Dairy Science & Technology · Dairy Chemistry · also Food Emulsions, Foams and Gels, Dairy Colloid Science
Milk as a colloidal system and the physical chemistry of the emulsions, foams and gels that make cream, ice cream, yoghurt and cheese.
In the exams: ICAR AICE subject 37 Unit 8 and subject 34 Unit 1 (emulsions, foams and gels)
Milk is not a simple solution. It is a colloidal system in which fat is emulsified as globules, casein is dispersed as micelles, whey proteins are dissolved as individual molecules, and lactose and salts are in true solution. Almost every dairy product is made by deliberately changing this colloidal state: cream is a concentrated emulsion; butter is made by inverting that emulsion; whipped cream and ice cream are foams held up by partially coalesced fat; yoghurt is an acid gel of casein; cheese is a rennet gel that has been cut, drained and pressed. Understanding emulsions, foams and gels is therefore understanding why dairy products look, feel and keep the way they do.
This manual builds that understanding from first principles. It starts with the colloidal state of milk and the forces that keep particles apart or bring them together, then covers interfaces and surfactants, emulsifiers and the hydrophilic-lipophilic balance, emulsion formation by homogenisation, and the ways emulsions break down. It then treats foams and overrun, protein and polysaccharide gels, acid and rennet milk gels, hydrocolloid stabilisers and their interactions with milk proteins, the microscopy and particle-sizing techniques used to study structure, and newer structures such as organogels, hydrogels and nanoemulsions.
Doctoral fellowship examinations in dairy chemistry and dairy technology include emulsions, foams and gels as a named unit, and postgraduate entrance examinations ask about homogenisation, creaming, overrun and coagulation. Typical questions involve Stokes' law, HLB calculations, overrun, the mechanisms of acid and rennet gelation, the action of carrageenan and pectin in dairy drinks, and the principles of light scattering and microscopy. Product developers use the same science daily to solve sedimentation, wheying-off, poor whipping and texture defects.
A colloid is a system in which particles between about 1 nanometre and 1 micrometre (and in practice up to tens of micrometres) are dispersed in a continuous phase. Milk contains several such dispersed phases at once. Fat globules, typically 0.1 to 15 µm in diameter with a volume-weighted mean of about 3 to 5 µm in cow milk, form an oil-in-water emulsion; each globule is surrounded by the milk fat globule membrane, a complex trilayer of phospholipids, proteins and glycoproteins derived from the secretory cell. Casein micelles, roughly 50 to 500 nm in diameter with an average around 150 to 200 nm, are association colloids of the four caseins (αs1, αs2, β and κ) held together with colloidal calcium phosphate; they are highly hydrated and occupy about 10% of the volume of milk. Whey proteins, chiefly β-lactoglobulin and α-lactalbumin, are globular molecules of a few nanometres in solution. Lactose, soluble salts and other small molecules form the true solution, the serum.
These phases explain many everyday observations. Milk is white because fat globules and casein micelles scatter light; skim milk looks bluish because the smaller micelles scatter short wavelengths more. Fat globules rise because they are less dense than the serum. Casein micelles stay dispersed through pasteurisation and even sterilisation, yet collapse rapidly when acid or rennet removes their stabilising forces. Whey proteins denature on heating and attach to micelles and globules. The dairy technologist manipulates each phase separately: separation removes the fat phase, ultrafiltration concentrates the micelles and whey proteins, acid or rennet gels the micelles, and homogenisation transforms the fat phase.
Whether particles stay apart or aggregate depends on the balance of attractive and repulsive forces. Van der Waals attraction acts between all particles and becomes strong at short distances. Electrostatic repulsion arises because particles carry charge, surrounded by a diffuse cloud of counter-ions, the electrical double layer. The DLVO theory, named after Derjaguin, Landau, Verwey and Overbeek, adds these two to give an energy barrier against aggregation. The thickness of the double layer, the Debye length, falls as ionic strength rises; for a 1:1 electrolyte at 25 °C it is about 0.304/√I nanometres, so at the ionic strength of milk, roughly 0.08 mol/L, it is only about 1 nm. Electrostatic repulsion in milk is therefore short-ranged, and the zeta potential of casein micelles is modest, around −20 mV at natural pH.
