Static and dynamic properties of proteins adsorbed at liquid interfaces

<p>The aim of the investigation described in this thesis was to increase the level of understanding of the role that proteins play in the preparation and subsequent stabilisation of foams and emulsions. One aspect of this role is facilitation of break-up, due to surface tension lowering. A second aspect is the formation of a viscoelastic interfacial layer, which affects both the short-term and long-term stability of the dispersion. Therefore, a systematic study of the changes in static and dynamic interfacial properties induced by proteins was carried out.</p><p>For part of this study, dealing with the interfacial rheology, several experimental techniques were used. These techniques were either properly modified existing techniques (Chapter 3, modified longitudinal wave set-up) or newly developed (Chapter 4, Dynamic Drop Tensiometer; Chapter 5, Concentric Ring Surface Shear Rheometer) to meet the requirements for measuring the rheology of adsorbed protein layers at liquid/liquid interfaces. These requirements are (i) isotropic deformation, without leakage of the interfacial layer, for the dilational modulus measurements at air/water and oil/water interfaces and (ii) shear modulus measurements at small oscillatory deformation.</p><p>The proteins chosen for this study wereβ-casein,β-lactoglobulin (BLG), bovine serum albumin (BSA), ovalbumin and lysozyme. This set of proteins was chosen, because they differ considerably in relevant aspects, such as molecular weight, molecular structure and iso-electric point.</p><p>In Chapter 1 the scope and context of this study are given including a brief introduction into (i) the molecular properties of these proteins, that are relevant to the adsorption, (ii) protein adsorption and interfacial rheology, and (iii) the relation between interfacial properties and the properties of emulsions and foams.</p><p>Chapter 2 deals with the adsorption of proteins at the air/water interface. The adsorption was determined by ellipsometry, a method by which not only the adsorbed amount but also the layer thickness and protein concentration in the adsorbed layer could be determined. The ellipsometric studies were combined with surface tension measurements at the same surface.</p><p>All proteins examined show high affinity adsorption, i.e. strong adsorption at low concentration in solution. The initial rate of adsorption of all proteins is well described by a simple diffusion equation. For all proteins examined, the value of the surface pressure (Π) are protein-specific, but otherwise unique, time-independent functions of the adsorption (Γ). Time independence of theΠ(Γ) curve was concluded from the finding thatΠandΓpairs measured at different bulk concentrations and at different stages of adsorption, all collapse into one single curve. In other words, each protein has a unique surface equation of state indicated by its measuredΠ(Γ) curve. This curve reflects the relative rigidity of the protein molecule. For flexible molecules likeβ-casein and PVA ,Γ <sub>min</sub> (=ΓwhereΠstarts to deviate measurably from zero) is low and from this point onward the surface pressure increases gradually with increasingΓ. For rigid globular proteins (BSA, ovalbumin and lysozyme)Γ <sub>min</sub> is higher and with further increase of the surface concentration the surface pressure increases steeply. At high protein concentration and long adsorption times, for most proteins multilayer adsorption takes place.</p><p>For ovalbumin, in the pH range 4-8 the effect of pH on theΠ-Γcurve is small, which indicates that electrostatic intermolecular forces do not contribute much to the surface pressure.</p><p>In Chapter 3 a longitudinal wave technique, modified to ensure isotropic surface deformation, was used to determine the dilational modulus,ε, of adsorbed protein layers, at the air/water interface. This modification fully eliminated the complicating shear effects that became apparent in dilational modulus measurements with adsorbed layers of proteins in a conventional set-up.</p><p>For all proteins examined at frequencies in the range from 0.01 to 1 rad/s, the initial part of theε(Π) plot is a straight line through the origin. The slope of this initial part ranges between +4 and +12 . No clear relationship between the slope and the rigidity of the protein molecule was found. However, the extent of this linear range is smaller for the flexible molecules (β-casein and PVA). From the fact that this slope significantly exceeds the ideal value of +1, it must be concluded that the behaviour of the adsorbed layer is far from ideal. In the linear range, the measured moduli coincide with the limiting moduli,ε <sub>0</sub> , calculated from theΠ(Γ) curve. This indicates that the surface pressure adjusts "instantaneously" to the changing adsorption during a compression-expansion cycle in time-scales ranging from 1 to 100 s. This also means that the modulus is purely elastic, i.e. the effect of relaxation processes is negligible. In this elastic range, differences between individual proteins are related to different degrees of non-ideality, reflected in the surface equation of state.</p><p>At higher surface concentrations a relaxation mechanism becomes operative, which is most probably not caused by diffusional exchange between surface and solution. This conclusion is based on calculations of the diffusional transport rate and the theoretical frequency spectrum of the modulus. Relaxation due to conformational changes is plausible. In the visco-elastic regionε≥ε <sub>0</sub> for all proteins examined. This is an extra argument against diffusional exchange.</p><p>The modulus increases in the order: PVA <β-casein <BSA < ovalbumin <κ-casein. For the first four molecules the flexibility of the molecule decreases in the same order. The high modulus in the case ofκ-casein cannot be attributed to the rigid molecular structure.