DISPERSION OF NANOPARTICU LATE SUSPENSIONS USING SELF-ASSEMBLED SURFACTANT AGGREGATES By PANKAJ KUMAR SINGH A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2002
Copyright 2002 By Pankaj K. Singh
This study is dedicated to my parents, who encouraged me to pursue my dreams; and to my wife, who taught me the importance of patience and perseverance.
ACKNOWLEDGMENTS I would like to acknowledge my advisor, Brij Moudgil, for his invaluable guidance and support. His deep insight into the fundamentals of applied science encouraged me to seek excellence in my research. Thanks are also due to other faculty members who helped me with discussions and suggestions. Those who deserve special recognition include Hassan El-Shall, Dinesh O. Shah, Wolfgang Sigmund, Rajiv K. Singh, E. Dow Whitney, Yakov Rabinovich, and Abbas Zaman. I would also like to thank the National Science Foundationâ€™s Engineering Research Center for Particle Science and Technology and our industrial partners, for financially supporting this research. I would like to gratefully acknowledge my group members including Joshua Adler, Bahar Basim, Scott Brown, Madhavan Esayanur, and Ivan Vakaralski for their help in carrying out my research, and also for their support and encouragement. Thanks go out to all my friends; Uday Mahajan, Kaustabh Rau, Raja Kalyanaraman, Mahesh Sreenivas, Prakash Thatavarthy, Sharad Mathur, Kim Christmas, and Gopalkrishna Subramanium, for making my stay in Gainesville enjoyable. Lastly, I would like to thank the staff of the Engineering Research Center for Particle Science and Technology including Rhonda Blair, Cheryl Bradley, Gill Brubaker, Shelley Burleson, Anne Donnelly, John Henderson, Lenny Kennedy, Sophie Leone, Kevin Powers, Byron Salter, Erik Sander, Gary Scheiffele, and Nancy Sorkin. iv
TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES...........................................................................................................ix ABSTRACT.......................................................................................................................xi CHAPTER 1 INTRODUCTION...........................................................................................................1 Characterization of the State of Dispersion/Suspension.................................................1 In-Situ Particle Size Measurement.............................................................................2 Rheology....................................................................................................................2 Sedimentation Rate....................................................................................................5 Outline of Dissertation....................................................................................................5 2 DISPERSION OF PARTICLES......................................................................................7 van der Waals Forces......................................................................................................8 Microscopic Approach...............................................................................................9 Macroscopic Approach..............................................................................................9 Electrostatic Forces.......................................................................................................11 Origin of Surface Charge.........................................................................................11 Ion adsorption......................................................................................................11 Ion dissolution.....................................................................................................12 Lattice imperfections and isomorphous substitutions.........................................12 Ionization/dissociation.........................................................................................12 Electrical Double Layers..........................................................................................13 Helmholtz model.................................................................................................13 Gouy-Chapman model.........................................................................................14 Stern/Graham model............................................................................................14 Zeta Potential (Electrokinetic Potential)..................................................................15 Measurement of zeta potential.............................................................................16 Calculation of Electrostatic Forces..........................................................................18 The linearized poisson-boltzmann approach.......................................................21 Analytical formulas.............................................................................................21 DLVO Theory and Colloidal Stability..........................................................................22 v
Polymeric Steric Forces for Colloidal Stability............................................................24 Solution Behavior of Polymers................................................................................24 Flory-Huggins theory..........................................................................................25 Free Volume theory.............................................................................................26 Adsorption of Polymer at the Particle Surface.........................................................27 Steric Barrier due to Adsorbed Polymer Molecules.................................................29 Dispersion of Nanoparticles Under Extreme Processing Conditions...........................31 3 SURFACTANTS AT THE SOLID/LIQUID INTERFACE.........................................35 Introduction...................................................................................................................35 Adsorption of Surfactants at the Solid-Liquid Interface...............................................36 Mechanisms of Adsorption......................................................................................36 Contributions to the Adsorption Energy..................................................................37 Electrical interactions..........................................................................................38 Lateral chain-chain interactions..........................................................................41 Chemical interactions..........................................................................................42 Characterization Techniques Involving Microscopy....................................................44 Scanning Tunneling Microscopy.............................................................................44 Atomic Force Microscope (AFM)............................................................................45 Brewster Angle Microscopy.....................................................................................49 Optical Characterization Techniques............................................................................51 Electron Spin Resonance Spectroscopy...................................................................51 Fluorescence Spectroscopy......................................................................................52 Fourier Transform Infrared Spectroscopy/Attenuated Total Internal reflection......54 Ellipsometry.............................................................................................................58 Optical Waveguide Lightmode Spectroscopy (OWLS)...........................................60 Non-Optical Characterization Techniques....................................................................61 X-Ray Reflectivity, and Grazing Angle Diffraction................................................61 Neutron Reflectivity.................................................................................................62 Quartz Crystal Microbalance (QCM)......................................................................63 4 MATERIALS AND METHODS...................................................................................66 Materials........................................................................................................................66 Methods.........................................................................................................................68 Surface Force Measurement.....................................................................................68 Suspension Stability.................................................................................................69 Viscosity...................................................................................................................69 Adsorption................................................................................................................70 Zeta Potential............................................................................................................70 Contact Angle...........................................................................................................71 FTIR/ATR................................................................................................................71 Measurement.......................................................................................................71 Calculations.........................................................................................................72 vi
5 DISPERSION OF NANOPARTICLES USING SELF-ASSEMBLED SURFACTANT AGGREGATES.......................................................................................................78 Introduction...................................................................................................................78 Results and Discussions................................................................................................80 Factors Controlling the Strength of Self-Assembled Surfactant Structures.............96 Chain Length.......................................................................................................96 Electrolyte Concentration..................................................................................100 Microstructure of Solid Substrate......................................................................102 Presence of Co-surfactants................................................................................102 Surface Charge..................................................................................................105 Temperature.......................................................................................................107 Role of Surface in Surfactant Self-Assembly........................................................108 Viscosity of Concentrated Suspensions.................................................................115 Summary.....................................................................................................................119 6 MECHANICAL AND THERMODYNAMIC PROPERTIES OF SURFACE SURFACTANT AGGREGATES..........................................................................123 Introduction.................................................................................................................123 Hertz Model for Interaction of Hard Sphere (Tip) with Soft Flat Bilayer..................124 Hertz Model for Interaction of Hard Sphere (Tip) with Soft Micelle-Like Spherical Surfactant Aggregates..........................................................................................127 Results and Discussion................................................................................................127 Mechanical Properties............................................................................................127 Energetics of Surface Structures............................................................................136 Summary.....................................................................................................................145 7 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK.............................147 Conclusions.................................................................................................................147 Suggestions for Future Work......................................................................................152 LIST OF REFERENCES.................................................................................................160 BIOGRAPHICAL SKETCH...........................................................................................171 vii
LIST OF TABLES Table Page 1-1: Different techniques for particle size measurement in suspension..............................3 3-1: Calculated values of for different values of the potential...............................41 0coulG 5-1: Properties of different crystal planes of alumina......................................................111 5-2: Effect of crystallographic orientation on self-assembly...........................................112 6-1: Mechanical, and thermodynamic properties of surface aggregates..........................130 viii
LIST OF FIGURES Figure Page 1-1: Relative viscosity as a function of the volume fraction...............................................4 2-1: Charge development in SiO 2 tetrahedra.....................................................................13 2-2: The Stern-Graham model of the electrical double layer............................................15 2-3: Variation of the Debye length with electrolyte concentration....................................19 2-4: The potential distribution between two approaching surfaces...................................20 2-5: The van der Waals attractive force, and electrostatic repulsive forces......................22 2-6: Total interaction energy as a function of separation distance....................................23 2-7: The three domains of sterically stabilized flat plates.................................................30 2-8: Polymeric steric repulsive forces................................................................................31 2-9: Surfactant as dispersant under high electrolyte concentration...................................34 3-1: Adsorption and zeta potential behavior of dodecylsulfate-alumina system...............40 3-2: Force-distance profile measured between the AFM tip, and a substrate with self-assembled surfactant aggregates...............................................................................47 3-3: Horizontal ATR sampling illustrating the parameters of significance.......................56 4-1: Electron micrograph of silicon nitride tip used for determining the tip radius..........67 4-2: FTIR spectra of C 12 TAB with polarized beam...........................................................75 4-3: Dependence of the order parameter S, on the dichroic ratio D..................................76 4-4: Surfactant structures and orientation, and the order parameter values.......................77 5-1: Measured interaction forces between the AFM tip and a mica surface.....................81 5-2: Correlation between suspension stability, and interparticle forces............................83 ix
5-3. Adsorption isotherm, zeta potential, and contact angle..............................................84 5-4. Possible self-assembled surfactant films at different concentration ranges...............86 5-5: Contact angle, maximum repulsive forces, and order parameter...............................90 5-6: Proposed self-assembled surfactant films at different concentrations.......................93 5-7: Stabilization mechanism in the presence of self-assembled surfactant aggregates....96 5-8: Maximum compressive force as a function of chain length.......................................97 5-9: Effect of chain length on surfactant self-assembly.....................................................99 5-10: Maximum compressive force on mica and silica...................................................101 5-11: Maximum compressive force as a function of SDS concentration........................103 5-12: Effect of micromolar SDS addition........................................................................104 5-13: Effect of pH on the steric repulsive forces.............................................................106 5-14: Effect of temperature on the maximum repulsive force.........................................107 5-15: Effect of substrate on surfactant self-assembly......................................................109 5-16: Different crystal planes of alumina used for force measurements.........................111 5-17: Surfactant aggregates on C, and R planes of alumina............................................113 5-18: Order parameter of adsorbed SDS on different crystal planes of alumina.............114 5-19: Viscosity as a function of shear rate for surfactant stabilized suspensions............116 5-20: Comparison between the viscosity of electrostatically stabilized, and surfactant stabilized suspension.............................................................................................117 5-21: Effect of temperature on the viscosity of a surfactant stabilized dispersion..........118 6-1: Hertzâ€™ model of interaction between hard sphere (tip) with soft flat layer..............125 6-2: Experimental force/distance curve...........................................................................128 6-3: Elastic modulus (E) as a function of chain length (n)..............................................135 6-4: Yield strength (Y) as a function of chain length (n).................................................136 6-5: Interaction of two hard flat planes through elastic bilayer, modeled as springs......137 6-6: Energetics of surface aggregate, and bulk micelle formation..................................143 x
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DISPERSION OF NANOPARTICULATE SUSPENSIONS USING SELF-ASSEMBLED SURFACTANT AGGREGATES By Pankaj Kumar Singh August 2002 Chairman: Dr. Brij M. Moudgil Major Department: Materials Science and Engineering The dispersion of particles is critical for several industrial applications such as paints, inks, coatings, and cosmetics. Several emerging applications such as abrasives for precision polishing, and drug delivery systems are increasingly relying on nanoparticulates to achieve the desired performance. In the case of nanoparticles, the dispersion becomes more challenging because of the lack of fundamental understanding of dispersant adsorption and interparticle force prediction. Additionally, many of these processes use severe processing environments such as high normal forces (> 100 mN/m), high shear forces (>10,000 s -1 ), and high ionic strengths (>0.1 M). Under such processing conditions, traditionally used dispersants based on electrostatics, and steric force repulsion mechanism may not be adequate. Hence, the development of optimally performing dispersants requires a fundamental understanding of the dispersion mechanism at the atomic/molecular scale. xi
This study explores the use of self-assembled surfactant aggregates at the solid-liquid interface for dispersing nanoparticles in severe processing environments. Surfactant molecules can provide a feasible alternative to polymeric or inorganic dispersants for stabilizing ultrafine particles. The barrier to aggregation in the presence of surfactant molecules was measured using atomic force microscopy. The barrier heights correlated to suspension stability. To understand the mechanism for nanoparticulate suspension stability in the presence of surfactant films, the interface was characterized using zeta potential, contact angle, adsorption, and FT-IR (adsorbed surfactant film structure measurements). The effect of solution conditions such as pH and ionic strength on the suspension stability, and the self-assembled surfactant films was also investigated. It was determined that a transition from a random to an ordered orientation of the surfactant molecules at the interface was responsible for stability of nanoparticulates. Additionally, the role of the surface in surfactant self-assembly was investigated. Mechanical and thermodynamic properties of the self-assembled layer at the solid-liquid interface were calculated based on experimental results, and compared to the corresponding properties in the bulk solution. xii
CHAPTER 1 INTRODUCTION Dispersion is defined as a uniform, and homogeneous mixture of two or more mutually immiscible phases, with one of the phases forming the continuous medium in which the other phases are dispersed. Several different types of dispersions are encountered in industrial applications. The most common ones include solid/liquid (suspension/slurries), liquid/liquid (emulsions), gas/liquid (foams), liquid/solid (gels), and liquid/gas (aerosols). These dispersions are encountered in almost every industry in some form or other during preparation or as end products. Examples of industrial applications of dispersions include paints, dyestuffs, printing ink, paper coatings, cosmetics, ceramics, microelectronics, agrochemical and pharmaceutical formulations, and various household products. Because, all dispersions involve an interface, the study of the interfacial properties, and their control and manipulation becomes important for the stability of dispersions. In this work, attention is focused on solid/liquid suspensions. However, the same concepts also would be valid for other kinds of dispersions. Characterization of the State of Dispersion/Suspension Coagulation/coalescence of dispersed particles leads to the formation of agglomerates or aggregates, which may adversely impact particle-processing technologies. The stability of a suspension is critical for several particulate processes. For example, in particulate separation processes using selective flocculation, the suspension must be well dispersed before the desired component can be flocculated out [MAT00]. 1
2 The presence of agglomerates adversely affects the rheological properties of particle slurries, and leads to severe problems in transport and pumping of particles [RUS89, REE95]. In the chemical mechanical polishing (CMP) of silicon wafers using particle suspensions as the polishing media, the presence of very small amount of aggregates can cause surface scratches and defects, leading to losses and lower yield [BAS01]. Hence, it is critical to characterize the stability of the dispersion being used. Some common techniques are discussed in the following sections. In-Situ Particle Size Measurement If the particleâ€™s primary particle size is known, then measurements of the size distribution in the suspension state can indicate whether the powder is dispersed or flocculated. Table 1-1 lists some of the in-situ particle sizing techniques, along with the size range they are capable of measuring and the concentration of the sample, which can be measured using these techniques. Caution should be exercised however when using the particle-sizing techniques, since no one technique is appropriate for all slurries. Most of the ensemble techniques (such as light scattering, photon correlation spectroscopy, referenced light scattering, acoustic spectroscopy, flow field fractionation, and sedimentation) are not sensitive to the tails of the particle-size distribution. There is definitely a need for developing sizing techniques that are sensitive to large particles, that are able to measure slurries without dilution, and that may be modified for in-line particle size analysis. Rheology Rheology is the science dealing with flow and deformation of materials. The rheological behavior of particle/liquid systems is important in most processing
3 operations, including powder and batch preparation, materials transport, coating and deposition, and shape forming. Table 1-1: Different techniques for particle size measurement in suspension and the applicable size range and concentrations. Particle size technique Size range Solids range (vol%) Electrical sensing zone 400 nm to 1.2 mm Dilute Light scattering 40 nm to 2 mm Dilute Photon correlation spectroscopy 2 nm to 5 m Dilute Referenced light scattering 30 nm to 5 m Dilute to 5.0 Acoustic spectroscopy 100 nm to 10 m 2.0 to 50.0 Flow field fractionation 15 nm to 2 m 2.0 to 20.0 Sedimentation 100 nm to 100 m Dilute Rheological properties are highly dependent on the physical structure of the particle/liquid system. Structure is governed by factors such as particle size and shape distribution, solid/liquid volume ratio, and interparticle forces. Rheological measurements often can be used to deduce information about the state of the particulate dispersion in the suspension. The viscosity, , is defined as the ratio of the shear stress () to the shear rate (D) D (1-1) At low solids loading, flocs/agglomerates may remain as individual flow units, which do not interact extensively. Flow behavior may be Newtonian, but viscosity will be higher than observed in corresponding dispersed suspension. This is due to the
4 immobilized liquid in the intra-floc (intra-agglomerate) voids, which results in an increase in the â€œeffectiveâ€ solids loading of the suspension (Figure 1-1). Figure 1-1: Relative viscosity as a function of the volume fraction for flocculated/ agglomerated, and dispersed suspensions. As the solids loading increases, agglomerated suspensions tend to show shear thinning behavior (i.e., the viscosity decreases as the shear rate is increased). Extensive structure (floc-aggregates) exists at low shear rates. Viscosities tend to be significantly higher than the corresponding dispersed suspensions due to immobilized liquid. As the shear rate is increased, structure is broken down, entrapped liquid is released, and viscosities are observed to decrease. Whether or not the flow units are broken down to primary particle size depends on the shear rate and the strength of the agglomerates.
5 Sedimentation Rate Monitoring the powder sedimentation rate allows us to assess the state of the dispersion. Poorly dispersed suspensions tend to undergo more rapid sedimentation compared to well-dispersed samples because of the large particle (i.e., floc/agglomerate) size. The simplest way to monitor the sedimentation rate would be to measure the sediment height as a function of time. A higher sediment height would indicate more unstable slurry. Another way to observe sedimentation would be to measure the turbidity of the suspension. Higher turbidity indicates higher scattering of incident light, which in turn indicates stable slurry. Settling however may not be feasible or practical for nano-size particles unless strong agglomeration is present. Settling is also influenced by other factors such as viscosity and structure formation. Turbidity measurements are most reliable when used on nonsettling, and translucent slurries. Outline of Dissertation To achieve desired results in applications involving particle dispersions, dispersion stability must be controlled and manipulated. In addition, the use of suspensions in emerging applications (such as CMP, controlled drug delivery systems, inks, and composites) under extreme conditions (high normal and shear forces, high ionic strength, extremes of pH and temperature) necessitates the development of robust dispersion schemes to achieve optimal performance. The present work is focused on developing the fundamentals of particulate dispersion under the conditions encountered in such processes. The discussion begins in Chapter 2, with a brief overview of different ways of dispersing colloidal suspensions (DLVO theory of colloidal stability, polymeric steric
6 forces), and the limitations of the traditionally used dispersing methods in colloidal processing under extreme environments. Additionally, initial results on the stabilization of nanoparticles using self-assembled surfactant aggregates are presented. Chapter 3 reviews several techniques used in the past to investigate the self-assembly of surfactants at the solid-liquid interface. Chapter 4 describes the various materials and methods/techniques used in this study. The self-assembly of surfactants at the solid-liquid interface and its relevance for the dispersion of nanoparticulate suspensions under extreme conditions is discussed in Chapter 5. We also discuss studies on ways to manipulate and control the self-assembly behavior. Mechanical and thermodynamic properties of the surface aggregates and aggregation phenomena at the surface are compared and contrasted to corresponding properties and phenomena in the bulk solution in Chapter 6. Finally, Chapter 7 summarizes the findings of this investigation and proposes possible avenues for further research.
CHAPTER 2 DISPERSION OF PARTICLES The fundamental criteria for colloidal stability (known as the DLVO theory) was developed independently by the teams of Derjaguin and Landau in Russia [DER41], and Verwey and Overbeek [VER48] in the Netherlands. Both groups came to the conclusion that the stability of similar particles (particles of one type) in suspension was a competition between van der Waals attraction and electrostatic repulsion. If particles dispersed in a liquid attract each other, they will stick together whenever they collide, forming aggregates, that grow heavy and sink, resulting in an unstable colloid. If, on the other hand, there are repulsive electrostatic forces between the particles, they stay apart. Mathematically, this may be expressed as the total energy, W tot , of interaction being equal to the sum of van der Waals attractive energy, W vdW , and electrostatic repulsive energy, W elect , as expressed in Eq. (2-1). (2-1) vdWelecttotWWW This simple relation has been used to predict the behavior of nearly ideal particulates and surfaces in solution and has been validated for a wide variety of conditions and surfaces. Many of the successful applications of DLVO theory have been reviewed in the textbooks by Israelachvili [ISR92], Lyklema [LYK95], and Hunter [HUN01]. The two forces (van der Waals and electrostatic interactions) are discussed briefly in the following sections. 7
8 van der Waals Forces Although van der Waals forces are used commonly for colloidal particles, the forces were originally developed for interaction between gas molecules. To explain the behaviors of gases and their deviation from ideality, J. D. van der Waals [VAN73] first suggested that the interactions could be expressed as a modification of the ideal gas law, PV = nRT, nRT)nbV)(VanP(2 (2-2) where P is pressure, n is the number of moles of gas, V is the volume, R is the gas constant, T is temperature, and a and b are constants specific to a particular gas. The constant b describes the finite volume of the molecules comprising the gas, and the constant a takes into account the attractive forces between the molecules. This modification significantly improved the model for the behavior of gases. The various types of attractive forces between molecules are now collectively termed van der Waals forces. The concept of an attractive force is also used to describe the properties of condensed matter. The commonly used Lennardâ€“Jones potential [LEN28] 126atom/atom r B r CW (2-3) where the net potential energy between atoms, W atom/atom , at distance, r, can be described by the competition of an attractive London dispersion forces [LON37], characterized by constant C and a Born repulsion term (arising from the overlap of electron clouds) characterized by constant B. The London constant, C, is primarily related to the synchronization of instantaneous dipoles created when the energy fields of neighboring atoms overlap.
9 Microscopic Approach In the microscopic approach, the summation of pairwise interactions between the atoms in one body and the atoms in another body is used to derive an expression for the energy of attraction between them. Using this approach, the interaction energy, W, between two spherical particles may be expressed in the following manner 22222sph/sphHR2R41lnHR2R2HR4HR26AW (2-4) where A is Hamakerâ€™s constant, R is the particle radius, and H is the separation distance between the particle surfaces. This relation is valid for all particle sizes. However, if it is considered that the radius of the two particles is much greater than the separation distance, the expression may be reduced to the more commonly used form represented by Eq. (2-5) [DER34]. H12ARWsph/sph (2-5) Derjaguin [DER34], de Boer [DEB36], and Hamaker [HAM37] all contributed significantly to the understanding of the van der Waals interactions between macroscopic bodies based on the individual interactions of London dispersion forces between atoms. Macroscopic Approach In the microscopic approach, the interaction between atoms or dipoles is calculated, assuming that a vacuum exists between the interacting atoms. However, for an atom at the core of a solid particle interacting with an atom in the core of another particle, the intervening media consists of other atoms of the solids, and hence the approach is not correct, since the intervening atoms can have a significant impact on the interaction. Additionally, in real materials several other interactions may exist that can contribute to the total interaction force. Hence, another method to determine and predict the attraction
10 between solids that takes into account the different types of interactions is required. However, accounting for the different types of interactions among different species in a solid is difficult and hence an alternative approach is required. In 1956 Liftshitz [LIF56], based on the assumption that both the static and oscillatory fields produced by the atomic components of solids should directly affect the absorption of electromagnetic energy by the material, derived a method to calculate the attraction between materials based on the differences in their dielectric spectra. This was the first attempt to calculate van der Waals forces based on the continuum or macroscopic approach. Although this approach clearly accounts for the different types of bonding and screening in a body, it is still a very difficult function to measure due to the wide range of frequencies and types of experiments needed to determine the entire function. To simplify this approach, Ninham and Parsegian [NIN70] proposed that major contributions to the overall attraction come from regions of dielectric relaxation, or regions where a specific atomic or molecular mechanism creates a resonant vibration. The characteristic absorption frequencies of some materials are relatively well characterized; for example the spectra of water has been characterized in the ultraviolet, infrared, microwave, and static frequency regimes [GIN72, DAG00]. However, it is still difficult to extract information for a wide variety of materials. Hough and White [HOU80] suggested that most contributions to the overall Hamaker constant come from dielectric relaxations in the UV and infrared regions, and hence dielectric spectra in this wavelength region could be used for the calculations. Considering a single UV relaxation frequency, an approximation for the Hamaker constant, A 131 , of Material 1 interacting with similar material through Medium 3 may be
11 written solely as a function of the differences in the static dielectric constants, (0), and indices of refraction in the visible range, n. The result is the Tabor-Winterton [TAB69] approximation 2/3232122321UV23131131nnnn232h3)0()0()0()0(4kT3A (2-6) where kT is the product of Boltzmannâ€™s constant and temperature and h is Plankâ€™s constant. Electrostatic Forces When particles are immersed in an aqueous medium, they often develop charge at the solid/water interface. If the charges are of the same sign, the particles will repel each other. This force often used to prevent coagulation in suspensions is called electrostatic repulsion. Origin of Surface Charge Most colloidal particles are charged by the following mechanisms: Ion adsorption Charge can develop at surfaces by unequal adsorption of oppositely charged ions. Ion adsorption may be positive or negative. Surfaces in contact with aqueous media are more often negatively charged because cations are more hydrated as compared to anions. Hydrocarbon oil droplets and air bubbles exhibit net negative charges because of negative adsorption of ions. Cations move away from the air-water and oil-water interface more than anions. Surface charge may also be established by the adsorption of charged surfactant molecules. Preferential adsorption of one type of ion on the surface can occur due to either London-van der Waals interactions or due to hydrogen or hydrophobic bonding.
12 Ion dissolution Unequal dissolution of ions in the case of ionic substances can lead to net negative charge on the substrate. A classical example of this mechanism is the silver iodide surface. When silver iodide is immersed in an aqueous environment, dissolution occurs as AgI Ag + + I . Since the solubility product for this equilibrium is relatively small (K sp = a Ag + . a I ~ 10 -16 ), the concentrations of Ag + and I in solution are small. The surface of the crystal consists of an array of Ag + and I ions in cubic close pack, and no net charge develops when the number of each ion is the same. However, equal numbers of each ion on the surface does not occur at the concentration where there are equal numbers of Ag + and I ions in the solution. Instead, due to their higher affinity for the surface, the iodide ions tend to be preferentially adsorbed at the surface. A detailed mathematical treatment of the AgI surface is given by Hunter [HUN01]. Lattice imperfections and isomorphous substitutions In the case of most clay mineral systems, lattice defects result in very large charge densities. The defect in such cases is in the form of isomorphous replacement of one ionic species by another of lower charge. For example, the replacement of a Si atom by an Al atom in the SiO 2 tetrahedra will yield a net negatively charged surface [Figure 2-1]. Ionization/dissociation Ionization of groups such as carboxyl (-COOH), sulphate (-O.SO 2 .OH), sulphonate (-SO 2 .OH), sulphite (-O.SO.OH), amine (-NH 2 ), and quaternary amine (-N + R 3 ) may take place at the surface, resulting in development of surface charges. The dissociation of these surface groups is pH dependent. This mechanism represents the charge development in proteins, and polymer latex systems that have carboxyl, sulphate and sulphonate groups on their surfaces. Also, this mechanism is applicable to the
13 behavior of oxide surfaces. For a detailed description of the dissociation models, the reader is referred to the textbook by Hunter [HUN01]. Figure 2-1: Charge development in SiO 2 tetrahedra, by substitution of Si atom by an Al atom. Electrical Double Layers Surface charge on a particle results in an unequal distribution of ions in the polar medium in the vicinity of the surface. Ions of opposite charge (counter ions) are attracted to the surface, and ions of like charges (co-ions) are repelled away from the surface. This unequal distribution gives rise to a potential across the interface. The exact distribution of the counter ions in the solution surrounding the charged surface is very important, since it determines the potential decay into the bulk from the charged surface. Electrostatic attraction, thermal motion, and forces other than electrostatic (specific adsorption) influence the counter ions in the vicinity of the surface. Several models have been proposed for the distribution of ions in the vicinity of the surfaces, and are summarized below: Helmholtz model In 1879, von Helmholtz proposed that all the counter ions are lined up parallel to the charged surface at a distance of about one molecular diameter. The electrical potential
14 decreases rapidly to zero within a very short distance from the charged surface in this model. This model treated the electrical double layer as a parallel plate condenser, and the calculations of potential decay were based on simple capacitor equations. However, thermal motion leads to the ions being diffused in the vicinity of the surface, and this was not taken into account in the Helmholtz model. Gouy-Chapman model This model proposed by Gouy (1910, 1917) and Chapman (1913) consists of a diffuse distribution of the counter ions, with the concentration of the counter ions falling off rapidly with distance near the surface because of the screening effect, and then falling off gradually. This model is accurate for planar charged surfaces with low surface charge densities, and distances far away from the surface. It is inaccurate for surfaces with high surface charge densities (especially at small distances from the charged surfaces) since it treats the ions as point charges, and neglects their ionic diameters. Stern/Graham model This model, shown in Figure 2-2, divided the double layer into two parts, a) a fixed layer of strongly adsorbed counter ions, adsorbed at specific sites on the surface; and b) a diffuse layer of ions similar to the Gouy-Chapman model. The fixed layer of ions is known as the Stern layer and the potential decays rapidly and linearly in this layer. The potential decay is much more gradual in the diffuse layer. In case of specifically adsorbing ions (multivalent ions, surfactants) the sign of the stern potential may be reversed.
