Search

Colloidal suspensions are an experimental model system for studying structural and mechanical properties of soft materials. These properties are manifested differently in colloidal solid-like phases such as crystals, glasses and gels. To further understand relations between structural and mechanical properties, it is necessary to develop well-defined colloids and employ techniques such as microscopy and rheology to study the structure and mechanics of their suspensions. This thesis presents five experimental chapters dealing with the synthesis and characterization of colloids and their suspensions. The first part of the thesis describes facile synthesis methods for latex, conjugated polymer and microgel colloids. In the second part, measurements of crystal-to-glass and glass-to-gel phase transformations in dense suspensions of microgel particles are presented.

Emulsions are dispersions of one liquid in another that are commonly used in various products, and methods such as high-pressure homogenisers and colloid mills are used to form emulsions. The size and size distribution of emulsion droplets are important for the final product properties and thus need to be controlled. Rapid coalescence of droplets during emulsification increases droplet size and widens the size distribution, and therefore needs to be prevented. To increase stability of emulsions, emulsifiers are added to adsorb at the oil-water interface before droplets collide. The time allowed for emulsifier adsorption is typically in the range of sub-milliseconds to seconds and to optimise emulsification processes, emulsifier adsorption and coalescence stability need to be measured in this time-scale, for which the microfluidic methods described in this thesis were developed. Chapter 2 provides an overview of existing literature on cross-flow microfluidic emulsification. The effects of various parameters such as microfluidic design, shear forces, and interfacial tension forces on droplet formation and the resulting droplet size are discussed, as well as the use of microfluidics to produce food-grade emulsions. Based on this evaluation, the methods to elucidate interfacial tension and coalescence stability are chosen, and these are presented in the next chapters. To measure emulsifier adsorption in the sub-millisecond time-scale, a tensiometric method was developed using a cross-flow microfluidic Y-junction, which is described in Chapter 3. This method is based on the relation between droplet size and interfacial tension at the moment of droplet formation, which is referred to as the acting interfacial tension. The acting interfacial tension of a system with hexadecane as the dispersed phase and sodium dodecylsulfate (SDS, a model surfactant) solutions as the continuous phase was successfully measured for droplet formation times ranging from 0.4 to 9.4 milliseconds and with high expansion rates (100-2000 s-1). Comparison of these results with data from a drop tensiometer (a conventional, static, and supra-second time-scale method) indicates that mass transport in the microfluidic Y-junction is fast and probably not limited by diffusion. Emulsifier mass transport conditions were further investigated in Chapter 4. The continuous phase viscosity and velocity were systematically varied and the effect on the acting interfacial tension in presence of water-soluble SDS was measured. We found that the acting interfacial tension was independent of the continuous phase viscosity, but was inversely dependent on continuous phase velocity. Both aspects led us to conclude that convective emulsifier transport in the continuous phase determines the acting interfacial tension in the Y-junction. When using oil-soluble surfactant Span 20 (dissolved in hexadecane), the acting interfacial tension also decreased with increasing continuous phase velocity, and we therefore concluded that convection also dominated mass transport of emulsifiers dissolved in the to-be-dispersed phase. The Y-junction method was used in Chapter 5 to elucidate the effect of the dispersed phase viscosity on adsorption of the food-grade emulsifiers Tween 20 (dissolved in the continuous water phase) and Span 20 (dissolved in the dispersed oil phase). A reduction in dispersed phase viscosity sped up adsorption of Tween 20, probably because the shorter hydrocarbon made intercalation of the hydrophobic surfactant tail at the interface easier. Dispersed phase viscosity had an even greater effect on adsorption of Span 20 because convective transport towards the interface was increased. Next to interfacial tension, also coalescence can be measured with microfluidics and a microfluidic collision channel was used in Chapter 6 to measure emulsion coalescence stability shortly after droplet formation under flow. Coalescence of emulsions stabilised with proteins was measured at various concentrations, pH values, and adsorption times. We found that protein concentrations just below the concentration needed for monolayer surface coverage may be used effectively. β-lactoglobulin-stabilised emulsions were most stable. Emulsions stabilised with whey protein isolate (with as main component β-lactoglobulin), were less stable and when these proteins were oxidised, this led to reduced stability, therewith indicating that also the oxidative state of proteins needs to be considered in emulsion formulation. The relevance of our work for microfluidic research and industrial emulsification processes is discussed in Chapter 7. Microfluidic devices can be used to study emulsion formation and stability under conditions relevant to industrial emulsification processes; at short time-scales and with convective mass transport. In this thesis we used various food-grade ingredients, and with that application in that field has come closer. We expect that the findings on emulsions can also be applied on foams. With the discussed microfluidic devices different aspects that are important for emulsion formation can be decoupled: for example interfacial tension during droplet formation and emulsion coalescence stability. Furthermore, microfluidic methods are available to for example gain insight in emulsion interface mobility and emulsion storage stability, and we envision that all these microfluidic methods will lead to faster ingredient screening, lower ingredient usage, and more energy efficient emulsion production.

We can learn from nature that, next to chemistry, surface structures can be used for tuning different functions of surfaces. In this thesis we present a novel fabrication method using colloidal templating on the air/water interface. Two distinct ways to obtain nanopatterned surfaces are described, namely (i) addition of PDMS on top of the colloidal monolayer and (ii) synthesis of polypyrrole around the particles of the monolayer. An increase in adhesion is found for the nanopatterned PDMS surfaces, and the contact angle of water on the nanopatterned polypyrrole surface is increased.