</p><p>Chapter 4 describes a new method, the Dynamic Drop Tensiometer, especially suitable for determining the dynamic properties of proteins adsorbed at oil/water interfaces. According to this method, a small drop is subjected to sinusoidal oscillations of its volume. The corresponding area changes produce interfacial tension changes, which are evaluated from measurements of the fluctuating shape of the drop, using the Young-Laplace equation. Compared to the conventional Langmuir trough set-up, this method is particularly suited for liquid/liquid interfaces, because (i) interfacial leakage is fully eliminated and (ii) uniform deformation is ensured even if one of the liquids is a viscous oil. An additional advantage of the method is its short response time. The dynamic properties of adsorbed protein layers at three interfaces (TAG (triacylglycerol)-oil/water, tetradecane/water and air/water) were compared. At the three interfaces, at low protein concentration, the conformation change upon adsorption is fairly fast, occurring within 1 min.. However, at high protein concentration (> 1g/l), during the first minutes after adsorption a situation exists that differs from the equilibriumΠ(Γ) curve. At low interfacial pressures, during a modulus measurement, the adaptation of the conformation is faster (< 1 s.). Non-ideality of the adsorbed layer increases in the sequence TAG-oil < tetradecane < air, which is probably related to a decrease of solution quality for the more hydrophobic amino acids, which decreases in the same sequence. At each of the different interfaces non-ideality increases with increasing rigidity of the protein molecule (β-casein<β-lactoglobulin<BSA<ovalbumin). Collapse-type effects and conformation-related relaxation are most pronounced at TAG-oil/water, especially with BSA, but absent at all interfaces with ovalbumin.</p><p>The surface shear properties of adsorbed protein layers are described in Chapter 5. These properties were determined with a newly developed concentric ring surface shear rheometer. The technique allows measurements over a wide range of frequencies and deformations. As the magnitude of the shear deformation markedly affects the shear modulus,μ <sub>s</sub> , an extrapolation to zero deformation is required to asses the shear properties of the undisturbed surface. Because the surface dilational modulus and the surface shear modulus both increase in the sequence PVA< Na-caseinate <BSA <ovalbumin, it is plausible that essentially the same molecular properties are responsible for the magnitude of both moduli. These properties are intermolecular interactions and the intramolecular structure. The low shear modulus for casein can be ascribed to relatively strong interactions between flexible, easily deformable molecules. With BSA and ovalbumin the interacting bonds are the weaker part of the network. So, for both moduli the rigidity of the molecule is the dominating characteristic. The finding that for most systemsε/μ <sub>s</sub> ≥3 indicates that the adsorbed protein layer can be modelled as a thin homogeneous gel layer. Such a model points to a significant ideal monolayer contribution toεat low to medium surface concentrations.</p><p>In Chapter 6 models describing the surface equation of state of adsorbed macromolecules were applied to the experimentalΠ(Γ) curves. These models were also applied to understand the dynamic behaviour of these layers. Statistical models, in which it is assumed that the macromolecules adsorb with all segments in direct contact with the surface, e.g. Singer equation, only explain the very low pressure part of the experimental curves of PVA andβ-casein. To explain the higher pressure part, progressive loop formation and molecular interaction must be accounted for. For rigid globular proteins, simple statistical models are unable to fit any part of the experimental curves, because such molecules only slightly change their conformation upon adsorption and consequently, will adsorb with only a small fraction of the segments at the surface, even at very low pressures.</p><p>A 2-D solution model, which accounts to first order for both entropy and enthalpy, is used to describe the non-ideal behaviour of adsorbed protein layers. This non-ideality was deduced from the highΓneeded to produce a measurableΠand the steep initial slopes of theε(Π) curves.</p><p>All above models need modification to describe the S-shaped part of theΠ(Γ) curves at high surface concentrations. This part of the curve can be described by the Soft Particle concept, which is a modification of the surface equation of state of a 2-D hard sphere fluid. The S-shape is attributed to a decrease of the molecular cross-sectional area with increasing surface concentration. This effect appears to be more pronounced for flexible molecules like PVA andβ-casein than for globular rigid molecules like BSA, ovalbumin and lysozyme. Experimentalε(Π) curves are within the limits that are predicted by this concept. A promising option is combining a molecular compressibility as used in the Soft Particle concept with the 2-D solution model.</p><p>In Chapter 7 it is shown that interfacial properties typical for proteins predict a larger drop size and a lower stability against recoalescence during production compared to low molecular weight (LMW) surfactants.</p><p>In the presence of both types of surfactant, concentrations and conditions can be chosen such that the LMW surfactant determines the dispersion efficiency, while the protein determines the long-term stability. A comparison between the different proteins reveals that, in the production stage, a higher dilational modulus at short times correlates with a faster build-up of stability against recoalescence. For a good long term stability a high dilational modulus of adsorbed protein layers at longer times is more important. In foams, retardation of Ostwald ripening, i.e. the growth of large bubbles at the expense of small ones, is probably the major factor. This mechanism depends on the ratio of the modulus to the surface tension, which ratio is considerably higher for proteins than for LMW surfactants in relevant cases.</p><p>For a measurable shear modulus a high surface concentration is required. Therefore, shear properties may only affect long term stability of emulsions and foams, but not break-up and stability against recoalescence during production.</p&gt