15 + + + 0 + solution + + + + + + + + + a) surface b) Figure 2-2: The Stern-Graham model of the electrical double layer. (a) Distribution of counterions in the vicinity of the charged surface, (b) Variation of electrical potential with distance from the surface. Zeta Potential (Electrokinetic Potential) Zeta potential is the potential of the surface at the plane of shear between the particle and the surrounding media as the particle and media move with respect to each other. In the presence of an applied electric field, the charged surface (and the attached material) tends to move in the appropriate direction, while the counterions in the mobile part of the double layer would have a net migration in the opposite direction. On the other
16 hand, an electric field would be created if the charged surface and the diffuse part of the double layer were made to move relative to each other. The plane of shear is beyond the Stern plane, and the zeta potential facilitates easy quantification of the surface charge. The pH at which the calculated zeta potential value is zero is known as the iso-electric point (IEP). Measurement of zeta potential The zeta potential of a particle is calculated from electrokinetic phenomena such as electrophoresis, streaming potential, electroosmosis, and sedimentation potential. Each of these phenomena and the determination of zeta potential from these techniques is discussed briefly in this section. Electrophoresis Electrophoresis is the movement of charged surface along with the adsorbed ions, in relation to a stationary liquid under the influence of an applied electric field. The electrophoresis cell consists of a horizontal glass tube with inlet and outlet taps and an electrode at each end. Platinum black electrodes are used for salt concentrations in the range of 10 -3 to 10 -2 mol dm -3 . Otherwise, appropriate reversible electrodes (such as Ag/AgCl or Cu/CuSO 4 ) must be used to avoid gas evolution. The mobility of the particle is viewed under the microscope at a â€œstationary planeâ€ in the cell, where the electroosmatic flow of the liquid caused by the charged surface of the cell is compensated by the return flow of the liquid. For a cylindrical cell the stationary plane is located at 0.2, and 0.8 of the total depth. The exact location depends on the width/depth ratio. Electrophoretic mobility is calculated from the time a particle takes to travel a fixed
17 distance. Zeta potential, (potential at the shear plane) is calculated using the Smoluchowski equation for spherical particles, which can be treated as point charges. E U (2-7) U E = mobility under applied potential (E) permittivity of the electrolyte media Viscosity of the medium Electrophoretic mobility measurement based instruments are more suited for measuring the zeta potential of fine particles. Streaming potential The liquid in the capillary of a porous plug carries a net charge given by the mobile part of the electrical double layer. When the liquid flows through the capillary or the plug, it gives rise to a streaming current and consequently a potential difference. The streaming potential can be measured using a micrometer instead of the electrometer. To minimize electrode polarization, alternating streaming current can be generated by forcing the liquid through the plug by a reciprocating pump. If E is the potential difference developed across a capillary of radius a, and length l, for an applied pressure difference p, then oKE (2-8) where, = permittivity of the medium, = zeta potential, = viscosity of the liquid, and K o = conductivity of the electrolyte solution.
18 Electroosmosis This technique involves the movement of liquid relative to a stationary charged surface (e.g., capillary or porous plug) by an applied electric field. The potential is supplied by electrodes, and the transport of liquid across the tube is observed through the motion of an air bubble in the capillary providing the return flow. For water at 25 o C, a field of about 1500 V/cm is needed to produce a velocity of 1 cm/sec if o is 100 mV. Sedimentation potential Sedimentation potential is the creation of an electric field when charged particles move relative to a stationary fluid. This technique is the least commonly used for measurement of zeta potential, because of several limitations associated with the measurement and calculation of the zeta potential. Calculation of Electrostatic Forces The presence of surface charges leads to the development of a potential gradient in the vicinity of the surface. The Poisson-Boltzmann distribution (PB), assuming ions as point charges, and non-interacting ions, defines the potential distribution as a function of distance from the surface as follows: kTZeZendxdrsinh2022 (2-9) where, Z is the electrolyte valency, e is the elementary charge (C), n is the electrolyte concentration (#/m 3 ), r is the dielectric constant of the medium, 0 the permittivity of vacuum (F/m), k is the Boltzmann constant (J/K), and T is the temperature (K). When solved, the PB equation gives the potential (), electric field (differential of the potential), and counterion density at any distance, x, away from the surface.
19 The Debye-Huckel parameter () describes the decay length of the electrical double layer, and the inverse of the parameter, -1 is known as the Debye length. -1 indicates the distance away from the surface where the distribution of ions in the solution is affected by the presence of a charged surface 2/10222kTenZr (2-10) with all of the symbols explained above. Figure 2-3 shows the variation of the Debye length as a function of the electrolyte concentration for electrolytes with different valence. The extent of the double layer decreases with increasing electrolyte concentration due to the shielding of surface charge, and ions of higher valence are more effective in screening surface charge. Electrolyte Concentration (M) 10-510-410-310-210-1100101 Debye Length ,-1, (nm) 0102030405060708090100 1-1 electrolyte 2-2 electrolyte 3-3 electrolyte Figure 2-3: Variation of the Debye length with electrolyte concentration for different z:z electrolytes.
20 The overlap of similar electrical double layers leads to a repulsive interaction, and is summarized in the next subsection. The overlap of electrical double layers is shown in Figure 2-4. Figure 2-4 b) shows the potential distribution after the overlap of the double layers. The dotted lines show the potential expected from non-overlapping double layers, and the solid line represents the potential due to the overlap. The expected potential due to the overlap is higher, leading to higher counterion concentration, which results in higher osmotic pressure, tending to push the particles apart. a) 0 + 0 + 0 + 0 + b) Figure 2-4: The potential distribution between two approaching surfaces, a) before overlap of electrical double layer, b) after overlap of the electrical double layer (overlap leads to higher potential between the planes, resulting in repulsion)
21 The linearized poisson-boltzmann approach To calculate the repulsive interaction due to the overlap of the EDLâ€™s, the PB equation (Eq. 2-9) needs to be solved numerically, which is difficult and computationally intensive. An easier approach is to produce analytical formulas using a series of approximations. The PB equation can be written in terms of series expansion terms. For low surface charge, and low ionic concentration condition, even the first term of the expansion results in more or less accurate estimate of the repulsive interaction. Analytical formulas Based on the simplifications, analytical formulas can be given for the calculation of the electrostatic energy of interaction between two flat surfaces. The formulas for two different models, constant potential and constant charge are given below: Constant Potential [DER87]: 2tanh112)(22/HkTekTnZHsifltflt W (2-11) Constant Charge [DER87]: 12coth12)(22/HkTekTnZHsifltflt W (2-12) where, W flt/flt is the energy between two flat plates (J/m 2 ), H separation distance (m), Z valency, n i ion concentration (#/m 3 ), k Boltzmann constant (J/K), T temperature (K), e electron charge (C), s stern potential (V), Debye-Huckel parameter (m -1 ). For large separation distances, W W and is given by the equation [DER87]: HspltpltpltpltekTZenkTHWH4tanh64)()(2// W (2-13)
22 Once the individual components of the DLVO forces can be computed, the DLVO theory can be used in order to understand and predict the stability of colloidal particles under different processing conditions. DLVO Theory and Colloidal Stability As mentioned in the beginning of this chapter, the DLVO theory assumed that the only force acting on particles in aqueous medium are the electrostatic and van der Waals forces. If the net interaction is repulsive, then the particles are expected to be dispersed and remain in suspension. The individual components of the DLVO interactions are plotted in Figure 2-5, for a given set of conditions (Hamaker constant A = 5 * 10 -20 J, Potential = 25 mV). As can be seen from the figure, electrostatic forces are repulsive in nature and the van der Waals forces are attractive. Figure 2-5: The van der Waals attractive force, and electrostatic repulsive forces as a function of separation distance -0.3-0.2-0.100.10.20.301020304050Separation Distance (nm)Interaction Force/Radius (mN/m) van der Waals Electrostatics A = 5.3 x 1 0 = 25 mV -20 J Figure 2-6 depicts the net interaction as a function of separation distance between two 0.5-micron particles under the condition shown in the previous figure. The point A
23 in the graph is referred to as the primary minima, point B is the primary maximum, and point C is the secondary minima. The primary maximum depicts the barrier to agglomeration between two particles. The higher the primary maximum, the better is the colloidal stability. Usually an agglomeration barrier of 5-10 kT is enough to keep particles dispersed when the only force moving the particles is the Brownian motion. The height of the primary maximum would increase with increasing charge on the particles, which leads to higher electrostatic repulsion. However, increasing the ionic strength of the solution will shield the surface charge, and thus result in lowering of the primary maximum. Figure 2-6: Total interaction energy as a function of separation distance -40-20020406080100120020406080100120140160180200Separation Distance (nm)Interaction Energy (kT) C B A A = 5.3 x 10 -20 J = 25 mV R = 0.5 m If the interaction is such that the primary maximum is overcome, then the particles coagulate in the primary minimum, and this coagulation is generally irreversible, and an unstable suspension is obtained. However, coagulation of particles can also take place in the secondary minima. The depth of the secondary minima increases with increasing particle size, and hence larger size particles may coagulate in
24 the secondary minima, even thought the primary maximum is very high. Coagulation in the secondary minima can be reversible, and the redispersion can be achieved by applying mechanical forces through agitation, or sonication. In industrial applications, electrostatic stabilization may not be the most adequate, and hence polymer molecules are used to stabilize particles by providing a steric hindrance to the agglomeration of particles. Polymeric Steric Forces for Colloidal Stability The word polymer is derived from Greek with â€œpolyâ€ meaning many and â€œmerâ€ meaning part. According to the IUPAC definition, â€œA polymer is a substance composed of molecules characterized by the multiple repetition of one or more species of atoms or groups of atoms (constitutional repeating units), linked to each other in amounts sufficient to provide a set of properties that do not vary markedly with the addition of one or a few of the constitutional repeating units.â€ Polymers are made from repeating units of chemical species known as monomers, whose typical molecular weight is between 50-100 daltons. Polymers at particle surfaces play an important role in a range of technologies such as stabilization, flocculation, enhanced oil recovery, and lubrication. In order to control and optimize these technologies, it is very important to understand the adsorption of polymers at the particle surfaces. Solution Behavior of Polymers In order to understand the role played by polymers at interfaces in particle processing, it is important to understand the solution behavior of polymers. The two major theories are briefly described below:
25 Flory-Huggins theory This theory calculates the free energy of mixing of pure amorphous polymers with pure solvent. The entropy and the enthalpy of mixing can be calculated separately [HUN01], and: MMMSTHG (2-14) where, is the total free energy change on mixing, MG M H is the enthalpy change on mixing, and is the change in entropy on mixing. MS The entropy of mixing was originally calculated by Flory using a lattice approach, but it can also be derived using a free volume approach. The entropy of mixing, S M can be denoted by Eq. (2-15), 2211lnlnvnvnkSM (2-15) where, k is the Boltzmann constant, n 1 is the number of solvent molecules, n 2 is the number of polymer molecules, v 1 is the volume fraction of the solvent in the polymer solution, and v 2 is the volume fraction of the polymer. The enthalpy of mixing is calculated by considering mixing to be a quasi-chemical reaction between the dissimilar solvent contacts and segment contacts: 1 â€“ 1 + 2 2 = 2 (1 2) (2-16) Eq. (2-16) depicts the formation of two solvent-polymer contacts (1-2) from a solvent-solvent contact (1-1), and a polymer segment-segment (2-2) contact. An interaction parameter 1 is defined where 1 kT is the difference in energy of a solvent molecule immersed in pure polymer, relative to that in pure solvent. For n 1 solvent molecules, the energy change is n 1 1 kT. The probability of a solvent molecule being in
26 contact with a polymer segment, in a polymer solution is v 2 , and hence the enthalpy of mixing, M H can be denoted by the following equation: (2-17) kTvnHM121 Therefore, combining Eqs. (2-14), (2-15), and (2-17), the free energy of mixing can be given by: 1212211lnlnvnvnvnkTGM (2-18) It can be seen that the dominating reason for dissolution of polymer is the increase in entropy of the solvent (e.g., water) molecules. Free Volume theory In the Flory-Huggins theory, 1 was considered primarily as an enthalpy term, it was later shown experimentally that for many non-aqueous polymer-solvent systems, positive values of 1 , that opposes mixing of solvent and polymer are determined by entropic considerations. The enthalpy terms are relatively small, and the sign can vary depending on the conditions [HUN01]. Also, it was found that 1 depends on the polymer concentration, and varies between 0.1-0.5, with 1 becoming more positive as the polymer concentration increases. The Flory-Huggins theory predicts that the mixing would be favored as the temperature increases, since the mixing is entropy driven. But, phase separation, has been observed for most polymer solutions, near the critical point of the solvent. (The critical point of the solvent is defined as the temperature where the transition takes place from good solvent to poor solvent) Most of the shortcomings of the Flory-Huggins theory have been overcome by the free volume theory [HUN01]. The entropic contribution to 1 was attributed to the difference in free volume between the solvent and the polymer. The increase in value of
27 1 with concentration was explained on the basis of the ordering of the solvent molecules on increasing the segment concentration. The phase separation near the critical temperature of the solvent was attributed to the decrease in entropy on mixing the solvent and polymer under these conditions. An important length scale associated with polymer in solutions, is the root mean square radius of the polymer coil in solution. For an unperturbed coil this is known as the unperturbed radius of gyration, R g , and is given by: 660MMlnlRg (2-19) where, n is the number of segments and l is the effective segment length, M is the molecular weight, and M 0 is the segment molecular weight. Eq. (2-19) is valid so long as the solvent is â€œidealâ€ for the polymer, i.e.; there are no interactions, either attractive or repulsive between the segments in the solvent. In real (non-ideal) solvents, the effective size of a coil can be larger or smaller than the unperturbed radius R g , and is sometimes referred to as the Flory radius, R F , where R F = R g , and is the intramolecular expansion factor, which depends on the nature of the solvent. Adsorption of Polymer at the Particle Surface The free energy of the overall process must be favorable, in order for a polymer to adsorb at a particle surface. The different factors contributing to the free energy change, when polymer adsorption takes place are [FLE93]: a) The adsorption energy, due to contacts of the polymer segments with the surface b) The conformational entropy of the chains c) The entropy of mixing of chains and solvent d) The polymer-solvent nearest-neighbor interactions
28 Adsorption takes place only if the energy of a segment-surface interaction is lower than that of solvent-surface interaction (factor a). A quantitative measure for this energy is the dimensionless parameter S , which is defined such that the net effect of the exchange of a solvent molecule on the surface and a segment in the bulk solution is S kT. For polymer adsorption to take place, S has to be positive and the adsorption energy is proportional to the number of adsorbed segments. Polymers bond with surfaces through a variety of mechanisms including electrostatic interactions, hydrogen bonding, hydrophobic interactions and specific chemical bonding, and the free energy for adsorption can be given by the following formula: .......GGGGGobonding-hochydrophobiochemoelecoads (2-20) Depending on the surface chemistry, and the nature and energetics of sites on the surface of the particle, different factors contribute to the adsorption energy. In order to control polymer adsorption on surfaces, the interplay between the different adsorption mechanisms should also be considered. Factors b), and c) represent the entropy loss occurring upon adsorption, and hence can be considered as the opposing forces for polymer adsorption. Factor b) accounts for the reduction of the internal degrees of freedom within the chains when they adsorb. Factor c) is related to the configurational entropy loss, which occurs when the homogeneous polymer solution is separated in a polymer-rich surface â€œphaseâ€ and a solution that becomes enriched with respect to the solvent. Factor d) results from the mutual interaction between segments and solvent molecules. In a poor solvent, the segment-solvent interaction is unfavorable. This forces the polymer out of the solution, promoting adsorption. In a good solvent, segment-solvent
29 interaction is favorable, and the aggregation of segments is unfavorable and as a result the adsorbed amount is less. At equilibrium, in order to minimize the energy, the adsorbed layer consists of polymer chains with several stretches of segments in the surface layer (trains), with the parts connecting the trains sticking out into the solution (loops). Moreover, at the chain ends, freely dangling tails may protrude in the solution. The relative amount of the polymer segments in trains, loops, and tails, depends on the energetics of the polymer adsorption process discussed earlier. Steric Barrier due to Adsorbed Polymer Molecules As two polymer-coated particles approach each other, the outer segments of the adsorbed polymer layers begin to overlap, leading to steric repulsive forces. This overlap usually takes place when the separation between the particles is less than a few R g (introduced earlier, Eq. (2-19)). The repulsive steric force is a repulsive osmotic force due to the unfavorable entropy because of the compression of the chains between the surfaces. This repulsive steric force is critical in particulate processing, because it can be used to stabilize dispersions, which are otherwise unstable. In order for the steric repulsion to be effective, the range of the repulsion should be greater than distances at which van der Waals forces become dominant. The steric repulsive force depends on the quantity or coverage of polymer on each surface, and on the quality of the solvent. Three domains of close approach for sterically stabilized particles can be defined [HUN01], based on the separation distance between the particles, d, and the thickness of the steric layers, L. (Figure 2-7). (i) d > 2L: This domain is referred to as the non-interpenetrational domain, since the adsorbed polymer layers cannot interact, and no steric repulsive forces are observed.
30 (ii) L < d < 2L: When the distance between the surfaces is less than twice the steric layer thickness, interpenetration occurs between the adsorbed polymer molecules on both the surfaces, and this domain is referred to as the interpenetrational domain. Due to the interpenetration, the solvent molecules are driven out of the overlap region, resulting in demixing. In good solvents, this demixing of segments and solvent leads to repulsion, due to the increase in the free energy of the system. (iii) d < L: When the separation distance is less than the adsorbed layer thickness, apart from interpenetration, compression of the adsorbed polymer molecules also takes place. In addition to demixing of solvent and segments, elasticity of the polymer segments being compressed also contributes to the free energy of the system. The elasticity arises from the reduction in the configurational entropy of the adsorbed polymer molecules, and irrespective of the quality of the solvent, leads to steric repulsion. For a detailed mathematical description of the three domains, the reader is referred to reference [HUN01]. i ) iii ) ii ) Figure 2-7: The three domains of sterically stabilized flat plates: (i) Non-interpenetration, (ii) Interpenetration, and (iii) Interpenetration and compression of the adsorbed layer.
31 A direct measure of the interaction energy can be obtained using the atomic force microscopy (AFM). Figure 2-8 depicts the force between a Si 3 N 4 AFM tip, and a flat alumina surface, both with and without ployacrylic acid (PAA) (MW 2,100) polymer. Under the conditions of the experiment (pH 9.8, 0.01 M NaCl), in the absence of the polymer, the electrostatic repulsive forces are not strong enough, and the van der Waals forces become dominant and lead to attractive interaction between the surfaces. In the presence of the polymer, steric repulsive forces are observed between the surfaces, and no net attractive force is measured. 01234051015Separation (nm) Force (nN) pH = 9.8 200 ppmPAA 0.01 MNaCl Figure 2-8: Polymeric steric repulsive forces. Interaction force between Si 3 N 4 AFM tip, and flat alumina substrate at pH 9.8, 0.01 M NaCl, both in the absence and presence of 200 ppm of polymeric dispersant PAA (2100 MW). Dispersion of Nanoparticles Under Extreme Processing Conditions Traditional and emerging technologies such as advanced structural ceramics, controlled drug delivery systems, abrasives for precision polishing, coatings, inks, and
32 nanocomposite materials are increasingly relying on nanoparticulate precursor materials to achieve optimum performance. Concurrently, environmental and safety issues have driven industrial processes towards synthesis and processing in aqueous media. The result has been increased attention to the role of particle size in the dispersion of concentrated (> 50 vol%) aqueous particulate suspensions [HIG90, HAC97, HIR97, HIM98, WU98]. Electrostatic repulsion is often applied to stabilize homogeneous, low ionic strength suspensions where pH can be controlled in order to provide sufficient surface charge. The energy imparted to particles during processing, or Brownian motion is relatively constant as a function of particle size, but the repulsive energy between particles due to electrostatics decreases proportionally with size, and hence a greater surface potential is needed to disperse nanoparticulates than larger particles. This effect may be great enough that pH adjustment or addition of inorganic dispersant, such as sodium silicate, may not always lead to suspension stability. Furthermore, many processes such as chemical mechanical polishing and crystallization, must operate in extremes of ionic strength (> 0.1 M), temperature (> 200 0 C, pH (<2, > 12), mechanical force (>100 mN/m), or high shear rate (> 10,000 s -1 ) where electrostatic stabilization alone may not be adequate to prevent agglomeration. Hence, polymeric reagents are more commonly utilized [TAD82, RUS89, REE95, CES88]. In concentrated suspensions, relatively large polymeric dispersant molecules may significantly increase the effective volume fraction of the suspension, and hence viscosity, due to the quantity of immobilized liquid in the dispersant coating and the relative high surface area of nanoparticles. To minimize this effect, dispersant molecules should be small and bind as little water as possible while still providing an adequate
33 barrier to agglomeration. Since most polymeric dispersants do not form compact layers at the interface, a relatively large molecule is often used (>10,000 MW) and the increase in effective volume fraction is likely to result in higher viscosity [CES88]. Additionally, many polymeric dispersants are polyacrylate based and may perform poorly under extremes of pH and ionic strength. These limitations have motivated the search for alternative dispersants to stabilize suspensions in extreme environments and maintain fluidity in concentrated suspensions. There are a number of necessary criteria that must be met for a molecule to act as a dispersant. The reagent must (i) adsorb to the surface under the given process conditions, (ii) not phase separate or otherwise adhere to the dispersant layer of an approaching particle, and (iii) provide adequate repulsion between particles to prevent agglomeration. Surfactant molecules adsorb at the solid/liquid interface through mechanisms such as electrostatic, hydrogen, hydrophobic, and specific bonding. Criteria two is fulfilled if the hydrophilic moieties of the surfactant molecules adsorbed on separate particles interact with each other as demonstrated by the ability of surfactants to stabilize oil-in-water emulsions. It remains, however, to be shown that surfactant layers can produce a barrier to agglomeration that is large enough to disperse inorganic particles in aqueous media. In the remainder of this dissertation, the utility of self-assembled surfactant films in the dispersion process will be investigated with specific emphasis placed on their effectiveness in extreme environments, such as high ionic strength. An initial result on the use of surfactants to stabilize nanoparticles under high electrolyte concentrations is being presented in Figure 2-9. The turbidity of the suspension after 60 minutes is plotted as a function of the added electrolyte
34 concentration. Higher turbidity means more scattering of light, indicating a stable suspension. It can be seen from Figure 2-9 that using a trimethylammonium bromide surfactant with twelve carbon atoms in the hydrocarbon tail, a suspension of 200 nm sol-gel silica particles can be stabilized under salt concentrations as high as 5 M NaCl. Thus, it is very important to investigate the unique dispersion properties of the surfactant molecule. The rest of the dissertation is dedicated to understanding the dispersion mechanism of nanoparticles in the presence of self-assembled surfactant aggregates by looking into the self-assembly process using several different interface characterization techniques. NaCl Concentration (M) 012345 Turbidity (after 60 minutes) 200400600800 Figure 2-9: Surfactant as dispersant under high electrolyte concentration. Stabilization of 200 nm silica particles using a trimethylammonium bromide surfactant.