The possibility of preparing 2D stable emulsions through mixing of homopolymers in a Langmuir monolayer is the core topic of this thesis. While colloid science has achieved well established results in the study of bulk dispersed systems, accounts on properties of mixed monomolecular films are fewer, and seldom systematic. The aim of this investigation is to contribute to a deeper understanding of the subject, in order to explore opportunities to apply the acquired knowledge to the fabrication of technologically relevant materials. In particular, this study focused on a possibly applicable, innovative strategy for the manipulation of the morphology and the patterning of mixed Langmuir monolayers: the possibility to stabilize and control a dispersion of homopolymers through the addition of a lineactant (the equivalent of a surfactant in three dimensional systems), able to adsorb preferentially at the interfacial contact line of polymer domains, thereby lowering the interfacial energy (line tension) in the system and favoring an effective dispersion of one component into the other. The state of the art of the preparation and investigation of 2D colloids is the subject of Chapter 2, which is a comprehensive review on several systems able to yield phase–separated Langmuir monolayers, and includes a general definition of the concept of a 2D colloid, the most relevant instrumental techniques and experimental tools available, a summary of several systems suitable for preparing 2D colloid dispersions, an introduction to the concept of lineactant, and several examples, both experimental and theoretical, in which compounds acting as lineactants have been investigated. This review clearly shows that the polymer–based mixtures are a poorly explored subject, when compared to amphiphiles of natural origin, and so the rest of the thesis has been devoted to the investigation of polymer–based Langmuir monolayers. This investigation has been carried out with two parallel approaches: classical experiments at the Langmuir trough and morphological characterization of the Langmuir monolayers with the Brewster Angle Microscope have been performed, along with Self–Consistent Field modeling of the same systems. The setup of the SCF model and comparison of SCF calculation with experimental data from the reference experiments are dealt with in Chapter 3. Surface pressure isotherms at the air/water interface were reproduced for four different polymers, poly–l–lactic acid (PLLA), poly (dimethylsiloxane) (PDMS), poly (methyl methacrylate) (PMMA), and poly (isobutylene) (PiB). The polymers are all insoluble in water, but display a different degree of amphiphilicity; therefore the four isotherms differed strongly. The polymers were described through a SCF model on a united atom level, taking the side groups on the monomer level into account. In line with experiments, the model shown that PiB spread in a monolayer which smoothly thickened at a very low surface pressure and area/monomer value. The monolayer made of PMMA had an autophobic behavior: a PMMA liquid did not spread on top of the monolayer of PMMA at the air/water interface. A thicker PMMA layer only formed after the collapse of the film at a relatively high pressure. The isotherm of PDMS had regions with extreme compressibility which were linked to a layering transition. Finally, PLLA wetted the water surface and spread homogeneously at larger areas per monomer. The classical SCF approach features only short–range, nearest–neighbor interactions. For the correct positioning of the layering and for the thickening of the polymer films, a power–law van der Waals contribution was taken into account in this model. Two–gradient SCF computations were performed to model the interface between two coexistent PDMS films at the layering transition, and an estimation of the length of their interfacial contact was obtained, together with the associated line tension value. The SF–SCF molecularly detailed modeling of PLLA, PDMS, PMMA, and PiB monolayers, spread at the air/water surface, has proven to be consistent with experimental data: the incorporation in the model of a detailed molecular description of the monomeric features of the four compounds examined has been crucial to reproducing the features of the adsorption and pressure/area isotherms. In Chapter 4, the same approach was applied to the description of polymer mixtures spread at the air/water interface. The aim of this chapter was to analyze topics such as 2D phase separation and partitioning in mixed polymeric Langmuir monolayers. Two of the four polymers studied in Chapter 3 were selected in order to obtain a mixed Langmuir monolayer. A system consisting of water–insoluble, spreadable, fluid–like polymers was prepared. The polymers were polydimethylsiloxane (PDMS) and polymethylmethacrylate (PMMA), combined, in some cases, with a minority of PDMS–b–PMMA copolymer. Both Langmuir trough pressure/area isotherm measurements and Brewster angle microscopy (BAM) observations were performed, and complemented with molecularly detailed self–consistent field (SCF) calculations. It was shown that PDMS undergoes a layering transition that is difficult to detect by BAM. Addition of PMMA enhanced contrast in BAM, showing a two–phase system: if this consisted of separate two–dimensional (2D) PMMA and PDMS phases, a PDMS–PMMA diblock should accumulate at the phase boundary. However, the diblock copolymer of PDMS–PMMA failed to show the expected “lineactant” behavior, i.e., failed to accumulate at the phase boundary. The calculations pointed to a non-trivial arrangement of the polymer chains at the interface: in mixtures of the two homopolymers, in a rather wide composition ratio, a vertical (with respect to the air/water interfacial plane) configuration was found, with PMMA sitting preferably at the PDMS/water interface of the thicker PDMS film, during the PDMS layering phase transition. This also explained why the diblock copolymer was not a lineactant. Both PMMA and PDMS–b–PMMA were depleted from the thin–thick PDMS film interface, and the line tension between the phases consequently increased in the binary mixtures, as well as in the ternary ones. The results shown in this chapter proved that gaining an accurate control over thin film structures at the microscopic level is a far from trivial task, and the acquisition of fundamental knowledge is necessary in order to interpret experimental data in an appropriate way. As a consequence, in Chapter 5 an investigation based solely on SCF modeling was carried out, in order to analyze which polymer blends could have the possibility to undergo lateral phase separation in two dimensions. Specifically, the model system investigated consisted of water–supported Langmuir monolayers, obtained from binary polyalkyl methacrylate mixtures (PXMA, where X stands for any of the type of ester side groups used: M, methyl–; E, ethyl–; B, butyl–; H, hexyl–; O, octyl–; L, lauryl–methacrylate). In particular, the conditions which determined demixing and phase separation in the two–dimensional system were addressed, showing that a sufficient chain length mismatch in the ester side group moieties is able to drive the polymer demixing. When the difference in length of the alkyl chain of the ester moieties on the two types of polymers was progressively reduced, from 11 carbon atoms (PMMA/PLMA) to 4 carbons only (POMA/PLMA), the demixing tendency was also reduced; it vanished, indeed, for POMA/PLMA. In the latter case the polymer/subphase interactions affected more the distribution of the polymer coils in the blend monolayer: mixing of the two polymers was observed, but also a partial layering along the vertical direction. Lineactancy was also considered, by selecting the mixture in which phase separation was best achieved: a third component, namely a symmetrical diblock copolymer of the type PLMA–b–PMMA, was added to a PMMA/PLMA blended monolayer. Adsorption of the diblock copolymer was observed exclusively at the contact line between the two homopolymer domains, together with a concomitant lowering of the line tension. The line tension varied with chemical potential of the diblock copolymer according to the Gibbs’ law, which demonstrated that PLMA–b–PMMA indeed acted as a lineactant (the two–dimensional analog of a surfactant) in the model system made of a binary demixed PMMA/PLMA Langmuir monolayer. In conclusion, the requirements needed to achieve polymer blend demixing in a Langmuir monolayer are the following: spreadable, insoluble polymers, with the same amphiphilicity degree, combined to a certain chemical mismatch of the side moieties are necessary in order to cause lateral demixing at the air/water interface. The polyalkyl methacrylate example investigated in the chapter represented a suitable model system, since the methacrylate backbone guarantees that the different polymers have the same affinity towards the water subphase, while the different ester moieties drive the occurrence of lateral demixing. The dependency of the lateral demixing on the difference in length between the two ester side groups chosen was demonstrated. A rather complex interplay of forces regulates the distribution of the polymer coils in the monolayer: subtle alterations of this complex balance might favor the dewetting of the mixture in a single domain, together with the layering of the blended polymers along the direction normal to the air/water interface, as well as accumulation of one polymer at the domain edge, instead of the occurrence of the lateral phase separation. Furthermore, the possibility to control emulsification of two–dimensional demixed polymer blends was proven. This was achieved by use of a diblock copolymer, which acted as a lineactant by adsorbing at the contact line of the polymer domains. The calculations demonstrated the possibility to extend the lineactant concept, first elaborated in the context of lipid membrane investigations, to the field of study of polymer thin films.

