polymers and Surfactants................................................................................644
13.3.1 Onset of Self-Assembly in Polymer–Surfactant Systems..........................................644
13.3.2 Available Techniques to Study the Polymer–Surfactant Interactions....................... 645
13.3.2.1 Surface Tension Measurements..................................................................646
13.3.2.2 Conductivity................................................................................................ 647
13.3.2.3 Gel Permeation Chromatography...............................................................648
13.3.2.4 Viscosity Measurements............................................................................. 649
13.3.2.5 Electromotive Force.................................................................................... 651
13.3.2.6 Isothermal Titration Calorimetry............................................................... 652
13.3.3 Interactions of Polymer and Surfactant with Opposite Charges............................... 654
13.3.3.1 Interactions of Anionic Polymers with Cationic Surfactants..................... 655
13.3.3.2 Interactions of Cationic Polymers with Anionic Surfactants..................... 659
13.3.4 Interactions of Uncharged Polymers and Charged Surfactants................................. 661
13.3.4.1 Interaction of Nonionic Polymers with Anionic Surfactants...................... 663
13.3.4.2 Interactions of Nonionic Polymer with Cationic Surfactant....................... 670
13.3.5 Interaction of Charged Polymers and Uncharged Surfactants.................................. 671
13.3.6 Interaction of Uncharged Polymers and Uncharged Surfactants............................... 672
13.3.7 Mechanisms of Polymer–Surfactant Associations.................................................... 673
References....................................................................................................................................... 677
13.1 INTRODUCTION
The combination of polymer and surfactant additives has been used in a variety of applications such
as drug delivery, oil recovery, and cosmetics (Bai et al. 2010; Dan et al. 2009; Harada and Kataoka
2006; Stoll et al. 2010; Villetti et al. 2011; Zhang et al. 2011). In general, polymers are introduced to
a surfactant system to control rheology, stability of the system, or to manipulate surface adsorption.
The addition of a polymer may help to remove a surfactant from a surface or to improve its adsorp-
tion at the surface. In many applications, we need a suitable rheology such as thickening of solution
or gelation of solution. The addition of polymer allows the manipulation and control of the solution
rheology. The polymer can also speed up the micellization process resulting in a decrease in the free
surfactant concentration. This property is exploited in skin formulations in which free surfactant
molecules can harm the skin and cause irritation. The combination of polymer and surfactant addi-
tives are also being exploited to intensify frictional drag reduction in the turbulent flow of liquids in
pipelines (Mohsenipour and Pal 2013).
639
640 Handbook of Surface and Colloid Chemistry
13.2.2 Surfactants
Surfactants consist of two parts: a hydrophilic head group and a hydrophobic tail group. The hydro-
philic head group (an ionizable polar group in case of ionic surfactants) can establish hydrogen bonds.
The hydrophobic tail group (a non-polar group) is typically a long chain alkyl group. Surfactants are
A Review of Polymer–Surfactant Interactions 641
known for their capability to lower the surface tension of liquids. When a surfactant is dissolved in
an aqueous solution, the hydrophobic groups are repelled by water while the hydrophilic groups are
attracted to the polar water molecules. This causes the hydrophobic groups to aggregate together
in a hydrocarbon phase in order to prevent contact with water. At the same time, the hydrophilic
polar groups surrounding the hydrocarbon phase are in contact with the water. Such aggregates of
surfactant molecules are called “micelles.” Micelles can have different sizes and shapes that are
highly dependent on the surfactant molecular structure and system conditions. Examples of differ-
ent shapes of the micelles are globular/spherical, disk-like, cylindrical or rod like, bilayer spherical
(vesicle), and hexagonal. The shape of the micelles can change from one form to another when the
solution conditions are changed (Zakin et al. 1998; Zhang 2005; Zhang et al. 2009).
Micelle shapes are determined to a large extent by a packing factor (p) defined as p = v/a 0 lc,
where v is hydrodynamic volume of surfactant molecule, lc is the length of the tail, and a 0 is the head
group cross-section area (see Figure 13.1). When p is equal to 1/3 or less, the surfactant will be cone
shape and the micelles will be spherical. This is the most commonly encountered micelle shape. For
p close to or equal to 1/2, the micelles are of cylindrical shape (rod like). Figure 13.1 shows different
micelle shapes and the corresponding p values.
Micelle shapes and sizes are also influenced by the surfactant concentration. With the increase in
the surfactant concentration, the micelles undergo a change in shape from spherical to thread-like
(Zakin et al. 1998), as shown in Figure 13.2. At high surfactant concentrations, entanglement of
thread-like micelles is often observed.
Figure 13.3 shows a simplified phase diagram of an aqueous surfactant system. The Kraft point
is the temperature at which the surfactant solubility is equal to the critical micelle concentration
(CMC). When the temperature is lower than the Kraft temperature, the surfactant exists in gel or
crystal form in the solution. When the temperature is above the Kraft temperature and the sur-
factant concentration is higher than the CMC, the micelles are formed in the surfactant solution.
The micelles are initially spherical in shape. With the increase in surfactant concentration or upon
a0
v
p=
v a0lc
lc
<1/3
<1/3 – 1/2
½–1
–1
>1
FIGURE 13.1 Different micelle shapes with corresponding packing factor (p) values. (From Zhang, Y.,
Correlations among surfactant drag reduction, additive chemical structures, rheological properties and micro-
structures in water and water/co-solvent systems, PhD thesis, The Ohio State University, Columbus, OH, 2005.)
642 Handbook of Surface and Colloid Chemistry
Hydrophilic head
Wormlike
micelle Increased
concentration
Spherical
micelle
FIGURE 13.2 Spherical and rod-like micelles. (From Rothstein, J.P., Rheol. Rev., 1, 2008.)
Solubility curve
Wormlike
micelles CMC II
Concentration
Spheroidal
micelles
CMC
Monomers
FIGURE 13.3 Simplified phase diagram of aqueous surfactant solution. (From Zhang, Y., Correlations
among surfactant drag reduction, additive chemical structures, rheological properties and microstructures in
water and water/co-solvent systems, PhD thesis, The Ohio State University, Columbus, OH, 2005.)
addition of counterions in the case of ionic surfactant, the micelle shape changes from spherical
to thread-like (Zakin et al. 1998). The concentration at which surfactants form rod-like micelles
is sometimes called CMCII. While the CMC is almost independent of temperature, the CMCII
increases with the increase in temperature.
Surfactants are categorized into four groups: anionic, cationic, nonionic, and zwitterionic. When
the hydrophilic part of a surfactant consists of a negatively charged group like a sulfate, sulfonate, or
carboxylate, the surfactant is called anionic. Most of the regular soaps consist of anionic surfactants.
Anionic surfactants are sensitive to water hardness, which can force them to precipitate. Cationic
surfactants have a positive charge on their hydrophilic part. Quaternary ammonium compounds are
the most common cationic surfactants. The cationic surfactants are mostly used as softening, anti-
static, soil repellent, and anti-bacterial agents. They are also used as corrosion inhibitors. They are
not sensitive to water hardness. These surfactants are expensive compared to anionic surfactants.
Nonionic surfactants consist of an uncharged hydrophilic part. This type of surfactants are not
capable of being ionized in aqueous solution as their hydrophilic group is of non-dissociable type.
Examples of nonionic surfactants are fatty alcohol polyglycosides and alcohol ethoxylates. They are
widely used in cleaning detergents. Zwitterionic or amphoteric surfactants consist of both negative
and positive charges on their molecules. The solution pH controls the charge on the hydrophilic
part of amphoteric surfactants. That is why they act as anionic surfactants in alkalic solution and as
cationic surfactants in acidic solution.
Table 13.1 gives examples (names and chemical formulae) of surfactants in different groups
(anionic, cationic, nonionic, and zwitterionic). The chemical structures of some of the commonly
A Review of Polymer–Surfactant Interactions 643
TABLE 13.1
Surfactant Classification
Class Examples Structures
Anionic Na stearate CH 3 (CH 2 )16 COO − Na +
Na dodecyl sulfate CH 3 (CH 2 )11 SO 4 − Na +
Na dodecyl benzene sulfonate CH 3 (CH 2 )11 C6 H 4SO3 − Na +
Cationic Laurylamine hydrochloride CH 3 (CH 2 )11 NH 3 + Cl −
Trimethyl dodecylammonium chloride C12 H 25 N + (CH 3 )3 Cl −
Cetyl trimethylammonium bromide CH 3 (CH 2 )15 N + (CH 3 )3 Br −
Nonionic Polyoxyethylene alcohol Cn H 2 n+1 (OCH 2 CH 2 )m OH
Alkylphenol ethoxylate C9H19–C6H4–(OCH2CH2)nOH
Polysorbate 80 HO(C2H4O)w (OC2H4)xOH
w +x+ y + z = 20
R=(C17H33)COO
O CH(OC2H4)yOH
CH2(OC2H4)zR
Propylene oxide-modified polymethylsiloxane ( CH3 )3 SiO((CH3 )2 SiO)x (CH3SiO)y Si(CH3 )3
(EO = ethyleneoxy, PO = propyleneoxy)
⏐
CH 2 CH 2 CH 2 O ( EO )m ( PO )n H
Zwitterionic Dodecyl betaine C12 H 25 N (CH 3 )2 CH 2 COO −
+
Source: Schramm, L.L. et al., Ann. Rep. Sect. C (Phys. Chem.), 99, 3, 2003.
O
H3C COO
C12H25 S O– Na+
O Abietic acid
Sodium dodecyl benzene sulfonate
CH3
H3C O CH(CH3)2
P C14H29
H3C O
O
Dimethyl ether of
tetradecyl phosphonic C8H17 O CH2 CH2 O H
n
Polyethoxylated octyl phenol
C11H29 C N CH2 CH2
OH
Lauryl mono-ethanol
O CH2OH
O
CH2 OOC R’ RC O OH
CH OH
HO OH
CH2
Glycerol diester (diglyceride) Sorbitan monoester
C12H25
Cl–
NH N+
C12H25
CH2 CH2 COOH
Dodecyl betaine N–dodecyl piridinium chloride
FIGURE 13.4 A few commonly used surfactants. (From Salager, J.-L., Surfactant’s types and uses,
In: Salager, J.L., ed., Firep Booket-E300-Attaching Aid in Surfactant Science and Engineering in English,
University of Andes, Merida, Venezuela, 2002, p. 3.)
644 Handbook of Surface and Colloid Chemistry
used surfactants are shown in Figure 13.4. Surfactants are used widely in different industries such as
food, paint, and petroleum. A rapid growth in the usage of surfactants is anticipated in the oil field in
the years to come. A large potential market exists for surfactants in enhanced oil recovery, but many
technical and economic problems still remain that need to be resolved before the surfactants can be
used at a large scale (Ahmadi and Shadizadeh 2013; Bachmann et al. 2014; Nelson 1982; Schramm
2000). Another important application of surfactants involves frictional drag reduction in pipeline
turbulent flows. Surfactants like polymers are known to cause a significant reduction in friction in
the pipeline turbulent flow of liquids. Drag reduction by surfactants did not receive any attention
until the work of Mysels (1949). More recent work in this area can be found in Qi and Zakin (2002),
Zakin et al. (2006, 2007), Zakin and Lui (1983), and Zhang et al. (2009). Although a good amount
of research work has been carried out on surfactants as drag reducers, they have been rarely used at
an industrial scale. This area needs more attention in the future work.
Further information about surfactants and their applications can be found in the works of Tadros
(2006), Salager (2002), and Schramm et al. (2003).
TABLE 13.2
Possible Combinations of Polymers and Surfactants for Studying Interactions
Surfactant Nonionic Polymer Anionic Polymer Cationic Polymer
Nonionic P0S0 P−S0 P+S0
Anionic P0S− P−S− P+S−
Cationic P0S+ P−S+ P+S+
A Review of Polymer–Surfactant Interactions 645
TABLE 13.3
CMC and CAC Values for Different Polymer/Surfactant Systems
Surfactant Polymer T (°C) CMC or CAC (mM) CAC/CMC C2 (mM)
SDS 25 8.0
SDS 0.1 wt% PEO 25 4.4 0.55
SDS 0.1 wt% PVP 25 2.1 0.26
CsDS 30 6.2
CsDS 0.1 wt% PEO 30 4.2 0.68 9.6
CsDS 0.1 wt% PVP 30 4.1 0.66 8.4
TMADS 25 5.4
TMADS 0.1 wt% PEO 25 4.6 0.85 8.0
TMADS 1 wt% PEO 25 4.7 0.87
TMADS 0.1 wt% PVP 25 4.6 0.85 8.6
TEADS 25 3.7
TEADS 0.1 wt% PEO 25 3.7 1.0
TEADS 40 3.8
TEADS 0.1 wt% PEO 40 3.8 1.0
TP ADS 25 2.2
TP ADS 1 wt% PEO 25 2.2 1.0
TP ADS 0.1 wt% PVP 25 2.25 1.01
TBADS 25 1.15
TBADS 0.1 wt% PEO 25 1.15 1.0
TBADS 1 wt% PEO 25 1.15 1.0
concentration (CAC).” In many studies published on the interactions of surfactants and polymers,
it is shown that CAC is lower than the CMC (critical micelle concentration) of the surfactant solu-
tion alone (Deo et al. 2007; Goddard and Ananthapadmanabhan 1993; Jönsson et al. 1998). In the
case of electrostatic interaction between ionic polymer with oppositely charged ionic surfactant,
CAC is generally found to be much lower than the CMC of the surfactant. The CAC is close to
CMC in the case of nonionic polymer and ionic surfactant. In addition to CAC, there is another
critical surfactant concentration, referred to as “polymer saturation point (PSP),” which is relevant
in polymer/surfactant systems. While CAC represents the onset of interaction, the polymer satu-
ration point (PSP) reflects the surfactant concentration where the polymer chains become satu-
rated with bound surfactant molecules or micelles (Diamant and Andelman 1999; Goddard and
Ananthapadmanabhan 1993; Hansson and Lindman 1996).