CHAPTER 3 SURFACTANTS AT THE SOLID/LIQUID INTERFACE Introduction A surfactant (surface active agent) is a substance that, when present at low concentration in a system, adsorbs onto the surface or interface, and alters to a marked degree the surface or interfacial free energies. Surface-active agents have a molecular structure consisting of two distinct groups, the head and tail. The tail is a structural group that has very little attraction for the solvent, and is known as a lyophobic group. The head has strong attraction for the solvent, and is called the lyophilic group. This kind of structure, consisting of lyophobic and lyophilic groups, is known as the amphipathic structure. When a surface-active agent is dissolved in a solvent, the presence of the lyophobic group may cause solvent molecules to form an ordered structure, thus reducing the entropy of the solvent molecules, and thus increasing the free energy of the system. As a result, the surfactant molecule has a tendency to preferentially partition out of the bulk at the interface. Hence, less work is required to create unit area of surface, resulting in the lowering of the interfacial energy or tension. The surfactant however cannot be completely removed from the system, since it has a lyophilic group, and complete expulsion of the surfactant would involve dehydration of those groups, which is energetically unfavorable. The chemical structures of groups suitable as lyophobic and lyophilic vary with the nature of the solvent, and the conditions of use. In a highly polar solvent, such as water, the lyophobic group may be hydrocarbon or fluorocarbon or siloxane chain of proper length, and the lyophilic groups may be an ionic or highly polar 35
36 group. Depending on the nature of the hydrophilic group, surfactants are classified as i) anionic (net negative charge), ii) cationic (net positive charge), iii) zwitterionic (both positive and negative charges may be present), and iv) nonionic (no apparent ionic charge). Surfactant adsorption is a phenomenon of critical importance to various industrial processes ranging from ore flotation, lubrication, paint technology, colloidal stability, detergency, pharmaceutics, agricultural soil conditioning, emulsion polymerization, to enhanced oil recovery [ROS89]. The process of and factors affecting surfactant micellization in bulk solutions leading to spherical or cylindrical micelles, bilayers, or bicontinuous phases are relatively well understood [ROS89]. At interfaces, however, the self-assembly process is influenced by additional surfactant-surfactant, surfactant-surface, surfactant-solvent, and solvent-surface interactions, including the free energy of adsorption, roughness, surface heterogeneity, charge, and crystallinity. Adsorption of Surfactants at the Solid-Liquid Interface Traditionally, adsorption isotherms combined with techniques such as contact angle and zeta potential have been used to delineate the self-assembly behavior of surfactants at the solid/liquid interface [FUE64, FUE75, SCA81a, SCA81b, MOU88, HOU83, SOM86, SOM87]. Mechanisms of Adsorption Some of the most common mechanisms for surfactant adsorption are: a) Ion Exchange: This mechanism involves the replacement of counterions adsorbed on the surface by similarly charged surfactant ions, b) Ion Pairing: Involves the adsorption of surfactant ions from solution onto oppositely charged sites unoccupied by counterions., c) Acid-Base Interaction: The adsorption can take place via hydrogen bond formation
37 between the surfactant and the surface, and d) Hydrophobic bonding: occurs when adsorption of surfactant molecule from the liquid phase takes place onto or adjacent to other surfactant molecule already adsorbed on the surface. Contributions to the Adsorption Energy The contributions to the total free energy of adsorption can be divided into several different components as follows [MOU88]: 0/000000desolvsolvscccchemhhelecadsGGGGGGG (3-1) where, represents the electrostatic contribution, is the contribution from hydrogen bonding, is the possible chemical interactions between the substrate and surfactant, such as covalent bond formation, and acid-base interactions, is the contribution from the lateral interaction between the adsorbed hydrocarbon chains. represents the interaction between hydrocarbon chains and nonpolar solids (hydrophobic bonding). is the change in energy due to solvation/desolvation of the surface and the surfactant species upon adsorption. It should be noted that the various contributions to the total free energy are not independent, and the interplay between them can be important. 0elecG 0hhG 0chemG0/desolvsolvG 0ccG 0scG Depending on the particle-surfactant system, one or more of the above contributions can be responsible for adsorption. The dominating contribution would depend on the nature and concentration of the surfactant, the surface chemistry of the particle, and solution properties such as pH and ionic strength. Electrostatic and lateral interaction forces are usually the major factors determining the adsorption of surfactants
38 on oxides and other nonmetallic minerals. Chemical interactions become more dominant for surfactant adsorption on salt type minerals such as carbonates and sulfides. Electrical interactions As discussed earlier, most solids in aqueous solution exhibit a net surface charge depending on the pH and ionic strength of the solution. The surface charge results in electrochemical potential in the vicinity of the surface. Oppositely charged surfactant molecules are attracted to the surface, and adsorb via electrostatic interactions. To maintain electroneutrality, the adsorption of ionic surfactants occurs through either exchange with co-ions in the double layer, or with an equivalent co-adsorption of counterions. The electrical interactions become important in the case of ionic surfactants, and can be broken down into two contributions [MOU88], columbic and dipole: (3-2) 000dipcoulelecGGG where, sjnjdipEG 0 (3-3) and (3-4) scoulzFG0 In the dipole term, nj is the change in the number of adsorbed molecules (dipoles) j; j is the dipole moment of j, and E s is the field strength at the Stern plane. For the adsorption of ionic surfactant, only the dipole moment of water needs to be considered. The dipole term arises due to the replacement of water dipoles by surfactants upon adsorption according to the following scheme [MOU88]:
39 (Surfactant) solution + n(H 2 O) adsorbed (Surfactant) adsorbed + n(H 2 O) solution (3-5) In the Coulombic term, z is the valency of the ionic surfactant, s represents the potential at the Stern plane in the electrical double layer, and F is the Faraday constant. Zeta potential can be used in place of s without introducing any significant errors in the calculations if there is no specific adsorption. At the iso electric point (IEP), the contribution from the coulombic term is expected to be negligible. In the absence of other contributions, Coulombic adsorption should proceed to the point where the adsorption reduces the zeta potential of the particle to zero. Any adsorption beyond that can be attributed to other types of interactions. Figure 3-1 [CHA87] shows the adsorption of dodecylsulfate on alumina. The adsorption of the negatively charged surfactant on initially positively charged alumina surface decreases the zeta potential, reducing it to zero. However, it is interesting to note that the adsorption continues to increase, although at a slower rate, even when the zeta potential is negative and the surface has the same sign as the oncoming surfactant molecules. This clearly indicates that even in systems where electrostatics is the dominant mechanism for adsorption, adsorption may proceed through other mechanisms such as chain-chain interaction between the adsorbed surfactant molecules. Experimental evidence suggests that the electrostatic term in Eq. (3-2) can be overridden by the other terms. In fact, flotation of most sulfide minerals is accomplished with anionic surfactants under conditions where the mineral is negatively charged, i.e., the surfactant and the surface have the same charge and electrostatic interaction is unfavorable for adsorption. For example, in the case of adsorption of fatty acids or alkylsulfonates on oxides or salt-type minerals, plays a dominant role in the 0chemG
40 adsorption processes, and adsorption can take place irrespective of the zeta potential of the surfaces. Figure 3-1: Adsorption and zeta potential behavior of dodecylsulfate-alumina system. (After reference [CHA87]) Assuming that Stern potential can be approximated by the zeta potential, can be calculated using Eq. (3-4). Table 3-1 [MOU88] lists the value of for different values of the zeta potential. 0coulG 0coulG
41 Table 3-1: Calculated values of from Equation (3-4) for different values of the potential. [MOU88] 0coulG Zeta Potential (mV) 0coulG (kcal/mole) 10 0.23 20 0.46 30 0.69 40 0.92 50 1.15 60 1.38 As depicted in Table 3-1, the energy involved is very small, and even at very high zeta potentials such as 60 mV, is only 1.38 (kcal/mole). This value is much smaller than the contribution (7 kcal/mole for hydrocarbon chain with 12 C atoms), hydrogen bonding (4-8 kcal/mole), and chemical interaction. Eq. (3-4) may underestimate the value of , because of the fact that the double layer theory assumes the surface to be uniformly charged. A better model would be to consider the presence of localized charges, which will increase the contribution of [LEV67]. 0coulG0coul 0ccG G 0coulG Lateral chain-chain interactions At a certain surfactant concentration, there is a sharp increase in the slope of the adsorption isotherm, clearly indicating an enhanced affinity of the surfactant for the surface. Increase in adsorption is accompanied by a sharp rise in the zeta potential, and in some cases, a reversal in the zeta potential is observed. These observations have been attributed to lateral interactions between the adsorbed surfactant tails, which lead to the formation of self-assembled two-dimensional structures known as â€œhemimicellesâ€. The formation of hemi-micelles due to hydrophobic interactions between the surfactant tails
42 was first proposed in 1955 by Gaudin and Fuerstenau [GAU55, FUE64, FUE75], and later verified by Somasundaran, and co-workers [CHA86, CHA87, WAT86]. In order to form hemimicelles, a certain critical amount of the adsorbed surfactant is required, which is known as the â€œcritical hemimicelle concentrationâ€. The hemimicelle concentration depends on the lateral interaction between the surfactant tails, which is dependent on the surfactant structure (number of CH 2 or aromatic groups, arrangement of atoms in the hydrocarbon chain). Additionally, the hemimicelle concentration is also dependent on the solution conditions (pH, ionic strength, and temperature). The energy for lateral chain-chain interactions, has been proposed to be proportional to the number of CH 0ccG 2 groups in the surfactant (n) according to the equation [MOU88]: (3-6) nGcc0 where, is the cohesive energy per mole of CH 2 groups. This factor has been calculated to be about 0.6 kcal/mole (~1 KT) per CH 2 group [SOM64], and hence for surfactants with chain lengths of 12 to 18 carbon atoms, the would be 7 to 10 kcal/mole. This energy contribution is significantly higher than the electrostatic contribution, and would explain the increased adsorption in some systems, even when the surfactant is similarly charged to the substrate. 0ccG Chemical interactions The chemical contribution may result from interactions such as covalent or complex bond formation, between the surfactants and the surface sites. Surfactants such as fatty acids, alkylsulfates, alkylsulfonates, amines, and alkylhydroxamate have been proposed to adsorb by means of chemical interaction on a variety of particles. Also, surfactants containing hydroxyl, phenolic, carboxylic, and amine groups can hydrogen
43 bond with the surface sites. Infrared spectroscopy has been used to understand the chemisorption of surfactants at the surface, by examining the shift in the characteristic peaks of the surfactants upon adsorption. Surfactant can react with the dissolved ionic species in the bulk to form insoluble complexes. It has been proposed that adsorption of the surfactant can take place through similar reaction with ions on the surface, or by surface precipitation, under conditions where, no bulk precipitation takes place, but the interfacial concentration is high enough to exceed the solubility product in the interfacial region [ANA85]. Over the past ten years, several new techniques have become available, and advances in more established techniques have been made, for the study of surfactants in bulk media, and at interfaces. Surfactant structures at the solid-liquid interface have been utilized to stabilize particulate dispersions [COL97, SOL99, KOO99, BRE99, EVA96], but in all these studies the dispersion in the presence of adsorbed surfactant has been attributed to electrostatic repulsion due to the formation of bilayer structures at the solid-liquid interface, at high surfactant concentrations. However, as illustrated in Figure 2-9, surfactants can be used to stabilize suspensions at salt concentrations as high as 5 M. Under these conditions, the charge due to adsorbed surfactant is negligible, indicating that under high electrolyte conditions, electrostatics is not the primary stabilization mechanism. The measurement of the structure of the adsorbed layers of surfactants is critical to developing an understanding of the stability of particulate dispersions. In this chapter, some the newer techniques for characterizing surfactants at the solid-liquid interface are discussed, in order to summarize the state of the art in terms of surfactant
44 structures, and self-assembly at the solid liquid interface. The techniques are divided into three categories: 1) Microscopic, 2) Optical, and 3) Non-optical techniques. Characterization Techniques Involving Microscopy Scanning Tunneling Microscopy Scanning tunneling microscopy (STM) allows the imaging of surfaces in-situ on the molecular scale, with a lateral resolution of 0.2 nm. The instrument is composed of a sharp metallic tip, and an electronically conducting solid, which may have an adsorbed layer of the surfactant molecules. The position of the tip, both normal, and parallel to the surface is controlled very accurately. If the electronic orbitals of the outermost atoms of the two solids overlap, a tunneling current flows, under the application of an electric potential difference between the two surfaces. Typically, the tunneling current decreases by about one order of magnitude for every angstrom increase in the gap. This high sensitivity of the current to the gap width enables the control of the width by a feedback loop. Scanning the tip across the surface allows its topography to be obtained in two ways. If the tip height is adjusted with the feedback loop maintaining a constant current, the image obtained is a map of the tip height z versus the lateral coordinates x and y. Alternatively, if the tip is scanned at constant height, the current is monitored as a function of x, and y. The tip material varies from tungsten to platinum/iridium alloys, while the solid substrates used so far include platinum, gold, or silver evaporated onto mica, and several cleaved layered conductors like highly oriented pyrolytic graphite (HOPG), and molybdenum selenide (MoSe 2 ). The technique has been used to investigate the structure and dynamics of adsorbed or deposited layers of materials, which include alkanes, alkanols, and carboxylic acids [MAT93, ZLA95, CUN96, ZHA96, STE97].
45 However, due to the need for conducting substrates, the use of this technique for materials such as oxides, which are used for industrial dispersions, is limited. Another microscopy technique, which does not suffer from these limitations, is the atomic force microscope (AFM). Atomic Force Microscope (AFM) In the last few years, the atomic force microscope (AFM) has become the technique of choice for directly imaging surfactants adsorbed at the solid/liquid interface. This technique can be used for a wide range of substrates, both conducting, and non-conducting, under a wide variety of environments. Apart from topography information, the AFM can also be used to directly measure the interaction forces between substrates. The technique involves controlling the interaction between a tip attached to a cantilever, and a flat substrate mounted on a piezoelectric crystal. The piezoelectric crystal enables the movement of the substrate both in the vertical, and the lateral direction. The cantilever is usually made from silicon, with a bypyramidal tip at the end. The typical radius of the tip is 10 nm. The measurement of forces is done in a mode referred to as the â€œcontact modeâ€, and this mode is also used for mapping surfaces. This involves bringing the substrate in contact with the tip at the end of the cantilever. Depending on the interaction forces between the tip, and the substrate, the cantilever either bends away (repulsive forces), or bends towards the substrate (attractive forces). The deflection of the cantilever is detected using position sensitive photodiode detectors. After the interaction is measured, the sample is advanced in the y direction, and the measurement repeated. The deflection of the cantilever is maintained constant by adjustment of the tip-sample force. This provides a three-dimensional set of x, y, and z displacements for the image. This imaging mode is
46 good for hard substrates such as mica, oxides, and graphite, where bringing the substrate in contact with the AFM tip does not alter the nature of he surface. Non-contact modes of imaging are commonly used for soft samples, such as adsorbed surfactant/polymer layers. One of the most commonly used modes for imaging surfactant structures at the interface is the â€œsoft-contact modeâ€ [MAN94, MAN95, AKS96, DUC96]. In this technique the interaction between the substrate coated with the surfactant structures, and the AFM tip is measured. A typical force profile is plotted in Figure 3-2. As seen for the figure, the interaction between the surfactant coated surface and the AFM tip is repulsive in nature. As the separation distance is decreased, an increase in the repulsive force is observed, till a maximum repulsive force, F max is obtained. If the separation distance is further decreased, the forces become attractive, and the tip jumps into the surface. In the soft-contact mode, the separation between the substrate, and the tip is adjusted near the separation distance shown by the dashed line on Figure 3-2, where the forces are still repulsive in nature. Unfortunately, in this method the forces at the point of measurement are weaker than when in contact, and hence the resolution of the images thus obtained is decreased. Over the last decade, AFM is increasingly being used for measuring the forces between surfaces covered with adsorbed surfactant layers [PAS81, PAS88, KEK89, RUT94]. Rutland, and Parker [RUT94] have measured the forces between glass surfaces in cetyltrimethylammonium bromide (CTAB) surfactant solutions at pH 10. They found that initially the interaction forces between the bare surfaces is repulsive, but as the concentration is increased, attraction due to hydrophobic interactions are observed. As
47 the surfactant concentration is further increased, repulsive forces are again observed. This was explained on the basis of the formation of surfactant bilayers at concentrations beyond the bulk critical micelle concentration, on the surface of glass, which resulted in electrostatic repulsive forces. It is interesting to note that in all the studies reported on force measurements, the repulsive forces at high surfactant concentrations have been attributed to electrostatic repulsion due the formation of bilayers on the substrates. Separation Distance (nm) 05101520 Interaction Force / R (mN/m) 050100150200250 Fmax Figure 3-2: Force-distance profile measured between the AFM tip, and a substrate with self-assembled surfactant aggregates One of the first papers on the imaging of surfactant aggregates at the solid-liquid interface at high surfactant concentrations was published by Manne and Gaub in 1995 [MAN95]. They studied the structure of the self-assembled aggregates of trimethlyammonium bromide surfactants on the surface of mica, and silica, and reported that there is a strong influence of the surface in controlling the aggregate structure. For example, in quaternary ammonium surfactant systems, full cylindrical micelles tend to form on mica, that meander across the surface with changes in direction corresponding to
48 the mica lattice structure. Amorphous silica, on the other hand, lacks these atomic rows and as such, spherical structures are observed. On a hydrophobic substrate such as graphite, half cylinders were observed. They also reported that the formation of surfactant structures at the interface is controlled by the structure, or the geometry of the surfactant molecule. For example, on mica, bilayers structures for a double-chained surfactant (didodecyl dimethlyammonium bromide) were observed, whereas single chain surfactants form cylindrical aggregates. Since then numerous papers have been published [MAN94, MAN95, DUC96, AKS96, WAN97, DUC99, PAT99, FIE99, WAL99, VEL00] on these structures and their dependence on solution conditions, surfactant type, and time. Partick et. al [PAT99] investigated the equilibrium structures of adsorbed films of quaternary ammonium surfactants on mica. It was found that depending on the chain of the surfactant, the headgroup, and the associated counterion, the shape of the aggregates could be either spherical, or cylindrical. For example, dodecyl trimethylammonium bromide surfactant forms cylinders on the surface of mica, and by replacing the methyl groups in the headgroup by ethyl groups, the structure of the aggregate changes to spherical. Also by changing the counterion from bromide to chloride, spherical aggregates of the dodecyl trimethylammonium surfactant could be observed. Wall, and Zukoski [WAL99] investigated alcohol-induced transformations of cetyl trimethylammonium bromide (CTAB) aggregates adsorbed on the surface of mica. In CTAB solutions of fixed concentration, increasing amounts of alcohol spread the lateral spacing the CTAB cylinders on the mica surface. At high alcohol concentrations, relative to the alcohol chain length, the surface aggregates changed from a cylinder to spherical aggregates. The structural transformation was explained by hydrophobic
49 interactions between the alcohol, and the surfactant micelles. Another study by Ducker, and Wanless [DUC99] has examined the effect of binding site competition, or the adsorption of salts such as HBr, KBr, and N(CH 2 CH 3 ) 4 Br on the adsorption of CTAB on mica surface. In the absence of salt, at twice the critical bulk micelle concentration, CTAB initially forms cylindrical surface aggregates. The cylinders transform to flat bilayer structure within 24 hours. The introduction of 10 mM K + produces cylindrical aggregates that are stable, and a further increase in the concentration of K + produces defects in the cylinders. More defects were observed by introducing H + ions than K + (at the same concentration). This effect was explained on the basis of the higher affinity of the H + ion for the mica surface. In the case of K + ions, at a critical repulsive force, the defective cylindrical structure changes to a spherical or flattened disklike structure. These investigations using the AFM reveal that the self-assembly of the surfactants at the solid-liquid interface is dependent on the nature of the surfactant molecule, and the nature of the underlying substrate. Surfactants with larger chain lengths, or smaller head groups tend to form more compact surfactant aggregates. Additionally, the nature of the substrate, and the density of the binding sites affect the self-assembled surfactant structure. Although, the AFM provides insight about the shape of the surfactant aggregates, it is limited to very high surfactant concentrations (> 2 CMC), because of the inherent instability of the surfactant aggregates at lower concentrations under the AFM experimental conditions. Brewster Angle Microscopy Since this technique is non-perturbative in nature, it is very useful for probing the surface of soft or fluid interfaces, both with respect to depth, and within their plane.
50 Light, X-ray, and neutrons can be used as the incident radiation, and two types of experiments can be carried out for each of these radiations. In a reflectivity experiment, the scattering vector (difference between the wave vectors of the incident, and reflected beam) is perpendicular to the interface, allowing information to be gained in the structure normal to the surface. In a scattering experiment, the scattering vector (difference between the incident, and the scattered beam) has a component parallel to the interface, so that the structure within the plane of the interface can be probed. The resolution is of the order of the wavelength of the radiation, being a few angstroms for X-rays, and neutrons, but only a few microns for light. The intensity of the radiation obtained in the experiments is strongly dependent on the properties of the interfacial layer adsorbed at the interface. If the optical anisotropy of the interface (due to molecules having a particular orientation), and interfacial roughness (due to thermal fluctuations, or inhomogeneities) are taken into account, this technique can be used to probe adsorbed surfactant films on scales larger than a few microns. Using a high-powered laser, the liquid-vapor interface is illuminated with light polarized in the perpendicular orientation. The reflected light, and part of the scattered light is collected using a microscope objective, and an image of the surface is formed using a sensitive video camera. Images in focus on a strip which is a few tens of microns wide with an in plane resolution of 5-10 m can be pieced together creating large images [TEE97, VOL97, MEL97, MEL98, VID98, LAU98]. The Brewster angle microscopy technique has been used for investigating phase transitions in monolayers of surfactants at the air-water surface. A recent study by Vollhardt and Melzer [MEL97] has evaluated the experimental conditions required for
51 the occurrence of a first order phase transition in a Gibbâ€™s monolayer. They showed that for the same surfactant, the morphological features of the condensed phase structures are very similar to those observed in Langmuir films. Optical Characterization Techniques Electron Spin Resonance Spectroscopy Although Electron Spin Resonance (ESR) spectroscopy has been used for some time for in situ characterization of polymer and surfactant assemblies in solution [TOW95, LI97], it has been employed more recently to investigate adsorbed layers on solid particles dispersed in liquids [MAL92, KRI96]. ESR provides information both about the polarity of the microenvironment, and on the orientation, and mobility of the probe being used for the studies. An ESR spectrum arises from transitions between spin levels of unpaired electrons in an external magnetic field by absorption of microwave radiation. Substances that show ESR spectra include free radicals, odd electron molecules, and paramagnetic transition metal ions and their complexes. In ESR spectra, the three important parameters in the spectral line are the intensity, width, and the fine structure. ESR absorption is proportional to the concentration of the free radical, or the paramagnetic material present. The probe is usually a molecule containing nitroxide radical, and is used at very low concentrations. The width of an ESR resonance depends on the relaxation time of the spin state under study. Since the relaxation times are short in solids, and relatively long in liquids, and are inversely related to the line width, their measurement provides information on the chemical nature of the environment.
52 A unique feature of the ESR spectra is that the line position and splitting depend on the direction of the magnetic field relative to the molecular axis of the probe. This phenomena, called spectral anisotropy, gives rise to a different spectra depending on whether the probe is immobilized (in a solid-like environment), or is rotating randomly in a liquid-like environment. The rotation mobility of the probe depends on the polarity, and the microviscosity of the surrounding media. Thus, this technique can be used to determine if the probe is near the polar headgroups in a micellar aggregate, or is embedded in the hydrophobic interior. Krishnakumar, and Somasundaran [KRI96] have utilized the ESR spectroscopy to probe the mechanism of the simultaneous adsorption of surfactant, and water onto alumina particles dispersed in oil, and in elucidating the relationship between the adsorbed layer structure, and the stability of the suspension to flocculation. Using a 7-doxyl octadecanoic acid, in which the ESR-sensitive moiety is attached to the seventh carbon atom of the C 18 alkyl chain, ESR spectroscopy was used to determine in situ, the changes in the organization of the adsorbed layer. Based on the ESR measurements they were able to determine that the alumina surface interacts primarily with a layer of water (of varying thickness) with the surfactant forming a second layer oriented with the headgroup in water, and the hydrophobic chains in oil. Fluorescence Spectroscopy Fluorescence probes provide a tool for determining the microenvironment of micelles in terms of their polarity. The dependence of the fluorescence response on the solvent properties enables the determination of the environment. Symmetrical aromatic molecules such as pyrene show marked solvent dependence in the fine structure of their emission spectra. The change in the relative intensities of the first, and third bands (I 3 /I 1 )
53 of the pyrene spectra shows the greatest sensitivity to solvent properties, and hence is utilized to determine the polarity of the surrounding microenvironment. Qualitatively, (I 3 /I 1 ) decreases as the solvent polarity increases. Pyrene fluorescence has been used to study surfactant (sodium dodecyl sulfate (SDS)) adsorption at the surface of alumina particles [SOM86] from an aqueous solution. It was found that the (I 3 /I 1 ) ratio exhibited a sharp increase, in the region corresponding to the formation of hemi-micelles (region II, Figure 3-1). At saturation adsorption (region 4, Figure 3-1) (I 3 /I 1 ) value is similar to that obtained within bulk micelles of SDS, indicating full aggregation of the surfactant species on the solid surface. However, at saturation, bulk micelles are also present, and can contribute to the signal. Fluorescence probes are also used to determine the microviscosity of the surrounding medium, and the concept is very similar to the ESR spectroscopy described in the above sub-section. The fluorescence method has also been used to determine the size of microenvironments such as micelles, and also surface micellar aggregates [SOM87, FAN97]. This essentially involves measurement of the extent and kinetics of specific photochemical reactions in the environment of interest. Photochemical reactions followed are usually diffusion controlled quenching reactions. When a fluorescence probe is excited by a short nanosecond pulse of light, its decay is enhanced in the presence of molecules that act as quenchers. The lifetime of the probe under quenching conditions is determined by the concentrations of the probe, as well as the rate at which they diffuse, and encounter each other. Kinetic analysis of the fluorescence decay profiles can therefore provide information on the local concentration of the reactants, and hence the
54 size of the local environment. The kinetic analysis was utilized [SOM87] to determine the aggregate sizes of SDS on alumina surface (Figure 3-1). The aggregates in Region II were of relatively uniform size, while in Region III, a marked increase in the aggregate size was observed. In region II, the surface is not fully covered, and enough adsorption sites remain. The fact that the aggregate size in region II is uniform indicates that adsorption takes place by the formation of more aggregates, but of the same size. Region III occurs through the growth of the existing aggregates, rather than through the formation of more aggregates. This is possible due to the hydrophobic interaction between the hydrocarbon tails of the surfactant molecules. Fourier Transform Infrared Spectroscopy/Attenuated Total Internal reflection (FTIR/ATR) Though the FTIR can be used in several different modes, the most relevant method for studying surfactants at the solid liquid interface is the internal reflection technique, and hence will be discussed in detail in this subsection. The FTIR/ATR technique has been used in this study to determine the orientation of the adsorbed surfactant molecules (for experimental details see Chapter 4). Internal reflection spectroscopy is the technique of recording the optical spectrum of a sample material that is in contact with an optically denser but transparent medium and then measuring the wavelength dependence of the reflectivity of this interface by introducing light into the denser medium [HAR67]. According to Harrick, who developed the fundamentals of the technique, â€œTotal internal reflectionâ€ TIR is a familiar phenomenon. It can be observed with a glass of water for example. If the side of the glass below the water level is viewed obliquely through the water surface, it appears to be completely silvered and one can no longer see objects behind it. The reason for this is
55 that light striking the glass surface is totally reflected and therefore does not pass through the surface to illuminate these objects. However; if one looks at ones fingers touching the glass, the skin patterns are clearly evident which indicates that total reflection has been destroyed where contact is made, viz., at the ridges of the skin but not at the valleys of the skin where no contact is made. This clarity is explained by the penetration of the electromagnetic field into the rarer medium a fraction of a wavelength beyond the reflecting surface and that when a suitable object is brought near enough to the surface to interact with this penetrating field, total reflection is destroyed. This phenomenon is explained by the formation of a standing wave in the denser medium perpendicular to the reflecting surface, and an evanescent, non-propagating field in the rarer medium, the electric-field amplitude of which decays exponentially with distance from the surface. So, in the case of internal reflection spectroscopy, reflectivity is a measure of the interaction of the evanescent wave with the sample material, with the resulting spectrum being characteristic of said sample material. The ATR technique developed simultaneously, and independently by Harrick [HAR60], and Fahrenfort [FAH61], involves placing the sample in contact with an internal reflection element (IRE) of high refractive index. Infrared radiation is focused onto the edge of the IRE, reflected through the IRE, and then directed towards a detector through infrared mirrors. A schematic of an experimental setup is shown in Figure 3-3.
56 Figure 3-3: Horizontal ATR sampling illustrating the parameters of significance (I = incident radiation, R = reflected radiation) Although complete internal reflection occurs at the sample/IRE interface, an evanescent wave is generated at the interface, which penetrates the sample to a short distance (d p ) depending on the wavelength, , of the incident light. This evanescent wave can be absorbed within the sample. An absorption spectrum of the sample in contact with the IRE can thus be generated. The spectrum depends on a number of factors such as the angle of incidence (), the refractive index of the IRE material (n 1 ), and the refractive index of the sample (n 2 ). The penetration depth is defined as the distance required for the amplitude of the electric field to fall to 1/e of its value at the surface: 212212nnSinndp (3-6) The incident infrared light passes through the optically denser IRE and reflects at the surface of the sample. The propagating light passing through the IRE then forms a standing wave perpendicular to the total reflecting surface. Thus, if the sample absorbs the radiation, the radiation becomes attenuated. The reflectance of the attenuated wave can be expressed as: R = 1 â€“ d e (3-7)
57 Where d e is the effective layer thickness and the absorption coefficient (absorptivity) of the layer. The energy loss in the refractive wave is termed attenuated total reflection. For multiple reflections (N), the total reflectance, R N , is expressed as: R N = 1 â€“ N d e (3-8) The effective layer thickness, d e is defined as the thickness required in transmission measurements to obtain the same absorbance as that from a single reflection at the phase boundary of the sample. d e = a/(3-9) where, the absorption parameter, a, is defined as the reflection loss per reflection: a = (100 â€“ R)% (3-10) Several investigators have used the FTIR/ATR technique to study surfactant adsorption at the solid-liquid interface. Jang, and Miller [JAN93] have used the ATR technique to study the in situ adsorption density of Langmuir-Blodgett films at the IRE surface. They also developed equations to quantify the adsorption density on the surface of the IRE. Sperline et. al [SPE87] have developed a method, and the necessary equations for quantitative spectral analysis of surfactants adsorbed onto the IRE. A recent paper published by Sperline et. al [SPE97] describes the temperature dependent structure of adsorbed sodium dodecyl sulfate at the alumina/water interface, using polarized beam infrared spectroscopy. Kung, and Hayes [KUN93] have examined the adsorption of cationic surfactants on silica, to identify the structural changes that occur as a function of surfactant surface coverage, and state of wetttability of the surface.
58 Ellipsometry The ellipsometer works on the principle of reflection of polarized light from a surface. Light can be polarized either in the plane of incidence or parallel to it, and the amount of reflection and phase shift given to the incident light depends upon its orientation. If the index of refraction (n) and extinction coefficient (k) for a material are known, one can measure the thickness of a film of the material on top of a bulk substrate or even another film. By using different angles of incidence, it is possible to determine the film thickness without knowing the index of refraction or extinction coefficients. When considering the effects of reflection on incident light, the magnetic portion of the wave is usually ignored. The incident light can have two different components of the electric field vector, Ep, and Es. The electric field labeled Ep oscillates in a plane parallel to the plane of incidence. The electric field labeled Es oscillates in a plane perpendicular to the plane of incidence. There are two phenomena that happen at the reflecting surface, which provide information about the surface layer. First, the amount of attenuation of a light beam at the surface depends on its orientation to the plane of incidence. If a linearly polarized light beam with equal Ep and Es components strikes the surface, the reflected beam will not have the same ratio of Ep and Es components. The second phenomenon that happens at the surface is a phase shift of the Es component. This phase shift depends on the nature of the surface and of the angle of incidence. Once the light beam hits the surface, its components will undergo attenuation and a phase shift. As a result, the reflected beam will not be linearly polarized. Instead, the two components, Ep and Es, will be out of phase such that, at any given point in time, the sum of the two components will not be
59 zero. For linearly polarized light, the electric field is zero in between any given maximum and minimum, but this is not true for the reflected light in this case. This light is referred to as elliptically polarized light. The electric field vector, instead of being confined to a single direction (as in the case of linearly polarized light), will rotate in the x-y plane (perpendicular to the direction of travel for the wave), and the tip of the vector will trace out an ellipse. In the case where the phase shift is exactly 90 degrees, and if the two components have the same magnitude, the light is referred to as circularly polarized light. The reflected light therefore carries information about the nature of the surface. Different materials and different film thicknesses will yield different amounts of attenuation and phase shifts for the incident beam. Furthermore, these effects also depend on wavelength. A detailed derivation of the equations for calculating the adsorbed mass from the refractive index, and the thickness of the adsorbed layer has been provided by Cuypers et. al [CUY83]. Wngernerud, and Olofsson [WN92] have compared the adsorption of tertadecyltrimethlyammoniumbromide surfactant on silica determined from Ellipsometry studies with the solution depletion method, and found a very good agreement between the two results. Tiberg, and Landgren [TIB93] have investigated the adsorption of nonionic surfactants at the silica/water interface using ellipsometry. Results obtained by them on the thickness of the adsorbed surfactant films, as well as the adsorbed amounts were in good agreement with those obtained by solution depletion methods. Adsorption of cationic surfactants using ellipsometry has been investigated by Eskilsson, and Yaminsky [ESK98], and also by Hutchison, and Klenerman [HUT99].