The possibility of preparing 2D stable emulsions through mixing of homopolymers in a Langmuir monolayer is the core topic of this thesis. While colloid science has achieved well established results in the study of bulk dispersed systems, accounts on properties of mixed monomolecular films are fewer, and seldom systematic. The aim of this investigation is to contribute to a deeper understanding of the subject, in order to explore opportunities to apply the acquired knowledge to the fabrication of technologically relevant materials. In particular, this study focused on a possibly applicable, innovative strategy for the manipulation of the morphology and the patterning of mixed Langmuir monolayers: the possibility to stabilize and control a dispersion of homopolymers through the addition of a lineactant (the equivalent of a surfactant in three dimensional systems), able to adsorb preferentially at the interfacial contact line of polymer domains, thereby lowering the interfacial energy (line tension) in the system and favoring an effective dispersion of one component into the other. The state of the art of the preparation and investigation of 2D colloids is the subject of Chapter 2, which is a comprehensive review on several systems able to yield phase–separated Langmuir monolayers, and includes a general definition of the concept of a 2D colloid, the most relevant instrumental techniques and experimental tools available, a summary of several systems suitable for preparing 2D colloid dispersions, an introduction to the concept of lineactant, and several examples, both experimental and theoretical, in which compounds acting as lineactants have been investigated. This review clearly shows that the polymer–based mixtures are a poorly explored subject, when compared to amphiphiles of natural origin, and so the rest of the thesis has been devoted to the investigation of polymer–based Langmuir monolayers. This investigation has been carried out with two parallel approaches: classical experiments at the Langmuir trough and morphological characterization of the Langmuir monolayers with the Brewster Angle Microscope have been performed, along with Self–Consistent Field modeling of the same systems. The setup of the SCF model and comparison of SCF calculation with experimental data from the reference experiments are dealt with in Chapter 3. Surface pressure isotherms at the air/water interface were reproduced for four different polymers, poly–l–lactic acid (PLLA), poly (dimethylsiloxane) (PDMS), poly (methyl methacrylate) (PMMA), and poly (isobutylene) (PiB). The polymers are all insoluble in water, but display a different degree of amphiphilicity; therefore the four isotherms differed strongly. The polymers were described through a SCF model on a united atom level, taking the side groups on the monomer level into account. In line with experiments, the model shown that PiB spread in a monolayer which smoothly thickened at a very low surface pressure and area/monomer value. The monolayer made of PMMA had an autophobic behavior: a PMMA liquid did not spread on top of the monolayer of PMMA at the air/water interface. A thicker PMMA layer only formed after the collapse of the film at a relatively high pressure. The isotherm of PDMS had regions with extreme compressibility which were linked to a layering transition. Finally, PLLA wetted the water surface and spread homogeneously at larger areas per monomer. The classical SCF approach features only short–range, nearest–neighbor interactions. For the correct positioning of the layering and for the thickening of the polymer films, a power–law van der Waals contribution was taken into account in this model. Two–gradient SCF computations were performed to model the interface between two coexistent PDMS films at the layering transition, and an estimation of the length of their interfacial contact was obtained, together with the associated line tension value. The SF–SCF molecularly detailed modeling of PLLA, PDMS, PMMA, and PiB monolayers, spread at the air/water surface, has proven to be consistent with experimental data: the incorporation in the model of a detailed molecular description of the monomeric features of the four compounds examined has been crucial to reproducing the features of the adsorption and pressure/area isotherms. In Chapter 4, the same approach was applied to the description of polymer mixtures spread at the air/water interface. The aim of this chapter was to analyze topics such as 2D phase separation and partitioning in mixed polymeric Langmuir monolayers. Two of the four polymers studied in Chapter 3 were selected in order to obtain a mixed Langmuir monolayer. A system consisting of water–insoluble, spreadable, fluid–like polymers was prepared. The polymers were polydimethylsiloxane (PDMS) and polymethylmethacrylate (PMMA), combined, in some cases, with a minority of PDMS–b–PMMA copolymer. Both Langmuir trough pressure/area isotherm measurements and Brewster angle microscopy (BAM) observations were performed, and complemented with molecularly detailed self–consistent field (SCF) calculations. It was shown that PDMS undergoes a layering transition that is difficult to detect by BAM. Addition of PMMA enhanced contrast in BAM, showing a two–phase system: if this consisted of separate two–dimensional (2D) PMMA and PDMS phases, a PDMS–PMMA diblock should accumulate at the phase boundary. However, the diblock copolymer of PDMS–PMMA failed to show the expected “lineactant” behavior, i.e., failed to accumulate at the phase boundary. The calculations pointed to a non-trivial arrangement of the polymer chains at the interface: in mixtures of the two homopolymers, in a rather wide composition ratio, a vertical (with respect to the air/water interfacial plane) configuration was found, with PMMA sitting preferably at the PDMS/water interface of the thicker PDMS film, during the PDMS layering phase transition. This also explained why the diblock copolymer was not a lineactant. Both PMMA and PDMS–b–PMMA were depleted from the thin–thick PDMS film interface, and the line tension between the phases consequently increased in the binary mixtures, as well as in the ternary ones. The results shown in this chapter proved that gaining an accurate control over thin film structures at the microscopic level is a far from trivial task, and the acquisition of fundamental knowledge is necessary in order to interpret experimental data in an appropriate way. As a consequence, in Chapter 5 an investigation based solely on SCF modeling was carried out, in order to analyze which polymer blends could have the possibility to undergo lateral phase separation in two dimensions. Specifically, the model system investigated consisted of water–supported Langmuir monolayers, obtained from binary polyalkyl methacrylate mixtures (PXMA, where X stands for any of the type of ester side groups used: M, methyl–; E, ethyl–; B, butyl–; H, hexyl–; O, octyl–; L, lauryl–methacrylate). In particular, the conditions which determined demixing and phase separation in the two–dimensional system were addressed, showing that a sufficient chain length mismatch in the ester side group moieties is able to drive the polymer demixing. When the difference in length of the alkyl chain of the ester moieties on the two types of polymers was progressively reduced, from 11 carbon atoms (PMMA/PLMA) to 4 carbons only (POMA/PLMA), the demixing tendency was also reduced; it vanished, indeed, for POMA/PLMA. In the latter case the polymer/subphase interactions affected more the distribution of the polymer coils in the blend monolayer: mixing of the two polymers was observed, but also a partial layering along the vertical direction. Lineactancy was also considered, by selecting the mixture in which phase separation was best achieved: a third component, namely a symmetrical diblock copolymer of the type PLMA–b–PMMA, was added to a PMMA/PLMA blended monolayer. Adsorption of the diblock copolymer was observed exclusively at the contact line between the two homopolymer domains, together with a concomitant lowering of the line tension. The line tension varied with chemical potential of the diblock copolymer according to the Gibbs’ law, which demonstrated that PLMA–b–PMMA indeed acted as a lineactant (the two–dimensional analog of a surfactant) in the model system made of a binary demixed PMMA/PLMA Langmuir monolayer. In conclusion, the requirements needed to achieve polymer blend demixing in a Langmuir monolayer are the following: spreadable, insoluble polymers, with the same amphiphilicity degree, combined to a certain chemical mismatch of the side moieties are necessary in order to cause lateral demixing at the air/water interface. The polyalkyl methacrylate example investigated in the chapter represented a suitable model system, since the methacrylate backbone guarantees that the different polymers have the same affinity towards the water subphase, while the different ester moieties drive the occurrence of lateral demixing. The dependency of the lateral demixing on the difference in length between the two ester side groups chosen was demonstrated. A rather complex interplay of forces regulates the distribution of the polymer coils in the monolayer: subtle alterations of this complex balance might favor the dewetting of the mixture in a single domain, together with the layering of the blended polymers along the direction normal to the air/water interface, as well as accumulation of one polymer at the domain edge, instead of the occurrence of the lateral phase separation. Furthermore, the possibility to control emulsification of two–dimensional demixed polymer blends was proven. This was achieved by use of a diblock copolymer, which acted as a lineactant by adsorbing at the contact line of the polymer domains. The calculations demonstrated the possibility to extend the lineactant concept, first elaborated in the context of lipid membrane investigations, to the field of study of polymer thin films.