Table 13.3 gives the CMC, CAC, and PSP values for a number of polymer/surfactant systems.
The variations of CAC with temperature and polymer concentration are also indicated. The ratio
CAC/CMC depends on the type of polymer/surfactant combination. For example, the ratio CAC/
CMC is 0.26 for SDS/0.1% of PVP solution and 0.55 for SDS/0.1% PEO solution. This shows that
even if the type of surfactant is the same, the CAC can be different for different polymers. Thus, the
structure of polymer has a strong effect on the CAC.
electromotive force (EMF), and isothermal titration calorimetry (ITC) are also used. In this section,
a brief discussion of these techniques is presented.
c
Pure surfactant
Surface tension
CMC
b
0 a
Surfactant concentration
(a)
c
Polymer/surfactant mixture
Surface tension
CAC PSP
b
0 Surfactant concentration a
(b)
FIGURE 13.5 Schematic plots of surface tension for (a) pure surfactant and (b) mixed surfactant–polymer
system. (From Mohsenipour, A.A., turbulent drag reduction by polymers, surfactants and their mixtures in
pipeline flow, PhD thesis, University of Waterloo, Waterloo, Ontario, Canada, 2011.)
A Review of Polymer–Surfactant Interactions 647
PVP-concentration
0 g/L
1 g/L
60 3 g/L
10 g/L
γ (dyn/cm)
50
T1
40 T2
FIGURE 13.6 Surface tension of aqueous solutions of anionic surfactant SDS/PVP versus total SDS concen-
tration. (From Lange, H., Kolloid Z. Z. Polym., 243, 101, 1971.)
smaller than the CMC. At PSP (shown as T2), the aggregation of polymer/surfactant molecules
stops and free surfactant micelles begin to form.
13.3.2.2 Conductivity
The interaction between polymer and surfactant could be investigated by the conductivity mea-
surements. The electrical conductivity measurements are usually used to detect any changes in the
solution behavior when an ionic surfactant is added to the aqueous solution. If there occurs any
interaction, the solution conductivity is expected to change. Figure 13.7 shows the typical conductiv-
ity plots for pure ionic surfactant solution and a mixed polymer/ionic surfactant system.
Curve A shows the trend for pure ionic surfactant solution whereas curve B demonstrates the con-
ductivity trend for a mixture of polymer/ionic surfactant. For pure surfactant solution (curve A), the
conductivity is a linear function of the surfactant concentration below the CMC. The ionic s urfactant
is completely dissociated below the CMC. Above the CMC, micelles are formed. The micellar
contribution to conductivity is less than that of the same number of free surfactant molecules.
2,000
1,600
Conductivity (µS/cm)
Curve B
1,200
Curve A
800
400
FIGURE 13.7 Schematic conductivity plots for pure surfactant and mixture of surfactant and polymer.
648 Handbook of Surface and Colloid Chemistry
0.65
κ
(S/m)
0.55
0.45
0.35
0.06 0.08 0.10 0.12 0.14 0.16
SDEC (m)
FIGURE 13.8 Specific conductivity of PEO–sodium decyl carboxylate (SDEC) solutions as a function
of SDEC concentration at a fixed PEO concentration of 5000 ppm. (From Blokhus, A.M. and Klokk, K.,
J. Colloid Interface Sci., 230(2), 448, 2000.)
Consequently the slope of the conductivity plot above CMC is lower than that below CMC although
the conductivity continues to increase with the increase in the surfactant concentration. For the
mixed surfactant and polymer systems, the conductivity plot shows two breakpoints. The first break
point is known to occur at CAC, the concentration of surfactant where surfactant monomers begin
to associate with the polymer chains. The second break point occurs at PSP (polymer saturation
point) where the polymer molecules are saturated with surfactant. Above the PSP, the addition of
surfactant results in the formation of free micelles. For systems where the interaction between the
surfactant and polymer molecules is weak, the CAC and PSP points are not easily detectable based
on the conductivity measurements (Mohsenipour 2011).
Figure 13.8 shows the conductivity data for a system consisting of PEO–sodium decyl carboxyl-
ates. Two breaks observed in the conductivity curves are attributed to CAC and PSP.
P(44.4)
P + S(37.6)
S(42.5)
P + S(39.2)
S(42.5)
0 10 20 30 40 50 60
Retention time (min)
FIGURE 13.9 GPC analysis of SDS/PEO system. The PEO sample was dissolved in the mobile phase; flow
rate, 0.5 mL/min; RI detector; injection volume, 50 mL: (a) 1000 ppm PEO (mobile phase—distilled water);
(b) 750 ppm PEO (mobile phase—0.017 M SDS); and (c) 750 ppm PEO (mobile phase—0.017 M SDS dis-
solved in 0.1 wt% NaCl solution). (From Veggeland, K. and Austad, T., Colloid Surf. A: Physicochem. Eng.
Aspects, 76, 73, 1993.)
(MW = 20,000 g/mol) in pure water using a column of Ultrahydrogel™ 2,000 (13 mm). Figure
13.9b shows the chromatogram of 750 ppm PEO dissolved in 0.017 M SDS. Figure 13.9c gives
the chromatogram of 750 ppm PEO dissolved in 0.017 M SDS with 0.1 wt% NaCl. These figures
show that although the flow rates are the same, the polymer exits the instrument at different times,
that is, the retention time of the peaks is altered. The retention times were 44.4, 37.6, and 39.2 min
in distilled water, in the presence of surfactant, and in the presence of both surfactant and salt,
respectively.
When SDS is added to PEO, the effective size of the PEO chains increases due to electrostatic
repulsion between the negatively charged micellar aggregates bound to the polymer chain. As the
size of species changes, the retention time of the species also changes (37.6 min in Figure 13.9a
compared to 44.4 min in Figure 13.9b). The addition of salt can cause suppression of electro-
static repulsion, and consequently the size of polymer–surfactant complex and its retention time
are affected (see Figure 13.9c). The negative or vacant peaks S in the chromatograms shown in
Figure 13.9b and c are related to the decrease in the surfactant concentration due to the formation
of the polymer–surfactant complexes.
tp
ηr = (13.1)
tw
650 Handbook of Surface and Colloid Chemistry
8
PEO and SDS
7
Relative viscosity
5
2
CAC
1
500 ppm PEO
0
0 5,000 10,000 15,000
SDS (ppm)
FIGURE 13.10 Relative viscosity of PEO solution as a function of SDS concentration. (From Ghoreishi, S.
et al., Langmuir, 15(13), 4380, 1999.)
tp − tw
ηs = (13.2)
tw
To measure the flow time of a solution, the Ubbelohde-type capillary viscometer is used. The flow
time of solution is compared with the flow time of base fluid (water). As an example, Figure 13.10
shows the relative viscosity for the PEO/SDS system. When surfactant is added to polymer solu-
tion, the hydrodynamic size of polymer is increased due to the attachment surfactant micelles to the
backbone of PEO chains. This causes an increase in relative viscosity. The relative viscosity begins
to rise at CAC as shown in Figure 13.10.When the surfactant concentration reaches PSP, the relative
viscosity begins to flatten.
The Ubbelohde-type capillary viscometer gives viscosity at a low shear rate. To obtain the full
viscosity–shear rate profile, a coaxial cylinder viscometer can be employed. The shear rate in a
coaxial cylinder viscometer is calculated using the following equations:
2S 2
Newtonian fluids: γ Ri = Ω0 (13.3)
S2 −1
2N
Non-Newtonian fluids: γ Ri = Ω0 (13.4)
1 − S −2 N
2π(rpm )
Ω= (13.5)
60
where
S in Equation 13.3 is the ratio of rotor radius to bob radius
N is the slope of ln Ω versus ln(torque) data
A Review of Polymer–Surfactant Interactions 651
0.01
500 ppm PEO Pure PEO
100 ppm SDS
600 ppm SDS
1400 ppm SDS
2000 ppm SDS
Apparant viscosity (Pa s) 2500 ppm SDS
3000 ppm SDS
5000 ppm SDS
0.001
100 1000
Shear rate (L/s)
FIGURE 13.11 Apparent viscosity versus shear rate of PEO/SDS mixtures. (From Mohsenipour, A.A.,
turbulent drag reduction by polymers, surfactants and their mixtures in pipeline flow, PhD thesis, University
of Waterloo, Waterloo, Ontario, Canada, 2011.)
The apparent viscosity at any shear rate is given as the ratio of shear stress to shear rate.
Figure 13.11 shows the apparent viscosity versus shear rate plots obtained by a coaxial cylin-
der viscometer for mixtures of PEO/SDS. The addition of SDS to pure PEO solution causes an
increase in apparent viscosity. The solution shows a Newtonian behavior at low SDS concentrations.
The solution becomes pseudo plastic shear-thinning at high SDS concentration.
⎛ 2.303RT ⎞
E = E0 − ⎜ ⎟ log(m1 ) (13.6)
⎝ F ⎠
where m1 is the concentration of free monomer surfactant. The slope (−2.303RT/F) and the
intercept E 0 are determined from the linear data in the initial region of the EMF curve, which
652 Handbook of Surface and Colloid Chemistry
40
20
0
–20 T1
FIGURE 13.12 EMF plot of the SDS electrode (reference Br−) as a function of the total SDS concentration
for the SDS/PVI system in 10 −4 mol dm−3 NaBr: diamond symbol—pure SDS; square symbol—SDS + PVI
(0.1% w/v). (From Ghoreishi, S. et al., Langmuir, 15(13), 4380, 1999.)
mostly overlaps with the data for pure surfactant. Note that R is the universal gas constant, T the
absolute temperature, and F the Faraday constant.
6.2
6.0
5.8
µ (cal/s)
5.6
Syringe 5.4
5.2
0 5 10 15 20 25 30 35 40 45 50
Time (min)
Output
Adiabatic shield
∆T
FIGURE 13.13 Schematic diagram of a power compensation ITC. (From Lewis, E.A. and Murphy, K.P.,
Isothermal Titration Calorimetry. Protein-Ligand Interactions, Springer, 2005, pp. 1–15.)
4.5
2
4
1
3.5
ΔH (kj/mol of injection)
CAC 0
3
–1
2.5 0 10 20 30
2
1.5
1 Water
0.1 wt% PEG-20
0.5 0.1 wt% PEG-32
0
0 5 10 15 20 25 30 35
SDS (mM)
FIGURE 13.14 The ITC curve of 0.2 M SDS titrating into 0.1 wt% PEG-20 ◻) and 0.1 wt% PEG-32 (Δ)
at 298 K. The open circle is the SDS dilution curve in water. The insert is a plot for the difference curves of
titrating 0.2 M SDS into PEGs and water. (From Dai, S. and Tam, K., J. Phys. Chem. B, 105(44), 10759, 2001.)
Above this point the deviation in the titration curves is observed. This point is CAC. The SDS con-
centration where the titration curves (pure SDS and SDS/PEG curves) meet each other again is the
PSP point. In the present SDS/PEG system, the endothermic reaction is attributed to the dehydration
of PEG segments from water phase and the association of these segments with the hydrophobic core
of the SDS micelles.
654 Handbook of Surface and Colloid Chemistry
100
CMC
10
CMC and CAC (mM)
1
CAC
0.1
0.01
10 12 14 16
No. of carbon atoms in the alkyl chain, nc
FIGURE 13.15 Relation between CMC and CAC for a system consisting of an ionic surfactant and an oppo-
sitely charged polymer. (From Jönsson, B. et al., Surfactant-Polymer Systems in Surfactant and Polymers in
Aqueous Solution, John Wiley & Sons, New York, 2003, Chapter 13; Thalberg, K. and Lindman, B., J. Phys.
Chem., 93(4), 1478, 1989.)
A Review of Polymer–Surfactant Interactions 655
and oppositely charged surfactants, the driving force for the interaction includes both electrostatic
and hydrophobic forces. However, the structure of the polymer, especially the distribution of the
hydrophobic modifier on the backbone of the polymer chain, has a strong influence on the polymer–
surfactant interaction. With microblocky distribution, the interaction is highly dependent on the
concentration of surfactant and in this case intermolecular association is improved. In the case of
polymers with random distributions, intramolecular associative behavior is increased (Winnik and
Regismond 1996). Viscosity measurement is often used to detect the conformational behavior of
polymer chains in a surfactant solution. The interaction between oppositely charged polymer and
surfactant can have a strong effect on the viscous behavior of polymer/surfactant solutions.
The oppositely charged polymers and surfactants may form a single phase without complex
formation, or may form a soluble surfactant–polymer complex, or may result in phase separation
(Piculell and Lindman 1992; Wang et al. 1999, 2000). If the interaction leads to liquid–liquid phase
separation, the process is called “coacervation.” If the interaction leads to a liquid–solid phase sepa-
ration, then the process is named “precipitation.” Many factors affect the interaction of oppositely
charged surfactants and polymers that consequently have an effect on the possible phase separation
phenomena. Extensive studies have been done on both theoretical and practical aspects of phase
separation. It is important to know the conditions under which coacervation occurs in applications
dealing with the formulation of cosmetics and pharmaceuticals.