60 The main advantage of this technique is the fact that no probe molecules are required, hence eliminating any artifacts due to the probe molecules themselves. Additionally, the measurement itself is very fast, and simple. However, the technique is limited by the fact that only reflecting substrates can be used for the experiments. Additionally, in cases where extremely detailed information is required, or systems with unknown optical parameters are being investigated, the theory becomes very complicated, and the need for fitting parameters limits the accuracy of the results obtained. Optical Waveguide Lightmode Spectroscopy (OWLS) In this technique, by incorporating a grating in a planar optical waveguide one creates a device with which the spectrum of guided lightmodes can be measured. When the surface of the waveguide is exposed to different solutions, the peaks in the spectrum shift due to molecular interactions with the surface. Optical Waveguide Lightmode spectroscopy (OWLS) is a highly sensitive technique that is capable of real-time monitoring of the interactions. Since this integrated optical method is based on the measurement of the polarizability density (i.e., refractive index) in the vicinity of the waveguide surface, radioactive, fluorescent or other kinds of labeling are not required. In addition, measurement of at least two guided modes enables the absolute mass of adsorbed molecules to be determined. The sensitivity of this technique is better than most of the techniques such as Ellipsometry, or scanning reflectometry, and is best suited for biological samples. The sensing technique is based on the generation of evanescent waves, which extends for few hundred nanometers within the sample. A grating serves to incouple light into a planar optical waveguide, in which the light then propagates, generating an evanescent wave. This evanescent wave is used to probe the optical
61 properties of the adsorbed molecules. The incident plane-polarized light is diffracted from the grating, and begins to propagate via internal reflections within the waveguide. As a result thousands of internal reflections take place, resulting in enhanced sensitivity. This technique has been used for biological applications, for protein-DNA interactions, biomembranes, interaction of surfaces with blood plasma and serum, and cell surface interactions [VR02]. However, the use of this technique for surfactants adsorbed at the solid-liquid interface is relatively unexplored. Non-Optical Characterization Techniques X-Ray Reflectivity, and Grazing Angle Diffraction The advantage with techniques using X-rays is that the probing beam can penetrate a solid, liquid, or vapor without much loss in the original intensity, and hence the technique is very sensitive to structure of the interface. Two types of experiments can be carried out using the X-rays beams. The first known as specular reflection provides information perpendicular to the interface. The second method, known as grazing angle diffraction uses a beam at grazing incidence to the surface, and provides structural data about the orientation in the plane of the monolayer. As mentioned earlier, since the wavelength of X-rays (1.5 A 0 ) is much smaller than that of light, it can be used to probe molecular layer thicknesses. In the specular reflection mode, a well-collimated monochromatic beam of X-rays is reflected of a flat surface, at glancing angles typically less than 5 0 , and the intensity of the reflected beam is measured as a function of angle (or as a function of wavelength at a fixed angle). The reflectivity depends on both the angle, and the wavelength. With increasing angle, the reflectivity falls off rapidly, and is eventually lost in the background scattering. The contrast between the adsorbed layer at the interface, and the subphase is
62 very low in case of this technique, and hence it is very rarely used alone to evaluate surfactant structures at the interface. In the grazing angle diffraction mode, an X-ray beam is tilted downwards towards a surface such that the angle of incidence is fixed at a value slightly smaller than the critical angle for total external reflection. Although the beam is almost totally reflected, an evanescent wave penetrates into the subphase to a depth of about 50 angstroms. A position sensitive detector records the diffracted radiation. The grazing angle diffraction provides information about lattice parameters, which can be used to calculate the area per molecule. Additionally, the technique provides information about the tilt angle of the surfactant tails with respect to the surface normal. Insoluble monolayers of the compound 1palmitoyglycerol have been investigated using this technique [DEW98]. It was shown that both the tilt angle, and the area per molecule decreased with surface pressure, tending eventually to an arrangement of untilted molecules in a hexagonal array. Neutron Reflectivity This technique relies on using a deutrated surfactant, and by adjusting the hydrogen (H): deuterium (D) ratio in the underlying water phase, the experiment can be setup such that, apart from a background signal, the signal is entirely from the surfactant layer adsorbed at the interface. The surface excess of the surfactant can then be easily determined with good accuracy. Further H/D labeling of the surfactant can be used to obtain signals form the selected parts of the monolayer. Selective labeling allows the determination of thickness of individual parts of the layer. Crystalline solids are sufficiently transparent to neutrons to do the neutron reflection experiment at the solid-liquid interface, with the neutron beam entering form the solid side of the interface. However, the size of the sample required to obtain a good
63 signal limits the choice of substrates to silicon, and crystalline, and amorphous quartz. Mica cannot be used for the experiments, since it diffracts neutrons strongly when in the required orientation, with the basal planes forming the surface. However, the actual surface may be coated with a wide range of materials, thus providing a number of possibilities for surfactant adsorption at the solid-liquid interface. The technique is more sensitive at higher surfactant coverageâ€™s, when the thickness of the adsorbed layer exceeds 4 nm. At this thickness, the interference effects start to become more significant, enabling realistic measurement of the thickness, and the coverage of an adsorbed layer. For example, the combination of the measurement of normal protonated C 12 E 6 at the quartz-water interface [LEE89, MCD92] using water matched to quartz (50:50 D 2 O: H2O), shows that the surfactant is in the form of bilayers (deduced form the thickness of the layer), and that it does not cover the surface completely (deduced from the large fraction of D 2 O in the fitted uniform bilayers). This shows that the surfactant is in the form of defective bilayers or flattened micelles, which have been imaged using AFM. Quartz Crystal Microbalance (QCM) The QCM technique has been used in vacuum physics for many years providing reliable technique for thickness monitoring. Only recently, the technique has been refined for use in liquid media. The QCM comprises a thin, piezo-electric quartz crystal sandwiched between a pair of electrodes, usually made of gold. It operates when the piezo-electric wafer is made to oscillate at a characteristic resonant frequency, by applying an alternating electric field across the metal electrodes. The crystal resonance is basically an acoustic wave mode operating at frequencies in the MHz region. Upon adsorption onto the crystal, the frequency of oscillation is changed, enabling the calculation of the change in mass, which is directly proportional to the change in
64 frequency. The resonant frequency of the QCM can be measured with conventional feedback loop oscillators, and frequency counters, which typically allow dynamic measurements at timescales approaching 10 ms. Keller, and Kasemo [KEL98] have measured the kinetics of adsorption of 12.5 nm radius vesicles onto silica, oxidized gold, and a self-assembled monolayer of methyl-terminated thiols, using a quartz crystal microbalance (QCM). The measurement of the shift in the resonant frequency, and the change in energy dissipation as a function of time, was used to measure the adsorbed mass, and the mechanical properties of the adsorbed layer as it formed. Another study by Hk et. al [H] reports the adsorption of proteins, and the viscoelastic properties of the surface layer adsorbed on a methyl-terminated thiol monolayer on a gold surface. In case of surfactants, the QCM technique can be used to study the amount of surfactants adsorbed at the interface, and also to determine the microviscosity of the adsorbed layer. Although the QCM technique is very sensitive to the adsorbed amount, and also provides information about the viscoelastic properties of the adsorbed layer, the change in mass reported includes the mass of the water layer coupled with the adsorbed surfactant, or polymer, and hence some correction has to be applied in order to get accurate values for the amount adsorbed. Although past investigations have revealed much about the structure of surfactant films adsorbed on surfaces, very little has been reported on the correlation between such structures, and their practical implications such as dispersion stability. The goal of the present study is to investigate the similarities between the formation and stability of bulk
65 micelles and self-assembled surfactant surface structures and to demonstrate the effectiveness of an adsorbed micelle layer in stabilizing particulate suspensions.
CHAPTER 4 MATERIALS AND METHODS Materials Interaction forces using AFM were measured between a silicon wafer with a 2 m thick coating of CVD silica (Motorola) or mica substrate (S & J Trading Inc.) and the silicon nitride tip of an atomic force microscope cantilever. The root mean square (RMS) roughness of the silica, as measured by atomic force microscopy, was 0.2-0.3 nm. Triangular oxide sharpened contact mode cantilevers (200 m long, thick legged) were obtained from Digital Instruments. The radius of the tip was determined from scanning electron microscope (SEM) images (Figure 4-1). To check the validity and consistency of the experiments, measurements were repeated using different tips, which exhibited similar ( 10%) interaction forces. The cantilever spring constant of 0.11 N/m was measured using the resonance frequency method [CLE93]. In some measurements, stiffer cantilever of 0.55 N/m were used, to investigate the area of the force/distance curves where the tip jumps into the surface. Water was produced by a Millipore filtration system and had an internal specific resistance of 18.2 M and less than 7 ppb carbon. Trimethylammonium bromide surfactants of varying chain length (C 10-18 TAB) of 99% purity were obtained from Aldrich Chemical Co. All other reagents were of at least 98% purity. Silica surface was cleaned by rinsing with acetone, and methanol in sequence, and subsequent boiling for several hours in water. Muscovite mica was freshly cleaved by peeling mica plates using 66
67 double sided tapes, prior to each experiment. The tips used were subjected to UV treatment for about 30 minutes to remove any organic contaminants. FT-IR/ATR experiments were carried out on silicon single-crystal parallelepiped internal reflection elements (IRE) (55 mm x 5 mm x 2 mm, 45 o incident angle) obtained from Spectra-Tech Inc. Silica powder of 99.9% purity used in this study was obtained from Geltech, Inc. The powder was used as received and had a nominal diameter of 200 nm, as stated by the manufacturer. The experimentally measured volume average (d 50 ) particle diameter (using a Coulter LS230 laser diffraction apparatus) of the powder was determined to be 250 nm. Powder density and surface area were measured to be 2.1 g/cm 3 and 14.62 m 2 /g respectively. Figure 4-1: Electron micrograph of silicon nitride tip used for determining the tip radius
68 Methods Surface Force Measurement Surface force was measured using a Digital Instruments Nanoscope III in a fused silica liquid cell. In a typical experiment, solutions of increasing surfactant concentration were injected sequentially into the cell. It was determined, for each of the surfactants used in this investigation and on each of the substrates, that the order of addition of surfactant did not affect the measured interaction force and that rinsing with water quickly produced a force profile equivalent to curves measured in the absence of surfactant. These observations indicated that equilibrium is established quickly in this system. A new cantilever was used for each experiment. Because small variations in spring constant and radius may occur between cantilevers, even from the same batch, the interaction force measurements between each tip and silica surface in the presence of a standard solution of 32 mM dodecyltrimethylammonium bromide at pH 6 and 0.0001 M NaCl were carried out for bench-marking, and reliability. The characteristic maximum repulsive force (magnitude of repulsion where the tip jumps into contact with the surface) under these conditions was used to normalize the interactions between different tip/surface combinations. The absolute values of the maximum repulsion was found to vary as much as 10% primarily due to small variations in the tip radius. The interaction forces measured are elastic contact forces, and hence cannot be normalized by the radius of the probe [SNE65]. Hence, the measured interaction forces as a function of distance are presented without normalizing them with the tip radius. AFM measurements in solutions of C 10 -C 16 TAB were carried out after 10-15 minutes of adsorption time, which was determined to provide stable forces, with no
69 further change with time. For C 18 TAB, approximately 2 hours of equilibration time was required, indicating slower adsorption kinetics, or structural re-orientation. Suspension Stability Suspension stability was determined by measuring the turbidity of a 0.02 vol% suspension of dense spherical monodisperse silica particles formed by the Stber process and obtained from Geltech Inc. Mean particle diameter was 250 nm as determined by laser light scattering (Coulter LS230) and surface area determined by nitrogen adsorption (Quantachrome Nova 1200) was 14.6 m 2 /g. Turbidity was measured using a Hach model 2100 turbidity meter. The optical system of the instrument consists of a tungsten filament lamp, lenses, and apertures to focus the light, a 90 0 detector to monitor scattered light, a forward-scatter light detector, a transmitted light detector, and a back-scattered light detector. For turbidity measurements, samples were prepared by adding 0.05 grams of silica particles to 50 ml of suspending fluid (surfactant solutions of different concentrations in 0.1 M NaCl solution). Samples were then sonicated for 10 minutes and the pH of the dispersion adjusted to the desired value using Fisher brand HCl. The suspensions were then allowed to shake for 2 hours on a wrist shaker and if needed the pH was re-adjusted prior to the turbidity measurements. Stable suspensions showed minimal decrease in turbidity after 60 minutes of settling while the turbidity of unstable suspensions decreased significantly due to the more rapid sedimentation of coagulates. Viscosity Viscosity measurements were carried out using a Paar Physica UDS 200 rheometer with a cone-and-plate geometry. All experiments were performed at 25 C and the sample temperature was controlled to within 0.1 C using water as the heat transfer
70 fluid. In all experiments, a cone of radius 4.30 cm with a cone angle of 0.5 (a gap size of 25 m) was used. Viscosity measurements as a function of time (at a fixed shear rate) indicated that sedimentation of the particles and water evaporation from the samples during experiments were negligible. To check for other possible errors (e.g, inertial and secondary flows, edge effects, etc.), some of the experiments were repeated using a cone of radius 3.75 cm with a cone angle of 1.0 (a gap size of 50 m). The results did not change over the time period of experiments and the viscosity values measured with two different cones agreed within experimental error of 3%. Adsorption The adsorption of surfactant on the silica particles was measured by the solution depletion method. A known concentration of surfactant is mixed for four hours with a 5 vol% particle suspension equilibrated and centrifuged to separate the supernatant from the surfactant coated particles. Residual surfactant concentration was then measured by total organic carbon analysis using a Tekmar-Dohrmann Phoenix 8000 analyzer. To prevent foaming during analysis, all samples were diluted to less than 50 mg/L concentration of surfactant. Known concentrations of surfactant were also measured before and after centrifuging to demonstrate the sensitivity, and reliability of the method. Zeta Potential The zeta potential of the particles was measured using a Pen-Kem Lazer Zee Meter, model 501, based on the principle of electrophoretic mobility. Differential voltage was kept below 200 V to minimize sample degradation. At least six measurements for each sample were carried out. For zeta potential measurements, silica slurries were
71 prepared by adding 0.1 grams of 250 nm (diameter) silica particles to 50 ml of suspending media following the same procedure as for turbidity measurements. Contact Angle Contact angle was measured between an air bubble and a CVD silica substrate using a Ram-Hart goniometer model A100. The angle of the substrate was adjusted to within 5 of the angle at which an approximately 10 L air bubble placed on the underside of the substrate would not leave the surface. Six values of the advancing contact angle were measured on each of at least three bubbles placed on the surface. The results presented in this study are the averages of these measurements. FTIR/ATR Measurement FT-IR experiments were conducted with nitrogen purged Nicolet Magna 760 spectrometer equipped with a DTGS detector. All the spectra were the results of 256 co-added scans at a resolution of 4 cm -1 . The background spectrum for all the experiments was the single-beam spectra of the dry silicon crystal. Spectrum of each sample was measured with 0 o and 90 o plane-polarized light, obtained using a wire grid polarizer, in order to calculate the dichroic ratio (dichroism). All the spectra were obtained five times and the results presented are the average of the measurements. Spectra of n-propanol was measured as a reference for randomly oriented molecules, and spectra of C 12 TAB were measured at various concentrations of C 12 TAB to investigate average orientation of hydrocarbon chains in the surfactant films adsorbed on the crystal. Asymmetrical CH 2 stretching mode peaks (near 2920 cm -1 ) were used for the calculation of dichroism, although the symmetrical peaks (near 2850 cm -1 ), have also been used in several cases, to verify the results. Instrumental noise was measured using the same number of scans and
72 resolution and the same gain as used in the experiments. The average peak-to-peak noise was approximately 2% of the peak intensity, which translates to approximately 2% error in the calculated values of the dichroism. Calculations The theory of attenuated total internal reflection spectroscopy (ATR) was proposed by Harrick [HAR60] and independently by Fahrenfort [FAH61] in the 1960s. This surface sensitive technique involves the internal reflection of electromagnetic wave (light) through an internal reflection element (IRE). As the light is totally reflected at the IRE/sample interface, an evanescent wave extends from the surface of the IRE into the sample, with the intensity decaying exponentially. The evanescent wave can be described by three orthogonal electric field vectors, each of which interacts with the sample adsorbed at the surface of the IRE. The electric field amplitudes have been derived by Harrick [HAR60], and in the final form by Haller and Rice [HAL70]. In order to determine molecular orientations within adsorbed surfactant films, several researchers [FRE91, RAB93, JAN95, BAN95] have used polarized beam infrared spectroscopy. The key parameter in determining molecular orientations is the dichroic ratio or dichroism, which is defined as psAAD (4-1) where, A s and A p are the experimentally measured absorption with perpendicular and parallel polarized beam respectively. (Reciprocal value R = (1/D) has also been used by several researchers for determining molecular orientation). Deriving values of A s and A p from absorption due to x, y, and z components of the electric field, dichroism can be written as,
73 zxyAAAD (4-2) where, A x , A y , and A z are the absorption due to the x, y, and z components of the electric field. Axes x, y, and z are directed along, across, and normal to the crystal (IRE) surface. The absorption of electromagnetic beam, A, is proportional to the square of the scalar product of the electric field vector of evanescent wave, E, and transition dipole moment of the adsorbed film, M. A (E . M) 2 (4-3) The values for A x , A y , and A z will depend on the model of the adsorbed layer structure used. Ulman [ULM90] has used a model with fixed angles of the alkyl chain from the normal to the surface. The other model proposed by Zbinden [ZBI64], which considers a uniaxial symmetric distribution of the transition dipole moments M, about the alkyl chain (c axis) with fixed angle between M and c, and a uniaxial symmetric distribution of the c axis about the z axis (surface normal), with fixed angle between the c and z axes, appears to represent the adsorbed film more realistically. Formulae for dichroism for this model was first developed by Frey and Tamm [FRE91], and Rabinovich et al. [RAB93], and used later in several studies [JAN95, BAN95]. Averaging A x , A y , and A z through rotation about the c and z axes, the expression for dichroism was obtained [FRE91, RAB93] from Eqs. (4-2) and (4-3) as: 1sin*sinsin*cos2sin22222222222zxyEEED (4-4) Experimentally, the absorption of the symmetric and asymmetric stretching modes of the methylene groups in the surfactant was measured. The transition dipole
74 moments of these modes are perpendicular to the C-C axis, i.e., the angle between M and the C-C axis is 90 degrees. Solving Eq. (4-4) for with = 90 o gives: )*(/*212arcsin222DEEDExyz (4-5) Eqs. (4-4), and (4-5) were obtained by neglecting reflection at the entrance and exit surface of the crystal [ULM90] The absorption of the evanescent wave in the bulk solution of the surfactant is negligible, due to the lower concentration of the surfactants in the bulk, as compared to their concentration in the adsorbed layer at the solid-liquid interface. This fact was proven by measuring absorption of internally reflected beam with diluted solutions of hexane in water. Figure 4-2 shows the spectra obtained using parallel and perpendicular plane polarized infrared beam, at two different bulk concentrations of the surfactant. The value of D was calculated from the experimental data using Eq. (4-1), and the different components of the electric field E x , E y , and E z were calculated from the formulae given by Haller and Rice [HAL70]. Refractive indices (n) used for these calculations were n silicon =3.42, n water = 1.42, and n surfactant film =1.46. Index of refraction for water at wavenumber = 2900 cm -1 was experimentally measured by Pinkley et al. [PIN77]. The refractive index for the surfactant film, at the appropriate wave numbers, was calculated, assuming that the film is composed of 50% water (n=1.42) and 50% pure hydrocarbon (n=1.50). This assumption impacts the magnitude of the order parameter. However, the relative changes in the order parameter as a function of the surfactant concentration, which is critical for determining the structural transitions in the self-assembled surfactant
75 film, will remain the same. Using the experimental values for D, and theoretical values of E x , E y , and E z , the values of the angle were calculated from Eq. (4-5) for increasing concentrations of the surfactant. -0.0010.0010.0030.0050.0070.0090.0110.0130.01527002800290030003100Wavenumber (cm-1)Absorbance p polarization s polarization a) -0.0010.0010.0030.0050.0070.0090.0110.0130.01527002800290030003100Wavenumber (cm-1)Absorbance p polarization s polarization b) Figure 4-2: FTIR spectra of C 12 TAB at a) 4 mM and b) 8 mM bulk concentration in the CH 2 streching region [assymetric (2920 cm -1 ), symmetric (2850 cm -1 )] with parallel (p) and perpendicular (s) polarized infrared beam [SIN01] Following standard definition used for liquid crystals, an order parameter S [STE58, STE61] can be defined as 5.0cos5.12S (4-6)
76 Figure 4-3 shows the dependence of the order parameter S on dichroism D. The graph was plotted using Eqs. (4-5), and (4-6), and using the equations given by Haller and Rice [HAL70] to calculate the values of E x , E y , and E z. (refractive index values used are mentioned above). Knowing the experimental value of D, the order parameter, S, can be estimated from Figure 4-3. -0.5-0.3-0.10.10.30.220.127.116.11.511.5Dichroic Ratio (D)Order Parameter (S) Figure 4-3: Dependence of the order parameter S, on the dichroic ratio D [SIN01]. Figure 4-4 schematically lists the order parameter for different surfactant structures and molecular orientations at the interface. For the alkyl chain axis normal to the surface (homeotropic orientation), = 0 o and S=1, indicating perfect order. For full disorder, i.e; random distributed orientation of molecules, 54.7 o and S=0. For perfect planar orientation, i.e., the alkyl chain axis parallel to the surface, 90 o and S = -0.5.
77 Order Parameter Surfactant Structure (Orientation) 1 Normal (Homeotropic) 0 Random -0.5 Perfect Planar Figure 4-4: Surfactant structures and orientation, and the order parameter values associated with the structures [SIN01].
CHAPTER 5 DISPERSION OF NANOPARTICLES USING SELF-ASSEMBLED SURFACTANT AGGREGATES Introduction As mentioned in Chapter 2, traditional and emerging technologies such as advanced structural ceramics, controlled drug delivery systems, abrasives for precision polishing, coatings, inks, and nanocomposite materials are increasingly relying on nanoparticulate precursor materials to achieve optimum performance. Concurrently, environmental and safety issues have driven industrial processes towards synthesis and processing in aqueous media. The result has been increased attention to the role of particle size in the dispersion of concentrated (> 50 vol%) aqueous particulate suspensions [HIG90, HAC97, HIR97, HIM98, WU98]. Furthermore, many processes such as chemical mechanical polishing and crystallization, must operate in extremes of ionic strength (>0.1 M), temperature (>200 0 C), pH (<2, >13), mechanical force (3-8 psi), or high shear rate (> 10, 000 s -1 ). Under these extreme conditions, traditionally used dispersing methods such as electrostatics, inorganic dispersants, and polymeric dispersants may not perform adequately (discussed earlier in Chapter 2). To achieve optimal performance in the processes, there is a need for a more robust reagent scheme for dispersing nanoparticles under these extreme conditions. Reproduced in part with permission from Langmuir , 2000, 16, 18, 7255-7262, and Langmuir, 2001, 17, 468-473, Copyright 2000, 2001 Am. Chem. Soc. 78
79 The necessary criteria that must be met for a molecule to act as a dispersant, include (i) the reagent must adsorb to the surface under the given process conditions, (ii) not phase separate or otherwise adhere to the dispersant layer of an approaching particle, and (iii) provide adequate repulsion between particles to prevent agglomeration. Surfactant molecules adsorb at the solid/liquid interface through electrostatic, hydrogen, hydrophobic, and specific bonding. Criteria two is fulfilled if the hydrophilic moieties of the surfactant molecules adsorbed on separate particles interact with each other as demonstrated by the ability of surfactants to stabilize oil-in-water emulsions. It remains, however, to be shown that surfactant layers can produce a barrier to agglomeration that is large enough to disperse inorganic particles in aqueous media. In the remainder of this chapter, the utility of self-assembled surfactant films in the dispersion process will be investigated with specific emphasis placed on their effectiveness in extreme environments, such as high ionic strength. The process of surfactant adsorption, and self-assembly at the solid liquid interface, and techniques used to quantify these processes has already been discussed in Chapters 3, and 4. This chapter focuses on a specific system in order to understand the dispersion behavior of nanoparticles in the presence of adsorbed self-assembled surfactant aggregates. Surfactant structures at the solid-liquid interface have been utilized to stabilize particulate dispersions [COL97, SOL99, KOO99, BRE99, EVA96], but in all these studies the dispersion in the presence of adsorbed surfactant has been attributed to electrostatic repulsion due to the formation of bilayer structures at the solid-liquid interface, at high surfactant concentrations. However, as illustrated in Figure 2-9,
80 surfactants can be used to stabilize suspensions at salt concentrations as high as 5 M. Under these conditions, the charge due to adsorbed surfactant is negligible, indicating that under high electrolyte conditions, electrostatics is not the primary stabilization mechanism. The following discussion seeks to understand the dispersion mechanism under high salt conditions, by investigating the surfactant self-assembly at the solid-liquid interface. Results and Discussions Figure 5-1 depicts the surface forces present between a silicon nitride tip of an AFM cantilever and a mica substrate without surfactant and in the presence of dodecyltrimethylammonium bromide solution (C 12 TAB â€“ where C 12 represents the number of carbon atoms in the alkyl chain) at 32 mM at pH 4 with 0.1 M NaCl. All experiments were performed at room temperature. The forces measured in this system are elastic contact forces (between the AFM tip, and the surface surfactant aggregates), and hence do not scale with the radius of the probe used [SNE65]. As a result all the forces in the present study are reported without normalization. In the absence of surfactant, no repulsive forces are observed and as the tip approaches the substrate, van der Waals forces cause the tip to jump into contact. This behavior was observed regardless of ionic strength or pH on both silica and mica substrates. However, in the presence of surfactant, mica has been shown to be coated with full cylindrical micelles, as mentioned in Chapter 3 [MAN94, MAN95]. This was also confirmed in this investigation by imaging the adsorbed surfactant layer under these conditions. As a result of this coverage, significant repulsive forces are observed. The variation in the measured forces was approximately 5%. Note that the average adhesion force between the tip, and the substrate, with and
81 without surfactant films was not statistically different, indicating that real contact between the tip and surface was made in both cases. The most striking aspect of the force interaction curve in the presence of C 12 TAB is that at relatively short separation distances, 2.6 to 7 nm, a strong repulsive force exists between the two surfaces. At around 2.6 nm, the repulsion seems to disappear allowing the tip to jump-in and contact with the substrate. Although this interaction may at first seem similar to the traditional electrostatic repulsion followed by van der Waals attraction, as described by DLVO theory, it is important to note that because the radius of curvature of the AFM tip is small (15 nm) the interaction forces measured in this case (~ 250 mN/m) are approximately an order of magnitude greater than electrostatic repulsion (~30 mN/m). Separation Distance (nm) 05101520 Interaction Force (nN) -0.50.00.51.01.52.02.53.03.54.0 MaximumRepulsiveForce 32 mM C12TABWithout Surfactant Figure 5-1: Measured interaction forces between the AFM tip and a mica surface in 0.1 M NaCl solution at pH 4, both with and without 32 mM of C 12 TAB surfactant. Dotted line indicates the region of mechanical instability where the cantilever â€œjumpsâ€ into contact with the surface [ADL00].
82 Hence, it is proposed that the large repulsion between the two surfaces may be attributed to the elastic deformation of the self-assembled surface surfactant layer followed by its collapse at a specific applied force. A similar process may be observed in bulk micelle systems, where application of hydrostatic pressure leads to annihilation of bulk micelles [KAN83]. Past analyses of the forces between surfactant coated surfaces [PAS81, PAS88, KEK89, RUT94], have focused primarily on the contributions of electrostatic repulsion and hydrophobicity. In this investigation, by measuring the interaction between a flat surface and a very small radius probe, normalized forces were accessed that are much greater in magnitude (~ 250 mN/m) than reported in previous investigations [PAS81, PAS88, KEK89, RUT94]. Hence, the maximum repulsive force, defined by the force at which the tip jumps into contact with the surface, may be determined. Furthermore, it is important to note that all of the experimental surface interactions described below have similar features at approximately the same distances for a given alkyl chain length, regardless of surface type or solution conditions. Figure 5-2 depicts both the suspension turbidity, a measure of stability of a suspension of silica particles, and the surface forces present between the AFM tip and silica substrate as a function of C 12 TAB concentration. Under these conditions, the silica suspension without dispersant is unstable as determined by turbidity measurements. A stable suspension of spherical silica particles showed a minimal decrease in turbidity after 60 minutes while the turbidity of unstable suspensions reduced by approximately a factor of four over the same time period. At high concentrations of surfactant, a large repulsive
83 barrier is observed via surface force measurement and, correspondingly, the surfactant produces a stable suspension. Stability under relatively high electrolyte concentrations further indicates that electrostatic repulsion, as proposed by previous authors [PAS81, PAS88, KEK89, RUT94, COL97, SOL99, KOO99, BRE99, EVA96], may not be the dominant stabilization mechanism at work in this system. As reported earlier (Figure 2-9), at 32 mM of C 12 TAB surfactant, silica suspensions were found to be stable even at 5 M NaCl concentration. At this salt concentration, zeta potential of silica was measured to be +2 mV. Instead, it is proposed that the steric repulsion arising from the elastic deformation, and the yield strength of the self-assembled surfactant films is the dominant repulsion mechanism. C12TAB Concentration (mM) 01020304050 Suspension Turbidity (NTU) 100200300400500600700 Maximum Repulsive Force (nN) 0.00.51.01.52.02.53.03.5 Figure 5-2: Correlation between suspension stability, and interparticle forces. Turbidity in NTU (nephelometric turbidity units) of silica particles after 60 minutes in a solution of 0.1 M NaCl at pH 4 as a function of C 12 TAB concentration, and the measured interaction forces between an AFM tip and silica substrate under identical solution conditions [ADL00].