Steric stabilisation adds a further barrier. Adsorbed or attached polymers, such as the flexible, negatively charged C-terminal 'hairs' of κ-casein on the surface of casein micelles, resist being compressed or interpenetrated when two particles approach. This hairy layer is the main reason micelles remain stable in milk. Remove it with chymosin, or collapse it by acidifying towards pH 4.6 or by adding ethanol, and the micelles aggregate.
Other forces matter too. Hydrophobic interactions drive aggregation of exposed non-polar groups; calcium bridges link negatively charged groups; and non-adsorbing polymers cause depletion attraction, because they are excluded from the gap between close particles and the resulting osmotic pressure pushes the particles together. Depletion flocculation explains why adding some polysaccharides to milk drinks causes separation.
An interface is the boundary between two phases, such as oil and water or air and water. Molecules at an interface have fewer favourable neighbours than those in the bulk, so creating interface costs energy. The interfacial tension, γ, is this energy per unit area, or equivalently the force per unit length along the interface. Pure water against air has a surface tension of about 72 mN/m at room temperature; triglyceride oil against water has an interfacial tension of about 25 to 30 mN/m. Making an emulsion or a foam multiplies interfacial area enormously, so it needs energy and, to keep the result, molecules that lower γ and protect the interface.
Surface-active molecules, surfactants in the broad sense, are amphiphilic: they have hydrophilic and hydrophobic parts and so accumulate at interfaces, lowering γ. The Gibbs adsorption equation links the fall in γ with concentration to the amount adsorbed per unit area, the surface excess. Small-molecule surfactants form micelles above a critical micelle concentration, after which surface tension changes little.
A curved interface carries a pressure difference, the Laplace pressure, ΔP = 2γ/r for a sphere of radius r. Small droplets therefore have high internal pressure: a droplet of radius 0.5 µm with γ = 10 mN/m has ΔP = 2 × 0.01/0.5 × 10⁻⁶ = 40 kPa. Breaking droplets requires stresses that exceed this pressure, which is why making fine emulsions needs intense energy and why surfactants that lower γ make it easier. Laplace pressure also drives Ostwald ripening, the growth of large droplets or bubbles at the expense of small ones.
Two broad kinds of surface-active agents operate in dairy systems, and they protect interfaces in different ways. Small-molecule surfactants, such as monoglycerides, phospholipids including lecithin, polysorbates and sucrose esters, adsorb quickly, pack closely and lower interfacial tension strongly. Their films are fluid: when a film is stretched, surfactant flows along the interface from regions of low tension to high tension, dragging liquid with it and healing thin spots. This Gibbs-Marangoni mechanism stabilises films against rupture.
Proteins adsorb more slowly, lower γ less, but form thick, viscoelastic films. On adsorbing, proteins partly unfold, exposing hydrophobic segments to the oil or air, and interact laterally to form a cohesive skin; flexible proteins such as β-casein form loose, thick layers with loops and tails, while globular proteins such as β-lactoglobulin form dense, rigid layers that may cross-link through disulphide bonds. These films stabilise droplets by steric and electrostatic repulsion and by their mechanical strength.
Proteins and small surfactants compete. Because small surfactants lower γ more, they can progressively displace proteins from an interface, a process called orogenic displacement, in which surfactant domains grow and compress the protein network until it is pushed off. Mixed films are weaker than either pure film because the two stabilising mechanisms interfere. Dairy technologists exploit this: in ice cream, emulsifiers such as mono- and diglycerides displace casein from fat globule surfaces during ageing, weakening the films so that partial coalescence can occur during freezing and whipping, which builds the fat network that gives dryness and shape retention.
