Verdikken en geleren : een fysisch chemisch onderzoek naar de invloed van polymeren op de reologie van waterige systemen als model voor levensmiddelen

The purpose of this study was to determine the influence of some molecular properties on the thickening and gelling ability of a number of hydrophilic polymers. <em>Thickening is</em> defined as enhancing the viscosity of a system either in the Newtonian or in the non-Newtonian sense. <em>Gelling</em> is defined as imparting a more or less solid shape to aqueous systems by adding polymers, resulting in elastic properties. The results have been interpreted with the aid of existing theories on the behaviour of concentrated polymer solutions, melts and networks. Mainly synthetic hydrophilic polymers have been used in this study, because these are better defined than polymers of natural origin.In Chapter I the importance of the role of polymers in improving the consistency of foods is described briefly. It is pointed out that, though many hydropolymers have been used for centuries, this application was based mainly on empiricism rather than on understanding. However, to control the consistency of existing products and to develop new products, such as those based on vegetable protein, it is very important to understand their mode of action.The materials and methods used in this study are described in Chapter 2. Most attention is paid to the properties of polyvinyl alcohol (PVA). PVA has the tendency to structurize itself to some extent at low temperatures, under influence of added compounds and by the application of a shear field. These structurized chains may lead to association and cross-linking. Compounds like NaCNS and the presence of acetate groups in the polymer chain strongly prevent this association.The cone-penetrometer has been introduced as a suitable instrument to estimate the elasticity modulus. This apparatus can be calibrated with an apparatus for coaxial creep experiments. An essentially new feature is also the method for calculating the various molecular weight averages <em>(M _n , M _v ,</em> M <em>_w</em> and <em>M _z )</em> of mixtures of fractions. With this method in conjunction with suitable experiments, it is possible to establish quickly which molecular weight average is responsible for a certain physical property of a polymer system.In Chapter 3 the influence of molecular weight and molecular weight distribution on the thickening action of a number of hydropolymers is treated. The influence of the concentration was investigated simultaneously. Finally a study was made of the influence of both the molecular variables mentioned on the temperature dependency of the viscosity and on the non-Newtonian flow behaviour.Viscosity was measured of the following solutions: polyvinyl alcohol, polyethylene glycol, polypropylene glycol, polyvinyl pyrrolidone, dextran and sodium-carboxymethylcellulose. From these measurements it could be concluded that the mixing method of section 2.8 allows a rapid assessment of the position of <em>M _v</em> relative to <em>M _n , M _w</em> and <em>M _z . M _v</em> deviates more from <em>M _w</em> than predicted by FLORY from the value of the constant <em>_n</em> in the MARK-HOUWINK equation.For dextran <em>a</em> = 0.5 and <em>M _v = M _n VM _w /M _n ,</em> for PVA <em>a = 0,67</em> and M <em>_v</em> ~ <em>M _w</em> , for NaCMC <em>a =</em> 1.5 <em/> and <em>M _v</em> ~ <em>M _z . M _v</em> appears to be independent of the concentration so that the dimensions of the isolated molecule also determine the viscosity in concentrated solutions. Thus, it is possible to arrive at a semi-empirical equation for the viscosity of PVA solutions as a function of molecular weight and concentration. This equation accounts for the dimensions of the isolated molecules and the increasing interaction with increasing concentration. For this purpose an extended HUGGINS-equation was used. Such a general equation can only be formulated if <em>a</em> and <em>K _H</em> (HUGGINS constant) are really constants over the whole range of molecular weights of the polymer involved. The same <em>M _v</em> which determines the Newtonian viscosity governs also the non-Newtonian flow behaviour. The width of the molecular weight distribution neither influences the temperature dependence of the viscosity nor the non-Newtonian behaviour.Investigations on the influence of molecular weight, molecular weight distribution, the degree of hydrolysis and the polymer concentration on the gel strength and syneresis of PVA gels are described in Chapter 4. For gels with cross- linkages consisting of physical bonds, <em>M _n</em> governs the gel strength, only if the gels exhibit elastic behaviour. However, if there is also a marked viscous component in the deformation, the large molecules contribute relatively more to the gel strength. Then an average between <em>M _n</em> and <em>M _w</em> is determinative. An important part of the PVA molecules do not play any role in the network. This can be concluded from the pronounced increase of the modulus with concentration at low polymer concentrations and from the large amount of PVA that is exudated with the syneresis liquid, especially at low polymer concentrations. Gel formation with PVA 88 and Congo red is less efficient than with PVA 98.5 and Congo red. The strength of the last mentioned gels increases strongly with time and sometimes syneresis occurs. The