84 At lower C 12 TAB concentrations, a different behavior is observed. As the concentration of surfactant is first increased, no change in the attractive force between the tip and surface is observed. The measured force profile does not deviate from the interaction with no surfactant, presented in Figure 5-1, up to a concentration of 8 mM of surfactant. As shown in Figure 5-3, at this concentration, the electrophoretic zeta potential is +53 mV (with a saturation level of +61 mV). At this concentration, the measured adsorption on the particles is 3.2x10 -6 mol/m 2 (with a saturation at 8.0x10 -6 mol/m 2 ) or 40% of saturation adsorption. Figure 5-3. Adsorption isotherm (squares), zeta potential (triangles) and contact angle (spheres) of the silica surfaces, in 0.1 M NaCl at pH 4 as a function of solution C 12 TAB concentration. In the region where the system becomes stabilized by the surfactant no significant change in the trends of the measured values are observed [ADL00, SIN01]. C12TAB Concentration (mM) 10-310-210-1100101102 Zeta Potential (mV) -100102030405060708090100 Adsorption Density (mol/m2) 10-910-810-710-610-510-4 Advancing Contact Angle 010203040506070 CMC0 ABCDEF The high values of the zeta potential in the presence of 0.1 M NaCl can be explained by the fact that the surfactant structures at the interface shift the shear plane away from the surface, and the shear plane is approximately 0.3 nm from the plane of the head group of the surfactant aggregates [DAV63]. Hence, the measured zeta potential is a property of the surfactant film. Considering that the zeta potential for spherical micelles
85 in the bulk solution at 0.1 M salt concentration has been measured to be 70 mV [BIN00], the surfactant structure at the solid-liquid interface could exhibit similar zeta potential behavior (~60-70 mV). At saturation, the average area per molecule, assuming a spherical self-assembled surface aggregate, is 42 2 /molecule, which compares favorably to 38 2 /molecule reported at the air water interface for similar solution conditions (0.1 M NaCl at 25 C) [ROS89]. Under the same conditions, the area per molecule in the bulk spherical micelles, as calculated from the micelle size (4 nm), and the aggregation number (80) is approximately 55 2 . This suggests that the surface aggregates are more compact compared to the bulk micelles, indicating the possible catalytic role of surface in surfactant adsorption, and self-assembly. By 10 mM C 12 TAB where zeta potential is +59 mV and surfactant adsorption is 4.0x10 -6 mol/m 2 (50% of saturation adsorption) a barrier to agglomeration is observed. Between the same surfactant concentrations, the initially unstable suspension becomes fully stabilized. The dispersion mechanism, and its dependence of the amount, nature, and mobility of the surface aggregates are discussed towards the end of this sub-section. The adsorption isotherm and zeta potential measured on the silica particles and the contact angle measured on the silica plate as a function of C 12 TAB concentration are presented in Figure 5-3. These measurements were performed in order to elucidate the structure of this surfactant film over the region of stability transition and determine its proximity to known concentrations of previously reported structural transitions. Figure 5-4 shows pictorially the possible structures existing on the silica surface at concentration ranges corresponding to letters (A) â€“ (F) in Figure 5-3.
86 A B C D E F Figure 5-4. Possible self-assembled surfactant films at concentration ranges corresponding to A â€“ F in Figure 3. A) individual surfactant adsorption; B) low coverage of hemi-micelles; C) higher coverage of hemi-micelles; D) coexisting hemi-micelles and bilayers, spherical surfactant aggregates, or monolayers with semi-spherical caps; E) bilayers, spherical surfactant aggregates, or monolayers with semi-spherical caps at less than full surface coverage; F) spherical surfactant aggregates or monolayers with semi-spherical caps at full surface coverage [ADL00]. As surfactant is introduced to the system, it has been shown to first adsorb as individual surfactant molecules [FUE64]. However, at the critical hemi-micelle concentration lateral interactions between the alkyl chains of the surfactant molecules begin to increase the affinity of the molecules for the surface and adsorption occurs more rapidly. In the present system, from 0.007 to 20 mM C 12 TAB, the adsorption isotherm is linear, indicating that at this salt concentration the critical hemi-micelle concentration is less than 0.007 mM and that hemi-micelles may be present on the surface at even at the lowest measured concentrations. This is depicted in Figure 5-4 as individual surfactant
87 adsorption at some concentration below 0.007 mM (A), and as a mixture of hemi-micelles and individually adsorbed surfactant molecules at concentrations around (B). Up to a concentration just below 0.1 mM of surfactant, these adsorbed hemi-micelles have little effect on the surface hydrophobicity, most likely due to their low concentration on the surface (less than 2% of the theoretical monolayer coverage, 4.37x10 -6 mol/m 2 ). The zeta potential changes from mV to 0 mV in this concentration range, and the amount of surfactant adsorbed (4.37x10 -6 mol/m 2 ) corresponds directly to the charge needed to neutralize the charge density (0.00768 C/m 2 ) of the bare silica. At and above 0.1 mM the adsorbed surfactant increases hydrophobicity and generates a positive zeta potential. This reversal of zeta potential is commonly explained by the formation of bilayer patches at the surface [MOU88, HOU83]. However, in this system the hydrophobicity of the surface continues to increase despite the now positive surface potential. Hence, zeta potential reversal appears to be more likely a result of enhanced adsorption due to either hydrophobic association during hemi-micellization or specific adsorption. The hemimicellar structure, which are expected to exist up to, the maximum value of hydrophobicity (approximately at 2.3 mM) is shown pictorially as (C) in Figures 5-3, and 5-4. At the maximum in hydrophobicity, indicating the concentration at which the maximum number of alkyl groups are oriented towards the solution, yields an adsorption of 10 -6 mol/m 2 or 23% of theoretical monolayer coverage. It is this relatively low coverage that results in a maximum contact angle (57) significantly less than expected for a fully hydrophobized surface (> 90). Between 2.3 and 6 mM (23-53% of theoretical monolayer coverage) hydrophobicity significantly decreases and after 6 mM, zeta potential begins to increase
88 at a more rapid rate. This significant increase in the rate of zeta potential rise and rapid decrease in hydrophobicity indicates that an increasing number of trimethylammonium moieties are getting oriented towards the solution (reverse surfactant orientation). Of critical importance, at these concentrations, is the self-assembled structure that allows this orientation. Traditionally, bilayers have been proposed [MOU88, HOU83], but from images of the surface at higher concentrations (32 mM), spherical self-assembled surface aggregates are observed. A recent series of papers [JOH00a, JOH00b], based on thermodynamic criteria only, has also suggested that the observed images could actually be semi-spheres on top of perfect monolayers. It should be noted that these thermodynamic predictions ignore the unfavorable contributions of the hydrophobic edges of the monolayers and the area between close-packed semi-spheres on the monolayer surface, both of which are exposed to the aqueous solution in this structure. Critical to understanding the onset of repulsion and suspension stability is determination of the intermediate structures that form between hemi-micelles and the aggregates observed at higher concentrations. In the concentration range corresponding to (D), there is a gradual decrease in hydrophobicity. This suggests that the transition to a structure with a reversed surfactant orientation is not sudden but that spherical self-assembled aggregates, bilayers, or monolayers with semi-spherical caps coexists with hemi-micelles in this concentration range. These possible structures are depicted in (D) in Figure 5-4. At the CMC under these conditions (8.0 mM) no significant change in the zeta potential, adsorption, or hydrophobicity trends occurs, indicating that it is the residual monomer bulk concentration which remains constant after formation of the bulk micelles,
89 that primarily influences the structure at the surfaces [SCA81a]. On silica, at 7 mM the contact angle is at its minimum value and at 10 mM the zeta potential has leveled off. These observations indicate that in this concentration range, spherical self-assembled aggregates, bilayers, or monolayers with semi-spherical caps are the only existing structures (Figure 5-4, (E)) even though the surface, at 10 mM for instance, would be only approximately 45% covered by such structures. As concentration is further increased, more of these aggregates are expected to form until concentrations are reached (starting at 32 mM) where they may be directly observed via atomic force microscopy (Figure 5-4, (F)). Since spherical structures are observed in this systems at these concentration ranges it may be assumed that the surface aggregates are either spherical or monolayers with semi-spherical caps at these concentrations. In order to determine the dispersion mechanism in the presence of the self-assembled surfactant aggregates, it is very critical to determine the ex act structural transitions in the surface surfactant structure. Of critical importance are the regions marked as D, and E where the formation of self-assembled surfactant aggregates takes place. In order to determine the exact structural transitions, the FTIR/ATR technique using polarized infrared beam was used. The calculations, and the information obtained from this technique have already been introduced in Chapter 4, under the methods sub-section. Figure 5-5 shows the variation in the order parameter (correlation between order parameter, and surfactant structure has been presented in Figure 4-4) at the silicon surface as the concentration of dodecyltrimethylammonium bromide is increased. The reproducibility of the results for dichroism (D), obtained in different experiments (under
90 the same experimental conditions), as well as coinciding values of dichroism for symmetrical and asymmetrical streching modes (with mutually orthogonal transitional dipole moments), proves the validity of the adsorbed surfactant model suggested by Zbinden [ZBI64], which was used in the present study to calculate the tilt angle and the order parameter S. C12TAB Conc. (mM) 0510152025 Order Parameter (S) -0.2-0.10.00.10.18.104.22.168.6 Maximum Repulsive Barrier (nN) 0.00.20.40.60.81.0 Advancing Contact Angle (o) 01020304050 ABCDEF Figure 5-5: Contact angle of silicon surface (triangle), measured maximum repulsive forces between an AFM tip and silicon surface (squares), and order parameter (circles) of the adsorbed surfactant structures on a silicon surface in 0.1 M NaCl at pH 4.0 as a function of solution C 12 TAB concentration [SIN01]. At very low surfactant concentrations (Region A), the value of the order parameter is close to zero, indicating the presence of random structures at the surface due to surfactant molecules adsorbing individually. This observation is supported by the contact angle data where no appreciable change in the contact angle is determined; indicating that hemi-micellization has not yet started. As the surfactant concentration
91 increases beyond approximately 0.1 mM, the order parameter increases and becomes positive, indicating the formation of hemi-micelles at the interface in regions B and C. The sharp decrease in the order parameter in region D, corresponds to the fall in the contact angle values between 4 and 7 mM of the surfactant concentration. Increasing value of the order parameter up to approximately 4 mM, indicates that increasing number of hemi-micelles are being formed at the interface in regions B and C. Considering that a decrease in the order parameter is observed beyond C, formation of bi-layers as the transition structure may be ruled out, since the order parameter should have increased or remained constant if hemi-micelles resulted in the formation of bi-layers. The decrease in the order parameter could also be explained if the interface structure consisted of a mixture of decreasing amount of hemi-micelles and an increasing amount of randomly oriented spherical or cylindrical aggregates. AFM imaging has demonstrated the presence of spherical aggregates of C 12 TAB on the surface of silica, which is similar in nature to silicon. Hence, it was assumed that the randomly oriented aggregates observed on silicon are spherical in shape. Beyond 7 mM (Regions E and F), the observed values of the order parameter is close to zero, indicating the presence of randomly oriented spherical aggregates at the interface. Johnson and Nagrajan [JOH00b], based on thermodynamic considerations, under conditions similar to those used in the present study, have proposed the formation of hemi-cylinders on top of perfect mono-layers for C 12 TAB on silica at concentrations much below the bulk CMC (measured value of bulk CMC = 8 mM under conditions used in the present study). The order parameter for these composite structures would be an average of the monolayer, and randomly oriented aggregate, and would be approximately
92 equal to 0.3. Formation of any such structure was predicted at low surfactant concentrations (0.096 mM). However, based on the measurements in the present study, at this concentration the contact angle is still increasing, which rules out the formation of structures with polar head of the surfactant molecule pointing out towards the solution. Additionally, the adsorption amount is very low (~10 -7 moles/m 2 ) which is only about 2% of monolayer coverage, which rules out the formation of structures with hemi-cylinders on top of perfect monolayers. Instead, the formation of hemi-micelles at these low surfactant concentrations, as postulated earlier [GAU55, CHA86, CHA87, WAT86], appears to be more likely. Furthermore, the FT-IR, contact angle, and zeta potential measurements do not indicate transitions proposed by Johnson and Nagarajan into composite hemi-spheres, composite disks and bi-layers. Additionally, transition to fully randomly oriented cylinders or spheres, seems to occur just below the CMC at approximately 7 mM, as compared to 10 times the CMC proposed by these investigators. Figure 5-5 also correlates the onset of repulsive force barrier in the presence of surfactants to the structural transitions taking place at the interface, as indicated by the order parameter obtained from FTIR/ATR experiments. A decrease in the order parameter beyond 4 mM indicates the presence of spherical aggregates at the interface above 4 mM, along with hemi-micelles. Leveling off of the order parameter at around 6 mM is indicative of a complete and complete transition to randomly oriented spherical structures between 6 mM and 8 mM. This is the same transition region where steric repulsive forces become dominant, suggesting that the spherical aggregates provide the steric barrier to agglomeration.
93 Based on the contact angle, order parameter, zeta potential, adsorption, and the repulsive forces measured using AFM, the structural transitions are summarized by the schematic shown in Figure 5-6, which indicates the structures present at the interface at the concentration regions A-F in Figures 5-3 and 5-5. A B C D E F Figure 5-6: Proposed self-assembled surfactant films at concentrations corresponding to A-F in Figures 5-3 and 5-5. A) Individual surfactant adsorption, B) Low concentration of hemi-micelles on the surface, C) higher concentration of hemi-micelles on the surface, D) hemi-micelles and spherical surfactant aggregates formed due to increased surfactant adsorption and transition of some hemi-micelles to spherical aggregates. E) randomly oriented spherical aggregates at onset of steric repulsive forces, F) surface fully covered with randomly oriented spherical aggregates [SIN01]. Between 8 and 10 mM, a repulsive surface force, as well as suspension stability is first observed. One of the possible mechanisms that could result in this behavior is the restructuring of the self-assembled surfactant film. However, according to the order parameter and contact angle measurements, the formation of spherical self-assembled surfactant aggregate is possible at concentrations as low as 2 mM, and the surface
94 structure before the onset of forces is predominantly in the form of spherical aggregates. Hence, structural transitions as the surfactant concentration changes form 8 mM to 10 mM, is not necessarily the reason for the repulsive forces, and stability. Hence alternative explanations need to be considered. One of the mechanisms could be a liquid-like to solid-like transition of the surface surfactant aggregates, or increased cohesion within the aggregates. This would be reflected in the aggregation number, and also the area per molecule within the aggregates. The aggregation number can be calculated based on the size of the aggregates (5.0 nm), and the adsorption density at a particular concentration. The aggregation number thus calculated is approximately 60 at 8 mM (just before onset of forces, and stability), and 75 at 10 mM (just beyond the onset of forces and stability). The aggregation number at saturation adsorption is equal to 120. The area per molecule changes from 85 2 at 8 mM, to 78 2 at 10 mM, with the saturation value being 42 2 . This decrease in the area/molecule in going from 8 mM to 10 mM is estimated to increase the hydrophobic force between each chain from 6.0 kT (0.5 kT/CH 2 group), to 7.23 kT (0.6 kT/CH 2 group) [CHR01], creating more cohesive, and mechanically resilient surface aggregates, thus leading to the onset of forces, and stability. A second possibility is that at lower concentrations the self-assembled surfactant aggregates are not present in large enough quantities and hence do not result in repulsive forces because they can possibly be laterally displaced by the energy associated with an approaching AFM tip. As saturation is approached, the mobility of these structures could be substantially reduced forcing the approaching tip to puncture the self-assembled surfactant aggregates resulting in repulsive forces and the stabilization of particulates in
95 suspension. This phenomenon is schematically shown in Figure 5-7. The mobility of the surface aggregates at low concentrations could be explained on the basis of the spacing between the surface aggregates. Based on the adsorption density data, below the transition in forces (~8 mM), the inter-aggregate spacing is calculated to be approximately 3.2 nm, whereas above the transition concentration (~10 mM), the spacing is less than 2 nm. This reduced spacing can significantly limit the mobility, or percolation of the surface aggregates, leading to steric repulsive forces, and stability. Additionally, in the transition region, the surface coverage (based on the adsorption density measurement) changes from 40% of saturation adsorption to 50%. This change in the surface coverage may lead to a transition from a liquid-like ensemble of surface structures, to a more solid-like ensemble. A similar transition is observed in the gelation process where a transition in concentration above 50% results in change of the structure from a jelly with a finite elasticity, to a rigid structure [STA85]. Additionally, as the two surfaces approach each other, the intervening fluid media should be displaced. This would involve the flow of the fluid media on the surface of the solid, or the thinning of the confined fluid film. It has been shown by Haidara et. al. [HAI00] that the dewetting behavior of confined thin fluid films is strongly affected by the surface fraction, and topological features on nanoheterogeneous surfaces. A similar phenomena could occur in the present system, where change of surface coverage from 40% to 50% by the aggregates can restrict the thinning of the water film associated with the aggregates. This could limit their mobility, and force the approaching tip to puncture the self-assembled surfactant aggregates, and thereby resulting in repulsive forces and the stabilization of particulates in suspension.
96 Below Transition Above Transition Figure 5-7: Stabilization mechanism in the presence of self-assembled surfactant aggregates. Below transition concentration, the mobility of the aggregates is high; hence no repulsive forces are present. Factors Controlling the Strength of Self-Assembled Surfactant Structures Chain Length It is expected that the strength and formation of these surfactant aggregates should be analogous to bulk micelles. The stability of micelles in bulk solution is directly related to the slow micellar relaxation time, 2 . Aniansson and coworkers [ANI76], via the pressure-jump technique, measured the stability of a series of sodium alkylsulfates and found an increase in micellar stability as the chain length was increased from C 10 to C 16 . This increase was attributed to increased hydrophobic bonding between alkyl chains that occurs with longer hydrocarbon chain length. The maximum repulsive forces as a
97 function of alkyl chain length for C x TAB (x = 10, 12, 14, and 16) on silica at twice the critical bulk micelle concentration of the respective surfactants are plotted in Figure 5-8. Alkyl Chain Length (# Carbon Atoms) 10111213141516 Maximum Repulsive Force (nN) 0123 Figure 5-8: Maximum compressive force between an AFM tip and silica surface in deionized water as function of alkyl chain length of C x TAB surfactant at twice the CMC of the surfactant molecules (68 mM, x=10; 16 mM, x=12; 3.6 mM, x=14; 0.92 mM, x=16). Repulsive force increases as the hydrophobic bonding between surfactant molecules increases. The maximum repulsive force increases linearly over the range of chain lengths measured, indicating increasing resistance to compression of the self-assembled surfactant surface structures due to increased hydrophobic attraction between the hydrocarbon chains. Additionally, further discussion in Chapter 6 illustrates, that the cohesion energy between the surfactant tails per CH 2 group increases as the chain length increases, leading to structures, which have higher resistance to elastic deformation (higher elastic modulus), and yielding. This also provides further evidence that it is the resistance to deformation of the surfactant structures themselves that provide the majority of the repulsive energy barrier to agglomeration.
98 Octyltrimethylammonium bromide (C 8 TAB) was also investigated. However, no repulsive force and no image of surface micelles could be observed at twice the bulk critical micelle concentration (280 mM). If the linear dependence of the maximum repulsive force is extrapolated back to a chain length of eight carbon units, the value of the maximum compressive force approaches zero. Zeta potential measurements of a silica slurry under these solution conditions showed that the net zeta potential increases from -86 mV without surfactant to +21 mV in the presence of C 8 TAB. This indicates that self-assembled surfactant aggregates do indeed form on the surface, but AFM measurements are unable to detect these structures. The effect of chain length on surfactant self-assembly was also investigated using the FTIR/ATR technique on silica substrates. The order parameter, S, was measured as function of concentration for C x TAB (x = 12, 14,16, and 18), and is illustrated in Figure 5-9. The surfactant concentrations have been normalized by the bulk critical micelle concentration. For increasing chain length, from twelve to sixteen carbons, the maximum in the order parameter increases. This indicates that the hemi-micelles being formed with the longer chain length surfactants are relatively more compact, and are more oriented towards the surface normal, due to the increased hydrophobic attraction between the surfactant tails. This observation follows the same trend reported by Wakamatsu, and Fuersternau [WAK68], where they reported that the critical hemi-micelles concentration for alkyl sulphonates on alumina decreases with increasing chain length of the surfactant. The other important thing to note is the concentration with respect to the bulk CMC where the maximum in the order parameter is observed. As mentioned earlier, this maximum corresponds to the concentration where transition from hemi-micelle structures
99 to randomly oriented micelle-like surface aggregates begins. This transition occurs at lower relative concentration as the chain length increases, indicating that the formation of surface micellar aggregates is more favorable as the chain length increases. The stronger hydrophobic interaction, with increasing chain length weakens the repulsive polar head interaction, thereby facilitating formation of the aggregates. -0.100.10.20.22.214.171.124.126.96.36.199.188.8.131.52.80.911.11.2C/CMCOrder Parameter (S) C12TAB C14TAB C16TAB C18TAB Figure 5-9: Effect of chain length on surfactant self-assembly The other observation from Figure 5-9 is the order parameter variation for C 18 TAB. In this case, the order parameter keeps on increasing with surfactant concentrations, indicating that the aggregate structure formed at high concentrations is bilayers, and not randomly oriented spherical aggregates. One of the possible explanations for this could be that the experiments were carried out at room temperature (25 0 C), which is lower than the Krafft point for C 18 TAB (34 0 C), where lamellar
100 crystalline phases (bilayers) are known to exist in the bulk solution [DAV98, HAY98], and at the solid-liquid interface [LI98]. These structures could form in the bulk solution, and then adsorb at the surface of silica, or they could form directly at the solid-liquid interface. Electrolyte Concentration Further correlation of the stability and formation of bulk micelles and the self-assembled surfactant surface structures is provided in Figure 5-10, where the maximum repulsive forces at pH 4 on mica with and without the addition of 0.1 M NaCl are presented as a function of C 12 TAB concentration. Similar to the increase in alkyl chain length, addition of electrolyte, due to salting out effect, decreases the bulk critical micelle concentration (CMC) and increases the slow micellar relaxation time. Under the present experimental conditions, the CMC of C 12 TAB, as determined by surface tension, was reduced from 16 mM with no additional electrolyte to 8.0 mM in the presence of 0.1 M NaCl. The difference in the onset of the barrier to agglomeration produced by the self-assembled surfactant films with and without electrolyte in solution mirrors this behavior. Without electrolyte, 10 mM of surfactant results in strong repulsive surface forces, whereas with electrolyte, 5 mM is sufficient. Additionally, the magnitude of the repulsive force at a given surfactant concentration is greater at the higher electrolyte concentration. It is to be noted that the onset of forces in the presence of electrolyte, is much sharper as compared to the case without any added electrolyte. Using the same surface characterization techniques as in the previous section, it was found that the transition of surface structure to micelle-like surface aggregates in the absence of any added electrolyte occurs over a much wider concentration range, explaining why the onset of forces is also much more gradual. These results indicate the similarity of bulk micelle
101 formation and stability with self-assembled surfactant films. Note that for the mica surface the strong repulsive force in both cases is observed at concentrations where there are no micelles present in the bulk. Hence, the surfactant structures on the surface are not just micelles adsorbed from the bulk but are structures that are actually catalyzed by the substrate. The effect of the surface may be further assessed by examining the barrier formation on silica compared with mica under equivalent solution conditions. C12TAB Concentration (mM) 05101520253035 Maximum Repulsive Force (nN) 0123456 Figure 5-10: Maximum compressive force as a function of C 12 TAB concentration in pH 4 solution between an AFM tip and a mica substrate without additional electrolyte (circles and solid line), with 0.1 M NaCl (squares and long dashed line), and a silica substrate with 0.1 M NaCl (triangles and short dashed line). The magnitude of the repulsive force and concentration at which they arise is dependent not only on solution conditions but also on the substrate material [ADL00].
102 Microstructure of Solid Substrate The maximum repulsive force on silica under the identical solution conditions of pH 4 and 0.1 M NaCl is also plotted in Figure 5-10. The effect of the two different substrates is significant. Mica not only has a lower critical concentration at which repulsive force is first observed but also produces significantly greater repulsive forces at a given concentration than the silica surface. This may be due to the different atomic structures of the two substrates, which in turn result in different surfactant surface structures (cylindrical surface micelles on mica compared to spherical on amorphous silica). The onset of the repulsive forces seems to occur very near or just after the CMC in the case of silica whereas on mica this was observed to occur at significantly lower concentrations than the bulk CMC. The effect of substrate will be further discussed in a later subsection. Presence of Co-surfactants In bulk micellization, the addition of co-surfactants has been shown to reduce bulk CMC and enhance the stability of micelles [SCA86, PAT98]. Figure 5-11 illustrates the dependence of the maximum repulsive force on mica in the presence of C 12 TAB at 32 mM (twice bulk CMC) as a function of anionic sodium dodecylsulfate (SDS) concentration. It has been proposed that oppositely charged surfactant incorporates itself into the micelles and by reducing the repulsion between the ionic groups, increases stability and lowers the bulk CMC [SCA81b]. A similar process can be expected to occur at the surface. A significant increase in the stability and maximum repulsive force of the surfactant film is observed in the presence of anionic surfactant (e.g. SDS, see Figure 5-11). It has also been suggested that the maximum in two dimensional film stability and
103 packing results from a 1:3 molecular ratio (33 mol % SDS) due to hexagonal close packing at the interface [SHA71]. SDS Concentration (mol %) 051015202530 Maximum Repulsive Force (nN) 024681012 Figure 5-11: Maximum compressive force between an AFM tip and mica substrate in deionized water with 32 mM C 12 TAB as a function of SDS concentration. Before precipitation occurs, the addition of a co-surfactant significantly increases the magnitude of the repulsive force interaction [ADL00]. The maximum repulsive force does seem to plateau near this value. However, precipitation occurred (~27 mol%) before the predicted optimum concentration was reached (33 mol% SDS). The increase in maximum repulsive force as a function of SDS concentration demonstrates the ability of a co-surfactant to further stabilize self-assembled surfactant surface structures. However, as depicted in Figure 5-12, very small additions of SDS were observed to have a dramatic effect on the formation of the surfactant surface structures. Figure 5-12 shows the correlation of suspension stability of silica particles with the maximum repulsive force measured against a silica plate in the presence of 3 mM C 12 TAB and 0.1 M NaCl at pH 4 as a function of micromolar addition of SDS. At 5 M SDS addition, no repulsive force is observed between the surfaces but at 10 M SDS, strong repulsive
104 force has developed which continues to increase with SDS concentration. Correspondingly, over an identical range of SDS concentration, the initially unstable suspension becomes stabilized. SDS Concentration (M) 020406080100 Maximum Repulsive Force (nN) 012345678910 Suspension Turbidity (NTU) 200300400500600700 Figure 5-12: Effect of micromolar SDS addition. The turbidity of a 0.02 vol% suspension of sol-gel derived 250 nm silica particles after 60 minutes in a solution of 0.1 M NaCl at pH 4 with 3 mM C 12 TAB concentration as a function of micromolar SDS addition, and the measured interaction forces between an AFM tip and silica substrate under identical solution conditions [ADL00]. The abnormally low concentration of SDS needed to form the surface surfactant structures may be explained by the preferential adsorption of SDS at the silica interface. Adsorption experiments, using total organic carbon analysis to determine the total amount of adsorbed surfactant, and inductively coupled plasma spectroscopy to determine the SDS concentration through the sulfur emission line, have shown that nearly all the SDS added, adsorbed at the solid/liquid interface. Hence, the molecular ratio of SDS: C 12 TAB at the interface was estimated to be on the order of 1:10 instead of 1:100 in bulk solution. This is particularly interesting because it was found that no measurable
105 quantity of SDS adsorbed on silica in absence of the C 12 TAB. The magnitude of the repulsive forces increase as further SDS is added, and levels off when the molecular ratio becomes approximately 1:3, as predicted by Shah [SHA71]. Additionally, since little SDS is present in solution, the system is far below bulk CMC and yet strong repulsive forces are once again observed. The use of co-surfactants or other co-adsorbing reagents is a critical factor in the utility of a surfactant dispersant in industrial processes. Not only may the concentrations for effective stabilization be reduced, but also many other options can become available to control the overall dispersion of single and multi-component suspensions. Availability of these engineered dispersants systems can enhance the processing of nanoparticulate suspensions for emerging specialized end uses. Surface Charge Figure 5-13 illustrates the role of surface charge on surfactant self-assembly. The zeta potential of silica surface at pH 4.0 is measured to be 0 mV, whereas at pH 9.0, the zeta potential is mV. The onset of force barrier at the two different pH values, without any additional electrolyte, is illustrated in Figure 5-13. It is seen that, the onset of forces takes place at the same surfactant concentration (well below the bulk CMC of 16mM) irrespective of the pH. Zeta potential, contact angle, and FTIR measurements indicate similar structural transitions over the same surfactant concentration range (mentioned in the previous section). However, the magnitude of the force barrier is higher at pH 9.0, indicating either that the surfactant aggregates formed at pH 9.0 are more compact, or larger in number than at pH 4.0.
106 C12TAB Concentration (mM) 5101520253035 Maximum Repulsive Force (nN) 0123 p H= 9.0 p H= 4.0 Figure 5-13: Effect of pH on the steric repulsive forces due to adsorbed surfactant aggregates The silica surface has higher surface charge, and thus higher surface charge density at pH 9.0 (~0.08 C/m 2 ), as compared to pH 4.0 (~0.03 C/m 2 ). These surface charges are also the adsorption sites for the positively charged surfactant molecule. Higher density of the adsorption sites leads to higher surfactant adsorption, thereby decreasing the molecular distance between the surfactant tails, thus leading to higher hydrophobic attraction. This increased hydrophobic attraction results in more cohesive surface aggregates, leading to higher steric repulsive forces. Additionally, as seen in the case of increasing chain length, increased cohesiveness of the nucleating sites leads to faster transition to micelle-like surface aggregates (see Figures 5-9, and 5-13).