An emulsifier is a surface-active substance that helps form an emulsion and keep it stable. In the strict food-additive sense the term covers small-molecule surfactants, while proteins such as caseins and whey proteins, and some polysaccharides such as gum arabic and modified starches, act as emulsifying agents too. Common food emulsifiers include lecithins (mixtures of phospholipids from soybean, sunflower, rapeseed or egg), mono- and diglycerides of fatty acids, their esters with acetic, lactic, citric and diacetyl tartaric acids, polyglycerol esters, propylene glycol esters, sorbitan esters, polysorbates, sucrose esters and stearoyl lactylates. Each is approved for particular foods at particular levels; the Codex General Standard for Food Additives and national lists set these.
Emulsifiers serve functions beyond emulsification. In ice cream they promote partial coalescence and dryness. In bread they strengthen dough and slow staling by complexing amylose. In chocolate, lecithin lowers viscosity, and polyglycerol polyricinoleate lowers yield value. In margarine and spreads, which are water-in-oil emulsions, lipophilic emulsifiers such as monoglycerides and lecithin stabilise water droplets and prevent spattering. In recombined milks and creams, emulsifiers improve stability during heating. In coffee whiteners, they help the fat disperse when added to hot coffee.
Choosing an emulsifier depends on the type of emulsion, the processing conditions, the other ingredients, legal permissions and labelling. Many products use blends because no single molecule offers every property required.
The hydrophilic-lipophilic balance, or HLB, introduced by Griffin, is a number that describes the balance between an emulsifier's water-loving and oil-loving parts. On Griffin's scale, values run from 0 to 20. Low-HLB emulsifiers, about 3 to 6, are more soluble in oil and favour water-in-oil emulsions; high-HLB emulsifiers, about 8 to 18, are more soluble in water and favour oil-in-water emulsions. Wetting agents fall around 7 to 9, detergents around 13 to 15 and solubilisers around 15 to 18. Typical values are about 3 to 4 for distilled monoglycerides, about 4 to 5 for sorbitan monooleate and about 15 for polysorbate 80; lecithins vary widely with composition.
For non-ionic emulsifiers, Griffin gave formulas: for polyoxyethylene derivatives HLB = E/5, where E is the weight percentage of ethylene oxide, so an ethoxylate with 75% ethylene oxide has HLB 15; for polyol fatty acid esters HLB = 20(1 − S/A), where S is the saponification number of the ester and A the acid number of the fatty acid. Davies later proposed adding group numbers for each chemical group.
HLB values of blends are approximately additive by mass. To obtain a required HLB of 12 from sorbitan monooleate (4.3) and polysorbate 80 (15.0), the fraction of polysorbate x satisfies 4.3(1 − x) + 15.0x = 12, so x = 7.7/10.7 ≈ 0.72, that is 72% polysorbate and 28% sorbitan ester.
Bancroft's rule complements HLB: the phase in which the emulsifier is more soluble tends to become the continuous phase. HLB is a useful first guide but does not account for temperature, oil type, salts or the strong effect of proteins, so final choice is made by trial.
An emulsion is a dispersion of droplets of one liquid in another with which it is immiscible. Oil-in-water emulsions include milk, cream, mayonnaise and dressings; water-in-oil emulsions include butter and margarine. Multiple emulsions, such as water-in-oil-in-water, are used experimentally to encapsulate ingredients.
Making an emulsion has two simultaneous tasks: breaking droplets and covering the new interface with emulsifier before droplets collide and merge. Droplet break-up requires the external stress to exceed the Laplace pressure of the droplet over a long enough time. In laminar flow, viscous shear deforms and splits droplets; in turbulent flow, inertial forces from eddies of about the droplet's size tear droplets apart; cavitation, the formation and collapse of vapour bubbles, contributes shock waves in high-pressure valves and ultrasonic devices. The resulting droplet size depends on the energy density applied, the interfacial tension, the viscosity of both phases, the amount of emulsifier and how quickly it adsorbs.