Fully atomistic analysis of polymeric systems is computationally very demanding because the time and length scales involved span over several orders of magnitude. At the same time many properties of polymers are universal in the sense that they do not depend on the chemical nature of the comprising monomers. This makes coarse-grained methods, such as self-consistent field (SCF) modeling, an ideal tool for studying them. In this thesis we employ SCF modeling to study intra- and intermolecular self-organization organization of polymers and ordering of polymers near interfaces. Where possible, the results are compared to experiments and predictions of analytical theories.

The research described in this thesis concerns the formation, solution properties, and adsorption of polyelectrolyte complexes composed of at least one diblock copolymer with a neutral and a charged block and either an oppositely charged homopolyelectrolyte or a diblock copolymer, with a neutral block and an oppositely charged polyelectrolyte block. Upon mixing the aqueous solutions of the different polymers, the oppositely charged polyelectrolytes associate, forming a polyelectrolyte complex. Polyelectrolyte complex micelles – called complex coacervate core micelles (C3Ms) in this thesis – are the main focus of this thesis, but the formation of smaller aggregates, soluble complex particles, is also investigated. The salt concentration, pH, and the chemical structure of the polyelectrolytes are important variables in the formation of these polyelectrolyte complexes. In chapter 2 C3Ms were made from multiple polymer species; a diblock copolymer with a polyelectrolyte block and a neutral block, poly(acrylic acid)-block-poly(acryl amide), an oppositely charged polyelectrolyte, poly(N,N-dimethyl aminoethylamide), and a second diblock copolymer species with a charged block and a neutral block, poly(N,N-dimethyl aminoethylamide)-block-poly(glyceryl methacrylate). The polyelectrolyte block of the second diblock copolymer species had charged blocks that were oppositely charged to that of the first diblock copolymer species and whose neutral block was different from that of the first diblock copolymer. The effect of systematically varying the ratio of the homopolyelectrolyte and second diblock copolymer (based on the number of chargeable groups), while keeping the mixing fraction f+ (that is the number of positively chargeable groups, divided by the total number of chargeable groups) constant, was studied with light scattering. It was shown that the size of the resulting C3Ms decreased with increasing percentage of the second diblock copolymer, from 25 nm hydrodynamic radius, to 16 nm. Using a simple geometrical model and the light scattering intensities, the aggregation numbers were estimated to be in the range of 20-70 polymers. In chapter 3 the used diblock copolymer, poly([4-(2-aminoethylthio)-butylene] hydrochloride)-block-poly(ethylene oxide), has a polyelectrolyte part with a rather hydrophobic backbone which slows down the formation of the aggregates and the subsequent rearrangements. It was mixed with the oppositely charged poly(acrylic acid). Using light scattering and cryogenic transmission electron microscopy, it was shown that the complexes formed at f+ = 0.3 are initially very large (> 140 nm) and network like (as there is relatively little neutral polymer to stop the growth of the complexes), and rearrange relatively quickly, compared to the complexes formed at f+ = 0.5 and 0.7 (80 nm), towards small micellar complexes. The very large transient complexes formed at f+ = 0.3 are called highly aggregated polyelectrolyte complexes (HAPECs). The complexes formed at f+ = 0.5 are apparently most stable; that is, their size remains the same in time. It was concluded that there are at least three factors which influence the rearrangement rate of polyelectrolyte complexes; (1) high neutral blocks content, (2) excess charge, and (3) the chemistry of the polyelectrolytes. Increasing the salt concentration has previously been determined to speed up the rate of rearrangements as well. Furthermore, the radius of the complexes at f+ = 0.5 (80nm) is too large for the complexes to have the typical core-corona structure. Apparently, these large complexes are HAPECs as well. However, with different preparation procedures micelles can be obtained; if the HAPECs are forced to disassemble by changing the pH to an extreme value (either 11 or 3) and are subsequently re-assembled by changing the pH back to normal (7), the resulting C3Ms have a radius of about 15 nm. This is probably the state of minimum free energy, the stable state, whereas the highly aggregated complexes are in a metastable state (as they do not spontaneously rearrange in time). In chapter 4 complex coacervate core micro-emulsions (C3-μEs) were obtained by mixing solutions of anionic polyelectrolytes (poly(acrylic acid)) and diblock copolymers with an anionic polyelectrolyte block and a neutral block (poly(acrylic acid)-block-poly(acryl amide)) with solutions of a cationic polyelectrolyte (poly(N,N-dimethyl aminoethylamide)). By varying the fraction of the anionic polyelectrolyte and anionic diblock copolymer species, while keeping f+ constant, C3-μEs with radii varying from about 15 to 100 nm were prepared. Basically, these are C3Ms of which the core is swollen with extra polyelectrolyte complex, composed of oppositely charged homopolyelectrolytes. The solvent was shown to have a pronounced effect upon the size of the obtained complexes; in NaNO3 larger complexes were obtained which are in a metastable state. In phosphate buffer (a salt known to weaken the attractive forces between the used polyelectrolytes), smaller complexes were obtained, which are probably in the stable state. The geometrical model introduced in chapter 2 was extended and predicted a linear growth of the C3-μEs. The experimentally observed growth was however, non-linear, probably due to a transition of the neutral polymers in the corona from more star-like to more crew-cut behaviour (shown by self consistent field calculations). In chapter 5 the ability of a layer of adsorbed C3Ms with a more glass-like core (composed of poly([4-(2-aminoethylthio)-butylene] hydrochloride)-block-poly(ethylene oxide) and poly([4-(2-carboxy-ethylthio)-butylene] sodium salt)-block-poly(ethylene oxide)), to prevent protein adsorption to either silica or cross-linked 1,2 polybutadiene was investigated. With atomic force microscopy it was shown that the layer consists of closely packed adsorbed complex coacervate core micelles. Protein adsorption to the coated surfaces was generally reduced by > 80 %. The different forces and many variable parameters of the investigated system cause the time scales on which SCPs and C3Ms rearrange to span a very wide range; they can be both reversible and irreversible systems.