107 Temperature In the bulk solution, it is known that the critical micelle concentration of ionic surfactants increases as the temperature is increased [FLO61]. To investigate the effect of temperature on the strength of the surface aggregates, force measurements were carried out on silica surface as a function of temperature. The concentration of surfactant studied was 11 mM, since this concentration is just beyond the onset of forces in the presence of 0.1 M NaCl, and would be most sensitive to changes in the barrier magnitude. Temperature (EC) 25303540 Maximum Repulsive Force (nN) 0.00.10.20.184.108.40.206.7 AFM Tip/SilicapH 4; 0.1 M NaCl11 mM C12TAB Figure 5-14: Effect of temperature on the maximum repulsive force in the presence of self-assembled surfactant aggregates The maximum repulsive force as a function of temperature is illustrated in Figure 5-14. The maximum repulsive force decreases as the temperature increases, indicating that the surface aggregates are less favorable, and weaker or are less stable as the temperature is increased. Increasing temperature increases the vibrational entropy of the surfactant tails, thus increasing the intermolecular spacing, resulting in lower cohesion of
108 the surface aggregates. This temperature dependence also mimics the bulk micelle formation behavior. As the temperature was raised from 25 0 C to 40 0 C, the repulsive force decreases, and become negligible. This lowering of the force barrier was also reflected in terms of suspension stability, when an initially stable suspension at 25 0 C became unstable at 40 0 C. Temperature measurements can be used to calculate the threshold cohesion which leads to stability, by correlating it to particles size distribution, also measured as function of the temperature. Role of Surface in Surfactant Self-Assembly The effect of substrate on surfactant self-assembly was introduced earlier in Chapter 3, where it was shown [MAN94] that C 14 TAB forms cylindrical aggregates on mica, and spherical aggregates on the surface of silica. Also, from Figure 5-10 it is seen that the onset of steric forces in the presence of C 12 TAB surfactant, occurs at concentrations well below the bulk cmc in the case of mica, whereas for silica the onset of forces occurs just beyond the bulk cmc. The effect of different substrates on the self-assembly of surfactants was investigated by force measurements on silicon, alumina, and mica, in addition to silica mentioned previously. Figure 5-15, shows the effect of different substrates on surfactant self-assembly under the same solution conditions of pH 4.0, and in the presence of 0.1 M NaCl.
109 0 1 2 3 4 5 6 7 1 2 3 4 Surfactant Concentration/Bulk CMCMaximum Repulsive Force (nN) Bulk CMC Silica Alumina Mica Silicon Figure 5-15: Effect of substr ate on surfactant self-assembly Crystalline substrates such as silicon and mica cause the structural transitions of surfactants to micelle like su rface aggregates at concentra tions well below the bulk CMC. On the other hand, the transition in silica, whic h is amorphous, such structures form at or beyond the bulk CMC. The formation of the su rface aggregates before the bulk CMC in the case of crystalline aggregat es indicates that the periodic arrangement of atoms (found in crystalline substrates) appears to aid in the formation of the surface aggregates, by increasing the molecular packing, or decrea sing the area per head group. Evidence for this is found in the case of mica, where the cylindrical surface aggregates follow crystal directions on the surface of mica [MAN95]. The magnitude of the forces would depend on the molecular packing, and also the aggreg ation number within the aggregates. For example, cylindrical aggregates on mica have a higher aggregation number, and packing, and hence exhibit higher repulsive forces. Howe ver, the energy of in teraction between the head group and the surface sites can also be a ltered depending on the crystallinity of the 0.4 0.6 0.9 1.1
110 surface, with the higher energy sites on amorphous silica having higher energy of interaction. The effect of the crystallinity, and the energetics of the surface sites on surfactant self-assembly need to be further investigated. The results indicate that the surface microstructure and morphology play an important role in the self-assembly of surfactants and can have implications for selective dispersion and separation. Since the main objective of the studies was to evaluate dispersion stability using self-assembled surfactant aggregates, alumina particles were selected for further investigation of stability of crystalline substrates. The surfactant used was sodium dodecyl sulfate (SDS), since it is oppositely charged with respect to the alumina particles at neutral pH. When the suspension stability tests were carried out at pH 4.0, in the presence of 0.1 M NaCl, as expected, the onset of stability took place at surfactant concentrations (1.0 mM) below the bulk CMC (1.5 mM). This result was expected, since onset of forces for alumina also takes place at concentrations below the bulk CMC (Figure 5-15). Next, the measurement of forces was carried out on the basal plane of alumina, which is the most commonly used plane for force measurements. Surprisingly, the onset of forces on the basal plane was found to occur at concentrations lower than the concentration at which onset of stability had occurred in the case of alumina particles. This was in contrast to the silica case (Figure 5-2) where a very good correlation between the onset of forces, and stability was observed. In order to explain this discrepancy, it was hypothesized that in the case of alumina, since the surface is crystalline in nature (as opposed to amorphous silica), the surface of the alumina particle most likely has a random distribution of surface crystal planes, and each plane would have different self-assembly behavior. To prove the
111 hypothesis, flat plates with different crystallographic orie ntation of alumina, shown schematically in Figure 5-16, were obtained, and force measurements were carried out on these planes. Figure 5-16: Different crystal planes of alumina used for force measurements. Table 5-1 lists the density of Al sites on the different planes, along with the isoelectric point of these planes. It is to be noted that, for SD S surfactant, Al sites are also the adsorption sites. Table 5-1: Properties of different crystal plan es of alumina [BEL97] Plane Crystallographic Orientation Density of Al3+ Sites Iso-Electric Point (IEP) M (1010) 2.2 pH 3-4 C (0001) 1.25 pH 4-5 A (1120) 1.1 pH 6-7 R (1102) 0.85 pH 9-10 As seen from Table 5-1, the density of adsorption sites is the highest for the M plane, and the lowest for R plan e. It should be noted that, due to the different distribution of atoms, the surface sites on the different crys tal planes might have dissimilar energetics, which can also impact the self-assembly pro cess. The forces in the presence of selfa1 a2 a3 [ 0001 ] C-axis C plane Aplane Rplane A plane (1 1 2 0) C plane (0 0 0 1) R plane (1 1 0 2) M plane (1 0 1 0)
112 assembled aggregates of SDS were measured on different planes. The magnitude of the maximum repulsive barrier, and the concentration for the onset of forces on the different planes, along with the concentration for the onset of stability on the alumina particles are presented in Table 5-2. As seen from Table 5-1, and 5-2, the magnitude of the maximum repulsive forces is highest on the M plane, which also has the highest density of the adsorption sites. Additionally, the onset of forces occurs at higher concentrations for planes with lower density of the adsorption sites. The results indicate that a higher density of adsorption sites leads to higher adsorption density, and the inter-molecular distance between the surfactant tails is decreased, leading to higher hydrophobic attraction. This increased hydrophobic attraction results in more cohesive surface aggregates, leading to higher steric repulsive forces. Additionally, increasing cohesiveness of the nucleating sites leads to faster transition to micelle-like surface aggregates. Table 5-2: Effect of crystallographic orientation on self-assembly. Summary of the maximum repulsive forces, and the onset concentration for different crystal planes of alumina, along with the concentration for onset of stability of alumina particles. Sample Surfactant concentration (mM) at onset of stability/forces Maximum Repulsive Force (nN) 300 nm Alumina Particle 1.0 M â€“ plane 0.5 2.2 C â€“ plane 0.6 1.6 A â€“ plane 1.1 1.2 R plane 1.2 0.8 These results follow the same trend as in the case of silica where varying the pH conditions varied the density of the adsorption sites. Interestingly, the stabilization for the
113 alumina particles lies in between the concentration at which onset of forces occurred for different planes, indicating that the surface of the particle is some kind of a weighted average of all the different crystal planes. In order to further understand the differences in the self-assembly behavior of SDS on different crystal planes of alumina, the C, and R plane which exhibited maximum difference in the force behavior were selected. Going back to the magnitude of the repulsive forces on the C, and R plane, it was observed that the C plane exhibits higher magnitude of the force. Two possible explanations could be given to explain this difference in the forces, and are depicted in Figure 5-17. Figure 5-17: Surfactant aggregates on C, and R planes of alumina. Schematic explaining the difference in force magnitude on the C, and R planes of alumina. A) Less compact aggregates on R plane, B) Less number of aggregates on R plane. B A R-plane C-Plane The first reason could be that the structures forming on the C plane are more compact than the structures on the R plane, and hence exhibit more resistance to elastic deformation, resulting in higher forces. The other possible reason could be that the number of surfactant aggregates on the C plane is higher than that on the R plane.
114 To further understand the differences in the self-assembly behavior, the FTIR/ATR technique was used to determine the surfactant ordering on the two different planes. Figure 5-18 shows the order parameter as a function of SDS concentration on the two planes. SDS Concentration (mM) 0.010.1110 C Plane R Plane Order Parameter (S) 0.000.050.100.220.127.116.110.35 Figure 5-18: Order parameter of adsorbed SDS on different crystal planes of alumina The maximum order parameter on the C plane is higher, indicating that the hemi-micellar aggregates are more compact, and hence more oriented towards the surface normal. This difference can be explained by the fact that the density of adsorption sites is higher on the C-plane, and hence the surfactant tails can come much closer to each other, resulting in higher hydrophobic attraction. Also, the concentration at which the maximum in the order parameter is observed is lower (0.18 mM) in the case of the C-plane, as compared to the R-plane (0.5 mM). The maximum corresponds to the transition in surface aggregate structure from hemi-micelles to micelle-like surface aggregates, which
115 results in the steric repulsive forces. This delayed transition in the case of R-plane, explains why the concentration at which repulsive forces are observed (Table 5-2) is higher for the R-plane, as compared to the C-plane. Based on these preliminary investigations, it is clear that the surface plays a very important catalytic role in the self-assembly of surfactants at the solid-liquid interface. However, a more detailed study correlating the self-assembly behavior with the nature, density, and the energetics of the adsorption sites on the surface is needed to delineate the role played by the surface. Viscosity of Concentrated Suspensions Most of the colloidal dispersions used in the industry are highly concentrated (~40-50 volume % solids), and the need for robust dispersants at high solids loadings is critical. In this subsection, the use of self-assembled surfactant aggregates for dispersing concentrated slurries will be discussed. Figure 5-19 is a plot of viscosity as a function of shear rate for suspensions of 250 nm silica particles at solids loading of 45 %vol and 50 %vol prepared in aqueous solutions of 0.1 M NaCl containing 32 mM bulk concentration of the surfactant (suspension pH=4). The viscosity shows changes with both shear rate and volume fraction of the particles. Both samples exhibit non-Newtonian behavior and there is no indication of shear thickening behavior over the entire range of shear rate investigated (1 s -1 up to 15,000 s -1 ). It was not possible to prepare a dispersion of 50 %vol silica particles in an electrolyte solution of 0.1 M NaCl in the absences of C 12 TAB, whereas addition of surfactant to the suspending media made it possible to prepare a well dispersed slurry of low viscosity at 50 vol% solids.
116 0.0010.010.11110100100010000100000shear rateviscosity, pa.s 45 volume % 50 volume % Figure 5-19: Viscosity as a function of shear rate for 250 nm (diameter) dispersions of silica particles of 45 %vol and 50 %vol stabilized with C 12 TAB (0.1 M NaCl, 25 C and suspension pH=4). Figure 5-20 is a comparison between the viscosities of dispersions of 250 nm silica particles at 45 %vol prepared at two different conditions. The first sample was prepared in an electrolyte solution of 0.001 M NaCl without the addition of surfactant and the second sample was prepared in an electrolyte solution of 0.1 M NaCl and 32 mM bulk equilibrium C 12 TAB concentration (the pH of both samples was adjusted to 4). While the first sample is stabilized through long range electrostatic repulsion (zeta potential = 35 mV) between the suspended particles, the second sample is stabilized through surfactant induced steric forces due to the adsorption of C 12 TAB molecules onto the surface of the particles. There is a significant difference between the viscosities of the two samples even at very high shear rates where hydrodynamic forces are dominant. Surfactant stabilized dispersion shows a lower and a more uniform resistance against flow than electrostatically stabilized dispersion. This indicates that C 12 TAB is a more effective
117 despersing agent for silica slurries under extreme conditions. The data presented in Figure 5-20 indicate that the differences between the viscosities of the samples are more significant at low shear rates 0.010.1110110100100010000Shear RateViscosity, pa.s 0.001 M, No C12TAB 0.1 M NaCl, 32 mMC12TAB Figure 5-20: Comparison between the viscosity of electrostatically stabilized (0.001 M NaCl) and surfactant stabilized (0.1 M NaCl, 250 mM C 12 TAB) dispersions of 250 nm (diameter) silica particles of 45 %vol (25 C and suspension pH=4). As shown in the previous sections, temperature can alter the cohesiveness of the surfactant dispersant layer, thus impacting the dispersion stability. Most of the concentrated dispersions used in the industry are subjected to varying temperatures either during processing, or during transportation, where the control of viscosity is important. Hence, it is critical to study the effect temperature will have on the viscosity of
118 concentrated slurries, stabilized using self-assembled surfactant aggregates. Figure 5-21 represents the effect of temperature on the viscosity of a 50 %vol dispersion of 250 nm silica particles prepared in an electrolyte solution of 0.1 M NaCl containing 12 mM bulk equilibrium concentration of C 12 TAB surfactant (concentration just beyond the onset of forces, and stability) (suspension pH=4) at a shear rate of 50 s -1 . Temperature (C) 35404550556065 Viscosity (Pa s) 0.200.250.300.350.400.45 200 nm Silica50 vol%; 50 s-1pH 4; no NaCl39 mg/g C12TAB Figure 5-21: Effect of temperature on the viscosity of a surfactant stabilized (0.1 M NaCl) dispersion of 250 nm (diameter) silica particles of 50 %vol at a shear rate of 50 s -1 (suspension pH=4). It can be seen that over the temperature range studied, the viscosity of the dispersion initially decreases with increasing temperature, reaches a minimum at approximately 55 C and then increases with further increase in temperature. This indicates that C 12 TAB dispersing agent can further improve the flowability of the dispersion at higher temperatures.
119 Increase in temperature will increase Brownian motion of the particles which is in favor of increasing the viscosity of the dispersion due to increase in number of collision between the suspended particles. On the other hand there will be a decrease in the viscosity of the suspending fluid with increasing temperature, which tends to decrease the viscosity of the dispersion over the entire range of temperature. When surfactants are present in the system, increasing temperature decreases the cohesiveness of the surface surfactant structures, which may result in an increase in the viscosity of the suspension. Summary The investigations carried out in the present study have revealed the feasibility of self-assembled surfactant aggregates as dispersants under extreme conditions such as high electrolyte concentration. It was also shown that the key to achieving dispersion stability is the control of the intermolecular packing, or the cohesion within the surfactant surface aggregates. The next step would be the development of a predictive methodology, for optimal slurry formulation, taking into account the concentration of the surfactant dispersant, and the properties of the specific substrate. As mentioned in the beginning of this chapter, the first criterion for any dispersant selection, which would be true for surfactants also, is that the dispersant molecule should adsorb on the surface of the particle. For a particular particle surface, the surfactant should be chosen such that it adsorbs on the surface through either electrostatic, hydrogen, hydrophobic, or specific bonding. The adsorption of the surfactant at the interface can be verified using techniques such as adsorption isotherms, zeta potential, and contact angle. Once the surfactant adsorbs, the next step is to achieve a repulsive agglomeration barrier, which depends on the concentration of the surfactant, and formation of cohesive micelle-like surface aggregates. Techniques such as
120 turbidity, particle size measurement, and the AFM can be used to determine the threshold concentration of the surfactant required to achieve stability. As illustrated in this chapter, the barrier to agglomeration depends on the molecular packing within the surface aggregates, which can be controlled using several methods such as surfactant structure (chain length), additives (electrolyte, co-surfactants), processing conditions (temperature, pH), and the surface microstructure (crystalline versus amorphous, density of the adsorption sites). Not only does the control of molecular packing provide a tool to tailor, and manipulate the agglomeration barrier, it can also be used to control the amount of dispersant needed to achieve stability. This was evident in Figure 5-12 where small amounts of the SDS co-surfactant reduced the dosage of the primary C 12 TAB surfactant from 10 mM or higher to 3 mM. By controlling the inter-molecular packing, not only the concentrations for effective stabilization can be reduced, but also many other options can become available to control the overall dispersion of single and multi-component suspensions. Availability of these engineered dispersants systems can enhance the processing of nanoparticulate suspensions for emerging specialized end uses. One of the applications of self-assembled surfactant stabilized slurries has been in the development of optimally performing slurries for Chemical Mechanical Polishing (CMP) [BAS02]. Silica slurries stabilized by C 12 TAB surfactant in the presence of 0.6 M NaCl were used to polish silica substrates. The surfactant stabilized slurries provided stability by introducing high particle-particle repulsion and improved the surface quality upon polishing. On the other hand, material removal rate was negligible. Two alternatives were suggested to explain the negligible material removal in the presence of C 12 TAB
121 self-assembled surfactant. One of the reasons was that the high repulsive force barrier, due to the mechanical strength of the surfactant structures, is not overcome during the polishing process, thus preventing direct contact between the abrasive particle, and the surface to be polished. The fundamentals discussed in this chapter, on the dependence of the aggregation barrier on the cohesiveness, or the molecular packing within the surfactant aggregates can be used to overcome this problem. Since the additives, processing conditions, and the nature of the substrate are fixed for a particular polishing process, the effect of chain length on molecular packing was used to manipulate the barrier to agglomeration. Further research revealed that the barrier was overcome during the polishing process, and that lubrication due to the adsorbed surfactant molecules was the dominant factor controlling the material removal rate. It was found that the lubrication properties of the surfactant layer depend on the adsorption strength, and also the cohesiveness of the adsorbed surfactant structures at the interface. The adsorption strength, and cohesiveness were manipulated by changing the surfactant structure (chain length), additives (monovalent, and bivalent salts), processing conditions (pH), in order to achieve high removal rate (> 400 nm/minute), along with good surface finish (RMS < 1 nm). It can be seen from the above example that the fundamentals developed in this study can be used to generate guidelines for designing optimally performing slurries stabilized using self-assembled surfactant aggregate concept. The stabilization mechanisms in the presence of self-assembled surfactant aggregates have been proposed in this chapter, and the factors controlling the cohesiveness of the surface aggregates have been discussed. However, the mechanical
122 properties of the surface aggregates, which are believed to be the origin of steric repulsive forces, still need to be determined, and quantified. To our knowledge, mechanical properties such as elastic modulus, and yield strength of surface aggregates, and their relevance to suspension stability, have not been reported in the literature. Therefore, a systematic study using the AFM was undertaken, and is discussed in the next chapter.
CHAPTER 6 MECHANICAL AND THERMODYNAMIC PROPERTIES OF SURFACE SURFACTANT AGGREGATES Introduction In the previous chapter, the use of self-assembled surfactant aggregates for dispersion of nanoparticles was discussed. It was shown that the rigidity, or the strength of the surface aggregates is very critical to creating steric repulsive forces, which lead to stabilization under extreme processing conditions such as high electrolyte, and high shear forces. Also, it was shown that the mechanical properties of the surfactant aggregates could be varied by manipulating the intermolecular cohesion in order to control the barrier to agglomeration. Additionally, it was shown that the surface plays a catalytic role in the self-assembly of surfactants, and depending on the nature of the substrate, the shape, and the mechanical properties of the surface aggregates could vary. In order to further understand the self-assembly process of the surfactants at the solid-liquid interface, and to delineate the role played by the surface, a systematic attempt was made to calculate, and predict the mechanical properties of the surface aggregates, and the associated energetics. As mentioned in the previous chapter, Johnson, and Nagarajan [JOH00a, JOH00b] have used thermodynamics of the surfactant adsorption process, to predict from the first principles, the aggregate geometry, and the concentration at which aggregates form at the interface. A limitation of their model is the fact that it requires parameters such as the distribution of the surface adsorption sites, interaction energy of the head 123
124 group with the surface sites, interaction between the adjoining head groups, the difference between the water-solid surface interfacial tensions. These parameters are difficult to measure experimentally, and hence have to be assumed. Additionally, the thermodynamic predictions ignore the unfavorable contributions of the hydrophobic edges of the monolayers and the area between close-packed semi-spheres on the monolayer surface, both of which are exposed to the aqueous solution. As a result, the concentrations, and the aggregate geometries predicted do not agree with corresponding experimental values. However, there have been no experimental or theoretical studies evaluating the mechanical properties of surface catalyzed micelle-like aggregates, and the energetics associated with them. The only exception is an article by Subramanian and Ducker [SUB01], where specific (per unit area) energy of dodecytrimethlyammonium bromide aggregates on silica surfaces is evaluated from experimental force/distance curves using glass particle (diameter = 8 m) as the probe. The present work is a systematic attempt to obtain the mechanical properties, and energy of the surface surfactant aggregates in a more precise, and reliable way, by using experimentally measured force/distance curves. The energy of the surface aggregates, thus calculated, is compared to the bulk micelles to delineate the role played by the surface in surfactant self-assembly. Hertz Model for Interaction of Hard Sphere (Tip) with Soft Flat Bilayer Figure 6-1 schematically illustrates the model for interaction of a hard sphere (for example, the AFM tip) with a soft flat layer (e.g., surfactant bilayer adsorbed on flat hard substrate). For the illustrated geometry, Hertz theory [HER96] of contact between elastic bodies is valid [HER96]. The Hertz relationship, for the applied force F and the radius of the contact zone x, is:
125 R KxF3 (6-1) Where, EK/)1(75.0/12 (6-2) E and are the elastic (Youngâ€™s) modulus, and Poisson ratio of the surfactant bilayer, respectively, and R is the radius of the probe (tip). H0 H Hard Sphere (AFM Tip) R F x Adsorbed Layer Figure 6-1: Hertzâ€™ model of interaction between hard sphere (tip) with soft flat layer (surfactant bilayer)
126 The radius of the contact zone, x, is related to the penetration depth as: (6-3) Rx/2 The value of is measured experimentally as: HH 0 (6-4) Where, H 0 is the minimal distance between the spherical probe and flat substrate, at which Hertz force starts to act, and H is the distance between the probe and the underlying substrate (silica or mica) under the given pressure. Value of H 0 should be close to the height of the surface surfactant structure. Combining Eqs. (6-1), (6-3), and (6-4): 5.05.10)(RHHKF (6-5) This equation allows calculation of the elastic modulus from experimentally measured dependence of force on distance. Experimental force/distance curves also allow the calculation of yield stress of surfactant aggregates. It is assumed that the layer breaks (yields) under the center of the tip where pressure is maximal. This pressure is equal to [HER96] R K xP 2/3 (6-6) At the point of yielding, the pressure in the center is equal to the yield stress Y. Substituting the expression for x from Eq. (6-6) into Eq. (6-1), the maximal force F max can be expressed as: 2323max/)3/2(KYRF (6-7) Experimental values of F max , and value of K calculated using Eq. (6-5), can be used to calculate the yield stress, Y.
127 Hertz Model for Interaction of Hard Sphere (Tip) with Soft Micelle-Like Spherical Surfactant Aggregates Eqs. (6-1) to (6-7) listed in the previous section, are also valid for interaction of two spheres; however, in the case of surfactant aggregates the radius R should be expressed as an effective radius, R ef , 11RRRRRttef (6-8) Where R t and R 1 are radii of tip and soft spherical aggregates on the substrate, respectively. The model for the interaction of two spheres may not be applicable for interaction of the tip with a densely packed array of spheres on the substrate. However, for the tip radius R t = 15 nm and aggregate radius R 1 = 3nm, the calculation shows that for penetration depth =3 nm (from H 0 = 6 nm to minimal distance H=3 nm), the tip only starts to touch the neighboring aggregates. Therefore, only the interaction of tip with one central aggregate may be taken into account. Hertzâ€™ theory is only applicable for small deformations, and itâ€™s applicability to the present system is discussed in the following section. Results and Discussion Mechanical Properties Experimental force/distance graphs (F vs H) for the interaction between AFM tip and self-assembled surfactant aggregates on silica substrate are plotted in Figure 6-2. Graphs shown in Figure 6-2a, were obtained for C n TAB aggregates on the silica surface in 0.1 M NaCl solution with C n TAB concentrations 168; 32; 7.2; and 1.8 mM for n=10, 12, 14 and 16, respectively. These concentrations correspond to twice the CMC for
128 surfactants in DI (and higher than twice the CMC for surfactants in 0.1 M NaCl solution), where micelle-like surface aggregates are known to exist on the surface [MAN94, MAN95, DUC96, AKS96, WAN97, ADL00, SIN01]. Figure 6-2b represents interaction forces for 0.4 mM C 18 TAB in 0.1 M NaCl solution. Similar measurements were also carried out for the surfactant aggregates on the mica surface. 0123051015Separation (nm) Force (nN) C14TAB C12TAB C16TAB C10TAB 0.1 M NaCl, Silica Figure 6-2a: Experimental force/distance curve (circles) measured between AFM tip and C n TAB aggregates on silica substrate at pH=5.8 in 0.1 M NaCl solution of C n TAB at 168; 32; 7.2 and 1.8 mM for n=10, 12, 14 and 16, respectively. The concentrations are equal to twice the CMC for C n TAB in DI water. Theoretical (solid) curves are plotted using Eq. (6-5) with fitting values of H 0 and E from Table 6-1a. Value of R in Eq. (6-5) is calculated using Eq. (6-8) for interaction of spherical tip with spherical surfactant aggregate.
129 036912051015Separation (nm) Force (nN) Figure 6-2b: Experimental force/distance curve (circles) measured between AFM tip and C 18 TAB aggregates on silica substrate at pH=5.8 in 0.1 M NaCl solution of C 18 TAB at 0.4mM concentration. Theoretical (solid) curves are plotted using Eq. (6-5) with fitting values of H 0 and E from Table 6-1a. Value of R=15 nm (radius of tip) was used for calculating the interaction of spherical tip with the flat bilayer. Small deviations for the Hertz fitting at separation distances greater than 5 nm could be due to additional electrostatic repulsion from the charged surfactant head groups. All parameters, calculated from Figure 6-2, and results for mica, using Eqs. (6-1) to (6-8) are listed in Table 6-1. The specific energy per each surfactant molecule within the aggregates is also listed in Table 6-1, and will be discussed in the next section.