The factors that affect the interaction of oppositely charged surfactants and polymers are the
surfactant type, micelles charge density, ionic strength, polymer/surfactant ratio, polymer molecu-
lar weight (MW), linear charge density, and temperature. However, the electrostatic factors, such
as macromolecular charge densities and ionic strength, are the most significant (Wang et al. 1999).
The subject of interactions between oppositely charged polymers and surfactants can be divided
into two categories. One category deals with anionic polymer and cationic surfactant and the other
deals with cationic polymer and anionic surfactant. The interactions are similar in principle in both
the categories.
–2.5
NaCMC
NaPA
–3 NaDxS
NaPSS
NaPSS
–3.5 NaPVS
log(CAC/M) –4
–4.5
–5
–5.5
–4 –3 –2 –1 0 1
log([Na+]/M)
FIGURE 13.16 Effect of salt on the CAC for solutions of C12TAB and different polyelectrolytes. CP =
0.5 mM. Added salt: NaBr (polymers NaCMC, NaPA, NaPVS), NaCl (polymers NaDxS, NaPSS). (From
Hansson, P. and Almgren, M., J. Phys. Chem., 100(21), 9038, 1996.)
The effect of linear charge density (ξ) on the interaction has been a part of many studies. These
studies have shown that with the decrease in ξ, there is a less effective screening of the charges on
the surfactant molecules attached to the backbone of polymer. This results in a larger repulsion and
consequently a smaller aggregation number. From a thermodynamic point of view, less work is
needed to move the polyelectrolyte from the micelle when the charge density is small. A decrease in
the aggregation number with the decrease in ξ is consistent with an increase in CAC.
Almost the same results were obtained by Wallin and Linse (1996a,b) using Monte Carlo simu-
lations to investigate the complexation of a charged micelle with an oppositely charged polymer.
They found that the critical micelle concentration (CMC) is greatly reduced in the presence of
polyelectrolyte. The most reduction in CMC occurs for the highest linear charge density along with
the most flexible polyelectrolyte (110 times smaller as compared with pure surfactant CMC). In the
case of lower linear charge density and the more rigid polymer structure, the reduction in CMC was
about 10 times. The increase in the linear charge density leads to the stabilization of the system by
a decrease in the electrostatic energy and an increase in the entropy due to an increase in the coun-
terions of the macro-ions.
A Review of Polymer–Surfactant Interactions 657
120
100
80
Ns
60
40
FIGURE 13.17 Relation between aggregation number and CAC for solutions of C12TAB and different poly-
electrolytes. (From Hansson, P. and Almgren, M., J. Phys. Chem., 100(21), 9038, 1996.)
Hansson and Almgren (1996) concluded that for almost all of the systems that they investigated,
there existed a relation between the aggregation number (NS) and CAC. They reported that for a
fixed salt concentration, a large NS generally corresponded to a low CAC (see Figure 13.17).
Prajapati (2009) investigated the interaction of CPAM (copolymer of polyacrylamide and sodium
polyacrylate) and OTAC-p (octadecyltrimethylammonium chloride) and observed that strong inter-
actions are present between anionic polymer CPAM and cationic surfactant OTAC-p based on the
conductivity, viscosity, surface tension, and turbidly measurements. When CPAM is introduced to
the surfactant solution, the CMC value is increased due to transfer of free OTAC-p molecules to
CPAM chains. The cationic OTAC-p molecules neutralize the charge on the backbone of CPAM
chains. At higher CPAM concentrations, parts of the CPAM chains become inaccessible to OTAC-p
molecules. The results from the conductivity measurements showed that the CMC value increases
from 5700 ppm for pure OTAC-p to 6800 and 7100 ppm in the presence of 500 and 1000 ppm
CPAM, respectively (see Figure 13.18).
The viscosity change was also monitored with the increase in surfactant (OTAC-p) concentration.
A sharp reduction in viscosity was observed reflecting a strong interaction between the surfactant
and the polymer. The electrostatic repulsive forces that cause the extension of polymer chains are
neutralized. This leads to the shrinkage of polymer chains resulting in a large reduction in the
6,000
5,000
Conductivity (μS/cm)
4,000
3,000
18
16 500 ppm PAM
250 ppm PAM
14 100 ppm PAM
Relative viscosity
12
10
8
6
4
2
0
0 200 400 600 800 1000 1200 1400 1600 1800
OTAC (ppm)
FIGURE 13.19 Relative viscosity for different concentration solutions of CPAM as a function of OTAC
concentration. (From Mohsenipour, A.A. et al., Can. J. Chem. Eng., 91(1), 181, 2013.)
1,000
0
0 2,000 4,000 6,000 8,000 10,000
OTAC-p concentration (ppm)
relative viscosity. Similar results were reported by Mohsenipour et al. (2013) (see Figure 13.19).
The interaction resulted in the insolubility and precipitation of polymer chains. Precipitation was
reflected in the turbidity data (see Figure 13.20). The turbidity data of the solutions showed that with
the addition of surfactant, the phase separation takes place. The phase separation point is very close
to the maximum turbidity shown in Figure 13.20.
Zhang et al. (2013) studied the interaction of long-chain imidazolium ionic liquid (C14mimBr)
and polystyrene sulfonate (NaPSS), which is an anionic polyelectrolyte. They used surface tension,
isothermal titration microcalorimetry (ITC), dynamic light scatting (DLS), and conductance meth-
ods to study the interactions. The surface tension exhibited a complicated behavior with the addition
of surfactant to the polymer solution when the surfactant concentration was below the CMC value.
The formation of the surfactant/polymer complexes was affected by the concentrations of surfactant
and NaPSS. They explained the influence of surfactant on surfactant–polymer interactions mecha-
nistically as shown in Figure 13.21. When cationic surfactant molecules are added to the solution of
an anionic polymer, the head groups of surfactant (positive) molecules attach to the negative sites on
the backbone of the polymer. This surfactant–polymer complex moves to the surface as more sur-
factant monomer becomes attached to the backbone of the polymer and the complex becomes more
A Review of Polymer–Surfactant Interactions 659
FIGURE 13.21 Precipitation of anionic polymer upon interaction with a cationic surfactant. (From Zhang, Q.
et al., Appl. Surf. Sci., 279, 353, 2013.)
hydrophobic. At the CAC point, micelles start to form on the polymer chains. With further increase
in the surfactant concentration, the surfactant–polymer complexes present at the surface move to
the bulk solution. At some surfactant concentration (indicated as T2), the polymer chains become
saturated with surfactant micelles. This point is the PSP (polymer saturation point). With further
addition of surfactant, the surfactant molecules go the surface. When the CMC concentration of
surfactant is reached, the surfactant molecules begin to form polymer-free micelles in the solution.
Surface
solution
Surface tension
With polymer
Without polymer
Precipitation
zone
Concentration surfactant
FIGURE 13.22 Conditions in the bulk solution and at the surface of the solution of polycation (fixed con-
centration) and anionic surfactant. Solid line: hypothetical surface tension curve of surfactant alone; dashed
line: surface tension curve for the mixture of surfactant and polycation. (From Goddard, E., Colloid Surf.,
19(2), 301, 1986.)
As mentioned earlier, when the polymer is a polyion with charges opposite to that of a s urfactant
a strong interaction is expected. The starting sign of such interaction is the formation of precipitates.
Figure 13.23 shows a simple qualitative phase diagram reflecting the interaction of cationic cellulosic
polymer JR* (Union Carbide, Danbury, CT) with triethanolamine lauryl sulfate (TEALS). Goddard
(1994) recorded the appearance of the system and developed the diagram shown in Figure 13.23. A
clear system is accessible at low and high concentrations of surfactant. This was also observed in
the works of Prajapati (2009) and Mohsenipour et al. (2013).
Based on the research work that has been carried out on the interaction of polyelectrolytes and
ionic surfactants, the CAC may be looked upon as a solubility parameter. CAC is the solubility
threshold when the two components (polymer and surfactant) are mixed. At the CAC, the solution
starts becoming turbid and the surfactant–polymer complex begins to precipitate. This process
takes place in all the systems containing ionic polymer and an oppositely charged surfactant and is
called colloid titration (Goddard and Ananthapadmanabhan 1993; La Mesa 2005). In colloid titra-
tion, the CAC can be expressed as:
where the first term indicates the polymer content and the second term surfactant concentration.
The exponent n is the number of ions need to reach the stoichiometry. It has been shown that
precipitation usually happens at a moderate concentration of surfactant and n is close to unity
(La Mesa 2005).
Mukherjee et al. (2011) studied the interaction of cationic polymer poly(diallyldimethylammonium
chloride) (PDADMAC) with the anionic surfactants sodium dodecyl sulfate, sodium dodecylben-
zenesulfonate, and sodium N-dodecanoylsarcosinate using tensiometry, turbidimetry, calorimetry,
viscometry, dynamic light scattering (DLS), and scanning electron microscopy (SEM). They found
that the morphologies of the surfactant–polymer complexes have different patterns depending on
A Review of Polymer–Surfactant Interactions 661
10
C C SP SP SP P SP SP SP C C
1.0
Polymer JR 400 (wt%)
C C SP SP SP P P SP C C
C C SP SP P P T C C
C C SP SP P P T CC
0.1
C = Clear
T = Turbid
P = Precipitate
SP = Slight precipitate
0.1 1.0 10
Teals (wt%)
FIGURE 13.23 Solubility diagram of mixed polymer (JR 400) and surfactant (triethanolamine lauryl
sulfate) system. (From Goddard, E., J. Am. Oil Chem. Soc., 71(1), 1, 1994.)
the composition and the solvent environment. The type of the surfactant head groups also had a
significant influence on the interaction process.
Liu et al. (2014) studied the effects of three sulfonate gemini surfactants with different hydrophobic
chain lengths (8, 10, and 12 carbon atoms) on the optical properties of a fluorene-based conjugated
cationic polymer poly{[9,9-bis(6′-N,N,N-trimethylammonium)hexyl]-fluorene-phenylene}bromide
(PFP), which was dissolved in either dimethyl sulfoxide (DMSO)/water solutions (4% v/v), or pure
water. They observed a decrease in the photoluminescence (PL) intensity when surfactant and PFP
were dissolved in DMSO/water solutions. They also reported a red shift of emission maxima at low
surfactant concentrations. With the addition of more surfactant the PL intensity was enhanced. A
mechanistic model was presented (see Figure 13.24) to explain their observations. In un-aggregated
polymer chains, the surfactant molecules get attached to the backbone of the polymer chains
and neutralize the charges present on the polymer chains. The aggregates of polymer chains are
formed with surfactant attached to each chain, as shown schematically in Figure 13.24a. For aggre-
gated polymer chains, the surfactant molecules get attached to each chain of the aggregate (see
Figure 13.24b). With the addition of more surfactant to the system, the aggregates are separated
resulting in charge-neutralized complexes of surfactant and single polymer chains. It would be
interesting to explore the surface tension behavior of such systems. Surface tension measurements
could bring in more information on the interaction of PFP/surfactants.
FIGURE 13.24 A mechanistic model describing the interaction of an anionic surfactant with a cationic
polymer: (a) 4% DMSO/water solution and (b) aqueous solution. (From Liu, X.-G. et al., Langmuir, 30(11),
3001, 2014.)
was varied. He reported two critical concentrations: CAC (critical aggregation concentration) and
PSP (polymer saturation point). His work has had a major impact in the area of polymer/surfactant
interactions. Jones formalized the important concepts and defined CAC and PSP when surfactant is
added to the polymer solution (Jones 1967). A good review on some of the early work on the interac-
tion of nonionic polymer and ionic surfactant is given by Goddard (1986a,b). A considerable effort
has been made in understanding the interactions between nonionic polymer and different types of
surfactants (Fishman and Elrich 1975; Jiang and Han 2000; Ma and Li 1989). Some of the recent
studies are discussed here.
Figure 13.25 shows schematically different regions of interaction between a nonionic polymer
and an ionic surfactant. At very low concentrations of surfactant (region I) no significant interaction
CAC Saturation
Free micelle
formation
Bound surfactant
Concentration of polymer
I II III IV
No binding
Free micelles
excess surfactant
Concentration of surfactant
FIGURE 13.25 Different regions of association between nonionic polymer and ionic surfactant. (From
Jönsson, B. et al., Surfactant-Polymer Systems in Surfactant and Polymers in Aqueous Solution, John Wiley &
Sons, New York, 2003, Chapter 13.)
A Review of Polymer–Surfactant Interactions 663
takes place between the polymer and the surfactant. In region II with surfactant concentration ≥ CAC
(critical aggregation concentration), the surfactant molecules begin to interact with the polymer
molecules resulting in the formation of polymer–surfactant complexes. At a certain concentration
of surfactant (corresponding to the end of region II), the polymer molecules become saturated with
the surfactant. In region III, the surfactant molecules tend to aggregate and form micelles on the
backbone of the polymer molecules. At high surfactant concentrations corresponding to region IV,
free micelles (not associated with polymer) are formed.
The small-angle neutron-scattering experiments have proven that the surfactant molecules attach
to the polymer chains in the form of micelles. The surfactant micelles on polymer chains are similar
to the micelles in the polymer-free surfactant solution. The only difference is the aggregation num-
ber, which is less for a surfactant–polymer system compared to the polymer-free solution. However,
the radius of gyration of the polymer molecule in a surfactant–polymer system is comparable to that
of a “free” macromolecule (Ruckenstein et al. 1987) in the absence of any surfactant.