The research described in this thesis concerns the formation, solution properties, and adsorption of polyelectrolyte complexes composed of at least one diblock copolymer with a neutral and a charged block and either an oppositely charged homopolyelectrolyte or a diblock copolymer, with a neutral block and an oppositely charged polyelectrolyte block. Upon mixing the aqueous solutions of the different polymers, the oppositely charged polyelectrolytes associate, forming a polyelectrolyte complex. Polyelectrolyte complex micelles – called complex coacervate core micelles (C3Ms) in this thesis – are the main focus of this thesis, but the formation of smaller aggregates, soluble complex particles, is also investigated. The salt concentration, pH, and the chemical structure of the polyelectrolytes are important variables in the formation of these polyelectrolyte complexes. In chapter 2 C3Ms were made from multiple polymer species; a diblock copolymer with a polyelectrolyte block and a neutral block, poly(acrylic acid)-block-poly(acryl amide), an oppositely charged polyelectrolyte, poly(N,N-dimethyl aminoethylamide), and a second diblock copolymer species with a charged block and a neutral block, poly(N,N-dimethyl aminoethylamide)-block-poly(glyceryl methacrylate). The polyelectrolyte block of the second diblock copolymer species had charged blocks that were oppositely charged to that of the first diblock copolymer species and whose neutral block was different from that of the first diblock copolymer. The effect of systematically varying the ratio of the homopolyelectrolyte and second diblock copolymer (based on the number of chargeable groups), while keeping the mixing fraction f+ (that is the number of positively chargeable groups, divided by the total number of chargeable groups) constant, was studied with light scattering. It was shown that the size of the resulting C3Ms decreased with increasing percentage of the second diblock copolymer, from 25 nm hydrodynamic radius, to 16 nm. Using a simple geometrical model and the light scattering intensities, the aggregation numbers were estimated to be in the range of 20-70 polymers. In chapter 3 the used diblock copolymer, poly([4-(2-aminoethylthio)-butylene] hydrochloride)-block-poly(ethylene oxide), has a polyelectrolyte part with a rather hydrophobic backbone which slows down the formation of the aggregates and the subsequent rearrangements. It was mixed with the oppositely charged poly(acrylic acid). Using light scattering and cryogenic transmission electron microscopy, it was shown that the complexes formed at f+ = 0.3 are initially very large (> 140 nm) and network like (as there is relatively little neutral polymer to stop the growth of the complexes), and rearrange relatively quickly, compared to the complexes formed at f+ = 0.5 and 0.7 (80 nm), towards small micellar complexes. The very large transient complexes formed at f+ = 0.3 are called highly aggregated polyelectrolyte complexes (HAPECs). The complexes formed at f+ = 0.5 are apparently most stable; that is, their size remains the same in time. It was concluded that there are at least three factors which influence the rearrangement rate of polyelectrolyte complexes; (1) high neutral blocks content, (2) excess charge, and (3) the chemistry of the polyelectrolytes. Increasing the salt concentration has previously been determined to speed up the rate of rearrangements as well. Furthermore, the radius of the complexes at f+ = 0.5 (80nm) is too large for the complexes to have the typical core-corona structure. Apparently, these large complexes are HAPECs as well. However, with different preparation procedures micelles can be obtained; if the HAPECs are forced to disassemble by changing the pH to an extreme value (either 11 or 3) and are subsequently re-assembled by changing the pH back to normal (7), the resulting C3Ms have a radius of about 15 nm. This is probably the state of minimum free energy, the stable state, whereas the highly aggregated complexes are in a metastable state (as they do not spontaneously rearrange in time). In chapter 4 complex coacervate core micro-emulsions (C3-μEs) were obtained by mixing solutions of anionic polyelectrolytes (poly(acrylic acid)) and diblock copolymers with an anionic polyelectrolyte block and a neutral block (poly(acrylic acid)-block-poly(acryl amide)) with solutions of a cationic polyelectrolyte (poly(N,N-dimethyl aminoethylamide)). By varying the fraction of the anionic polyelectrolyte and anionic diblock copolymer species, while keeping f+ constant, C3-μEs with radii varying from about 15 to 100 nm were prepared. Basically, these are C3Ms of which the core is swollen with extra polyelectrolyte complex, composed of oppositely charged homopolyelectrolytes. The solvent was shown to have a pronounced effect upon the size of the obtained complexes; in NaNO3 larger complexes were obtained which are in a metastable state. In phosphate buffer (a salt known to weaken the attractive forces between the used polyelectrolytes), smaller complexes were obtained, which are probably in the stable state. The geometrical model introduced in chapter 2 was extended and predicted a linear growth of the C3-μEs. The experimentally observed growth was however, non-linear, probably due to a transition of the neutral polymers in the corona from more star-like to more crew-cut behaviour (shown by self consistent field calculations). In chapter 5 the ability of a layer of adsorbed C3Ms with a more glass-like core (composed of poly([4-(2-aminoethylthio)-butylene] hydrochloride)-block-poly(ethylene oxide) and poly([4-(2-carboxy-ethylthio)-butylene] sodium salt)-block-poly(ethylene oxide)), to prevent protein adsorption to either silica or cross-linked 1,2 polybutadiene was investigated. With atomic force microscopy it was shown that the layer consists of closely packed adsorbed complex coacervate core micelles. Protein adsorption to the coated surfaces was generally reduced by > 80 %. The different forces and many variable parameters of the investigated system cause the time scales on which SCPs and C3Ms rearrange to span a very wide range; they can be both reversible and irreversible systems.

Fully atomistic analysis of polymeric systems is computationally very demanding because the time and length scales involved span over several orders of magnitude. At the same time many properties of polymers are universal in the sense that they do not depend on the chemical nature of the comprising monomers. This makes coarse-grained methods, such as self-consistent field (SCF) modeling, an ideal tool for studying them. In this thesis we employ SCF modeling to study intra- and intermolecular self-organization organization of polymers and ordering of polymers near interfaces. Where possible, the results are compared to experiments and predictions of analytical theories.