130 Table 6-1: Mechanical, and thermodynamic properties of surface aggregates of C n TAB. Distance H 0 , at which force starts to act, and elastic modulus E are fitting parameters in Eq. (6-5), obtained from experimental force/distance curves shown in Figure 6-2, and results for mica. Yield stress Y is calculated from the same curves by using Eq. (6-7). For C 10 -C 16 TAB, values of R = R eff are calculated using Eq. (6-8), and R=R tip =15 nm was used for C 18 TAB. Energy per chain of the surface aggregate is calculated using Eq. (6-18) for C 10 C 16 TAB and Eq. (6-16) for C 18 TAB. Value of = 8.010 -6 moles/m 2 was used in Equation (6-16), and n ag = 115 in Eq. (6-18) a. Silica Chain Length (n) Aggregate Shape H 0 (nm) E (Mpa) Y (Mpa) W 1 (10 -20 J/tail) 10 Spherical 6.1 47 42 1.3 12 Spherical 7.5 46 44 3.4 14 Spherical 7.6 61 60 4.9 16 Spherical 8.2 73 74 8.2 18 Bi-layer 6.9 135 63 7.7 b. Mica Chain Length (n) Aggregate Shape H 0 (nm) E (Mpa) Y (Mpa) W 1 (10 -20 J/tail) 10 Cylinder 5.3 127 107 1.97 12 Cylinder 5.6 197 149 2.03 14 Cylinder 5.8 196 157 3.05 16 Cylinder 6.8 181 149 4.70 18 Bi-layer 5.2 444 144 8.20 From Table 6-1a and 6-1b, it is observed that the thickness (H 0 ) of the surfactant aggregate layers on silica and mica lie in the range of 5 (for C 10 TAB on mica) to 8 nm (for C 16 TAB on silica). For example, for C 14 TAB, the values of H 0 = 7.6 nm for silica and H 0 = 5.8 nm for mica were calculated. The calculated values of H 0 are close to the
131 experimentally measured diameter of micelle-like surface aggregates (5.3 + 0.3 nm on mica, and 5.7 + 0.7 nm on silica for C 14 TAB [MAN95]. The slightly larger values of H 0 for silica can be attributed to the difference in the shape of the surface structures (cylindrical aggregates on mica, and spherical aggregates on silica) [MAN95]. Experimental values of K were obtained by fitting experimental force/distance curves using Eq. (6-5). Values of elastic modulus were calculated by using Eq. (6-2) with an assumed value of = 0.5, assuming the surface aggregates to be similar to soft polymeric materials [ = 0.4 for plastics, and 0.49 for rubber] [SPE92], and the experimentally obtained value of K. It should be noted that the assumed values of do not affect the calculations significantly, since the equation consists of a 2 term. Spherical aggregates are known to form for C 10 TAB to C 16 TAB on silica [MAN95, VEL00], and hence the Hertz model for interaction between hard and soft sphere was used. The effective radius was calculated from Eq. (6-8), and experimentally measured values of R 1 =H 0 /2 (Table 6-1). Calculation for C 18 TAB was carried out using radius R=15 nm (equal to radius of the tip), because AFM imaging suggests that C 18 TAB forms bilayers [LI98] rather than spherical or cylindrical aggregates. Note, that even for the cylindrical surfactant aggregates, which form on mica [MAN95], the formulae for spherical aggregates had to be used due to the unavailability of Hertz formulae for interaction between cylindrical (surface aggregates) and spherical (AFM tip) surfaces. The experimentally measured K values were used to calculate the elastic modulus of the surface aggregates. It is observed that elastic modulus of the aggregates formed on mica is greater than that on silica, e.g., for C 12 TAB on silica E = 46 MPa, while on mica, the aggregates have an elastic modulus of 197 MPa. These values are of the same order
132 of magnitude as Youngâ€™s modulus of low-pressure polyethylene (E=200 MPa) and rubber (E= (10-100) Mpa) [SPE92]. The difference in elastic modulus of the surfactant aggregates on silica and mica could be due to different structure of these aggregates (spherical and cylindrical, respectively), and larger density of cylindrical aggregates. Additionally, the calculated values of E for silica are more reliable, since the formulae used for the calculation is valid only for spherical aggregates. In case of mica, the E values are approximate due to the extrapolation of the formulae to cylindrical aggregates. Values of the yield stress, Y, calculated from Eq. (6-7), and experimental values of maximal force F max are listed in Table 6-1a and 6-1b, respectively. For C 12 TAB on silica, the measured F max =1 nN (see Figure 6-2), and the yield stress was calculated to be Y= 44 MPa. For comparison, the yield stress of low-pressure polyethylene is reported to be approximately 6 to 20 MPa [SPE92]. The effect of chain length on the elastic modulus, and yield stress will be discussed later. From Figure 6-2, and similar results for mica, it appears that for both silica and mica, good agreement of Hertz theory with experimental force/distance curves is achieved over a wide range of separation distances, until a relative deformation of /H 0 0.5. Strictly speaking, the Hertz model should be applicable for deformations significantly smaller than the tip radius (R tip ) and the elastic layer thickness (H 0 ), i.e., tipR (6-9) and, 0H (6-10) Eq. (6-9) is satisfied, but Eq. (6-10) is definitely not fulfilled in the present experiments, where R tip =15 nm, max 3 nm, and H 0 = 5-8 nm. An alternative to the Hertz
133 model, the Sneddonâ€™s [SNE65] model, for interaction of an elastic flat layer and a rigid tip, might be more appropriate in the case of C 18 TAB [LI98]. This model coincides with Hertz model at small deformations, and is also valid for relatively large penetration, for which Eq. (6-9) doesnâ€™t hold good. For the experimental data, fitting values of elastic modulus and yield stress, have been calculated by using either Hertz model or Sneddonâ€™s model. Values obtained for both models, were found to be close to each other (~5% difference), hence the Hertz model was used due to its simplicity. The limitation imposed by Eq. (6-10) is more rigorous. It suggests that the flat soft layer is of infinite thickness, through which the deformation is propagating. For a thin elastic layer on a rigid substrate, the layer deformation is restricted. As a result, the calculated mechanical properties from the experimental force/distance curves with the Hertz Eqs. (6-5), and (6-7) are effective values, which will lead to larger than actual elasticity and yield stress of the surface aggregates [HEU96, DER98]. The correct values can only be obtained for thick aggregate layers. To avoid the effect of thickness of the layer on elasticity and yield stress, it can be assumed that at high pressure, and large penetration, the contact area, S cont , is almost independent of penetration, and lateral stresses are negligible. As a result, the simplest Hookâ€™s equation can be used [VAK01]: 0maxmaxmaxmax2/)2/(/HERFSFcont (6-11) Note, that Eq. (6-11) takes into account the average penetration of the sphere (tip) in the micellar layer (or bilayer), which is equal to max /2. Solving this equation for E, we obtain: )/(2max0maxRHFE (6-12)
134 Calculation using Eq. (6-12) for C 12 TAB on mica yields elastic modulus E = 140 MPa, which is approximately 1.5 times less than the value of E = 197 MPa obtained from Eq. (6-5). As mentioned above, this difference can be explained on the basis of small layer thickness of the surfactant aggregates. The values reported are the effective values of parameters E and Y, obtained for the thin aggregate layers (using Eqs. (6-2), (6-5), and (6-7)). Dependencies of elastic modulus E and yield stress Y on the number â€œnâ€ of carbons atoms in the hydrocarbon chain are presented in Figures 6-3, and 6-4. The elastic modulus of surface aggregates on silica appears to increase linearly with â€œnâ€, due to the formation of more compact structures with increasing chain length (more lateral chain-chain interactions). For surfactant structures on mica, the elastic modulus increases initially when the chain length is increased from ten to twelve carbon atoms in the tail. With further increase in chain length, E is constant up to C 16 TAB. This indicates that in the case of mica, maximum packing is achieved when the chain length increases from ten to twelve, and on further increase of chain length the lateral interaction per chain does not change significantly. A sharp increase of elastic modulus is observed for C 18 TAB, which is attributed to the change in shape of surfactant structures (bilayer for C 18 TAB, as compared to cylindrical aggregates for lower chain lengths). This is due to the fact that bilayers will have a higher packing density as compared to spherical or cylindrical aggregates [ROS89]. The elastic modulus for a bilayer on mica is higher than on silica, indicating denser packing in the case of mica. One of the reasons for this observation could be that the surface of mica under these conditions is slightly more negative (-40
135 mV) compared to silica (-30 mV), which would result in higher density of adsorption sites, thus leading to more cohesive structures. E (MPa) 0.1 M NaCl, pH 6.0 Mica Silica n (# Carbon Atoms) 18 16 14 12 10 450 350 250 150 50 Figure 6-3: Elastic modulus (E) as a function of chain length (n). Points are calculated from experimental force/distance curves for interaction between silicon nitride tip and silica (circles) or mica (sqaures) in solution of C n TAB in the presence of 0.1M NaCl at pH 5.8. The yield stress, Y, increases slightly with increasing chain length for spherical aggregates on silica. For cylindrical aggregates on mica, Y increases between C 10 and C 12 TAB and remains constant thereafter.
136 01002001012141618n (# Carbon Atoms)Y (Mpa)SilicaMica Figure 6-4: Yield strength (Y) as a function of chain length (n) Points are calculated using Eq. (6-7), and experimental force/distance curves for interaction between silicon nitride tip and silica (circles) or mica (squares) in solution of C n TAB in the presence of 0.1M NaCl at pH 5.8. The elastic modulus and yield stress for surfactant aggregates on mica are higher as compared to silica. This difference could possibly arise due to the different shape of the aggregates, and different packing densities within the aggregates [MAN95]. Another possibility could be that the calculations for cylindrical aggregates on mica are not precise, because the formulae used for the calculations are strictly valid only for spherical aggregates. Energetics of Surface Structures The energy required to puncture the surface aggregates can be, in principle, calculated by integrating the force/distance curves. To avoid graphical integration, which is less precise, and tedious, the following procedure was used. The interaction of two
137 hard flat layers, through a soft elastic interlayer (Figure 6-5) was considered. Hookâ€™s force F 1 per unit area of the flat elastic layer (bilayer) under pressure from the hard, non-deformable solid can be written as: tEF/1 (6-13) Where = H 0 -H is penetration (deformation) and t is the original thickness of the layer. F1 = E (H0 -H)/H0 H H0 Figure 6-5: Interaction of two hard flat planes through elastic bilayer, modeled as springs. The elastic energy W, per unit area of this layer, compressed till pressure p equals the yield strength, is given by: 02min0min0012/)(HHHEdHF W HHH (6-14) where, thickness of the bilayer is equal to H 0 , and H min corresponds to the distance, at which maximal force F max is achieved.
138 To investigate force/distance curves more carefully in the area of â€œjump-inâ€, harder cantilevers (stiffness k=0.6 N/m) were used for force measurements. However, tips with these cantilevers jump-in at the same distance as with softer cantilevers and no equilibrium points on the force/ distance curves could be measured after the maximum force. This indicates that after the maximum force is achieved, the â€œjump-inâ€ of the tip is very fast, resulting in an almost vertical force/distance profile. This also indicates that no additional energy is spent in further breakup of the surface aggregate, and the energy calculated from Eq. (6-14) is an accurate estimate. The indentation of the surface structure with the tip, results in increased intermolecular spacing, and decreased cohesion. With further indentation, and increase in intermolecular spacing, the hydrophobic attraction between the tails decreases rapidly, and finally the aggregates disintegrate catastrophically. This behavior mimics the propagation of cracks in a rigid material with high modulus of elasticity, such as ceramics. By using Eq. (6-15) to calculate the energy required to break one surfactant aggregate, and knowing the adsorption of surfactant in moles per unit area, , it is possible to calculate energy per hydrocarbon chain. Area per adsorbed head group, A, is: )/(2aNA (6-15) Where N a is Avogadroâ€™s number. Eq. (6-15) assumes that the adsorbed layer is a bilayer or spherical aggregate, so only half of all the molecules within an aggregate are adsorbed onto the substrate. The energy per unit hydrocarbon chain W 1 can be obtained from Eqs. (6-14), and (6-15) as follows:
139 )/()()/2(02min0min0011aHHHaNHHHEdHFNWA W (6-16) Eq. (6-16) assumes only half of the surfactant molecules in the surface aggregate are desorbed/detached, while another half (possibly in the form of monolayer of surfactant) remains adsorbed on the surface. This assumption is based on the fact that even under high loading pressures, which are sufficient to destroy the surface aggregates, the friction coefficient of silica remains much smaller (0.015) than the coefficient for bare silica surfaces (0.1) [ADL02]. Additionally, the adhesion force, measured after the surface aggregates have been destroyed (50 mN/m ), is much higher as compared to bare silica surfaces (10 mN/m ), indicating the presence of monolayers, which lead to higher adhesion due to hydrophobic attraction between the tip and the substrate coated with surfactant monolayer (both are hydrophobic in nature). Experimentally measured adsorption density of C 12 TAB on silica at bulk concentration of 16 mM (twice the CMC with 0.1M NaCl), is equal to = 7.510 -6 moles/m 2 [ADL00, SIN01]. The area per adsorbed head group, based on the adsorption density, is approximately equal to 0.4 nm 2 . The area per head group at the air-water interface at maximum packing is reported to be between 0.36-0.38 nm 2 [ROS89]. This indicates that at a bulk concentration of twice the CMC, saturation adsorption, and hence maximum packing of the surfactant molecules is achieved. As a result, it can be assumed that the adsorption at twice the CMC is independent of the substrate material (mica or silica), and the number of carbon atoms in the surfactant tail. Evidence for this is provided by Strm et. al. [STR00] where they report the adsorption of C 12 -, C 14, and C 16 TAB at twice the CMC (without electrolyte) on silica to be equal to approximately 5*10 -6 moles/m 2 . Hence, adsorption value of = 8*10 -6 moles/m 2 in the presence of 0.1
140 M NaCl was used for all the calculations, based on the experimentally measured value for C 12 TAB in the present study. Eqs. (6-14), and (6-16) assume the surfactant aggregates to be in the form of a flat layer (bilayer). This assumption is correct for C 18 TAB [LI98], but not for C 10 -C 16 TAB, which form layer of spherical (silica) or cylindrical (mica) aggregates. As mentioned above, for spherical aggregates of 6 nm diameter, even for a maximum penetration of 3 nm, it can be approximated that a tip with 15 nm in radius interacts with only one (central) spherical aggregate. This approximation allows the use of a simple Hertz model for interaction of a hard sphere (tip) with a soft sphere (surfactant aggregate), without taking into account the interactions with the neighboring surfactant aggregates, which would necessitate the development of a model to take care of the inter-aggregate interactions. This approximation is valid since the inter-micellar distances are same or even larger than the size of the micellar aggregates themselves [SCH02]. In this case, under pressure with deformation near 2 nm the spherical micelles are stretched along a horizontal axis, forming an ellipsoid/pancake, so other (adjoining) micelles donâ€™t play any role in the interaction. If it is correct, then the following simple formula allows the calculation of energy per hydrocarbon chain: agHnFdHW/2max1 (6-17) Where F is the interaction force between tip and surface aggregate (i.e., force measured by AFM), and n ag is the aggregation number of the surface aggregate. As explained above, coefficient â€ in Eq. (6-17) signifies that only half of the surfactant molecules in the surface aggregate are desorbed/detached, while another half remains adsorbed on the surface. In the following calculations, a value of n ag =115 for the surface structures was
141 used. This value of n ag was calculated on the basis of the size of surface aggregates, and the adsorption of C 12 TAB on silica, reported earlier [ADL00, SIN01]. It should be noted that n ag will be different for different substrates and chain lengths, and also vary with the surfactant concnetration. Combining Eqs. (6-5), and (6-17), and integrating the force/distance curve, the following formula is obtained for energy per chain in the spherical surface aggregate: agefnHHKRW5.2)(25.2min05.01 (6-18) Where R ef is calculated using Eq. (6-8). Energies of surfactant aggregates on silica and mica for different chain lengths are given in Table 6-1a and 6-1b, and are also plotted in Figure 6-6. Each point was obtained from experimental force/distance curve for interaction of silicon nitride tip with mica or silica substrate. Experimental values of W 1 were calculated from experimental force/distance curves using Eq. (6-18) for C 10 -C 16 TAB and Equation (6-16) for C 18 TAB. Figure 6-6 shows the calculated values of W 1 with increasing â€œnâ€. It is interesting to note that extrapolating the curve to n=8, approximately zero energy for surface aggregates of C 8 TAB is obtained, which indicates that either these aggregates do not exist, or are too unstable to be detected in the present experimental conditions. This fact was confirmed by measuring forces in solutions of C 8 TAB, and as predicted from the extrapolation, no measurable forces were observed irrespective of surfactant concentration up to 280 mM (twice the bulk CMC without any added electrolyte). As mentioned earlier, Subramanian, and Ducker [SUB01] have used experimental force/distance curves between glass particle (diameter = 8 m), and dodecytrimethlyammonium bromide aggregates on silica surfaces to calculate specific
142 (per unit area) energy of the surfactant aggregates. The values obtained by them are about ten times smaller compared to those calculated in the present study. This difference can be explained by the fact that the measured forces in the presence of self-assembled surface aggregates do not scale with the radius of the probe used, and hence the values obtained by them using a 8 m probe are much smaller compared to the values calculated in the present study for the given AFM tip (diameter 15 nm). As mentioned earlier, the forces being measured in this system are contact forces, i.e., forces generated after contact between the AFM tip, and the surfactant layer, and hence will not scale with the radius of probe. The energy per chain of surface aggregates, calculated using Eqs. (6-16), or (6-18) from experimental force/distance curves, were further compared with energy per chain in bulk micelles W 1, bulk . For ionic surfactants, the free energy of micellization (demicellization) can be estimated using the following expression derived on the basis of a mass action model [MOR92]: 5.55ln1,1CM C k T W bulk (6-19) Where CMC is in mole/l and is the fraction of charges of micellized surfactant ions neutralized by micelle-bound counter-ions. The term accounts for the counter-ion binding energy. In 0.1 M NaCl solution is close to unity [MAJ98]. When the micelle is destroyed, the surfactant monomer goes back into solution, and then dissociates, thus decreasing the entropy of the surrounding water molecules. Hence, this part of the micellization energy should not be taken into account in calculating the energy to break the micelles. Eq. (6-19) will thus reduce to the following expression, which has the same form as the equation for nonionic micelles:
143 5.55ln,1CMCkTWbulk (6-20) 02468108101214161820n (# Carbon Atoms)W1 (10-20J/chain)SilicaMicaBulk Figure 6-6: Energetics of surface aggregate, and bulk micelle formation. Dependence of micellization energy per hydrocarbon chain (W 1 ) of bulk micelles (triangles) and the surface surfactant aggregates (circles for silica and squares for mica) on the length of hydrocarbon chain. Another strategy for calculating the bulk micellization energy, proposed by Mohanthy et. al [MOH01], using the free energy approach introduced by Puvvada, and Blankschtein [PUV90, PUV92] estimates W 1, bulk 4.510 -20 J/tail for C 16 TAB. This agrees with the value obtained using Eq. (6-20) (~ 4.8 x 10 -20 J/tail). Comparison of the effect of chain length on experimental energy per chain for surface aggregates, W 1 , with theoretical values, W 1, bulk , calculated for bulk micelles using Eq. (6-20) is also plotted in Figure 6-6. The absolute values of energy (within the experimental variations) for surface aggregates on both silica, and mica are very similar to those for bulk micelles, indicating
144 similarities between the formation of self-assembled surfactant surface structures and the formation and stability of bulk micelles. The slope of the lines for surface aggregates on silica and mica in Figure 6-6 are greater as compared with that of bulk micelles. This indicates that the preference to form surface aggregates over bulk micelles increases as the chain length of the surfactant increases, i.e., the ratio of critical surfactant concentration for surface structures, CMC s (surfactant concentration where onset of steric repulsive forces takes place) to critical bulk micelle concentration CMC b should decrease with increasing chain length. This prediction was confirmed by experimental measurements of CMC s on mica. It was found that the ratio, (CMC s /CMC b ), was highest for C 10 TAB (~ 0.85), and decreased with increasing chain length, with the value for C 16 TAB being equal to 0.35. In the case of silica, the effect was less pronounced, with (CMC s /CMC b ), for C 10 TAB being 1.0, and the value for C 16 TAB being equal to 0.8. It should be noted that the CMC s reported here is based on the force measurements, and that the formation of micelle-like surface aggregates may occur at lower concentrations. On the other hand, present experiments have shown that surface structures on mica form at lower concentrations than on silica (e.g for C 12 TAB (CMC s /CMC b ) = 0.5 for mica, and (CMC s /CMC b ) = 1.0 for silica), i.e., the curve W 1 vs. n for mica should lie above that of silica, whereas the opposite is observed based on our calculations. As mentioned earlier, this indicates that the calculations for cylindrical micelles using Eq. (6-18), which was developed for spheres, gives rise to erroneous energy values in the case of cylindrical surface structures formed on mica.
145 Summary Hertz model of interaction between a hard sphere (tip) and a soft flat layer (bilayer-C 18 TAB) or soft sphere (surfactant aggregate-C 10 -C 16 TAB) can be used to fit experimentally measured interaction forces between the AFM tip, and self-assembled surfactant aggregates on solid substrates. Values of layer thickness (H 0 ), elastic modulus (E), and yield strength (Y), obtained from the fitting procedure, are comparable with corresponding parameters of some soft polymeric materials. Elastic modulus of surface aggregates increases slightly or remains constant with number of carbon atoms per chain for C 10 -C 16 TAB. Sharp increase of E occurs for C 18 TAB aggregates on mica, which may be related to the formation of a bilayer type of structure. The values of energy per chain within the surface aggregates, W 1 , on the solid/liquid interface, calculated by integrating the force/ distance curves, are of the same order as the energy W 1, bulk for bulk micelles, and increase linearly with the chain length, demonstrating similar behavior as bulk micelles. The dependence of energy on chain length indicates that the preference to form surface aggregates over bulk micelles increases as the chain length of the surfactant increases. Additionally, it is indicated that the theoretical relationships developed for interaction energies of spherical surface aggregates are not necessarily valid for other shapes such as cylindrical surface aggregates. In this chapter it was shown that the AFM measurements could be used to determine the mechanical properties (elastic modulus, and yield strength) of the surface surfactant aggregates, and the associated energetics. The mechanical properties are dependent on the intermolecular packing within the surface aggregates. An indication of the intermolecular packing is the energy per CH 2 group in the surfactant tail. As described in this chapter, higher energy per CH 2 group indicates higher cohesion, and
146 lower intermolecular spacing within the surface aggregates. Additionally, it can also be used to predict the critical surfactant concentration required in order to achieve surface aggregation, as compared to bulk aggregation. This will provide important guidelines in determining the optimal dosage of the surfactant dispersant required to produce the desired barrier to agglomeration. Additionally the difference in the energetics between different surfaces can be used as guidelines when designing selective dispersants in a multi-component system of particles.
CHAPTER 7 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK Conclusions Traditional and emerging technologies such as advanced structural ceramics, controlled drug delivery systems, abrasives for precision polishing, coatings, inks, and nanocomposite materials are increasingly relying on nanoparticulate precursor materials to achieve optimum performance. In order to achieve optimal performance in these processes, it is important to produce well-dispersed nanoparticulate suspensions. The dispersion of nanoparticles becomes challenging due to the fact that dispersant adsorption, stabilization mechanism, and interparticle force prediction for nanoparticles is not very well understood. Theories such as the DLVO theory for colloidal stability may not be valid for smaller (nm size particles) [DER87]. Additionally, nanoparticles are increasingly being used under extreme processing conditions such as high salt (chemical mechanical polishing; CMP, biofluids in medical applications), high pressure (CMP, high speed coatings), and presence of complex additives (CMP, nanocomposites, bio applications). Therefore, to develop suitable dispersants for nanoparticulates, the underlying fundamental mechanisms need to be understood. In the present study, the use of self-assembled surfactant aggregates at the solid-liquid interface was investigated for dispersing nanoparticles under extreme environments. Adsorbed surfactant aggregates were found to disperse 200 nm silica particles under high electrolyte (upto 5 M NaCl), high normal forces (200 mN/m), and high shear rate (> 20,000 s -1 ) conditions. It was also determined that electrostatic 147
148 repulsion due to adsorbed surfactants was not sufficient to produce a stable suspension under the present experimental conditions. Dispersion due to steric forces was thought to be the primary dispersion mechanism, and the origin of the steric repulsion was due to the mechanical strength of self-assembled surfactant structures at the solid-liquid interface. The interparticle forces in the presence of self-assembled surfactant aggregates were measured using atomic force microscope (AFM). The magnitude of the forces in the presence of surface aggregates was at least an order of magnitude higher, as compared to pure electrostatic forces, indicating that resistance to elastic deformation of surfactant structures was the primary stabilization mechanism at high ionic strength. The transition from unstable to stable suspension, and from no repulsive forces to repulsive forces occurs over a narrow concentration range, and a good correlation between onset of forces, and stability was observed. In the case of silicaC12TAB system, this transition was observed at surfactant concentrations just beyond the bulk critical micelle concentration (cmc) (8 mM at 0.1 M NaCl). To further understand the dispersion mechanism, the interfacial surfactant structure was investigated using techniques such as adsorption (amount of surfactant at the surface), zeta potential (charge), contact angle (measure of hydrophobicity of the surfactant structures), and FTIR/ATR (average surfactant orientation within the surface aggregates). It was found that initially, at very low concentrations, surfactants adsorb individually, and randomly. As the concentration increases, 2-D surfactant structures known as hemi-micelles form at the interface, due to increasing hydrophobic attraction between the surfactant chains. At a specific surfactant concentration, direct transition from hemi-micelles to spherical surface aggregates was observed, without the formation
149 of bilayers. At even higher surfactant concentrations, the surface is fully covered with randomly oriented spherical surface aggregates. The transition in surface aggregate structure to predominantly spherical aggregates occurs before the transition in forces. This indicates that the restructuring of the self-assembled surface surfactant layer was not a sufficient condition to achieve stability through steric repulsive forces. Proposed possible reasons for the transition in stability are as follows. In the transition region, the aggregation number of the surface aggregates increases, and hence the area per surfactant molecule decreases by about 7 A o2 . The reduced separation between the molecules will lead to increased hydrophobic attraction, thus rendering the surface aggregates to be more mechanically stronger. Alternatively, it was proposed that below transition concentration, the surface coverage by the aggregates is low (40% of saturation), and the estimated distance between the aggregates (~3 nm) is large, enabling their escape from the region between the two approaching surfaces. On the other hand, above the transition concentration, the surface coverage is high (50% of saturation), and the separation between the aggregates decreases to less than 2 nm, perhaps limiting the mobility of the aggregates, leading to the observation of steric repulsive forces due to the elastic deformation of the aggregates. Additionally, the increased surface coverage may also result in a transition from liquid-like ensemble to a solid-like ensemble of the surface aggregates, thus limiting their mobility. Increased surface coverage can also hinder the thinning of the intervening fluid film, further reducing the mobility of the aggregates. The effect of alkyl chain length, electrolyte concentration, co-surfactant addition, and temperature on the magnitude of the repulsive force was also investigated. Behavior of the self-assembled surfactant surface structures was noted to mimic the bulk micelles.
150 The role of surface in the surfactant self-assembly process was also investigated. It was observed that for crystalline substrates such as mica, silicon, and alumina, the onset of forces occurs at concentrations below the bulk cmc, whereas for silica, which is amorphous in nature, the transition occurs at concentrations just beyond the bulk cmc. In order to further understand the catalytic role of the surface, self-assembly process was investigated using different crystal planes of alumina. It was shown that the magnitude, and concentration for onset of repulsive forces depends on the crystallographic orientation, and appear to be directly dependent on the density of the adsorption sites on the different crystal planes. It should be noted that the energetics of the sites might also vary due to different distribution of atoms on the crystal planes, and can influence the self-assembly process. The role of the nature, density, and energetics of the surface adsorption sites need to be further investigated. Additionally, a methodology to calculate, and predict the mechanical properties, and the associated energetics of the surface aggregates was developed. Hertz model of interaction between a hard sphere (tip) and a soft flat layer (bilayer-C 18 TAB) or soft sphere (surfactant aggregate-C 10 -C 16 TAB) was used to fit experimentally measured interaction forces between the AFM tip, and self-assembled surfactant aggregates on solid substrates. Values of layer thickness (H 0 ), elastic modulus (E), and yield strength (Y), were obtained from the fitting procedure. The mechanical properties of the self-assembled surfactant structures are found to be comparable with those of some soft polymeric materials such as low-density polyethylene, and rubber (e.g., E = 40-150 MPa, Y = 10-50 MPa). It was determined that for C 10 -C 16 TAB, the elastic modulus of surface aggregates increases slightly or remains constant with number of carbon atoms per chain.
151 However, a sharp increase in E occurs for C 18 TAB aggregates on mica, which may be related to the formation of a bilayer type structures, which were observed using AFM imaging [LI98]. The values of energy per chain within the surface aggregates, W 1 , on the solid/liquid interface, calculated by integrating the force/ distance curves, are of the same order as the energy W 1, bulk for bulk micelles (~ (1-9)*10 -20 J/CH 2 group), and increase linearly with the chain length, demonstrating similar behavior as bulk micelles. The dependence of energy on chain length indicates that the preference to form surface aggregates over bulk micelles increases as the chain length of the surfactant increases. In the bulk solution, the driving force for micellization is the presence of the hydrophobic chain which causes solvent molecules to form an ordered structure, thus reducing the entropy of the solvent molecules, and increasing the free energy of the system This driving force increases as the chain length increases. However, on the surface, apart from the entropic driving force, the formation of nucleating sites, or seeds due to hydrophobic attraction between the neighboring adsorbed surfactant molecules, can play a key role in surfactant aggregation, and self-assembly. For lower chain length, the adsorption at the solid surface may take place randomly with little or no hydrophobic attraction between the neighboring chains, thus eliminating the additional driving force for surface aggregation. However, as the chain length increases, the surfactant adsorption is in the form of ordered patches (seeds) formed due to the increased hydrophobic attraction between the chains. These seeds are the nucleating sites for the surface aggregates, and provide an additional driving force for surfactant adsorption. Hence, as the chain length increases, the formation of surface aggregates is more favorable compared to bulk micelles.
152 The research carried out in the present work has revealed the feasibility of self-assembled surfactant aggregates to act as dispersants under extreme conditions such as high electrolyte concentration, and high normal and shear forces. However, the most important scientific contribution is the role of intermolecular spacing, and the energetics of surface aggregate formation in the mechanical resiliency of the surface aggregates, which leads to steric repulsive forces. The intermolecular spacing can be controlled using several methods such as surfactant structure (chain length), additives (electrolyte, co-surfactants), processing conditions (temperature, pH), and the surface microstructure (crystalline versus amorphous, density of the adsorption sites). Suggestions for Future Work The mechanisms of dispersion of nanoparticles in the presence of self-assembled surfactant aggregates has been delineated in the present study, by investigating the process of self-assembly of surfactant molecules at the solid-liquid interface. The transition in the surface aggregate structure, especially at low surfactant concentrations (< bulk CMC) has been determined based on measurement of interface properties, such as adsorption density, zeta potential, contact angle, and FTIR/ATR. Only at high surfactant concentrations (~twice the bulk CMC), the aggregate structure can be directly imaged at the interface using the AFM technique. However, the structures inferred on the basis of the experiments are the equilibrium structures, and do not provide information about the dynamics of the structure formation. The dynamics of structure formation can reveal the mechanisms of surfactant self-assembly, and can explain the transition in surface structure, which are not very well understood. For example, the direct transition from hemi-micelle type structures to micelle-like surface aggregates, without the formation of bilayers, may be explained by studying the dynamics.