As the interaction of nonionic polymer and ionic surfactant is controlled by the hydrophobic forces
between these two components, any factor that increases the hydrophobicity will have a strong
effect on the interaction. Polymers and surfactants with long chains and tails are more susceptible
to interaction than those with shorter chains or tails. In the case of a polymer, longer chains can
promote the hydrophobicity of the polymer. The hydrophobicity of polymer molecules could be
increased by the addition of a small number of long alkyl chains to the backbone of the polymer
molecules. Such polymers are called “hydrophobically modified (HM)” polymers.
Breuer and Robb (1972) studied the effect of polymer hydrophobicity on the interaction of
polymer with surfactant. SDS was used as the anionic surfactant along with several different poly-
mers: polyvinyl acetate (PVAc), polyvinyl alcohol (PVA), polypropylene oxide (PPO), PEO, and
methylcellulose (MeC).
They found that the interaction of polymer with SDS was in the following order of strength:
Thus, there occurs an increase in the level of interaction between polymer and surfactant with the
increase in hydrophobicity.
The length of the hydrocarbon chain can affect the hydrophobicity of the polymer. As the
length increases the hydrophobicity increases. The chain length can also affect some other f actors
leading to a decrease in the interaction between polymer and surfactant. For example, poly-
saccharide Dextran has a long and flexible chain but has a little tendency to interact with SDS
(Goddard 1986a,b).
664 Handbook of Surface and Colloid Chemistry
0
0 5 10 15 20 25
Total SDS concentration in cell (mmol/kg)
FIGURE 13.26 Surfactant binding isotherms for SDS with nonionic polymers: at 20°C for (—) CST-103;
(⋯⋯) HPMC; (— · —) MC; (— —) HPC, and (— · · —) HEC. (From Singh, S.K. and Nilsson, S., J. Colloid
Interface Sci., 213, 152, 1999.)
Figure 13.26 shows another example of the effect of hydrophobicity on the interaction of poly-
mers and surfactants (Singh and Nilsson 1999). The figure shows the binding curves for the cellulose
ethers with SDS at 20°C. The polymers investigated are ethyl hydroxyethyl cellulose (EHEC), also
called CST-103, hydroxypropyl methyl cellulose (HPMC), hydroxypropyl cellulose (HPC), methyl
cellulose (MC), and hydroxyethyl cellulose (HEC). Among them, CST-103 is the most hydrophobic
and HEC is the most hydrophilic. The binding curves show a clear trend in the critical aggrega-
tion concentration. The CAC decreases with the increase in hydrophobicity. The most hydrophobic
polymer (CST-103) has the lowest CAC value. The least hydrophobic polymer (HEC) shows the
highest CAC value.
Hydrophobically modified (HM) nonionic polymers exhibit a strong tendency to interact with
ionic surfactant. These polymers show a different behavior compared to regular nonionic polymers
when they interact with ionic surfactants. The rheology of these fluids has been investigated in great
detail. In most of the studies, it is observed that the addition of surfactant causes an increase in
viscosity up to a certain concentration and then a reduction in viscosity with further addition of sur-
factant. Figure 13.27 shows a simple mechanistic model proposed by Winnik and Regismond (1996)
for the interaction of ionic surfactant such as SDS, TTAC, and CTAC with a hydrophobically modi-
fied polymer such as poly(N-isopropylacrylamides) (HM-PNIPAM), a copolymer of NIPAM, and
N-n-alkylacrylamides (n-alkyl = decyl, tetradecyl, and octadecyl groups). The increase in viscosity
is attributed to the formation of complex on the alkyl substituents, which promotes interlinking and
networking of the polymeric chains. At a low surfactant concentration (C < CMC), the surfactant
micelles act as a bridge between different chains of polymer. With the increase in surfactant concen-
tration, the amount of polymer chains available to interact with the surfactant molecules is reduced.
In this case, the density of micelles on the backbone of polymer chains is increased to a point where
the repulsion of polymer chains becomes significant. Consequently, the network of the polymer
chains is destroyed resulting in a reduction of viscosity. Experiments with Py-labeled (pyridine)
PNIPAM and Py-labeled HM-PNIPAM samples confirmed this conclusion.
In another study, Dualeh and Steiner (1990) reported a strong interaction between C12-grafted
hydroxyethylcellulose (HM-HEC) in the presence of SDS. They found that this hydrophobically
modified polymer acts differently from the unmodified HEC. The hydrophobic characteristics of
HM-HEC was responsible for the strong interaction (Winnik and Regismond 1996).
A Review of Polymer–Surfactant Interactions 665
Polymeric Mixed
micelle cluster
Surfactant
addition
PNIPAM-C18/200 Surfactant
micelle
PNIPAM-C18Py/200
PNIPAM-Py/200
Key
NIPAM unit
Octadecyl chain
Pyrene
Surfactant
FIGURE 13.27 The interactions between surfactant and HM-PNIPAM, PNIPAM-C18/Py, and PNIPAM-P.
(From Winnik, F.M. and Regismond, S.T., Colloid Surf. A: Physicochem. Eng. Aspects, 118(1), 1, 1996.)
The influence of the size of the surfactant head group on the interaction between polymer and
surfactant has been investigated in several studies. Blokhus and Klokk (2000) studied the interac-
tions between PEO and the surfactants sodium alkylcarboxylates (octyl, decyl, and dodecyl) and
SDS by means of conductivity measurements and gel permeation chromatography (GPC). From the
conductivity measurements, the critical aggregation concentration (CAC), the ionization degree,
and the binding ratios were determined. The binding ratio was also determined from GPC. The
PEO–surfactant interactions were observed in most of the surfactants studied (except sodium
octanoate). For the polymer–surfactant complexes, the ionization degree was observed to be about
0.2 higher than the ionization degree for the corresponding aqueous micelles. Further, the binding
ratio decreased somewhat with decreasing the chain length of the alkylcarboxylate. The Gibbs free
energy analysis showed that the polymer–surfactant interaction decreases with decreasing the chain
length of the alkylcarboxylates and is weaker for alkylcarboxylate compared with alkylsulfate of
similar chain length (Blokhus and Klokk 2000). Brackman and Engberts (1992) also found out that
sodium alkylphosphates interact less with PEO than sodium alkylsulfates of similar chain length.
666 Handbook of Surface and Colloid Chemistry
Benrraou et al. (2003) studied the interaction between nonionic polymers and anionic surfac-
tants. The polymers used were poly(ethylene oxide) and poly(vinylpyrrolidone) and the surfactants
used were cesium and tetraalkylammonium (tetramethyl to tetrabutyl ammonium) dodecylsulfates.
They used the electrical conductivity method to determine the CAC for different polymer–sur-
factant combinations. They concluded that the value of the ratio CAC/CMC increases with the
increase in the radius of the counterion. Thus, the strength of interaction decreases upon increas-
ing the counterion radius. The sequence in which the ratio CAC/CMC varied is as follows: Na+ <
Cs+ < tetramethylammonium+ < tetraethylammonium+ = tetrapropylammonium+ = tetrabutylam-
monium+ = 1.0. The ratio CAC/CMC is unity in the case of the last three large-size ions indicating
the absence of any interaction. For small ions such as Na+, the CAC/CMC ratio is less than unity
indicating significant interaction.
Yan and Xiao (2004) carried out an investigation on the effect of anionic surfactant headgroup
on interaction between surfactants (alkyl sulfate and alkyl sulfonate) and nonionic polymer (PEO).
Their study revealed that C12SO3 and C12SO4 interact differently with PEO. They also concluded
that both hydrophobic and electrostatic interactions play important roles in the C12SO3/PEO and
C12SO4/PEO interactions. The headgroup of CnSO4 has an extra oxygen compared to CnSO3, which
makes it bigger. Furthermore, the charge on the CnSO4 headgroup and −CH2 (first group on alkyl
chain of surfactant) are opposite, while the CnSO3 headgroup and −CH2 have the same charge. This
makes the hydration cosphere of the hydrophobic tail of CnSO4 very different from that of the cor-
responding tail of CnSO3. As a result, the hydration effect of CnSO3 is stronger than that of CnSO4,
which means the hydration shell of the headgroup may extend over a larger part of the alkyl chain
in CnSO3. This effect “shortens” the effective hydrophobic tail of CnSO3 as compared with CnSO4.
As hydrophobicity is the most important factor in the interaction between nonionic polymer and
surfactant, the difference in the hydrophobic nature of CnSO3 and CnSO4 makes the interaction of
CnSO3–PEO weaker than that of CnSO4 –PEO.
The effect of polymer molecular weight (MW) on the interaction between nonionic polymer
and ionic surfactant has been studied in several papers. Schwuger (1973) investigated the effect of
PEO molecular weight on the surface tension of SDS solutions. He found that in the case of low
MW (600) PEO, the interaction was weak. The interaction was moderate when the MW of PEO
was 1550. However, for PEO molecular weights higher than 4000, the interaction was strong and
independent of the molecular weight.
Dai and Tam (2001) studied the interaction of a series of poly(ethylene glycol) (PEG) with
anionic surfactant sodium dodecyl sulfate (SDS) by mean of an isothermal titration calorimetry
(ITC). Figure 13.28 shows the thermograms for 0.2 M SDS titrating into 0.1 wt% PEG of different
molecular weights at 298 K. The CAC is not sensitive to molecular weight of PEG; it decreases only
slightly by increasing the molecular weight. This decrease was attributed to a decrease in the hydro-
phobocity of PEG. They also reported that PSP decreases with increasing PEG molecular weight. It
was concluded that the binding between PEG and ionic surfactants (the ratio of EO/SDS) increases
with the increase in the polymer molecular weight.
Some investigators have considered the effect of salt on the interaction between nonionic p olymer
and anionic surfactant. Mandal et al. (2013) investigated the interaction of sodium dodecyl sulfate
(SDS) with aqueous polyvinylpyrrolidone (PVP) under the influence of different salts including
NaCl, NaBr, NaI, KCl, LiCl, NH4Cl, Na2SO4, and Na3PO4. The concentration of SDS was varied
and the interactions were detected using techniques such as tensiometry, viscometry, conductom-
etry, and calorimetry. They found that the surfactant SDS binds to the PVP chains. Reverse micelles
of surfactant are formed on the backbone of the polymer chains. The presence of salt has a signifi-
cant influence on the interaction between the surfactant and the polymer.
Minatti and Zanette (1996) reported that the critical aggregation concentration (CAC) and poly-
mer saturation point (PSP) in PEO/SDS mixtures were affected by the presence of salt (NaCl).
Masuda et al. (2002) studied the swelling behavior of poly(ethylene oxide) (PEO) gels in aqueous
solutions of sodium dodecyl sulfate (SDS). They observed that in the absence of salt, PEO gels start
A Review of Polymer–Surfactant Interactions 667
8
7 Water
0.1 wt% PEG-08
6
0.1 wt% PEG-20
ΔH (kJ/mol of injection)
5 0.1 wt% PEG-32
0.1 wt% PEG-75
4 0.1 wt% PEG-100
3 0.1 wt% PEG-150
2
1
0
–1
–2
0 5 10 15 20 25 30 35
SDS (mM)
FIGURE 13.28 Thermograms for 0.2 M SDS titrating into 0.1 wt% PEG of different molecular weights
at 298 K. The open circle is the SDS dilution curve in water. (From Dai, S. and Tam, K., J. Phys. Chem. B,
105(44), 10759, 2001.)
to swell at a surfactant concentration lower than the CMC of SDS. This concentration was in agree-
ment with CAC value reported for the aqueous PEO solution.
Solutions of sodium dodecyl sulfate (SDS) and a variety of nonionic polymers listed in Table 13.4
were studied by Ghoreishi et al. (1999). The EMF method utilizing an SDS membrane selective
electrode was applied to investigate the interaction between the polymers and SDS.
They evaluated the CMC, CAC(T1), and PSP(T2) from the experimental data. The CAC and
PSP and the related information is given in Table 13.5. Almost all the solutions showed a clear
interaction. To calculate m1 (the free monomer surfactant concentration) at any SDS concentration
in the interaction zone, they used Equation 13.6 where the slope (~−(2.303RT/F)) and intercept E 0
are derived from the linear data in region I, which also overlaps with the data for pure SDS (see
Figure 13.12).
Ghoreishi et al. (1999) reported that at low salt concentrations different polymers show some
selectivity toward SDS. A variation in the maximum amount of SDS that could be bound to the
polymer was observed from one polymer to another. At high salt concentrations, all the polymers
behaved in a similar way.
Mesa and his team (Capalbi and La Mesa 2001; Gasbarrone and La Mesa 2001) studied the inter-
action of PVP and SDS in a dilute solution. They presented two graphs (see Figure 13.29) showing
the phase diagrams of ionic surfactant/nonionic polymer in water systems. The figures show three
subregions in which CAC and CMC play the main role. The CMC acts as a “triple” point, where
free surfactant, surfactant–polymer complex, and free micelles coexist (Gasbarrone and La Mesa
2001). Referring to Figure 13.29a, free polymer and free surfactant molecules co-exist in the region
on the left of the CAC line. In the center portion of the diagram (region in between the CAC and
CMC curves), micelles begin to form on the backbone of the polymer molecules. At high surfactant
concentrations in a region on the right side of the CMC curve, free micelles co-exist with surfactant/
polymer complex.
