Formation and structure of acidified milk gels
Introduction
Though yogurts are a growing segment of the dairy foods market, there is surprisingly little information available on the subject of gel formation in milk as a result of fermentation or acidification. This is even more surprising when it is recalled that perceived texture in these products has been identified as an important sensory property determining consumer acceptability (Muir & Hunter, 1992). The majority of research work carried out on these products has been technological in nature, mainly directed at process optimisation rather than at achieving any fundamental understanding of the underlying mechanisms of gel formation through which manipulative control of product texture might be achieved.
This is not to say that this dearth of understanding or lack of information exists in other fields of industry and science where gels are important. This paper will review the modern ideas of gelation drawn from these sources and consider their applicability to gel formation by acidification of milk. It will start from the very beginning by asking `What is a gel?’ It will progress from definition to description of yogurt as a particle gel; it will identify the nature of those particles; it will describe the kinetics of formation of these gels; by rescaling the gelation profiles, it will show that the mechanism of aggregation is modified by pre-heat treatment of the milk; and, finally, it will consider the kinetic information in the light of three theoretical models, namely the adhesive sphere model, percolation and fractals, drawing conclusions on their applicability and failings in this area.
Almost 50 years ago, Bungenberg de Jong (1949) defined a gel as a colloidal system of solid character, in which the colloidal particles somehow constitute a coherent structure, the latter being interpenetrated by a liquid system. This sounds more like an attempt to describe a drawing by Escher rather than a scientific definition. Whilst, as we shall see, the statement is undoubtedly correct, it is quite difficult to fully grasp its significance unless you already have some appreciation of the nature of a gel. Recognising that attempts to define the term `gel’ were becoming more and more obscure, with excessive preoccupation with terminology, Flory (1974) recalled the quotation attributed to D. Jordan Lloyd in the 1920s that `the colloidal condition, the gel, is one which is easier to recognise than to define'.
In more modern terminology, we describe a gel as a viscoelastic solid. That is a system which, depending on circumstances, can flow like a viscous liquid and in others behave like an elastic solid. We can see in this, the Bungenberg de Jong description with the network being able to respond to the applied mechanical force in the way that a solid would. We can also see here the problem in gel formation of D. Jordan Lloyd, echoed by Hermans (1949) and re-echoed by Flory (1974), of when do you recognise that this load-bearing network has been created, at what point do the elastic properties dominate those of the viscous.
This gel point is most easily defined by considering the rheological properties of the gel. Rheology is concerned with the interaction of three variables: applied force or stress; measured response or strain; and the time during which these events take place. In small deformation rheometry, an oscillating force is applied at a fixed frequency and an oscillating response is measured. This response lags on the applied force and is out of phase by an angle, δ. Defining shear modulus as the ratio of applied stress to shear strain, we can treat the measured complex shear modulus as a vector quantity and decompose it into its two components. These are an in-phase component, , known as the elastic or storage modulus which reflects the solid-like properties of the gel, and an out-of-phase component, , called the viscous or loss modulus which tells you how fluid-like is the behaviour of your system. The ratio of the two moduli, G″/G′ then defines the tangent of the phase angle. Thus when viscous properties dominate over elastic, tan δ is greater than 1 and the phase angle is greater than 45°. Conversely, when we have a gel and the elasticity dominates, phase angles of less than 45° are recorded. When you start with a liquid suspension and induce gelation, one definition of the gel point is that time when the phase angle transits 45° (Ross-Murphy, 1995).
More complicated definitions of a gel and the gel point have been proposed by other workers. In the above discussion, all of the measurements were considered to have been made at a single oscillation frequency. Almdal, Dyre, Hvidt and Kramer (1993) not only demanded that the storage modulus, G′, should dominate over the loss modulus, G″, but that the gel should show a flat mechanical spectrum in an oscillatory sweep experiment, i.e. that G′ and G″ should be independent of oscillation frequency, ω. Since, for a solution, G′ should be proportional to ω2 and G″ proportional to ω, the transition from sol to gel thus involves changing the exponent in the double-log plot of ω vs. G′ or G″ from 2 to approximately zero for G′ and from 1 to approximately zero for G″. Winter and Chambon (1986) then suggested that a better criterion for the gel point is when, in a frequency sweep experiment, both G′ and G″ show power-law behaviour in frequency with the same positive exponent. As Winter and Chambon (1986) have shown, this time does not necessarily correspond with that for the simple G′ and G″ crossover in a single frequency experiment. Even the ability to measure the simple crossover may depend on the sensitivity of the measuring instrument. Milk, for example, is of low initial viscosity and measurable values of G″ may be unobtainable with some rheometers. In such circumstances, gel point is often judged to be when the instrument response from the gelling sample becomes greater than the background noise. In the majority of studies discussed in this paper, gel point will be defined by this latter limiting criterion.