Equipment ranges from simple stirred tanks and colloid mills, which give droplets of several micrometres, through high-shear rotor-stator mixers, to high-pressure valve homogenisers, microfluidisers and ultrasonic homogenisers, which give sub-micrometre droplets. In dairy plants the high-pressure valve homogeniser is standard. A sufficient emulsifier must be present: if there is too little to cover the new interface, newly formed droplets recoalesce, and droplet size stops falling however much pressure is applied.
A milk homogeniser is a high-pressure positive-displacement piston pump that forces milk through a narrow adjustable gap between a valve seat and a valve, at velocities of the order of 100 m/s or more. Turbulence and cavitation in and after the gap break fat globules from a mean of about 3 to 4 µm to well under 1 µm. Typical pressures are 10 to 25 MPa in the first stage. A second stage at lower pressure, roughly a fifth of the first, breaks up clusters of newly formed globules that share adsorbed casein micelles; such clusters would otherwise raise viscosity and promote creaming.
Homogenisation is done with the fat liquid, typically at 55 to 80 °C, because crystalline fat does not break up properly. It enormously increases the fat surface area; at constant fat volume, area is inversely proportional to diameter, so reducing the mean from 3.5 to 0.5 µm raises area sevenfold. The original membrane material is far too little to cover this, so the new surface is covered mainly by adsorbed casein micelles and micelle fragments, with some whey proteins. The homogenised globules therefore behave partly like casein particles: they take part in acid and rennet gels, they are destabilised by the same treatments as casein, and they interact with denatured whey proteins.
Two side effects matter. Raw milk must not be homogenised and then held, because the disrupted membrane allows rapid lipolysis by milk lipase and a rancid flavour; homogenisation is therefore combined with pasteurisation. And homogenised milk does not cream, but its efficiency must be checked: one common test compares the fat content of the top layer with the rest after storage, and the United States milk ordinance, for example, requires that after 48 hours of storage the fat in the top 100 mL of a quart differs by no more than 10% from that in the remainder.
Emulsions are thermodynamically unstable: the system always loses free energy by reducing interfacial area. They can, however, be kinetically stable for long periods. Several processes break them down: gravitational separation (creaming or sedimentation), flocculation, coalescence, partial coalescence, Ostwald ripening and phase inversion. These often occur together and reinforce one another.
Creaming is the upward movement of droplets less dense than the continuous phase. For an isolated sphere in a Newtonian liquid, Stokes' law gives the velocity v = d²(ρc − ρd)g/(18η), where d is the diameter, ρc and ρd the densities of the continuous and dispersed phases, g gravitational acceleration and η the viscosity of the continuous phase. For a globule of 4 µm, a density difference of 100 kg/m³ and a serum viscosity of 2 mPa s, v = (4 × 10⁻⁶)² × 100 × 9.81/(18 × 0.002) ≈ 4.4 × 10⁻⁷ m/s, about 1.6 mm per hour. Because v depends on d², halving the diameter cuts creaming rate to a quarter; this is the principle of homogenisation. Raising viscosity or reducing the density difference also slows creaming.
Raw cow milk creams much faster than Stokes' law predicts for single globules, because cold agglutinins, immunoglobulin M, cluster globules at low temperature into large flocs that rise quickly. This cold agglutination gives the distinct cream line of unhomogenised milk. Heating milk more severely than minimum pasteurisation inactivates the agglutinins and weakens creaming, and homogenisation prevents it. Milks of species such as goats lack this agglutination and cream slowly.
Flocculation is the aggregation of droplets into clusters in which each droplet keeps its identity, separated by a thin film. It may be weak and reversible, as in depletion flocculation, or strong, as in bridging flocculation, where a polymer or protein adsorbs to two droplets at once. Flocs cream faster and raise viscosity.