Chari et al. (1994) reported the effect of SDS on both local polymer chain motions and long range
dimensions of the polymer coil. They showed that when a PEO coil is saturated with SDS micelles,
it is more swollen compared to free coils when it is in a good solvent. They also revealed that these
swollen chains are not fully stretched. Their results showed that the polymer coil at saturation is
more like a swollen cage rather than a necklace.
Romani et al. (2005) investigated the effect of addition of different combinations of polyethyl-
ene glycol (PEG, MW = 8,000 g/mol) and polyvinylpyrrolidone (PVP, MW = 55,000 g/mol) on
668 Handbook of Surface and Colloid Chemistry
TABLE 13.4
Nonionic Polymers Used in the Research Work of Ghoreishi et al. (1999)
Polymer Abbreviation Molecular Weight Supplier
Methyl cellulose MC >50,000 Aqualon
Hydroxypropyl cellulose HPC 100,000 Aldrich
Ethyl hydroxy ethyl cellulose EHEC 120,000 Bernol Noble AB, Sweden
Hydrophobically modified EHEC HM-EHEC 120,000 Bernol Noble AB, Sweden
Hydroxybutylmethyl cellulose HBMC 120,000 Aldrich
Hydroxypropylmethyl cellulose HPMC >50,000 Aqualon
Hydroxyethyl cellulose HEC 250,000 Aldrich
Poly(vinylpyrrolidone) PVP 360,000 Sigma
Poly(vinylpyrrolidone) PVP 40,000 Unilever
Poly(vinylpyrrolidone) PVP 24,000 Unilever
poly(vinylpyrrolidone) PVP 15,000 Unilever
Poly(vinylpyrrolidone) PVP 10,000 Sigma
Poly(vinylpyrridine) nitrogen oxide PVPy-N-0 200,000 Unilever
Poly(vinylpyrrolidone)/poly(vinylimidazole) PVP/PVI 40,000 Unilever
copolymer
Poly(propylallylamine) PPAA Unilever
Poly(vinylimidazole) PVI Unilever
Poly(ethylene oxide) PEO 4,000 BDH
Poly(ethylene oxide) PEO 6,000 BDH
Poly(propylene oxide) PPO 1,000 Aldrich
Poly(vinyl methyl ether) PVME 27,000 Aldrich
Methylvinylimidazole/vinylpyrrolidone copolymer MVI/VP 40,000 Unilever
N-vinylacryloylpyrrolidine and vinylpyridine PAPR 10,000
dicyanomethylide copolymer
Water wt.
Polymer wt%
3 CAC
Polymer wt.
2
CMC*
1 CMC*
CAC
0
0 0.5 1 1.5 2 2.5 Surfactant wt.
C/CMC°
FIGURE 13.29 Phase behavior of a system consisting of nonionic polymer and ionic surfactant. (From
Gasbarrone, P. and La Mesa, C., Colloid Polym. Sci., 279(12), 1192, 2001.)
the CMC of surfactant SDS. They measured the electrical conductivity, the zeta potential, and the
viscosity. They also used the fluorescence spectroscopy and small-angle X-ray scattering (SAXS).
The results showed that the SDS−polymer interaction begins at a low surfactant concentration,
and the CAC is dependent on the polymer composition. The average aggregation number varied
with surfactant concentration and was highly unstable compared to pure SDS micelles. The zeta
TABLE 13.5
Critical Concentration Associated with Binding in 1 × 104 mol dm−3 NaBr
T1 × 103 T2 × 103 Tf × 103 m1 at T2 × 103 T2 − m1 × 103 (T2 − m1)/CP × 103 (mol of
Polymer (mol dm−3) (mol dm−3) (mol dm−3) (mol dm−3) (mol dm−3) bound SDS/g dm−3 polymer) a
MC (0.5%) 0.003 0.025 0.025 0.0069 0.0224 0.00448 N/A
HPC (0.5%) 0.0015 0.023 0.022 0 0033 0.0197 0.00394 6.06
HPC (0.5%) 0.001 0.0232 0.023 0.006 0.0172 0.00344 5.3
EHEC (0.5%) (120,000) 0.002 0.0224 0.02 0.00436 0.0180 0.00361 6.65
HMEHEC (0.5%) 0.002 0.020 0 020 0.0068 0.013 0.0026 4.8
HBMC (0.5%) 0.0035 0.03 0.0103 0.0031 N/A N/A N/A
HPMC (0.5%) 0.004 0.022 0.013 0.0032 N/A N/A N/A
HEC (0.5%) 0.0024 0.0333 0.0182 0.0030 N/A N/A N/A
PVP (1%) (3,600,000) 0.0015 0.045 0042 0.0038 0.0412 0.00412 22.8
PVP (1%) (360,000) 0.002 0.050 0.050 0.0047 0.0453 0.00453 25.1
A Review of Polymer–Surfactant Interactions
PVP (0.5%) (10,000) 0.0009 0.048 0.045 0.0035 0.0445 0.0089 5.5
PVP (1%) (24,000) 0.0012 0.055 0.015 0.005 N/A N/A N/A
PVP (1%) (15,000) 0.0013 0.052 0.01 0.0045 N/A N/A N/A
PVP (1%) (10,000) 0.002 0.05 0.012 0.0047 N/A N/A N/A
PV1 (0.1%) 0.0004 0.022 0.014 0.0033 0.0167 0.0167 10.3
PVPY-NO (0.1%) <0.00002 0.017 0.014 0.0075 0.0095 0.0095 29.2
PPAA (0.1%) <0.00002 0.025 0.02 0.004 0.021 0.021 N/A
PVP/PVI (0.1%) 0.0007 0.016 0.015 0.0050 0.011 0.011 6.8
PVME (0.5%) 0.001 0.043 0.043 0.004 0.039 0.0078 3.24
PEO (1%) (4,000) 0.0045 0.038 0.006 0.0033 N/A N/A N/A
PEO (1%) (6,000) 0.002 0.038 0.008 0.0035 N/A N/A N/A
PAPR (0.5%) (10,000) 0.0025 0.055 0.033 0.00555 N/A N/A N/A
PPO (0.5%) 0.001 0.14 0.056 0 00456 N/A N/A N/A
MVI/VP (0.5%) 0.0005 0.04 0.038 0.005 0.035 0.007 4.3
potentials were found to increase linearly with the fraction of PVP at constant SDS concentration.
The results of the SAXS indicated that the PVP/PEG/SDS system consists of cylindrically shaped
structures with an anisometry ratio of about 3.0.
Meszaros et al. (2005) carried out a thermodynamic analysis of the interaction between 14 differ-
ent molar mass poly(ethylene oxide) (PEO) and sodium dodecyl sulfate (SDS) based on the measured
surfactant-binding isotherms. They measured the surfactant-binding isotherms using the potentio-
metric method in the presence of 0.1 M inert electrolyte (NaBr). When the molecular weight of PEO
was lower than 1000, no PEO/SDS complex formation was detected. In the molecular weight range
of 1000 < PEO < 8000, the critical aggregation concentration (CAC) and the surfactant aggregation
number decreased with the increase in the polymer molecular weight. The polymer concentration
had a direct effect on the amount of bound surfactant at saturation. They further observed that when
the molecular weight exceeded ~8000, the CAC no longer depended on the polymer molar mass,
and the bound amount of the surfactant at saturation became proportional to the mass concentration
of the polymer.
Burman et al. (2014) studied the interaction of a variety of anionic surfactants including sodium
dodecyl sulfate (SDS), sodium cholate (NaC), sodium deoxycholate (NaDC), and sodium taurode-
oxycholate (NaTDC) with a nonionic polymer of hydroxy propyl cellulose (HPC). They applied
the microcalorimetric, conductometric, and fluorimetric methods to study the interactions. Using
calorimetric and conductometric techniques, they could obtain some of the solution properties such
as the critical aggregation concentration (CAC), critical micelle concentration (CMC), polymer
saturation concentration (PSP), and the extent of binding of the surfactants with polymer. They
concluded that the hydrophobicity and charge density of surfactant have a strong effect on micelliza-
tion. The fluorescence results showed that increasing the surfactant concentration decreased the
micro-polarity.
the same copolymer and anionic surfactant sodium dodecyl sulfate (SDS). Surface tension, conduc-
tivity, and dynamic laser light scattering techniques were employed. Using the surface tension mea-
surements, the critical micelle concentration (CMC), free energy of adsorption (ΔGads), free energy
of micellization (ΔGm), surface excess concentration (Γ), and minimum area per molecule (A) were
calculated. The critical micelle concentration (CMC) and critical aggregation concentration (CAC)
at different temperatures were determined by means of the conductivity measurements. The changes
in the physicochemical properties of the micellized block copolymer were studied using dynamic
laser light scattering. The results showed that the physiochemical properties of diblock copolymer
are strongly influenced by the addition of surfactant. Due to the formation of polymer–surfactant
complex, the hydrodynamic radius of the diblock copolymer changed significantly. The interaction
was observed at a low concentration of surfactant indicating a low CAC. The CAC decreased with
the increase in the temperature.
Gemini surfactants are more effective compared to regular surfactants in lowering the surface
tension. Sardar and Kamil (2012a,b) studied the interaction between a nonionic polymer, (hydroxypro-
pyl)methyl cellulose (HPMC), and cationic gemini surfactants, bis(hexadecyldimethylammonium)
hexane dibromide (16-6-16), bis(hexadecyldimethylammonium)pentane dibromide (16-5-16), and
their corresponding monomeric counterpart cetyltrimethylammonium bromide (CTAB). Electrical
conductometry, fluorescence, and viscometry methods were used to characterize the interactions.
They reported that the gemini surfactants interact strongly with HPMC as compared with conven-
tional monomeric surfactant CTAB. The fluorescence measurements showed that the aggregation
number was higher in the case of CTAB as compared with the gemini surfactant. Upon the addition
of gemini surfactant to the polymer solution, the viscosity increased profoundly due to the forma-
tion of a network between the surfactant micelles and the polymer molecules. The results of this
study appear to be in contradiction with the work of Muzzalupo et al. (2007) who did not observe
any interaction between cationic gemini surfactants and nonionic polymer PVP.
Peng et al. (2013) studied a thermo-sensitive copolymer based on oligo(ethylene glycol)methac-
rylates with three types of cationic alkyltrimethylammonium bromide surfactants (RTAB with R
equal to C12, C14, and C16): dodecyltrimethylammonium bromide (DoTAB), tetradecyltrimethyl-
ammonium bromide (TTAB), and cetyltrimethylammonium bromide (CTAB). They used isother-
mal titration calorimetry (ITC), surfactant selective electrode (SSE), and dynamic light scattering
(DLS) methods to study the interaction. They reported a strong interaction between the polymer and
surfactants. The PSP was dependent on the polymer concentration and was independent of the tem-
perature. The binding affinity of different surfactants for polymer varied in the following sequence:
CTAB > TTAB > DoTAB.
Mohsenipour and Pal (2013) studied the behavior of polyethylene oxide (PEO) and octadecyl-
trimethylammonium chloride (OTAC) mixtures and reported that the relative viscosity increased
compared with pure PEO and surfactant alone. The increase in viscosity was attributed to the for-
mation of network of polymer molecules and surfactant micelles. Similar patterns were observed by
Prajapati (2009) for the same system using a different molecular weight PEO.
2
1
0
–1
FIGURE 13.30 Calorimetric titration curves for the addition of SB 3–14 (6.0 mmol kg−1) at 15°C to water
(◻) and to a 0.1 wt% polymeric solution of PEO (⦁), PPO (∆), and PAA (★) at pH 3.5. (From Brinatti, C. et al.,
Langmuir, 30(21), 6002, 2014.)
Using isothermal titration calorimetry (ITC) and differential scanning calorimetry (DSC), micellar
aggregates of polymer and surfactant were detected in all cases. They also measured the CMC’s of
the F127-rich micelles.
Nambam and Philip (2012) studied the interaction of a nonionic surfactant (nonylphenolethoxyl-
ate [NP9]) with a triblock copolymer of PEO–PPO–PEO in aqueous solution. The micellar size, zeta
potential, and the rheological properties were measured. They also compared the results with the
interactions of the same copolymer with an anionic surfactant (sodium dodecyl sulfate [SDS]) and
a cationic surfactant (cetyltrimethylammonium bromide [CTAB]). They reported that the addition
of NP9 does not have any effect on the viscoelastic properties. The same result was observed in the
case of CTAB whereas a drastic reduction was reported in the case of SDS.
The interaction of a series of zwitterionic surfactants (sulfobetaines) with some nonionic polymers
were studied by Brinatti et al. (2014) using an isothermal titration calorimetry (ITC). Polymers were
poly(ethylene oxide), PEO, M = 100,000 g/mol, poly(propylene oxide), PPO, M = 1,000 g/mol, and
poly(acrylic acid), PAA, M = 2,000 g/mol. The results showed that PAA could induce the f ormation
of micellar aggregates at low surfactant concentrations indicating significant interaction. The ITC
curves are presented in Figure 13.30 for the polymers PAA, PEO, and PPO. The curves are similar
for PEO and PPO and they follow the same curve as that of the surfactant sulfobetaine in water.
The introduction of PEO and PPO in solution changes the endothermic enthalpy value slightly at
low surfactant concentration. However, the CMC value is the same. The PEO and PPO curves are
different from each other at high surfactant concentrations above the CMC region. As PPO is more
hydrophobic compared to PEO, its enthalpy curve lies slightly above the dilution curve of surfactant.
The upward shift in the enthalpy curve for PPO relative to the dilution curve is about 1.5 kJ/mol.