In this thesis we study the kinetics of fibrilar aggregation of two model proteins widely used in the food industry -b-lactoglobulin (b-lg) and hen egg white lysozyme (HEWL). The kinetics of protein aggregation is studied mostly experimentally and, when possible, theoretically. The process of fibrilar (or linear) protein aggregation is the process of formation of elongated structures from otherwise compact (globular) proteins. Studying the kinetics of this process for different proteins can lead to a better understanding of the mechanism of the process and to a possible generalisation of this mechanism. The investigation of the morphology of the formed aggregates at different stages of the process of aggregation could also lead to a more complete picture of the detailed mechanism of the process. Last, but not least, is the influence of the protein stability on the type of the formed aggregates and the kinetics of fibrilar aggregation.The specific aims of this thesis are the following: 1) To investigate the kinetics of heat-induced fibrilar aggregation of two model proteins, bovineb-lg and HEWL, in as much detail as possible; 2) To study the morphology of the fibrils formed from both proteins; 3) To study the influence of the environment such as temperature, pH, and ionic strength on the kinetics of fibrilar aggregation and the morphology of the formed fibrils.The heat-induced fibrilar aggregation ofb-lg is investigated at pH 2.0, 80°C, and at various ionic strengths. Fibril formation is followed in situ using static (SLS) and dynamic light scattering (DLS), small angle neutron scattering (SANS), and proton NMR techniques. The fibrils that form after short heating periods (up to a few hours) disintegrate upon slow cooling, whereas fibrils that form during long heating periods do not disintegrate upon subsequent slow cooling. Even after prolonged heating, an appreciable fraction of the protein molecules is incorporated into fibrils, only when theb-lg concentration is above some critical concentration that is ionic strength dependent.The linear aggregation ofb-lg upon prolonged heating at pH 2.0 at80°Cappears to be a multistep process. Competing reactions lead to two products: long linear aggregates and low molecular weight "dead end" species. The "dead end" species comprises monomeric non-native protein molecules and cannot form fibrils. Fibril formation involves at least two steps: the reversible formation of linear aggregates, followed by a slow process of "consolidation" after which the fibrils no longer disintegrate upon subsequent slow cooling.Based on the obtained experimental data we have derived a kinetic model for the heat-induced aggregation ofb-lg at pH 2.0. The model involves a nucleation step and a simple addition reaction for the growth of the fibrils as well as a side reaction leading to the complete denaturation and inactivation of a part of the protein molecules. An analytical solution of the model for the early stages of the aggregation is obtained. The model describes very well the experimental data obtained by in situ SLS. It allows us to obtain molecular parameters for the kinetics of fibrilar aggregation ofb-lg as a function of the ionic strength. It gives us an expression for the apparent critical concentration for fibril formation due to the competition between the complete denaturation of the protein molecules and the formation of long fibrils. We also obtain the size of the critical nucleus for the fibril formation as a function of the ionic strength. In the case of a 13 mM ionic strength the critical nucleus consists of ca. 4 monomers; for all the other ionic strengths studied it is a dimer. This shows the important role that the non-specific electrostatic interaction has for the fibrilar aggregation ofb-lg at pH 2.0. It affects the rate of aggregation: the higher the ionic strength, the faster the aggregation. It also affects the detailed mechanism by which the aggregation takes place: the size of the critical nucleus increases when decreasing the ionic strength from 50 mM to 13 mM.We have also shown that time-resolved SANS can be used with success in studying protein aggregation and that with enough additional information for the aggregation process one can in practice obtain complete information about the aggregation kinetics of the process.Tapping mode atomic force microscopy results indicate that the fibrils formed at pH 2.0 upon heating at 80°Chave a periodic structure with a period of about 25 nm and a thickness of one or two protein monomers. The main difference between the fibrils observed at different ionic strengths is their length and curvature. Fibrils obtained at higher ionic strength are shorter and more curved as opposed to longer and straighter fibrils obtained at lower ionic strengths. In case of higher ionic strength the fibril formation is faster, more fibrils are formed and as a result the mean length of the fibrils is shorter. Fibrils obtained at all ionic strengths exhibit similar type of periodic morphology, which suggests that the detailed mechanism of fibril formation might be independent of the ionic strength, but specific forb-lg.In the case of HEWL we study the effect of pH and temperature on the fibril formation. Fibril formation is promoted by low pH and temperatures close to the midpoint temperature for protein unfolding (detected using far-ultraviolet circular dichroism (CD)). The stability of HEWL toward heat treatment is greatly influenced by the pH. The lower the pH, the lower the stability of the protein is. The conditions at pH 2.0 are unique in promoting the fibrilar aggregation of HEWL since heating of solutions at pH 3.0 and 4.0 to temperatures just above the midpoint of the unfolding transition of the molecule does not lead to the appearance of fibrilar aggregates.HEWL fibrils are formed after a lag time that is practically concentration independent. This means that the governing process for the fibril formation is the change in the structure of single protein molecules caused by a prolonged exposure to a temperature close to the midpoint of the unfolding transition. Nucleation presumably involves a change in the conformation of individual lysozyme molecules. Indeed, long term CD measurements at pH 2.0, T = 57°C show a marked change of the secondary structure of lysozyme molecules after about 48 h of heating.The fibril morphology is complex. The fibrils formed at pH 2.0 are long and straight with a length of the order of 5mm and predominant thickness of about 4 nm and consist of stiff rod-like subunits with length either 124 or 157 nm. On smaller scale the fibrils consist of a coiled structure with a period of ca. 30 nm that gives the appearance of the rod-like subunits probably because of defects occurring every 4 or 5 turns.The fibrils consist mostly of full-length HEWL, although, some fragments due to hydrolysis at pH 2.0 and 57°C are probably incorporated into the fibrils. At any rate the hydrolysis of the protein is not the cause of the aggregation since at pH 3.0 no hydrolysis is detected but fibrils do form.In conclusion we can say that for a full and general description of the processes of fibrilar aggregation of globular proteins the type of specific interaction responsible for the aggregation must be identified. The interacting parts of the protein must also be identified. The last and most difficult task is to characterise the conformation of the protein in solution at conditions suitable for aggregation.