occurrence of syneresis depends on the gel composition and <em>M _w</em> , whereas the steric hindrance of acetate groups in PVA 88 prevents syneresis. Syneresis is a consequence of continuing gelation because both gelation and syneresis proceed with <em>t</em>^1/2. FLORY's equation for swelling predicts also a decrease of the extent of swelling, and an increase of syneresis, with increasing cross-link density. It appears that at every molecular weight there is a region of PVA and Congo red concentrations to which syneresis is confined. With the molecular weight <em>M _w</em> as the third variable, it is possible to construct a three dimensional <em>'syneresis bag',</em> in which syneresis occurs. The existence of a limited syneresis region is explained with the following points:- rapid decrease of <em>M _c</em><em>(molecular weight between cross links)</em> as a function of concentration at lowpolymer concentrations- at high polymer concentrations a strong decrease of the entropy of mixing if syneresis takes place- increase of the dry polymer volume at high polymer concentrations, resulting in an increase of swelling and a decrease of syneresis- the chance of <em>freezing-in</em> of non-equilibrium is greater at higher concentrations for topological reasons.The gelling mechanism and gelling kinetics are further explored in Chapter 5. The most important tool is a dynamic rheologic monitoring of the gelling process which gives information especially on gelling kinetics. However, at the same time some results are obtained which help us to understand the mechanism of cross linking. From differential thermal analysis and estimation of the melting points of the gels, the nature of the cross links in the various PVA-gels became clearer. It was concluded that both PVA-Congo red gels and PVA-resorcinol gels have structurized PVA chain segments bound together physically in a multiple cross-linkage, whereas PVA-borax gels have random cross-linkages between arbitrary pairs of hydroxyl groups on the two chains. The difference between the cross-linkages of PVA- Congo red gels and PVA-resorcinol gels is that the former consist of hydrogen bridges between structurized chain segments and the latter of hydrophobic bonds.The arguments leading to these conclusions are:a. PVA-Congo red and PVA-resorcinol gels have both a constant <em>G' (storage modulus)</em> and a very low <em>G" (loss modulus)</em> over the entire frequency range (10 ^-4-10 ²rad s ^-1), whereas PVA-borax gels exhibit a transition from the rubber elasticity behaviour to flow behaviour within this frequency range. The lifetime of a cross-link in PVA borax gels is only about 1/8 s at 298 °K whereas those in PVA-Congo red gels are virtually permanent. These differences suggest a difference in the energy content of the cross-linkages.b. Reacting PVA-Congo red and PVA-resorcinol mixtures show an <em>'anti-Arrhenius' </em> behaviour and especially PVA-resorcinol shows very long induction periods before incipient gelling starts. This is an argument in favour of theformation of structurized chain segments before gelling starts.c. The degree of hydrolysis has a strong influence on the gel formation. <em>G</em> (elastic shear modulus) decreases by 80 % upon increasing the acetate content by 10 % for PVA-Congo red and <em>G</em> decreases by 25 % upon increasing the acetate content by 1 % for PVA-resorcinol. From these figures one can estimate that about 15 vinyl units are involved in one PVA-Congo red cross-linkage and about twice as many in a PVA-resorcinol cross-linkage.d. The energy content of a cross-link in a PVA-Congo red gel is estimated to be 80 kJ/mole by means of the melting point method of ELDRIDGE and FERRY, for PVA-resorcinol 200 kJ/mole was found. These data may be compared with literature data on PVA-borax gels, viz. 22 kJ/mole corresponding to 1 or 2 hydrogen bridges.e. Methanol has a strong negative influence on the gelling of PVA-resorcinol, but not on other gels. This points to a hydrophobic bond in PVA-resorcinol cross-links.f. PVA-resorcinol gels melt much more suddenly than PVA-Congo red gels. This agrees with the course of the energy content with temperature for a hydrophobic bond. This increases first with temperature and decreases at about 50°C.Congo red as well as resorcinol play an indirect role in cross-link formation. They structurize a fraction of the polymer chain segment before it can associate with another one. Resorcinol probably structurizes PVA in a zig-zag conformation after which two of such chains are bound together through the CH2 groups with a hydrophobic bond. Congo red probably induces helix formation. Two of such helices are bound together with hydrogen bridges. A large number of cross-links do not play an effective role in the network; especially at low concentrations most of the cross links are wasted. This can be concluded from the gelling kinetics, which has been investigated rheologically From the fourth order rate equation one can arrive at the fraction <em>f</em> of effective cross-links; <em>f</em> increases with the square of the polymer concentration.Chapter 5 illustrates the potentialities of the application of rheology in conjunction with theories of rubber elasticity on the gelling process. Rheological measurements can be advantageously exploited to complement other physical measurements on gels and gelling and thus contribute to a better insight.