Coalescence is the merging of two droplets into one when the thin film between them ruptures. It leads to ever-larger drops and finally to a separate oil layer, called oiling off. Coalescence is resisted by thick, viscoelastic or strongly charged interfacial films and by keeping droplets apart.
Partial coalescence is the key instability of dairy products. It happens when globules contain both liquid oil and fat crystals. A crystal protruding from one globule can pierce the film between it and a neighbour; oil flows round the crystal, joining the globules, but the crystal network inside prevents them from rounding up into one sphere. The result is an irregular clump that keeps the shapes of its parent globules. Partial coalescence is promoted by an optimum ratio of solid to liquid fat, by shear and by incorporation of air, where globules are pushed together at bubble surfaces, and by emulsifiers that weaken protein films. It is essential to whipped cream and ice cream, where clumped fat stabilises air bubbles, and to butter making, where churning causes so much partial coalescence that the emulsion inverts. It is a defect in other products, causing fat plugs in cream and fat clumps in UHT milk.
Ostwald ripening is the growth of large droplets at the expense of small ones by diffusion of dispersed phase through the continuous phase, driven by the higher Laplace pressure and solubility of small droplets. It is negligible for triglyceride oils, which are almost insoluble in water, but important for flavour oils, nanoemulsions and foams.
A foam is a dispersion of gas bubbles in a liquid or solid continuous phase. Food foams include whipped cream, ice cream, mousses, meringues, the head on beer, milk foam on coffee, and aerated chocolate. Bubbles are usually far larger than emulsion droplets, from tens of micrometres to millimetres, and the air-water interface has a higher tension than the oil-water interface, so foams are intrinsically less stable than emulsions.
Foams are formed by whipping or beating, by injecting gas through a sparger or porous plate, by releasing dissolved gas when pressure falls, as in aerosol cream, or by gas production in fermentation. Surfactants and proteins adsorb to the new air-water interface; whey proteins, caseins and egg-white proteins are good foaming agents because they adsorb, unfold and form cohesive films.
Three processes destroy foams. Drainage is the flow of liquid out of the films and channels between bubbles under gravity; it is slowed by raising the viscosity of the continuous phase or gelling it. Disproportionation, the foam equivalent of Ostwald ripening, is the diffusion of gas from small bubbles, at higher Laplace pressure, through the liquid into large bubbles, so that small bubbles shrink and vanish; it is resisted by gases of low solubility and by rigid interfacial films. Coalescence occurs when films between bubbles rupture. Lipids, especially liquid fat spreading at the air-water interface, are strong antifoams, which is why traces of yolk prevent egg whites from whipping, and why milk with lipolysed fat foams poorly.
Whipped cream is a foam stabilised not mainly by proteins but by partially coalesced fat. During whipping, protein adsorbs to the new bubbles, and fat globules attach to the bubble surfaces. Partly crystalline globules collide at the interface and in the serum and partially coalesce into a continuous fat network that encases the bubbles and links them, making the foam stiff and stable. For this the cream must contain enough fat, typically 30% or more and usually 35 to 40%, and the fat must be partly solid, so cream is aged and whipped cold, at about 5 °C. Warm cream, with liquid fat, whips poorly; over-whipping produces so much partial coalescence that butter grains separate. Homogenised cream whips poorly because its small, protein-coated globules resist partial coalescence, which is why emulsifiers are added to some aerosol and recombined creams.
Overrun expresses how much air has been incorporated. On a volume basis it is the increase in volume as a percentage of the original volume: overrun = (volume of foam − volume of mix)/volume of mix × 100. It is usually measured by weight: overrun = (mass of a given volume of mix − mass of the same volume of foam)/mass of the same volume of foam × 100. If 1 L of mix weighs 1,100 g and 1 L of ice cream weighs 550 g, overrun is (1,100 − 550)/550 × 100 = 100%.
Typical overruns are about 100% or more for whipped cream and for standard ice cream, and lower for premium ice cream; values depend on product and legal definitions, which in some countries set a minimum weight per volume. Overrun affects texture, melting, cost and the legal description of the product.