TABLE 13.6
Table of Surfactant Reactivity with Polymer
Type of Polymer Degree of Reactivity
P0 uncharged polymer S− > S+ ≫ S0
P+, polycation S− ≫ S0 ≫ S+
P−, polyanion S+ ≫ S0 ≫ S−
most active systems in terms of polymer–surfactant interactions are those consisting of oppositely
charged polymer/surfactant system and nonionic polymer/ionic surfactant system.
The interaction of polymers and surfactants is still an ongoing research. The mechanisms and
thermodynamics of interactions are not yet fully understood. In this section, we review some of the
literature related to the mechanisms of interaction between polymer and surfactant molecules.
The mechanism responsible for the aggregation of surfactant molecules in aqueous solution can
be extended to the association of polymer and surfactant. When surfactant is dissolved in water, the
headgroups (which are hydrophilic) point toward the water side, and the tails (which are hydropho-
bic) tend to escape the water phase. To create a hydrophobic environment, the tails form aggregates
(micelles). When surfactant molecules are introduced to a polymer solution, the tails of the surfac-
tant molecules are attracted to the hydrophobic sites available on the backbone of polymer chains.
Thus, aggregation of surfactant molecules often occurs on the backbone of the polymer chains.
This is the main mechanism for interaction and association of polymer and surfactant molecules.
The other mechanism for interaction between polymer and surfactant molecules is the electrostatic
attraction between oppositely charged polymer and surfactant molecules.
The polymer–surfactant interaction is highly dependent on the type of each component or reac-
tant (polymer and surfactant). The shape and format of complexes formed (if there is any interac-
tion) will be different with different types of components. Nagarajan (2001) has categorized the
different types of interactions between polymers and surfactants and has introduced the possible
shapes of polymer–surfactant complexes. These formats are presented in Figure 13.31. When both
the polymer and the surfactant have the same electrical charge, no interaction is expected. This
situation is represented as structure A in Figure 13.31. For the case where the polymer and the sur-
factant carry opposite charges, the structures shown as B and C in Figure 13.31 are observed. The
electrostatic interaction is the main driving force for the formation of such structures. The bonding
of the polymer and the surfactant molecules occurs in both structures. However, in the case of C
structure, the surfactant molecules lead to the formation of a network between different chains of
the macromolecules. Random copolymer or multiblock copolymer with short blocks interact with
the surfactant molecules such that the complexes formed are of structural format D. Structures E,
F, and G are formed in the case of hydrophobically modified polymers. Many factors can affect the
interaction between hydrophobically modified polymers and surfactants. Factors such as the size of
the hydrophobic modifier, the grafting density on the backbone of the polymer molecules, and the
ratio of polymer to surfactant may change the structure of polymer–surfactant complexes. Structure
E is possible at low concentration of surfactant. The individual surfactant molecules associate with
the hydrophobic sites on the polymer chains. No conformational change is expected in this case.
Increasing the surfactant concentration leads to a change in the structure from E to F where a
single cluster or micelle of surfactant molecules may interact with multiple hydrophobic modifier
sites present on the backbone of the same polymer chain. This association changes the polymer
conformation profoundly. When the surfactant concentration is high, the surfactant molecules form
clusters or micelles around each hydrophobic modifier site present on the polymer chain, as shown
A Review of Polymer–Surfactant Interactions 675
A. Polymer molecule does not interact with surfactants E. Polymer is hydrophobically modified. Individual
for electrostatic or steric reasons. No surfactant is bound surfactant molecules associate with one or more of the
to the polymer. For example, the surfactant and the hydrophobic modifiers on a single polymer molecule or
polymer are both anionic or both cationic multiple polymer molecules.
B. The polymer and the surfactant are oppositely F. Polymer is hydrophobically modified. Clusters of
charged. Single surfactant molecules are bound surfactant molecules associate with multiple
linearly along the length of the polymer molecules. hydrophobic modifiers on a single polymer molecule.
D. The polymer is an uncharged random or multiblock H. The polymer segments partially penetrate and wrap
copolymer. The surfactant molecules orient themselves around the polar head group region of the surfactant
at domain boundaries separating the polymer segments micelles. A single polymer molecule can associate with
of different polarities. one or more surfactant micelles.
FIGURE 13.31 Different types of polymer-surfactant complexes. (From Nagarajan, R., Polymer-surfactant
interactions, New Horizons: Detergents for the New Millennium Conference Invited Paper, 2001, Fort Myers, FL.)
in structure G. When a nonionic polymer is interacting with the ionic surfactant, the most probable
structure is H shown in Figure 13.31. In this case, the polymer molecule wraps around the surfactant
micelles and at the same time some portion of the polymer chain penetrates into the core of surfac-
tant micelles to change the core environment and make it more hydrophobic (which is preferred by
the surfactant tails).
676 Handbook of Surface and Colloid Chemistry
The early model proposed for the interaction of nonionic polymer with ionic surfactant by
Shirahama et al. (1974), referred to as the “string of beads” model, postulated the formation of
surfactant micelles (beads) on the backbone of the polymer chains. Later on two additional models
were proposed by Nagarajan (1980) and Ruckenstein et al. (1987). Nagarajan proposed the so-called
necklace model. According to this model, the polymer–surfactant complex consists of polymer
chain wrapped around one or more surfactant micelles. The polymer segments tend to partially pen-
etrate into the polar head group region of the micelles causing a reduction in the micelle core-water
contact area. Structure H of Figure 13.31 describes the necklace model closely. This model suggests
that the polymer–surfactant complex brings the micelles in a more stable position from the point of
view of free energy. Using this model, one can also explain the increase in the interaction tendency
between the polymer and the surfactant when the hydrophobicity of the polymer is increased.
The model by Ruckenstein et al. (1987) describes the formation of polymer–surfactant complex
based on the adsorption of micelles on the polymer chains. The presence of the polymer molecules
in water changes the micro-environment of the surfactant molecules and the surface free energy
between micellar hydrocarbon core and the solvent in the “free space” of the coiled macromolecule.
If the surfactant head group is small (such as anionic surfactant head group), the overall surface
free energy is decreased in the presence of polymer, and consequently the micellization process is
amplified. In such cases, the free aggregation of surfactant molecules starts only after the polymer
molecules are saturated with the bound micelles. If the head group is large (such as cationic surfac-
tant head group), the surface free energy is increased and the presence of the polymer has no effect
on the micellization of surfactant.
Wallin and Linse (1996a,b) proposed a model for the interaction between a surfactant and an oppo-
sitely charged polymer based on Monte Carlo simulation taking into consideration the effects of chain
flexibility, linear charge density, and surfactant tail length. They emphasized the importance of elec-
trostatic interaction and polyelectrolyte rigidity. Figure 13.32 shows some of the snapshots produced
by Monte Carlo simulation. The figure shows the changes in the structure of the polymer–micelle
complex with the increase in polymer rigidity. Clearly the flexibility of the polymer chain can have
(d) (e)
FIGURE 13.32 Snapshots of Monte Carlo simulation for interaction between oppositely charged polymer
and surfactant with different angles between consecutive chain segments: (a) 90°, (b) 135°, (c) 150°, (d) 165°,
and (e) 175°. (From Wallin, T. and Linse, P., Langmuir, 12(2), 305, 1996a.)
A Review of Polymer–Surfactant Interactions 677
(a) (b)
FIGURE 13.33 Schematic illustration of TX-100/PEG complexes formed with low molecular weight
(MW < 2000 Da) PEG (a) and high molecular weight (MW > 2000 Da) PEG (b). (From Ge, L. et al., Polymer,
48(9), 2681, 2007.)
a great effect on the structure of the complex. When the polymer molecules are flexible, the polymer
molecules wrap around the micelles and are completely adsorbed. The rigidity of the polymer mol-
ecule makes it difficult for the polymer chain to come close to the micelles and become attached.
Ge et al. (2007) investigated the interaction of nonionic surfactant TX-100 and polyethylene gly-
col (PEG) using fluorescence resonance energy transfer. Based on their research, they proposed two
possible models for the interaction between nonionic polymer PEG and nonionic surfactant TX-100.
They concluded that for low molecular weight PEG, the complexes are sphere-like (Figure 13.33a)
clusters where the polymer molecule wraps around a single micelle. For high molecular weight
PEG, the longer polymer chains mostly tend to form coral-like clusters (Figure 13.33b) where the
same polymer molecule is wrapped around multiple micelles.
REFERENCES
Ahmadi, M. A. and S. R. Shadizadeh (2013). Experimental investigation of adsorption of a new nonionic sur-
factant on carbonate minerals. Fuel 104: 462–467.
Anthony, O. and R. Zana (1996). Interactions between water-soluble polymers and surfactants: Effect of the
polymer hydrophobicity. 2. Amphiphilic polyelectrolytes (polysoaps). Langmuir 12(15): 3590–3597.
Asnacios, A., D. Langevin, and J.-F. Argillier (1996). Complexation of cationic surfactant and anionic polymer
at the air-water interface. Macromolecules 29(23): 7412–7417.
Aubry, T. and M. Moan (1996). Influence of a nonionic surfactant on the rheology of a hydrophobically associ-
ating water soluble polymer. Journal of Rheology 40(3): 441–448 (1978 present).
Bachmann, R. T., A. C. Johnson, and R. G. Edyvean (2014). Biotechnology in the petroleum industry: An
overview. International Biodeterioration & Biodegradation 86: 225–237.
Bai, G., M. Nichifor, and M. Bastos (2010). Cationic polyelectrolytes as drug delivery vectors: Calorimetric
and fluorescence study of rutin partitioning. The Journal of Physical Chemistry B 114(49):
16236–16243.
Banerjee, R., S. Dutta, S. Pal, and D. Dhara (2013). Spontaneous formation of vesicles by self-assembly of
cationic block copolymer in the presence of anionic surfactants and their application in formation of
polymer embedded gold nanoparticles. The Journal of Physical Chemistry B 117(13): 3624–3633.
Benrraou, M., B. Bales, and R. Zana (2003). Effect of the nature of the counterion on the interaction between
cesium and tetraalkylammonium dodecylsulfates and poly(ethylene oxide) or poly(vinylpyrolidone).
Journal of Colloid and Interface Science 267(2): 519–523.
Bibi, I., A. Khan, N. Rehman, S. Pervaiz, K. Mahmood, and M. Siddiq (2012). Characterization of
surfactant-diblock copolymer interactions and its thermodynamic studies. Journal of Dispersion Science
and Technology 33(6): 792–798.
Blokhus, A. M. and K. Klokk (2000). Interactions between poly(ethylene oxide) and sodium alkylcarboxylates
as studied by conductivity and gel permeation chromatography. Journal of Colloid and Interface Science
230(2): 448–451.
678 Handbook of Surface and Colloid Chemistry
Bonnaud, M., J. Weiss, and D. J. McClements (2010). Interaction of a food-grade cationic surfactant (lauric
arginate) with food-grade biopolymers (pectin, carrageenan, xanthan, alginate, dextran, and chitosan).
Journal of Agricultural and Food Chemistry 58(17): 9770–9777.
Brackman, J. C. and J. B. F. N. Engberts (1992). Effect of surfactant charge on polymer-micelle interactions:
n-Dodecyldimethylamine oxide. Langmuir 8(2): 424–428.
Breuer, M. and I. Robb (1972). Interactions between macromolecules and detergents. Chemistry & Industry
1972(13): 530–535.
Brinatti, C., L. B. Mello, and W. Loh (2014). Thermodynamic study of the micellization of zwitterionic surfac-
tants and their interaction with polymers in water by isothermal titration calorimetry. Langmuir 30(21):
6002–6010.
Burman, A. D., S. Ghosh, and A. Das (2014). Behavior of cationic and anionic surfactants including bile salts
in presence of hydroxy propyl cellulose (HPC). Journal of Nanofluids 3(2): 140–153.
Capalbi, A. and C. La Mesa (2001). Polymer-surfactant interactions. Journal of Thermal Analysis and
Calorimetry 66(1): 233–241.
Chandar, P., P. Somasundaran, and N. Turro (1988). Fluorescence probe investigation of anionic polymer-
cationic surfactant interactions. Macromolecules 21(4): 950–953.
Chari, K., B. Antalek, M. Lin, and S. Sinha (1994). The viscosity of polymer–surfactant mixtures in water. The
Journal of Chemical Physics 100: 5294.
Cooke, D., C. Dong, R. Thomas, A. Howe, E. Simister, and J. Penfold (2000). Interaction between gelatin
and sodium dodecyl sulfate at the air/water interface: A neutron reflection study. Langmuir 16(16):
6546–6554.
Couderc, S., Y. Li, D. Bloor, J. Holzwarth, and E. Wyn-Jones (2001). Interaction between the nonionic sur-
factant hexaethylene glycol mono-n-dodecyl ether (C12EO6) and the surface active nonionic ABA block
copolymer pluronic F127 (EO97PO69EO97) formation of mixed micelles studied using isothermal titration
calorimetry and differential scanning calorimetry. Langmuir 17(16): 4818–4824.
Dai, S. and K. Tam (2001). Isothermal titration calorimetry studies of binding interactions between polyethyl-
ene glycol and ionic surfactants. The Journal of Physical Chemistry B 105(44): 10759–10763.
Dan, A., S. Ghosh, and S. P. Moulik (2009). Physicochemistry of the interaction between inulin and alkyltri-
methylammonium bromides in aqueous medium and the formed coacervates. The Journal of Physical
Chemistry B 113(25): 8505–8513.