In this thesis we study the kinetics of fibrilar aggregation of two model proteins widely used in the food industry -b-lactoglobulin (b-lg) and hen egg white lysozyme (HEWL). The kinetics of protein aggregation is studied mostly experimentally and, when possible, theoretically. The process of fibrilar (or linear) protein aggregation is the process of formation of elongated structures from otherwise compact (globular) proteins. Studying the kinetics of this process for different proteins can lead to a better understanding of the mechanism of the process and to a possible generalisation of this mechanism. The investigation of the morphology of the formed aggregates at different stages of the process of aggregation could also lead to a more complete picture of the detailed mechanism of the process. Last, but not least, is the influence of the protein stability on the type of the formed aggregates and the kinetics of fibrilar aggregation.The specific aims of this thesis are the following: 1) To investigate the kinetics of heat-induced fibrilar aggregation of two model proteins, bovineb-lg and HEWL, in as much detail as possible; 2) To study the morphology of the fibrils formed from both proteins; 3) To study the influence of the environment such as temperature, pH, and ionic strength on the kinetics of fibrilar aggregation and the morphology of the formed fibrils.The heat-induced fibrilar aggregation ofb-lg is investigated at pH 2.0, 80°C, and at various ionic strengths. Fibril formation is followed in situ using static (SLS) and dynamic light scattering (DLS), small angle neutron scattering (SANS), and proton NMR techniques. The fibrils that form after short heating periods (up to a few hours) disintegrate upon slow cooling, whereas fibrils that form during long heating periods do not disintegrate upon subsequent slow cooling. Even after prolonged heating, an appreciable fraction of the protein molecules is incorporated into fibrils, only when theb-lg concentration is above some critical concentration that is ionic strength dependent.The linear aggregation ofb-lg upon prolonged heating at pH 2.0 at80°Cappears to be a multistep process. Competing reactions lead to two products: long linear aggregates and low molecular weight "dead end" species. The "dead end" species comprises monomeric non-native protein molecules and cannot form fibrils. Fibril formation involves at least two steps: the reversible formation of linear aggregates, followed by a slow process of "consolidation" after which the fibrils no longer disintegrate upon subsequent slow cooling.Based on the obtained experimental data we have derived a kinetic model for the heat-induced aggregation ofb-lg at pH 2.0. The model involves a nucleation step and a simple addition reaction for the growth of the fibrils as well as a side reaction leading to the complete denaturation and inactivation of a part of the protein molecules. An analytical solution of the model for the early stages of the aggregation is obtained. The model describes very well the experimental data obtained by in situ SLS. It allows us to obtain molecular parameters for the kinetics of fibrilar aggregation ofb-lg as a function of the ionic strength. It gives us an expression for the apparent critical concentration for fibril formation due to the competition between the complete denaturation of the protein molecules and the formation of long fibrils. We also obtain the size of the critical nucleus for the fibril formation as a function of the ionic strength. In the case of a 13 mM ionic strength the critical nucleus consists of ca. 4 monomers; for all the other ionic strengths studied it is a dimer. This shows the important role that the non-specific electrostatic interaction has for the fibrilar aggregation ofb-lg at pH 2.0. It affects the rate of aggregation: the higher the ionic strength, the faster the aggregation. It also affects the detailed mechanism by which the aggregation takes place: the size of the critical nucleus increases when decreasing the ionic strength from 50 mM to 13 mM.We have also shown that time-resolved SANS can be used with success in studying protein aggregation and that with enough additional information for the aggregation process one can in practice obtain complete information about the aggregation kinetics of the process.Tapping mode atomic force microscopy results indicate that the fibrils formed at pH 2.0 upon heating at 80°Chave a periodic structure with a period of about 25 nm and a thickness of one or two protein monomers. The main difference between the fibrils observed at different ionic strengths is their length and curvature. Fibrils obtained at higher ionic strength are shorter and more curved as opposed to longer and straighter fibrils obtained at lower ionic strengths. In case of higher ionic strength the fibril formation is faster, more fibrils are formed and as a result the mean length of the fibrils is shorter. Fibrils obtained at all ionic strengths exhibit similar type of periodic morphology, which suggests that the detailed mechanism of fibril formation might be independent of the ionic strength, but specific forb-lg.In the case of HEWL we study the effect of pH and temperature on the fibril formation. Fibril formation is promoted by low pH and temperatures close to the midpoint temperature for protein unfolding (detected using far-ultraviolet circular dichroism (CD)). The stability of HEWL toward heat treatment is greatly influenced by the pH. The lower the pH, the lower the stability of the protein is. The conditions at pH 2.0 are unique in promoting the fibrilar aggregation of HEWL since heating of solutions at pH 3.0 and 4.0 to temperatures just above the midpoint of the unfolding transition of the molecule does not lead to the appearance of fibrilar aggregates.HEWL fibrils are formed after a lag time that is practically concentration independent. This means that the governing process for the fibril formation is the change in the structure of single protein molecules caused by a prolonged exposure to a temperature close to the midpoint of the unfolding transition. Nucleation presumably involves a change in the conformation of individual lysozyme molecules. Indeed, long term CD measurements at pH 2.0, T = 57°C show a marked change of the secondary structure of lysozyme molecules after about 48 h of heating.The fibril morphology is complex. The fibrils formed at pH 2.0 are long and straight with a length of the order of 5mm and predominant thickness of about 4 nm and consist of stiff rod-like subunits with length either 124 or 157 nm. On smaller scale the fibrils consist of a coiled structure with a period of ca. 30 nm that gives the appearance of the rod-like subunits probably because of defects occurring every 4 or 5 turns.The fibrils consist mostly of full-length HEWL, although, some fragments due to hydrolysis at pH 2.0 and 57°C are probably incorporated into the fibrils. At any rate the hydrolysis of the protein is not the cause of the aggregation since at pH 3.0 no hydrolysis is detected but fibrils do form.In conclusion we can say that for a full and general description of the processes of fibrilar aggregation of globular proteins the type of specific interaction responsible for the aggregation must be identified. The interacting parts of the protein must also be identified. The last and most difficult task is to characterise the conformation of the protein in solution at conditions suitable for aggregation.

Mellema, M. (Michel), Scaling relations between structure and rheology of ageing casein particle gels , PhD Thesis, Wageningen University, 150 + 10 pages, references by chapter, English and Dutch summaries (2000).The relation between (colloidal) interactions, structure and rheology of particle gels is discussed, especially the properties and the spontaneous ageing behavior of rennet-induced casein(ate) or skim milk gels.Methods involved were Brownian dynamics simulations, confocal microscopy, permeametry and rheometry (large- and small deformations). A categorization of relevant (fractal) scaling models and types of structural rearrangements (particularly those affecting the rheological properties) for particle gels has been made, and applied to experimental data of rennet-induced casein gels.Using Brownian dynamics simulations, the relation between colloidal particle interactions and gel structure was obtained. The simulation model used, included a repulsive barrier between the particles; the bonds formed were irreversible. The aggregation was delayed by a long-range repulsive barrier, and a high fractal dimensionality of the gels (2.4-2.5) resulted. This value was independent of the details of the interactions and volume fraction of particles in the range of 3-10 vol% particles. The simulation results agreed well with experimental results on rennet-induced aggregation and gelation in skim milk.The (fractal) structure of rennet-induced casein gels, as studied by confocal scanning laser microscopy and permeametry was monitored during aging, at various pH values (5.3-6.65) and temperatures (20-30 °C). At low pH, a gradual coarsening of the structure was observed; the size of the pores and compact structures increased. This was reflected in a decrease in apparent fractal dimensionality (from 2.4 to 1.7), and in an increase in pore size and lower cut-off length of the fractal regime (from 0.5 to 1.5μm).A scaling model was developed for the rheological behaviour of particle gels as a function of structure and particle volume fraction. The main structural parameters are the fractal dimensionality, the size of the compact building blocks and two parameters describing the number of deformable links in the strands and the dominating type of deformation of these links. Application to rheological data (storage modulus, maximum linear strain, yield stress) as a function of volume fraction (5-9 vol%) showed that rennet-induced casein gels contain straight, elastic strands.Application of the model to results on measurements of the storage modulus during ageing, showed that at low pH rearrangements induce a dramatic increase in compact building block size and early disappearence of the fractal structure. It is concluded that the main types of rearrangements in rennet-induced casein gels after gelation, are particle fusion and stretching and breaking of strands.Keywords : colloid, structure, rheology, casein, rennet, gel, fractal scaling, computer simulation, confocal microscopy.