A gel is a continuous three-dimensional network of connected molecules or particles that traps a large amount of liquid, giving a material that behaves mechanically as a solid. In rheological terms a true gel has a storage modulus G′ larger than its loss modulus G″, both nearly independent of frequency, and shows a yield stress. Gels form when molecules or particles link together above a critical concentration; the network may be held by covalent bonds, such as disulphide bonds, or by physical junctions such as hydrogen bonds, hydrophobic interactions, ionic bridges and helix associations. Physical gels are often thermoreversible: gelatin gels melt on warming and reform on cooling.
Food gels are broadly polymer gels, in which long flexible or semi-rigid molecules form junction zones, and particle gels, in which aggregated particles form fractal clusters that connect into a network. Casein gels in yoghurt and cheese are particle gels; gelatin and most polysaccharide gels are polymer gels. Gels may contain dispersed particles such as fat globules; these are active fillers if they bind to the network and strengthen it, and inactive fillers if they do not.
Syneresis, the spontaneous expulsion of liquid from a gel as the network contracts, is a key property: it is a defect in yoghurt and desserts but is essential in cheesemaking, where it removes whey from the curd.
Whey protein gels form on heating. β-Lactoglobulin and other globular whey proteins unfold, expose hydrophobic groups and the free thiol of β-lactoglobulin, and aggregate by hydrophobic interactions and thiol-disulphide exchange. The structure depends on conditions: far from the isoelectric point (about pH 5.2) and at low ionic strength, strong repulsion produces fine-stranded, transparent gels made of thin chains; near the isoelectric point or at high salt concentration, rapid aggregation produces particulate, opaque gels that hold water less well. Pre-heated whey proteins can also be gelled cold by adding salt or acid, which is useful for heat-sensitive products. Gelatin, from collagen, forms thermoreversible gels through triple-helix junctions that melt near body temperature, giving the melt-in-the-mouth texture of jellies and some desserts.
Polysaccharide gels differ in their gelling mechanism. Agar forms helices that aggregate, giving gels that set around 35 to 40 °C but melt only near 85 °C, a large hysteresis. Kappa-carrageenan forms firm, brittle gels with potassium ions and shows syneresis; iota-carrageenan forms soft, elastic gels with calcium; lambda-carrageenan does not gel and serves as a thickener. Alginates form gels with calcium through the 'egg-box' binding of guluronic acid blocks, and these gels set cold and do not melt on heating. High-methoxyl pectin gels at high sugar and low pH; low-methoxyl pectin gels with calcium by the egg-box mechanism. Gellan forms gels with cations. Xanthan does not form a true gel alone but gives very shear-thinning, weak-gel solutions, and gels with locust bean gum through synergistic interaction. Starch thickens by gelatinisation and sets as it retrogrades.
Acid milk gels are the basis of yoghurt, fermented milks, quark, fresh acid cheeses and acid casein. Acidification is usually by lactic acid bacteria, or in model studies and some products by glucono-δ-lactone, which hydrolyses slowly to gluconic acid and gives uniform acidification. As pH falls from about 6.7, colloidal calcium phosphate progressively dissolves out of the micelles, and the negative charge on casein falls. Near pH 5.0 to 5.2 in unheated milk, the κ-casein hairs collapse, steric and electrostatic repulsion are lost, and micelles begin to aggregate into a particle gel. Gelation is complete by about pH 4.6, the isoelectric point of casein, at which yoghurt and acid curds are finished.
Heat treatment of the milk before acidification changes the gel profoundly. At yoghurt heat treatments, commonly about 85 °C for 30 minutes or 90 to 95 °C for several minutes, most whey proteins denature and many attach to the micelle surface through complex formation with κ-casein. Because denatured whey proteins have a higher isoelectric point, gelation starts at a higher pH, around 5.3, and the gel is firmer, more viscous and less prone to syneresis. This is one reason yoghurt milk is heated much more severely than market milk. Increasing milk solids, by adding milk powder or by membrane concentration, further firms the gel.