Dan, A., S. Ghosh, and S. P. Moulik (2010). Interaction of cationic hydroxyethylcellulose (JR400) and cationic
hydrophobically modified hydroxyethylcellulose (LM200) with the amino-acid based anionic amphi-
phile Sodium N-Dodecanoyl Sarcosinate (SDDS) in aqueous medium. Carbohydrate Polymers 80(1):
44–52.
Deo, P., N. Deo, P. Somasundaran, A. Moscatelli, S. Jockusch, N. J. Turro, K. Ananthapadmanabhan, and M.
F. Ottaviani (2007). Interactions of a hydrophobically modified polymer with oppositely charged surfac-
tants. Langmuir 23(11): 5906–5913.
Diamant, H. and D. Andelman (1999). Onset of self-assembly in polymer-surfactant systems. Europhysics
Letters 48: 170.
Dualeh, A. J. and C. A. Steiner (1990). Hydrophobic microphase formation in surfactant solutions containing
an amphiphilic graft copolymer. Macromolecules 23(1): 251–255.
Dubief, C., J. F. Grollier, and J. Mondet (1989). Cosmetic compositions containing a cationic polymer and an
anionic polymer as thickening agent.
Feitosa, E., W. Brown, and P. Hansson (1996a). Interactions between the non-ionic surfactant C12E5 and
poly(ethylene oxide) studied using dynamic light scattering and fluorescence quenching. Macromolecules
29(6): 2169–2178.
Feitosa, E., W. Brown, M. Vasilescu, and M. Swanson-Vethamuthu (1996b). Effect of temperature on the inter-
action between the nonionic surfactant C12E5 and poly(ethylene oxide) investigated by dynamic light
scattering and fluorescence methods. Macromolecules 29(21): 6837–6846.
Fishman, M. and F. Elrich (1975). Interactions of aqueous poly(N-vinylpyrrolidone) with sodium dodecyl
sulfate. II. Correlation of electric conductance and viscosity measurements with equilibrium dialysis
measurements. The Journal of Physical Chemistry 79(25): 2740–2744.
Gasbarrone, P. and C. La Mesa (2001). Interactions of short-chain surfactants with a nonionic polymer. Colloid
and Polymer Science 279(12): 1192–1199.
Gasljevic, K., G. Aguilar, and E. Matthys (2001). On two distinct types of drag-reducing fluids, diameter scal-
ing, and turbulent profiles. Journal of Non-Newtonian Fluid Mechanics 96(3): 405–425.
Ge, L., X. Zhang, and R. Guo (2007). Microstructure of Triton X-100/poly(ethylene glycol) complex investi-
gated by fluorescence resonance energy transfer. Polymer 48(9): 2681–2691.
A Review of Polymer–Surfactant Interactions 679
Ghoreishi, S., Y. Li, D. Bloor, J. Warr, and E. Wyn-Jones (1999). Electromotive force studies associated with
the binding of sodium dodecyl sulfate to a range of nonionic polymers. Langmuir 15(13): 4380–4387.
Goddard, E. (1986a). Polymer–surfactant interaction part II. Polymer and surfactant of opposite charge.
Colloids and Surfaces 19(2): 301–329.
Goddard, E. (1994). Polymer/surfactant interaction—Its relevance to detergent systems. Journal of the
American Oil Chemists’ Society 71(1): 1–16.
Goddard, E. and R. Hannan (1976). Cationic polymer/anionic surfactant interactions. Journal of Colloid and
Interface Science 55(1): 73–79.
Goddard, E. D. (1986b). Polymer–surfactant interaction part I. uncharged water-soluble polymers and charged
surfactants. Colloids and Surfaces 19(2): 255–300.
Goddard, E. D. and K. P. Ananthapadmanabhan (1993). Interactions of Surfactants with Polymers and Proteins.
CRC Press, Boca Raton, FL.
Goldraich, M., J. Schwartz, J. Burns, and Y. Talmon (1997). Microstructures formed in a mixed system of a
cationic polymer and an anionic surfactant. Colloids and Surfaces A: Physicochemical and Engineering
Aspects 125(2): 231–244.
Han, J., F. Cheng, X. Wang, and Y. Wei (2012). Solution properties and microstructure of cationic cellulose/
sodium dodecyl benzene sulfonate complex system. Carbohydrate Polymers 88(1): 139–145.
Hansson, P. and M. Almgren (1996). Interaction of CnTAB with sodium (carboxymethyl) cellulose: Effect of
polyion linear charge density on binding isotherms and surfactant aggregation number. The Journal of
Physical Chemistry 100(21): 9038–9046.
Hansson, P. and B. Lindman (1996). Surfactant-polymer interactions. Current Opinion in Colloid & Interface
Science 1(5): 604–613.
Harada, A. and K. Kataoka (2006). Supramolecular assemblies of block copolymers in aqueous media as nano-
containers relevant to biological applications. Progress in Polymer Science 31(11): 949–982.
Harwigsson, I. and M. Hellsten (1996). Environmentally acceptable drag-reducing surfactants for district heat-
ing and cooling. Journal of the American Oil Chemists’ Society 73(7): 921–928.
Hayakawa, K. and J. C. Kwak (1982). Surfactant-polyelectrolyte interactions. 1. Binding of dodecyltrimethyl-
ammonium ions by sodium dextransulfate and sodium poly(styrenesulfonate) in aqueous solution in the
presence of sodium chloride. The Journal of Physical Chemistry 86(19): 3866–3870.
Hayakawa, K. and J. C. Kwak (1983). Study of surfactant-polyelectrolyte interactions. 2. Effect of multivalent
counterions on the binding of dodecyltrimethylammonium ions by sodium dextran sulfate and sodium
poly(styrene sulfonate) in aqueous solution. The Journal of Physical Chemistry 87(3): 506–509.
Hayakawa, K., J. P. Santerre, and J. C. Kwak (1983). Study of surfactant-polyelectrolyte interactions.
Binding of dodecyl- and tetradecyltrimethylammonium bromide by some carboxylic polyelectrolytes.
Macromolecules 16(10): 1642–1645.
Hellsten, M. (2002). Drag-reducing surfactants. Journal of Surfactants and Detergents 5(1): 65–70.
Hormnirun, P., A. Sirivat, and A. Jamieson (2000). Complex formation between hydroxypropylcellulose and
hexadecyltrimethylamonium bromide as studied by light scattering and viscometry. Polymer 41(6):
2127–2132.
Jiang, W. and S. Han (2000). Viscosity of nonionic polymer/anionic surfactant complexes in water. Journal of
Colloid and Interface Science 229(1): 1–5.
Jones, M. N. (1967). The interaction of sodium dodecyl sulfate with polyethylene oxide. Journal of Colloid and
Interface Science 23(1): 36–42.
Jönsson, B. (1998). Surfactants and Polymers in Aqueous Solution, John Wiley & Sons, New York, NY.
Jönsson, B., K. Holmberg, B. Kronberg, and B. Lindman (2003). Surfactant-Polymer Systems in Surfactant and
Polymers in Aqueous Solution. John Wiley & Sons, New York, Chapter 13.
Jönsson, B., B. Lindman, K. Holmberg, and B. Kronberg (1998). Surfactants and Polymers in Aqueous Solution.
John Wiley & Sons, Chichester, U.K.
Kogej, K. and J. Škerjanc (1999). Fluorescence and conductivity studies of polyelectrolyte-induced aggrega-
tion of alkyltrimethylammonium bromides. Langmuir 15(12): 4251–4258.
Kwak, J. C. (1998a). Polymer-Surfactant Systems. CRC Press, Boca Raton, FL.
Kwak, J. C. T. (1998b). Polymer-Surfactant Systems. Surfactant Science Series, Vol. 77. Marcel Dekker, New York.
La Mesa, C. (2005). Polymer–surfactant and protein–surfactant interactions. Journal of Colloid and Interface
Science 286(1): 148–157.
Lange, H. (1971). Interaction between sodium alkyl sulfates and polyvinyl pyrrolidone in aqueous solutions.
Kolloid-Zeitschrift und Zeitschrift für Polymere 243: 101–109.
Lewis, E. A. and K. P. Murphy (2005). Isothermal Titration Calorimetry. In: Nienhaus, G.U., ed. Protein-
Ligand Interactions. Springer, New York, NY, pp. 1–15.
680 Handbook of Surface and Colloid Chemistry
Li, D., M. S. Kelkar, and N. J. Wagner (2012). Phase behavior and molecular thermodynamics of coacervation
in oppositely charged polyelectrolyte/surfactant systems: A cationic polymer JR 400 and anionic surfac-
tant SDS mixture. Langmuir 28(28): 10348–10362.
Liu, X.-G., X.-J. Xing, Z.-N. Gao, B.-S. Wang, S.-X. Tai, and H.-W. Tang (2014). influence of three anionic
gemini surfactants with different chain lengths on the optical properties of a cationic polyfluorene.
Langmuir 30(11): 3001–3009.
Ma, C. and C. Li (1989). Interaction between polyvinylpyrrolidone and sodium dodecyl sulfate at solid/liquid
interface. Journal of Colloid and Interface Science 131(2): 485–492.
Magny, B., I. Iliopoulos, R. Zana, and R. Audebert (1994). Mixed micelles formed by cationic surfactants and
anionic hydrophobically modified polyelectrolytes. Langmuir 10(9): 3180–3187.
Malovikova, A., K. Hayakawa, and J. C. Kwak (1984). Surfactant-polyelectrolyte interactions. 4. Surfactant
chain length dependence of the binding of alkylpyridinium cations to dextran sulfate. The Journal of
Physical Chemistry 88(10): 1930–1933.
Mandal, B., S. P. Moulik, and S. Ghosh (2013). Physicochemistry of interaction of polyvinylpyrrolidone (PVP)
with sodium dodecyl sulfate (SDS) in salt solution. Journal of Surfaces and Interfaces of Materials 1(1):
77–86.
Masuda, Y., K. Hirabayashi, K. Sakuma, and T. Nakanishi (2002). Swelling of poly (ethylene oxide) gel in
aqueous solutions of sodium dodecyl sulfate with added sodium chloride. Colloid and Polymer Science
280(5): 490–494.
Mészáros, R., I. Varga, and T. Gilányi (2005). Effect of polymer molecular weight on the polymer/surfactant
interaction. The Journal of Physical Chemistry B 109(28): 13538–13544.
Min, T., J. Yul Yoo, H. Choi, and D. D. Joseph (2003). Drag reduction by polymer additives in a turbulent chan-
nel flow. Journal of Fluid Mechanics 486: 213–238.
Minatti, E. and D. Zanette (1996). Salt effects on the interaction of poly(ethylene oxide) and sodium dodecyl
sulfate measured by conductivity. Colloids and Surfaces A: Physicochemical and Engineering Aspects
113(3): 237–246.
Mohsenipour, A. A. (2011). Turbulent drag reduction by polymers, surfactants and their mixtures in pipeline
flow. PhD thesis, University of Waterloo, Waterloo, Ontario, Canada.
Mohsenipour, A. A. and R. Pal (2013). Drag reduction in turbulent pipeline flow of mixed nonionic polymer
and cationic surfactant systems. The Canadian Journal of Chemical Engineering 91(1): 190.
Mohsenipour, A. A., R. Pal, and K. Prajapati (2013). Effect of cationic surfactant addition on the drag reduc-
tion behaviour of anionic polymer solutions. The Canadian Journal of Chemical Engineering 91(1):
181–189.
Mukherjee, S., A. Dan, S. C. Bhattacharya, A. K. Panda, and S. P. Moulik (2011). Physicochemistry of interac-
tion between the cationic polymer poly(diallyldimethylammonium chloride) and the anionic surfactants
sodium dodecyl sulfate, sodium dodecylbenzenesulfonate, and sodium N-dodecanoylsarcosinate in water
and isopropyl alcohol−water media. Langmuir 27(9): 5222–5233.
Muzzalupo, R., M. R. Infante, L. Pérez, A. Pinazo, E. F. Marques, M. L. Antonelli, C. Strinati, and C. La
Mesa (2007). Interactions between gemini surfactants and polymers: Thermodynamic studies. Langmuir
23(11): 5963–5970.
Mya, K. Y., A. M. Jamieson, and A. Sirivat (2000). Effect of temperature and molecular weight on binding
between poly(ethylene oxide) and cationic surfactant in aqueous solutions. Langmuir 16(15): 6131–6135.
Mysels, K. J. (1949). Napalm. Mixture of aluminum disoaps. Industrial & Engineering Chemistry 41(7):
1435–1438.
Nagarajan, R. (1980). Thermodynamics of surfactant-polymer interactions in dilute aqueous solutions.
Chemical Physics Letters 76(2): 282–286.
Nagarajan, R. (1989). Association of nonionic polymers with micelles, bilayers, and microemulsions. The
Journal of Chemical Physics 90: 1980–1994. doi: 10.1063/1.456041.
Nagarajan, R. (2001). Polymer-surfactant interactions. New Horizons: Detergents for the New Millennium
Conference Invited Paper, Fort Myers, FL.
Nambam, J. S. and J. Philip (2012). Effects of interaction of ionic and nonionic surfactants on self-assembly of
PEO–PPO–PEO triblock copolymer in aqueous solution. The Journal of Physical Chemistry B 116(5):
1499–1507.
Nelson, R. (1982). Application of surfactants in the petroleum industry. Journal of the American Oil Chemists
Society 59(10): 823A–826A.
Peng, B., X. Han, H. Liu, and K. C. Tam (2013). Binding of cationic surfactants to a thermo-sensitive copoly-
mer below and above its cloud point. Journal of Colloid and Interface Science 412: 17–23.