Mellema, M. (Michel), Scaling relations between structure and rheology of ageing casein particle gels , PhD Thesis, Wageningen University, 150 + 10 pages, references by chapter, English and Dutch summaries (2000).The relation between (colloidal) interactions, structure and rheology of particle gels is discussed, especially the properties and the spontaneous ageing behavior of rennet-induced casein(ate) or skim milk gels.Methods involved were Brownian dynamics simulations, confocal microscopy, permeametry and rheometry (large- and small deformations). A categorization of relevant (fractal) scaling models and types of structural rearrangements (particularly those affecting the rheological properties) for particle gels has been made, and applied to experimental data of rennet-induced casein gels.Using Brownian dynamics simulations, the relation between colloidal particle interactions and gel structure was obtained. The simulation model used, included a repulsive barrier between the particles; the bonds formed were irreversible. The aggregation was delayed by a long-range repulsive barrier, and a high fractal dimensionality of the gels (2.4-2.5) resulted. This value was independent of the details of the interactions and volume fraction of particles in the range of 3-10 vol% particles. The simulation results agreed well with experimental results on rennet-induced aggregation and gelation in skim milk.The (fractal) structure of rennet-induced casein gels, as studied by confocal scanning laser microscopy and permeametry was monitored during aging, at various pH values (5.3-6.65) and temperatures (20-30 °C). At low pH, a gradual coarsening of the structure was observed; the size of the pores and compact structures increased. This was reflected in a decrease in apparent fractal dimensionality (from 2.4 to 1.7), and in an increase in pore size and lower cut-off length of the fractal regime (from 0.5 to 1.5μm).A scaling model was developed for the rheological behaviour of particle gels as a function of structure and particle volume fraction. The main structural parameters are the fractal dimensionality, the size of the compact building blocks and two parameters describing the number of deformable links in the strands and the dominating type of deformation of these links. Application to rheological data (storage modulus, maximum linear strain, yield stress) as a function of volume fraction (5-9 vol%) showed that rennet-induced casein gels contain straight, elastic strands.Application of the model to results on measurements of the storage modulus during ageing, showed that at low pH rearrangements induce a dramatic increase in compact building block size and early disappearence of the fractal structure. It is concluded that the main types of rearrangements in rennet-induced casein gels after gelation, are particle fusion and stretching and breaking of strands.Keywords : colloid, structure, rheology, casein, rennet, gel, fractal scaling, computer simulation, confocal microscopy.

Peptides derived fromβ-casein (βCN) andβ-lactoglobulin (βLg) were analysed for their foam- and emulsion-forming and -stabilising properties (further denoted functional properties) and for their structural characteristics in order to elucidate structure-function relationships.βCN was hydrolysed by plasmin and subsequent fractionation of the hydrolysate resulted in various hydrophilic, amphipathic and hydrophobic peptide fractions with clear differences in functional properties. The highly-charged N-terminal part of the amphipathic peptides appeared to be important for the emulsion-stabilising properties ofβCN peptides. The main secondary structure element ofβCN(-peptides) in solution was the unordered random coil, but upon adsorption onto an hydrophobic interfaceα-helix was induced. The hydrophobic C-terminal part ofβCN accounted for the high maximum surface load on the interface, while the N-terminal part ofβCN seemed to be responsible for theα-helix induction upon adsorption. No clear relation between the secondary structure and the functionality was observed in this system but a relation between a high surface load and good stabilising properties seemed to exist.BovineβLg was hydrolysed by the action of trypsin, plasmin and Staphylococcus aureus V8 protease. Overall, the plasmin hydrolysate had the best functional properties at pH 6.7, compared to the other hydrolysates and was investigated further. DuringβLg/plasmin hydrolysis significant SH/SS-exchange has taken place yielding a large number of different peptides. The peptides present were (1) peptides composed of a single amino acid chain lacking a cysteine residue, (2) peptides composed of a single amino acid chain containing intramolecular disulphide bonds and (3) peptides composed of 2 amino acid chains linked by an intermolecular disulphide bond. The occurrence of the SH/SS exchange and the homogeneous distribution of charge and hydrophobicity hinder an efficient fractionation of the hydrolysate.In conclusion, the production of specific peptides and peptide fractions is more complicated forβLg than forβCN, mainly because of the differences in primary structure (such as the distribution of charge and hydrophobicity) between the proteins. The foam- and emulsion-forming properties of peptides can be superior to those of intact proteins, as long as they have both charged and hydrophobic areas. The foam- and emulsion-stabilising properties of peptides depend highly on the amount of repulsion they can produce (either by a strong amphipathicity or by a high surface load).

Peptides derived fromβ-casein (βCN) andβ-lactoglobulin (βLg) were analysed for their foam- and emulsion-forming and -stabilising properties (further denoted functional properties) and for their structural characteristics in order to elucidate structure-function relationships.βCN was hydrolysed by plasmin and subsequent fractionation of the hydrolysate resulted in various hydrophilic, amphipathic and hydrophobic peptide fractions with clear differences in functional properties. The highly-charged N-terminal part of the amphipathic peptides appeared to be important for the emulsion-stabilising properties ofβCN peptides. The main secondary structure element ofβCN(-peptides) in solution was the unordered random coil, but upon adsorption onto an hydrophobic interfaceα-helix was induced. The hydrophobic C-terminal part ofβCN accounted for the high maximum surface load on the interface, while the N-terminal part ofβCN seemed to be responsible for theα-helix induction upon adsorption. No clear relation between the secondary structure and the functionality was observed in this system but a relation between a high surface load and good stabilising properties seemed to exist.BovineβLg was hydrolysed by the action of trypsin, plasmin and Staphylococcus aureus V8 protease. Overall, the plasmin hydrolysate had the best functional properties at pH 6.7, compared to the other hydrolysates and was investigated further. DuringβLg/plasmin hydrolysis significant SH/SS-exchange has taken place yielding a large number of different peptides. The peptides present were (1) peptides composed of a single amino acid chain lacking a cysteine residue, (2) peptides composed of a single amino acid chain containing intramolecular disulphide bonds and (3) peptides composed of 2 amino acid chains linked by an intermolecular disulphide bond. The occurrence of the SH/SS exchange and the homogeneous distribution of charge and hydrophobicity hinder an efficient fractionation of the hydrolysate.In conclusion, the production of specific peptides and peptide fractions is more complicated forβLg than forβCN, mainly because of the differences in primary structure (such as the distribution of charge and hydrophobicity) between the proteins. The foam- and emulsion-forming properties of peptides can be superior to those of intact proteins, as long as they have both charged and hydrophobic areas. The foam- and emulsion-stabilising properties of peptides depend highly on the amount of repulsion they can produce (either by a strong amphipathicity or by a high surface load).