Defects of acid gels include whey separation (syneresis), caused by too little heat treatment, too low solids, too high incubation temperature, too low final pH or mechanical disturbance; and graininess or lumpiness from uneven acidification or poor mixing. Stirred yoghurt is broken after gelation and smoothed, so its texture depends on how the gel fragments are handled and cooled.
Rennet coagulation is the basis of most cheese. It proceeds in two overlapping phases. In the primary, enzymatic phase, chymosin, the principal enzyme of calf rennet, or other coagulants, specifically cleaves κ-casein at the Phe105–Met106 bond. This releases the hydrophilic C-terminal part, the caseinomacropeptide (residues 106 to 169), into the whey and leaves para-κ-casein on the micelle. The micelles lose their steric stabilisation. In the secondary, non-enzymatic phase, when most of the κ-casein, roughly 80 to 90%, has been hydrolysed, the destabilised para-casein micelles aggregate with the help of calcium ions and form a gel. The gel then strengthens as bonds multiply and particles rearrange.
The secondary phase depends strongly on temperature and calcium. Below about 15 to 20 °C aggregation is very slow, even though the enzyme acts, which is the basis of cold renneting methods; cheese milk is renneted at about 30 to 35 °C. Adding calcium chloride shortens coagulation time and firms the curd, especially in milk that has been cooled for storage or pasteurised. Lower pH speeds both phases. Milk heated so severely that whey proteins are extensively denatured, as in UHT treatment, coagulates poorly or not at all, because denatured whey proteins attached to κ-casein interfere with aggregation.
Rennet gels differ from acid gels. They form at the natural pH of milk, retain colloidal calcium phosphate and so are more elastic, and above all they synerese readily when cut, especially when stirred, heated and acidified. That controlled syneresis is how cheesemakers decide the moisture of each cheese. Gel firmness and the right cutting time can be followed by rheometers or by optical sensors in the vat.
Hydrocolloids are water-soluble or water-dispersible polymers, mostly polysaccharides plus gelatin, used to thicken, gel, stabilise and control water. In dairy products they are generally called stabilisers. Common ones are locust bean gum, guar gum, carboxymethylcellulose, carrageenans, alginates, pectins, xanthan, gellan, starches and gelatin, and microcrystalline cellulose in some drinks.
In ice cream, stabilisers such as locust bean gum, guar and carboxymethylcellulose raise mix viscosity, help smooth texture, slow meltdown and, most importantly, limit the growth of ice crystals during temperature fluctuations in storage. Locust bean gum and guar are incompatible with milk proteins and on their own cause wheying-off, a visible serum separation in the mix; a small amount of κ-carrageenan is added to prevent this. In chocolate milk, κ-carrageenan at a few hundredths of a per cent forms a weak network with casein micelles that suspends cocoa particles and prevents sedimentation without noticeable thickening. In yoghurt and desserts, gelatin, pectin, starches and carrageenans increase firmness and reduce syneresis. In processed cheese and cream cheese, locust bean gum and other hydrocolloids help water binding.
Stabiliser levels are low, usually well under 1%, and must be matched to the product, the process and permitted additive lists.
When proteins and polysaccharides are mixed, three outcomes are possible. If the two do not interact specifically and occupy different volumes, they may be co-soluble at low concentrations but separate into two phases at higher concentrations, one rich in protein and the other in polysaccharide. This segregative phase separation, or thermodynamic incompatibility, explains wheying-off in ice cream mixes and serum separation in milk drinks with added neutral gums. Depletion flocculation of casein micelles by non-adsorbing polysaccharides is a related effect.
If the protein and polysaccharide carry opposite charges, they attract and can form soluble complexes or separate into a dense complex coacervate phase; this associative separation occurs typically at pH values between the protein's isoelectric point and the pKa of the polysaccharide's acid groups. Complex coacervation of whey protein or gelatin with gum arabic or pectin is used for microencapsulation.