153 Techniques such as Ellipsometry [TIB93], and neutron reflectivity [MCD92] have been used to study the kinetics of structure formation at the solid-liquid interface. The minimum time for data collection in these techniques is on the order of (10-30) seconds. However, the relaxation time of micellar aggregates can be in the milli seconds range, and individual monomer adsorption time can be as low as pico-seconds. Hence, as mentioned above, only equilibrium information about the surface structures can be obtained using experimental techniques. However, to understand the mechanism of structure development at the interface, it is imperative to study the dynamics of the self-assembly process. Therefore, the development of molecular dynamic (MD) simulations [SHI99] to study the dynamics of the surfactant self-assembly at the pico-second time scale is essential. This will enable the visualization of the self-assembly process at a molecule-by-molecule level, at low surfactant concentrations, where experimental tools will not work. Also, the molecular cohesion, which results in the mechanical resiliency of the surface aggregates, can be predicted for different conditions such as surfactant concentration, surfactant structure (chain length), solution conditions (pH, ionic strength), and different substrates (site density, amorphous versus crystalline). Additionally, by incorporating the Monte Carlo method into the MD simulations, the shape, and concentration for surfactant self-assembled structures could be predicted. This will help in determining the threshold concentration of surfactants required to achieve optimal stability. Additionally, the simulations can be used to investigate the mobility of the surface aggregates as two surfaces approach each other. The reasons for transition of the interface structure from a liquid-like ensemble to a solid-like ensemble, which leads to repulsive forces, and stability, can also be investigated. The simulations can shed light on
154 why a critical amount of rigid self-assembled surfactant aggregates are needed at the interface to observe steric repulsive forces, and hence suspension stability under high electrolyte concentration. Although dynamics are essential to understand the mechanisms of self-assembly, thermodynamic models can be used to predict the surfactant concentrations needed for self-assembly. Based on these models, the cohesiveness of the structures as a function of the surfactant concentration can be predicted. These models can thus be used as predictive tools for designing optimally performing slurries. Thermodynamic model developed by Johnson, and Nagarajan [JOH00a, JOH00b] can be used as a starting point to develop more robust models, which can predict the shape, and surfactant concentration at which aggregates form. In the model proposed by them, the formation of the most energetically favored structure is predicted, without taking into account the amount of surfactant adsorbed at the interface. For example, Johnson, and Nagarajan have predicted the formation of structures resembling semi-spheres on top of perfect monolayers at low concentrations (0.096 mM), where the adsorption density is only 2% of theoretical monolayer coverage, thus ruling out the formation of such structures. By incorporating the surfactant adsorption density, which can be measured reliably, such inconsistencies in the model can be avoided. Additionally, interfacial properties such as zeta potential, and contact angle, which can be measured accurately, can be used as input parameters to identify plausible surfactant structures. The interaction parameters (headgroup-surface, chain-chain) used by Johnson, and Nagarajan are based on theoretically derived potentials, which may not be accurate. Instead, the direct measurement of the adsorption energy using calorimetry at different concentrations can be used as an input to determine
155 the energetics of the structure formation. One of the major drawbacks of the Johnson, Nagarajan model is that the unfavorable effect of hydrophobic edges (exposed to water) of the proposed structures has been ignored. This should be taken into account in predicting the interfacial structures. The work done in the present study to calculate the mechanical, and thermodynamic properties of the surface aggregates has provided useful insight into the origin of the steric repulsive forces, which result in stability under extreme processing conditions. But due to limitations of the models used, the results predicted are not always accurate (as depicted in the calculation of energy of surface aggregates on mica). Development of models for the interaction of the AFM tip with surface aggregates of different shapes is required, to accurately differentiate the role played by different substrates. One of the most important conclusions in this study is the role of molecular level cohesion in controlling particulate dispersion stability in the presence of surfactant dispersants. However, further studies are needed to quantify the stability, the nature, and orientation of the surfactant molecules within the surface aggregates, in order to fully understand the role played by molecular level cohesion which controls the mechanical strength of surface aggregates. The stop-flow technique, which has been used to study micellar stability in the bulk solution, could be modified by using electrodes impervious to particulate contamination, and dilute particulate suspension of known conductivity as a background reference, to determine the stability of the surface aggregates. However, this technique will only be sensitive for systems where surface aggregation is known to occur before the bulk micelle formation (mica, alumina) because once bulk micellization
156 occurs, the majority of the signal will come from the breaking of the bulk micelles. Additionally, techniques such as electron spin resonance (ESR), and fluorescence (discussed earlier in chapter 3) can be used to determine the aggregation number, hydrophobic environment, and the microviscosity within the surface aggregates. With information on the aggregation number, and the microviscosity within the same aggregate, the cohesiveness per monomer can be calculated. This information will provide a tool to predict the barrier to agglomeration, or the mechanical resiliency of the surface aggregates, based on a molecular level understanding. Though it has been clearly indicated that the molecular level cohesion controls the steric repulsive barrier, reliable tools to determine, and calculate the molecular level cohesion are not available. One of the simplest ways to look at molecular level cohesion is to calculate the area per surfactant molecule within the surface aggregates. This can be precisely calculated by knowing the adsorption density, and the shape, and size of the surface aggregates. AFM provides a tool to directly measure the shape, and size of the aggregates, but the availability of reliable techniques to quantify surfactant adsorption on flat surfaces (used for AFM experiments) is limited. Techniques such as ellipsometry, and neutron reflection have been used for estimating the adsorption of surfactants on flat plates. However, these techniques require fitting parameters such as refractive index, and density of the adsorbed layer, which are difficult to measure, and vary, with surfactant concentration. Hence, other experimental techniques need to be considered. As discussed in Chapter 3, one of the techniques being used for measuring adsorption of large molecular weight proteins is quartz crystal microbalance (QCM). This measurement is based on the change in frequency with changing mass, and hence is a direct measurement
157 of the adsorbed amount. However, the mass thus calculated includes also the mass of the water associated with the adsorbing moieties, which can lead to substantial errors in measuring the adsorption amount of surfactants, which have a low molecular weight (200-400 grams). If corrections for the adsorbed water can be developed, this technique can provide direct adsorption measurements, without the need for fitting parameters. Another promising technique is the optical waveguide lightmode spectroscopy (OWLS). Since this optical method is based on the measurement of the polarizability density (i.e., refractive index) in the vicinity of the waveguide surface, radioactive, fluorescent or other kinds of labeling are not required. In addition, measurement of at least two guided modes enables the absolute mass of adsorbed molecules to be determined. However, the theoretical basis for estimating the adsorbed mass still needs to be developed. Hence, the development of the above-mentioned techniques such as quartz crystal microbalance (QCM), optical waveguide lightmode spectroscopy (OWLS), and ellipsometer can accurately quantify surfactant adsorption on flat plates, leading to more precise values of intra molecular cohesion. As was seen in the present study, not only does the self-assembly, and molecular cohesion depend on the density of the adsorption sites, but they also depend on the nature (amorphous versus crystalline), and energetics of the surface sites. Tools to quantify the density of the adsorption sites, their nature, and their energetics need to be developed. Some of the techniques that could be used for this are x-ray crystallography, Raman spectroscopy, and the NMR. Additionally, a more detailed study is required in order to correlate the density, and energetics of the adsorption sites to the self-assembly of surfactants.
158 From an industrial point of view, stability of the suspensions over a long time period (the shelf life) is very critical in designing optimal, and commercially viable slurries. Hence, the stability of the surfactant stabilized slurries over long period of times (several days to months) needs to be evaluated. The key in this case would be the stability of the self-assembled surfactant aggregates at the solid-liquid interface, which can be achieved by engineering the strength, and cohesiveness of the surface films. In the bulk micelles, the stability is determined by measuring the slow relaxation time ( 2 ). Patist et. al [PAT00] have measured 2 values for several surfactants using the stop flow technique, and have illustrated the effect of additives (such as alcohols, co-surfactants), surfactant concentration, and chain length on the slow micellar relaxation time. As mentioned earlier, the stop flow technique with some modifications could be used for surface aggregates to determine the slow relaxation time of the surface aggregates. In this study, the factors controlling the magnitude of the repulsive forces (such as co-surfactant, chain length, temperature, alcohols) have been discussed. All the factors that increase the cohesion within the surface aggregates, leading to higher steric repulsive forces, will also lead to increased surface aggregate stability. Work needs to be carried out to correlate the stability of the surface surfactant structures with the long-term stability, or shelf life of slurries dispersed using surfactants. Apart from the use of self-assembled surfactants for dispersion, basic understanding of the process can provide significant advances in areas such as mesoporous materials for electronic, magnetic, and optical applications, where self-assembled surfactant aggregates can act as templates for functionalizing surfaces at the nano level. These templates can be used to generate controlled patterns, and porosity in
159 the inorganic materials by depositing the inorganic materials onto the organic templates. Once the patterns have been formed, the organic template (self-assemble surfactant structure) can be easily burned off, leaving a mesoporous material with controlled nanoporosity. Organic surfactant molecules are already being used [KIM96] for designing two-dimensional templates for fabrication of inorganic nanostructures such as macroscopically oriented composite films, polysiloxane networks, and multi-layered thin films of oriented metal oxide particles. The packing pattern of the hydrophilic head groups, and the intermolecular spacing between the organic structures play an important role in the formation of such structures. Additionally, the manipulation of the cohesiveness of the surfactant aggregates can be used in controlled drug delivery systems. Not only can the self-assembled surfactant aggregates provide stability to the drug particles within the human body â€“ a complex environment, they can also be used to modulate the packing at the drug-body fluid interface, thus controlling the rate of diffusion of drug molecules into the body fluids. If fast release of the drug is desired, the cohesiveness of the aggregates can be tailored to be low, allowing faster diffusion, and release rates. On the other hand, if a slow, and sustained release of the drug is desired, the higher cohesiveness of the aggregates will result in reduced diffusion of drug molecules into the body fluids.
LIST OF REFERENCES ADL00 Adler, J. J., Singh P. K., Patist, A., Rabinovich, Y. I., Shah, D. O., and Moudgil, B. M., Langmuir , 16 , 7255 (2000). ADL02 Adler, J. J., Rabinovich, Y. I., Esayanur, M. S., Vakarelski, I. U., and Moudgil, B. M., J. Colloid Interface Sci. (in review) AKS96 Aksay, I. A., Trau, M., Manne, S., Honma, I., Yao, N., Zhou, L., Fenter, P., Eisenberger, P. M., and Gruner S. M., Science , 273 , 892 (1996). ANA85 Ananthapadmanabhan, K., and Somasundaran, P., Colloids and Surfaces , 13 , 151 (1985). ANI76 Aniansson, E. A. G., Wall, S. N., Almgren, M., Hoffmann, H., Kielmann, I., Ulbricht, W., Zana, R., Lang, J., and Tondre, C., J. Phys. Chem. , 80 , 905 (1976). BAN95 Banga, R., Yarwood, J., and Morgan, A.M., Langmuir , 11 , 618 (1995). BAS01 Basim, G. B., Adler, J. J., Mahajan, U., Singh, R. K., and Moudgil, B. M., J. Electrochem. Soc. , 147 , 3523 (2000). BAS02 Basim, G. B., Formulation of Engineered Particulate Systems for Chemical Mechanical Polishing Applications , Ph.D. Dissertation, University of Florida, Gainesville, FL, 2002. BEL97 Bell, N. S., Interfacial Aspects of Glycothermally Synthesized Alpha Alumina , Ph.D. Dissertation, University of Florida, Gainesville, FL, 1997. BIN00 Binks, B. P., Cho, W-G., Fletcher, P. D. I., and Petsev, D. N., Langmuir , 16 , 1025 (2000). BRE99 Bremmell, K.E., Jameson, G.J., and Biggs, S., Colloids Surf. , 146 , 75 (1999). CES88 Cesarano, J. III, and Aksay, I. A., J. Am. Ceram. Soc. , 71 , 1062 (1988). CHA86 Chandar, P., Somasundaran, P., Waterman, K. C., and Turro, N. J., J. Phys. Chem., 91 , 150 91986). CHA87 Chandar, P., Somasundaran, P., and Turro, N. J. J. Colloid Interface Sci. , 117 , 31 (1987). 160
161 CHR01 Christenson, H. K., and Claesson, P. M., Adv. Colloid and Interface Sci. , 91 , 391 (2001). CLE93 Cleveland, J. P., Manne, S., Bockec, D., and Hansma, P. K., Rev. Sci. Instrum. , 64 , 1 (1993). COL97 Colic, M., and Fuerstenau, D. W., Langmuir , 13 , 6644 (1997). CUN96 Cunha, F., Tao, N. J., Wang, X. W., Jin, Q., and Duong, B., Langmuir , 12 , 6410 (1996). CUY83 Cuypers, Y. A., Corsel, J. W., Janssen, M. P., Kop, JM. M., Hermens, T., and Hemker, H. C., J. Biological Chem. , 258 , 2426 (1983). DAG00 Dagastine, R. R., Prieve, D. C., and White, L. R., J. Colloid and Interface Sci. , 231 , 351 (2000). DAV63 Davies, J. T., and Rideal, E. K., Interfacial Phenomena , Academic Press, New York, 1963. DAV98 Davey, D. W., Ducker, W. A., Hayman, A. R., and Simpson, J., Langmuir , 14 , 3210 (1998) DEB36 de Boer, J. H., Trans. Faraday Soc. , 32 , 10 (1936). DER34 Derjaguin, B. V., Kolloid Zeitschrift , 69 , 155 (1934). DER41 Derjaguin, B. V., and Landau, L. D., Acta Phys. Chim. USSR , 14 , 633 (1941). DER87 Derjaguin, B. V., Churaev, N. V., and Muller, V. M., Surface Forces , Plenum Publishing Corp., New York, 1987. DER98 Deruelle, M., Hervet, H., Jandeau, G., and Leger, L., J. Adhesion Sci. Thechnol. , 12 , 225 (1998). DEW98 DeWolf, C., Brezesinski, G., Weidemann, G., Mhwald, H., Kjaer, K., and Howes, P. B., J. Phys. Chem. B , 102 , 3238 (1998). DUC96 Ducker, W. A., and Grant, L. M., J. Phys. Chem. , 100 , 11507 (1996). DUC99 Ducker, W. A., and Wanless, E. J., Langmuir , 15 , 160 (1999). ESK98 Eskilsson, K., and Yaminsky, V. V., Langmuir , 14 , 2444 (1998). EVA96 Evanko, C. R., Dzombak, D. A., and Novak, J. W., Colloids Surf. , 110 , 219 (1996). FAH61 Fahrenfort, J., Spectrochim. Acta , 17 , 698 (1961).
162 FAN97 Fan, A., Somasundaran, P., and Turro, N. J., Langmuir , 13 , 506 (1997). FIE99 Fielden, M. L., Claesson, P., and Verrall, R. E., Langmuir , 15 , 3924 (1999). FLE93 Fleer, G., Cohen Stuart, M., Cosgrove, T., and Vincent, B., Polymers at Interfaces , Chapman & Hall, London, 1993. FLO61 Flockhart, B. D., J. Coll. Int. Sci. , 16 , 484 (1961). FRE91 Frey, S., and Tamm, L.K., Biophys. J. , 60 , 922 (1991). FUE64 Fuerstenau, D. W., Healy, T. W., and Somasundaran, P., Trans. AIME , 229 , 321 (1964). FUE75 Fuerstenau, D. W., and Wakamatsu, T., Faraday Discuss. Chem. Soc. , 59 , 157 (1975). GAU55 Gaudin, A.M., and Fuerstenau, D.W., Trans. AIME , 202 , 958 (1955). GIN72 Gingell, D. and Parsegian, V. A., J. Theor. Biol. , 36 , 41 (1972). HAC97 Hackley, V. A., Paik, U., Kim, B.-H., and Malghan, S. G., J. Am. Ceramic Soc. , 80 , 1781 (1997). HAI00 Haidara, H., Mougin, K., and Schultz, J., Langmuir , 16 , 7773 (2000). HAL70 Haller, G. L., and Rice, R. W., J. Phys. Chem. , 74 , 4386 (1970). HAM37 Hamaker, H. C., Physica , 4 , 1058 (1937). HAR60 Harrick, N. J., J. Phys. Chem., 64 , 1110 (1960). HAR67 Harrick, N. J., Internal Reflection Spectroscopy , Wiley, New York, 1967. HAY98 Hayes, W. A., and Schwartz, K., Langmuir , 14 , 5913 (1998). HER96 Hertz, H., in Miscellanious papers , McMillan&Co., London, 1896. HEU96 Heurberger, M., Dietler, G., and Schlapbach, L., J Vac. Sci. Technol. B , 14 , 1250 (1996). HIG90 Higashitani, K., Kondo, M., and Hatade, S., J. Colloid Interface Sci. , 142 , 204 (1990). HIM98 Himics, R., and Pineiro, R., Process Finishing , 63 , 8 (1998). HIR97 Hirata, Y., Ceramics Int. , 23 , 93 (1997).
163 H Hk, F., Rodahl, M., Kasemo, B., and Brzezinski, P., Proc. Natl. Acad. Sci. USA, Biophysics , 95 , 12271 (1998). HOU80 Hough, D. B., and White, L. R., Adv. Colloid and Interface Sci. , 14 , 3 (1980). HOU83 Hough, D. B., and Rendall, H. M., in Adsorption from Solution at the Solid/Liquid Interface , (G. D. Parfitt, C. H. Rochester Eds.), p. 247, Academic Press, London, 1983. HUN01 Hunter, R. J., Foundations of Colloid Science , 2 nd ed., Oxford University Press, Oxford, 2001. HUT99 Hutchinson, J., Klenerman, D., Langmuir , 15 , 7530 (1999). ISR92 Israelachvili, J. N., Intermolecular and Surface Forces , Academic Press, London, 1992. JAN93 Jang. W., and Miller, J.D., Langmuir , 9 , 3159 (1993). JAN95 Jang. W., and Miller, J.D., J. Phys. Chem. , 99 , 10272 (1995). JOH00a Johnson, R.A., and Nagarajan, R., Colloids Surf. , 167 , 21 (2000). JOH00b Johnson, R.A., and Nagarajan, R., Colloids Surf. , 167 , 31 (2000). KAN83 Kaneshina, S., Shibata, O., Nakamura, M., and Tanaka, M., Colloids Surf.: A, 6 , 73 (1983). KEK89 Kekicheff, P., Christenson, H. K., and Ninham, B. W., Colloids Surf.: A , 40 , 31 (1989). KEL98 Keller, C. A., and Kasemo, B., Biophysical J. , 75 , 1397 (1998). KIM96 Kimizuka, N., and Kunitake, T., Adv. Mater. , 8 , 89 (1996). KOO99 Koopal, L.K., Goloub, T., de Keizer, A., and Sidorova, M.P., Colloids Surf. , 151 , 15 (1999). KRI96 Krishnakumar, S., and Somasundaran, P., Colloids Surf. A , 117 , 37 (1996). KUN93 Kung, K-H. S., and Hayes, K. F., Langmuir , 9 , 263 (1993). LAU98 Lautz, C., Fischer, T. M., Weygand, M., Losche, M., Howes, P. B., and Kjaer, K., J. Chem. Phys. , 108 , 4640, 1998. LEE89 Lee, E. M., Thomas, R. K., Cummins, P. G., Staples, E. J., Penfold, J., and Rennie, A. R., Chem. Phys. Let. , 162 , 196 (1989). LEN28 Lennard-Jones, J. E. and Dent, B. M., Trans. Faraday Soc. , 24 , 98 (1928).
164 LEV67 Levine, S., Mingins, J., and Bell, G. M., J. Electroanal. Chem. Interface Electrochem. , 13 , 280 (1967). LI97 Li, F., Li, G-Z., Xu, G-Y., Wang, H-Q., and Wang, M., Colloids Surf. A , 127 , 89 (1997). 1997). LI98 Li, B., Fujii, M., Fukada, K., Kato, T., and Seimiya, T., Thin Solid Films , 312 , 20 (1998). LIF56 Lifshitz, E. M., Sov. Phys. JETP , 2 , 73 (1956). LON37 London, F., Trans. Faraday Soc. , 33 , 8 (1937). LYK95 Lyklema, J., Fundamentals of Interface and Colloid Science, Vol. II: SolidLiquid Interfaces , Academic Press, London, 1995. MAJ98 Majhi, P. R., and Moulik, S. P., Langmuir , 14 , 3986 (1998). MAL92 Malbrel, C. A., and Somasundaran, P., Langmuir , 8 , 1285 (1992). MAN94 Manne, S., Cleveland, J. P., Gaub, H. E., Stucky, G. D., and Hansma, P. K., Langmuir , 10 , 4409 (1994). MAN95 Manne, S., and Gaub, H. E., Science , 270 , 1480 (1995). MAT93 Matsuda, H., Kishi, E., Kuroda, R., Albrecht, O., Eguchi, K., Hatanaka, K., and Nakagiri, T., Thin Solid Films , 224 , 248 (1993). MAT00 Mathur, M., Singh, P. K., and Moudgil, B. M., Int. J. Min. Proc. , 58 , 201 (2000). MCD92 McDermott, D. C., Lu, J. R., Lee, E. M., Thomas, R. K., and Rennie, A. R., Langmuir , 8 , 1204 (1992). MEL97 Melzer, V., Weidemann, G., Vollhardt, D., Brezesinski, G., Wagner, R., Struth, B., and Mohwald, H., Supramol. Sci. , 4 , 391 (1997). MEL98 Melzer, V., Vollhardt, D., Weidemann, G., Brezesinski, G., Wagner, R., and Mohwald, H., Phys. Rev. E , 57 , 901 (1998). MOH01 Mohanty, S., Davis, H. T., Toritani, A., and McCormick, A. V., Langmuir , 17 , 7160 (2001). MOR92 Moroi., Y., Micelles, Theoretical and Applied Aspects , Plenum Press, New York, 1992. MOU88 Moudgil, B. M., Soto, H., and Somasundaran, P., in Reagents in Mineral Technology , (P. Somasundaran, B. M. Moudgil Eds.), p. 79, Marcel Dekker, New York, 1988.
165 NIN70 Ninham, B. W., and Parsegian, V. A., Biophys. J. , 10 , 646 (1970). PAS81 Pashley, R. M., and Israelachvili, J. N., Colloids Surf.: A , 2 , 169 (1981). PAS88 Pashley, R. M., McGuiggan, P. M., Horn, R. G., and Ninham, B. W., J. Colloid Interface Sci. , 126 , 569 (1988). PAT98 Patist, A., Huibers, P. D. T., Deneka, B., and Shah, D. O., Langmuir , 14 , 4471 (1998). PAT99 Patrick, H. N., Warr, G. G., Manne, S., and Aksay, I. A., Langmuir , 15 , 1685 (1999). PAT00 Patist, A., Oh, S. G., Leung, R., and Shah, D. O., Colloids Surf.: A , 176 , 3 (2000). PIN77 Pinkley, L. W., Sethna, P. P., and Williams, D., J. Opt. Soc. Am. , 67 , 494 (1977). PUV90 Puvvada, S., and Blankshtein, D., J. Chem. Phys , 92 , 3710 (1990). PUV92 Puvvada, S., and Blankshtein, D., J. Phys Chem., 96 , 5567 (1992). RAB93 Rabinovich, Y.I, Guzonas, D.A., and Yoon, R.H., J. Colloid Interface Sci ., 155 , 221 (1993). REE95 Reed, J. S., Principles of Ceramic Processing : 2nd ed. , John Wiley & Sons, New York, 1995 ROS89 Rosen, M. J., Surfactants and Interfacial Phenomena; 2nd ed. , John Wiley & Sons, New York, 1989. RUS89 Russel, W. B., Saville, D. A., and Schowalter, W. R., Colloidal Dispersions , Cambridge University Press, Cambridge, 1989. RUT94 Rutland, M. W., and Parker, J. L., Langmuir, 10 , 1110 (1994). SCA81a Scamehorn, J. F., Schechter, R. S., and Wade, W. H., J. Colloid Interface Sci. , 85 , 463 (1981). SCA81b Scamehorn, J. F., Schechter, R. S., and Wade, W. H., J. Colloid Interface Sci. , 85 , 479 (1981). SCA86 Scamehorn, J. F., in Phenomena in Mixed Surfactant Systems , (J. F. Scamehorn, Ed.), Chap. 1, American Chemical Society, Washington DC, 1986. SCH02 Schulz, J. C., and Warr, G. G., Langmuir , 18 , 3191 (2002).
166 SHA71 Shah, D. O., J. Colloid Interface Sci., 37 , 744 (1971). SHI99 Shinto, H., Tsuji, S., Miyahara, M., and Higashitani, K., Langmuir, 15 , 578 (1999). SIN01 Singh, P. K., Adler, J. J., Rabinovich, Y. I., and Moudgil, B. M., Langmuir, 17 , 468 (2001). SNE65 Sneddon, J. N., Int. J. Eng. Sci. , 3 , 47 (1965). SOL99 Solomon, M. J., Saeki, T., Wan M., Scales, P. J., Boger, D. V., and Usui, H., Langmuir , 15 , 20 (1999). SOM64 Somasundaran, P., Healy, T. W., and Fuerstenau, D. W., J. Phys. Chem. , 68 , 3562 (1964]. SOM86 Somasundaran, P., Turro, N. J., and Chander, P., Colloid Surf. , 20 , 145 (1986). SOM87 Somasundaran, P., Chander, P., and Turro, N. J., J. Coll. Inter. Sci. , 117 , 31 (1987). SPE87 Sperline, R. P., Muralidharan, S., and Freiser, H., Langmuir , 3 , 198 (1987). SPE92 Sperling, L. H., Introduction to Physical Polymer Science , Wiley-Interscience Publication, New York, 1992. SPE97 Sperline, R. P., Song, Y., and Freiser, H., Langmuir , 13 , 3727 (1997). STA85 Stauffer, D., Introduction to Percolation Theory , Taylor and Francis, London and Philadelphia, 1985. STE58 Stein, R.S., J. Polym. Sci. , 31 , 327 (1958). STE61 Stein, R.S., J. Polym. Sci. , 50 , 339 (1961). STE97 Stevens, F., Buehner, D., and Beebe, T. P., J. Phys. Chem. B , 101 , 6491 (1997). STR00 Strm, C., Hansson, P., Jnsson, B., and Sderman, O., Langmuir , 16 , 2469 (2000). SUB01 Subramanian, V., and Ducker, W., J. Phys. Chem. B , 105 , 1389 (2001). TAB77 Tabor, D., J. Colloid and Interface Sci. , 58 , 2 (1977). TAD82 Tadros, Th. F., in The Effect of Polymers on Dispersion Properties , (Th. F. Tadros, Ed.), p. 1, Academic Press, London, 1982.
167 TEE97 Teer, E., Knobler, C. M., Lautz, C., Wurlitzer, S., Kildae, J., and Fischer, T. M., J. Chem. Phys , 105 , 1913 (1997). TIB93 Tiberg, F., and Landgren, M., Langmuir , 9 , 927 (1993). TOW95 Towey, T. F., Baglioni, P., Martini, G., and Ristori, S., J. Phys. Chem. , 99 , 3939 (1995). ULM90 Ulman, A., in Fourier Transform Infrared Spectroscopy in Colloid and Interface Scinece , (S. R. Scheuing, Ed.), p 144, American Chemical Society, Washington DC, 1990. VAK01 Vakarelski, I. U., Toritani, A., Nakayama M., and Higashitani, K., Langmuir , 17 , 4739 (2001). VAN73 van der Waals, J. D., Over de Continuteit van de GasEn Vloeistoftoestand (On the Continuity of the Gaseous and Liquid States), Ph.D. Dissertation, Leiden University, Netherlands, 1873. VEL00 Velegol, S.B., Fleming, B. D., Biggs, S., Wanless, E. J., and Tilton, R. D., Langmuir , 16 , 2548 (2000). VER48 Verwey, E. J. ., and Overbeek, J. Th. G., Theory of the Stability of Lyophobic Colloids , Elsevier, Amsterdam, 1948. VID98 Vidon, S., and Leblanc, R. M., J. Phys. Chem B , 102 , 1279 (1998). VOL97 Vollhardt, D., and Melzer, V., J. Phys. Chem. B , 101 , 3370 (1997). VR02 Vrs J., Ramsden, J. J., Sccs, G., Szendr, I., De Paul, S. M., Textor, M., and Spencer, N. D., Biomaterials , 23 , 3699 (2002). WAK68 Wakamatsu, T., and Fuerstenau, D. W., Adv. Chem. Ser. , 79 , 161 (1968). WAL99 Wall, J. F., and Zukoski, C. F., Langmuir , 15 , 7432 (1999). WN92 Wngernerud, P., and Olofsson, G., J. Coll. Interf. Sci. , 153 , 392 (1992). WAN97 Wanless, E. J., Davey, T. W., and Ducker, W. A., Langmuir , 13 , 4223 (1997). WAT86 Waterman, K. C., Turro, N. J., Chandar, P., and Somasundaran, P., J. Phys. Chem. , 90 , 6829 (1986). WU98 Wu, G., and Matijevic, E., J. Dispersion Sci. Tech. , 19 , 903 (1998). ZBI64 Zbinden, R.Z., in Infrared Spectroscopy of High Polymers , Academic Press, New York, 1964.
168 ZHA96 Zhang, Y., Li, Q., Xie, Z., Hua, B., Mao, B., Chen, Y., and Tian, Z., Thin Solid Films , 274 , 150 (1996). ZLA95 Zlatkin, A., Yudin, S., Simon, J., Hanack, M., and Lehman, H., Adv. Mat. Opt. Electr. , 5 , 259 (1995).
BIOGRAPHICAL SKETCH Pankaj was born in Jamshedpur, India on January 24, 1973. He finished his high school education from Rajendra Vidyalaya, India. In August 1992 he was admitted into the Materials and Metallurgical Engineering program at the Indian Institute of Technology (IIT), Kanpur. He earned a Bachelor of Technology degree from IIT in May 1996, and was ranked third in his class. In August 1996, he came to the United States of America in pursuit of graduate education, and joined the Department of Materials Science and Engineering at the University of Florida (UF). While at UF, he obtained a masterâ€™s degree in the Spring of 1999, and expects to obtain a Doctor of Philosophy degree in the Summer of 2002. After graduation he plans to join Unilever HPC, Chicago, Illinois. 171