Another way of classifying gels is in terms of the structural elements of the continuous gel network. Following Flory (1974), such a scheme sub-divides networks into four categories as follows.
- 1.
Well-ordered lamellar structures.
- 2.
Covalent polymer networks.
- 3.
Polymer networks formed through physical aggregation.
- 4.
Particulate, disordered structures.
Into Class 1 would fall soap gels, but also inorganic gels from clays or other minerals. In polymer gels (Class 2), continuity of structure is provided by a ramified three-dimensional network comprising structural units covalently linked to each other. Some of the units must be polyfunctional, i.e. capable of connecting to more than two other units. Polyacrylamide gels or vulcanised rubbers would be a good example of these. Then in Class 3 we have physical gels of (possibly) entangled polymers. Gelatin gels would be a good example of these where primary molecules, usually of a linear structure but of finite size, come together to form junction zones at particular points along the protein chain. The liaisons thus established fulfil the role of the polyfunctional cross-links in gels of Class 2. Class 4 type gels, particle gels, are built of clusters of aggregated (spherical) particles which network to form a continuous structure extending throughout the enclosing volume. Acidified milk gels fall into this particular class.
Section snippets
Acidified milk gels are particle gels
Prior to the mid-1980s, the justification for this statement relied on electron microscopy of fully formed acid gels that indicated that the networks were composed of strands of particulate material (Heertje, Visser & Smits, 1985; Modler & Kalab, 1983). But the question could be asked `Were these simply reformed from individual caseins as they passed through their isoelectric point and precipitated or were they the result of a particle–particle aggregation process?'
Roefs, Walstra, Dalgleish and
Kinetics of gel formation in acidified milks
Previous studies (Horne & Davidson, 1993) of the kinetics of acid gel formation in skim-milk induced by hydrolysis of glucono-δ-lactone (gdl) showed that pre-heating the milk at 90°C for 10 min shifted the pH at the gelation time upwards from ∼5.1 in raw milk to ∼5.5 in heat-treated milk. Using a series of pre-heating temperatures, Horne and Davidson (1993) showed that the transition between these two limiting gelation pH's centred around a treatment temperature of 75°C, coinciding with the
Adhesive sphere model
De Kruif (1997) and De Kruif, Jeurnink and Zoon (1992) have developed the application of this model to the aggregation reactions of destabilized casein micelles. An extensive review of this approach appears earlier in this volume (DeKruif, 1999) and only points of relevance to the development of gel properties and their time dependence are considered here.
Native casein micelles normally behave as hard spheres (DeKruif et al., 1992; Griffin, Price & Griffin, 1989) stabilised by κ-casein protein
Conclusions
Each of these theories considers the gelation process from one particular viewpoint. The adhesive sphere model considers bonding, the fractal model describes the geometric structure of the network, and percolation looks to the critical level of bond formation. All are successful in their own spheres of influence but can be drastically in error outside these.
Their major failing is their inability to describe the kinetics of the gelation process. The experimental results presented here have shown
Acknowledgements
Core funding for the Hannah Research Institute is provided by the Scottish Office Agriculture, Environment and Fisheries Department. Some of the experimental results included were obtained in the European Commission Framework IV Programme project CT96-1216 `Structure, rheology and physical stability of particle systems containing proteins and lipids’ and their financial support is gratefully acknowledged.
References (41)
Skim milk acidification
Journal of Colloid and Interface Science
(1997)Casein micelle interactions
International Dairy Journal
(1999)- et al.
Variation of the viscosity of a concentrated, sterically stabilized colloid—effect of ethanol on casein micelles of bovine milk
Journal of Colloid and Interface Science
(1989) - et al.
Microstructure of yogurt stabilized with milk proteins
Journal of Dairy Science
(1983) On the stability of casein micelles
Journal of Dairy Science
(1990)- et al.
Towards a phenomenological description of the term `gel'
Polymer gels and Networks
(1993) - Bremer, L. B. G. (1992). Ph.D. thesis, Agricultural University, Wageningen, The...
- et al.
Theoretical and experimental study of the fractal nature of the structure of casein gels
Journal of the Chemical Society: Faraday Transactions
(1989) - Brown, W. D. (1987). Ph.D. thesis, University of Cambridge,...
A survey of the study objects in this volume