A Review of Polymer–Surfactant Interactions 681
Piculell, L. and B. Lindman (1992). Association and segregation in aqueous polymer/polymer, polymer/surfac-
tant, and surfactant/surfactant mixtures: Similarities and differences. Advances in Colloid and Interface
Science 41: 149–178.
Prajapati, K. (2009). Interactions between drag reducing polymers and surfactants. MSc thesis, University of
Waterloo, Waterloo, Ontario, Canada.
Ptasinski, P., F. Nieuwstadt, B. Van Den Brule, and M. Hulsen (2001). Experiments in turbulent pipe flow with
polymer additives at maximum drag reduction. Flow, Turbulence and Combustion 66(2): 159–182.
Qi, Y. and J. L. Zakin (2002). Chemical and rheological characterization of drag-reducing cationic surfactant
systems. Industrial & Engineering Chemistry Research 41(25): 6326–6336.
Romani, A. P., M. H. Gehlen, and R. Itri (2005). Surfactant-polymer aggregates formed by sodium dodecyl
sulfate, poly(N-vinyl-2-pyrrolidone), and poly(ethylene glycol). Langmuir 21(1): 127–133.
Rothstein, J. P. (2008). Strong flows of viscoelastic wormlike micelle solutions. Rheology Reviews 1–42.
Ruckenstein, E., G. Huber, and H. Hoffmann (1987). Surfactant aggregation in the presence of polymers.
Langmuir 3(3): 382–387.
Saito, S. (1987). Polymer-surfactant interactions. In: Schick., M., ed. Nonionic Surfactants Physical Chemistry.
Surfactant Science Series, Vol. 23. Marcel Dekker Inc., New York, pp. 881–892.
Saito, S. and T. Taniguchi (1973). Effect of nonionic surfactants on aqueous polyacrylic acid solutions. Journal
of Colloid and Interface Science 44(1): 114–120.
Salager, J.-L. (2002). Surfactant’s types and uses. In: Salager, J.L., ed. Firep Booket-E300-Attaching Aid in
Surfactant Science and Engineering in English, University of Andes, Merida, Venezuela, p. 3.
Sardar, N. and M. Kamil (2012a). Interaction between nonionic polymer hydroxypropyl methyl cellulose
(HPMC) and cationic gemini/conventional surfactants. Industrial & Engineering Chemistry Research
51(3): 1227–1235.
Sardar, N. and M. Kamil (2012b). Solution behavior of anionic polymer sodium carboxymethylcel-
lulose (NaCMC) in presence of cationic gemini/conventional surfactants. Colloids and Surfaces
A: Physicochemical and Engineering Aspects 415: 413–420.
Scheuing, D. R., T. Anderson, W. L. Smith, E. Szekeres, and R. Zhang (2014). Cationic micelles with anionic
polymeric counterions compositions thereof. U.S. Patent 8933010.
Schramm, L. L. (2000). Surfactants: Fundamentals and Applications in the Petroleum Industry. Cambridge
University Press, Cambridge, U.K.
Schramm, L. L., E. N. Stasiuk, and D. G. Marangoni (2003). Surfactants and their applications. Annual Reports
Section C (Physical Chemistry) 99: 3–48.
Schwuger, M. (1973). Mechanism of interaction between ionic surfactants and polyglycol ethers in water.
Journal of Colloid and Interface Science 43(2): 491–498.
Shimizu, T., M. Seki, and J. C. Kwak (1986). The binding of cationic surfactants by hydrophobic alternating
copolymers of maleic acid. Colloids and Surfaces 20(4): 289–301.
Shirahama, K., K. Tsujii, and T. Takagi (1974). Free-boundary electrophoresis of sodium dodecyl sulfate-
protein polypeptide complexes with special reference to SDS-polyacrylamide gel electrophoresis.
Journal of Biochemistry (Tokyo, Japan) 75(2): 309.
Shubin, V. (1994). Adsorption of cationic polymer onto negatively charged surfaces in the presence of anionic
surfactant. Langmuir 10(4): 1093–1100.
Singh, S. K. and S. Nilsson (1999). Thermodynamics of interaction between some cellulose ethers and SDS
by titration microcalorimetry: II. Effect of polymer hydrophobicity. Journal of Colloid and Interface
Science 213: 152–159.
Smith, G. L. and C. L. McCormick (2001). Water-soluble polymers. 79. Interaction of microblocky twin-tailed
acrylamido terpolymers with anionic, cationic, and nonionic surfactants. Langmuir 17(5): 1719–1725.
Sreenivasan, K. R. and C. M. White (2000). The onset of drag reduction by dilute polymer additives, and the
maximum drag reduction asymptote. Journal of Fluid Mechanics 409: 149–164.
Stoll, M., H. Al-Shureqi, J. Finol, S. Al-Harthy, S. Oyemade, A. de Kruijf, J. Van Wunnik, F. Arkesteijn, R.
Bouwmeester, and M. Faber (2010). Alkaline-surfactant-polymer flood: From the laboratory to the field.
SPE EOR Conference at Oil & Gas West Asia, Muscat, Oman. doi: 10.2118/129164-MS.
Tadros, T. F. (2006). Applied Surfactants: Principles and Applications. John Wiley & Sons, New York.
Thalberg, K. and B. Lindman (1989). Interaction between hyaluronan and cationic surfactants. The Journal of
Physical Chemistry 93(4): 1478–1483.
Thalberg, K., B. Lindman, and G. Karlstroem (1990). Phase diagram of a system of cationic surfactant and
anionic polyelectrolyte: Tetradecyltrimethylammonium bromide-hyaluronan-water. Journal of Physical
Chemistry 94(10): 4289–4295.
682 Handbook of Surface and Colloid Chemistry
Thuresson, K., B. Nystroem, G. Wang, and B. Lindman (1995). Effect of surfactant on structural and thermo-
dynamic properties of aqueous solutions of hydrophobically modified ethyl (hydroxyethyl) cellulose.
Langmuir 11(10): 3730–3736.
Thuresson, K., O. Söderman, P. Hansson, and G. Wang (1996). Binding of SDS to ethyl (hydroxyethyl)
cellulose. Effect of hydrophobic modification of the polymer. The Journal of Physical Chemistry
100(12): 4909–4918.
Touhami, Y., D. Rana, G. Neale, and V. Hornof (2001). Study of polymer-surfactant interactions via surface
tension measurements. Colloid & Polymer Science 279(3): 297–300.
Trabelsi, S. and D. Langevin (2007). Co-adsorption of carboxymethyl-cellulose and cationic surfactants at the
air-water interface. Langmuir 23(3): 1248–1252.
Veggeland, K. and T. Austad (1993). An evaluation of gel permeation chromatography in screening s urfactant/
polymer interaction of commercial products in saline aqueous solutions. Colloids and Surfaces A:
Physicochemical and Engineering Aspects 76: 73–80.
Veggeland, K. and S. Nilsson (1995). Polymer-surfactant interactions studied by phase behavior, GPC, and
NMR. Langmuir 11(6): 1885–1892.
Villetti, M. A., C. I. D. Bica, I. T. S. Garcia, F. V. Pereira, F. I. Ziembowicz, C. L. Kloster, and C. Giacomelli
(2011). Physicochemical properties of methylcellulose and dodecyltrimethylammonium bromide in
aqueous medium. The Journal of Physical Chemistry B 115(19): 5868–5876.
Volden, S., J. Genzer, K. Zhu, M.-H. G. Ese, B. Nyström, and W. R. Glomm (2011). Charge-and temperature-
dependent interactions between anionic poly(N-isopropylacrylamide) polymers in solution and a cationic
surfactant at the water/air interface. Soft Matter 7(18): 8498–8507.
Wallin, T. and P. Linse (1996a). Monte Carlo simulations of polyelectrolytes at charged micelles. 1. Effects of
chain flexibility. Langmuir 12(2): 305–314.
Wallin, T. and P. Linse (1996b). Monte Carlo simulations of polyelectrolytes at charged micelles. 2. Effects of
linear charge density. The Journal of Physical Chemistry 100(45): 17873–17880.
Wang, C. and K. Tam (2002). New insights on the interaction mechanism within oppositely charged polymer/
surfactant systems. Langmuir 18(17): 6484–6490.
Wang, Y., K. Kimura, P. L. Dubin, and W. Jaeger (2000). Polyelectrolyte-micelle coacervation: Effects of
micelle surface charge density, polymer molecular weight, and polymer/surfactant ratio. Macromolecules
33(9): 3324–3331.
Wang, Y., K. Kimura, Q. Huang, P. L. Dubin, and W. Jaeger (1999). Effects of salt on polyelectrolyte-micelle
coacervation. Macromolecules 32(21): 7128–7134.
Winnik, F. M. and S. T. Regismond (1996). Fluorescence methods in the study of the interactions of surfactants
with polymers. Colloids and Surfaces A: Physicochemical and Engineering Aspects 118(1): 1–39.
Winnik, F. M., S. T. Regismond, and E. D. Goddard (1997). Interactions of an anionic surfactant with a
fluorescent-dye-labeled hydrophobically-modified cationic cellulose ether. Langmuir 13(1): 111–114.
Witte, F. M. and J. B. F. N. Engberts (1987). Perturbation of SDS and CTAB micelles by complexation with
poly(ethylene oxide) and poly(propylene oxide). The Journal of Organic Chemistry 52(21): 4767–4772.
Yan, P. and J. X. Xiao (2004). Polymer-surfactant interaction: Differences between alkyl sulfate and alkyl
sulfonate. Colloids and Surfaces A: Physicochemical and Engineering Aspects 244(1–3): 39–44.
Yang, S.-Q. (2009). Drag reduction in turbulent flow with polymer additives. Journal of Fluids Engineering
131(5): 051301.
Zakin, J., Y. Zhang, and W. Ge (2007). Drag reduction by surfactant giant micelles. Surfactant Science Series
140: 473.
Zakin, J. L., Y. Zhang, and Y. Qi (2006). Drag reducing agents. In: Lee, S., ed. Encyclopedia of chemical pro-
cessing, Vol. 2. Taylor and Francis, Boca Raton, FL, pp. 767–786.
Zakin, J. L., B. Lu, and H. W. Bewersdorff (1998). Surfactant drag reduction. Reviews in Chemical Engineering
14(4–5): 67.
Zakin, J. L. and H. L. Lui (1983). Variables affecting drag reduction by nonionic surfactant additives. Chemical
Engineering Communications 23(1): 77–88.
Zakin, J. L., J. Myska, and Z. Chara (1996). New limiting drag reduction and velocity profile asymptotes for
nonpolymeric additives systems. AIChE Journal 42(12): 3544–3546.
Zana, R. and M. Benrraou (2000). Interactions between polyanions and cationic surfactants with two unequal
alkyl chains or of the dimeric type. Journal of Colloid and Interface Science 226(2): 286–289.
Zana, R., W. Binana-Limbelé, N. Kamenka, and B. Lindman (1992). Ethyl (hydroxyethyl) cellulose-cationic
surfactant interactions: Electrical conductivity, self-diffusion and time-resolved fluorescence quenching
investigations. The Journal of Physical Chemistry 96(13): 5461–5465.
A Review of Polymer–Surfactant Interactions 683
Zhang, H., D. Wang, and H. Chen (2009). Experimental study on the effects of shear induced structure in a
drag-reducing surfactant solution flow. Archive of Applied Mechanics 79(8): 773–778.
Zhang, Q., W. Kang, D. Sun, J. Liu, and X. Wei (2013). Interaction between cationic surfactant of 1-methyl-3-
tetradecylimidazolium bromide and anionic polymer of sodium polystyrene sulfonate. Applied Surface
Science 279: 353–359.
Zhang, X., D. Taylor, R. Thomas, and J. Penfold (2011). The role of electrolyte and polyelectrolyte on the
adsorption of the anionic surfactant, sodium dodecylbenzenesulfonate, at the air-water interface. Journal
of Colloid and Interface Science 365(2): P656–P664.
Zhang, Y. (2005). Correlations among surfactant drag reduction, additive chemical structures, rheological prop-
erties and microstructures in water and water/co-solvent systems. PhD thesis, The Ohio State University,
Columbus, OH.
Zhao, G. and S. B. Chen (2006). Clouding and phase behavior of nonionic surfactants in hydrophobically
modified hydroxyethyl cellulose solutions. Langmuir 22(22): 9129–9134.
CHEMISTRY
HANDBOOK OF
SURFACE AND
COLLOID CHEMISTRY
FOURTH EDITION
This new edition of the Handbook of Surface and Colloid Chemistry informs you of significant recent
developments in the field. It highlights new applications and provides revised insight on surface and colloid
chemistry’s growing role in industrial innovations. The contributors to each chapter are internationally
recognized experts. Several chapters represent new research areas while others provide updates on important
areas of the field.
Reduced in length, the new edition presents a more concise volume for quicker understanding of the physical
principles necessary for application. It includes extensive references for understanding related phenomena,
providing a reference point to broaden knowledge of theoretical and practical functions. It also illustrates
surface and colloid chemistry’s relevance in the struggle against global issues such as energy resources,
environmental control, transportation, housing, biotechnology, health, medicine, drinking water, and
food production.
The Handbook of Surface and Colloid Chemistry, Fourth Edition is an invaluable resource for staying
informed on progress in the field. It keeps you current with theories and their applications to the
development of technology so that you can find more effective solutions to vital problems facing us today
and tomorrow.
K20789