Specific adsorption can stabilise. In acidified milk drinks, such as drinking yoghurts and acidified milk beverages at pH about 3.8 to 4.2, high-methoxyl pectin adsorbs onto the positively charged casein particles through its negatively charged galacturonic acid blocks, while loops and tails of pectin extend into the serum and provide steric and electrostatic repulsion; this prevents aggregation and sedimentation even after heat treatment. Too little pectin leaves particles bare and leads to bridging flocculation; too much non-adsorbed pectin can cause depletion flocculation. Carboxymethylcellulose and soybean polysaccharides are used similarly. κ-Carrageenan's stabilising action in neutral milk relies on a specific interaction with κ-casein on the micelle surface, combined with its own weak gelation.
Microscopy shows how the components of a dairy product are arranged. Light microscopy, with resolution limited to about 0.2 µm by the wavelength of light, shows fat globules, air bubbles, ice crystals, lactose crystals and protein aggregates; selective stains help, such as fat-soluble dyes for fat and protein stains for casein. Polarised light reveals crystalline material such as fat and lactose crystals.
Confocal laser scanning microscopy uses fluorescent dyes, for example Nile red for fat and fast green or rhodamine for protein, and a scanning laser with a pinhole to collect light from a single focal plane. It produces sharp optical sections that can be stacked into three-dimensional images, with little sample preparation, so it can follow structure in yoghurt, cheese, cream and ice cream almost in its natural state and even during processing.
Electron microscopy resolves far smaller features. Transmission electron microscopy, at nanometre resolution, shows the internal structure of casein micelles, fat globule membranes and protein strands, but needs fixation, dehydration and thin sectioning or replication, which can introduce artefacts. Scanning electron microscopy images surfaces and fracture faces in three-dimensional appearance; cryo-preparation reduces artefacts by freezing samples rapidly. Atomic force microscopy probes surfaces at near-molecular resolution and measures local mechanical properties. Each technique has trade-offs between resolution, sample preparation, field of view and the risk of artefacts, so researchers usually combine several.
The size distribution of fat globules, micelles and aggregates is measured by several methods. Static laser light diffraction passes a laser beam through a dilute sample and measures scattered light intensity against angle; Mie theory, using the refractive indices of particle and medium, converts the pattern to a volume-based size distribution, covering roughly 0.02 to 2,000 µm. Common outputs are the volume-weighted mean diameter d43 and the surface-weighted Sauter mean diameter d32 = Σnd³/Σnd². The specific surface area of an emulsion follows from d32: area per unit volume of emulsion = 6φ/d32, where φ is the volume fraction of the dispersed phase. For 4% fat by volume and d32 = 0.5 µm, area = 6 × 0.04/0.5 × 10⁻⁶ = 4.8 × 10⁵ m² per m³, or 0.48 m² per mL.
Dynamic light scattering, also called photon correlation spectroscopy, measures fluctuations in scattered light caused by Brownian motion. Small particles move faster; the diffusion coefficient obtained from the fluctuations gives the hydrodynamic diameter through the Stokes-Einstein equation. It suits particles from about a nanometre to a few micrometres, such as casein micelles and nanoemulsion droplets, and reports a z-average diameter and a polydispersity index; samples must be dilute and free of large contaminants.
Zeta potential is measured by electrophoretic light scattering: an electric field moves charged particles, the velocity is measured by Doppler shift, and the electrophoretic mobility is converted to zeta potential using the Henry or Smoluchowski equation. Other techniques include electrical sensing zone counting, nuclear magnetic resonance for solid fat content and droplet size, ultrasonic spectroscopy for concentrated systems, and rheology for the mechanical consequences of structure.
New structuring approaches extend classical colloid science. An organogel, called an oleogel when the liquid is edible oil, is a liquid oil structured into a solid-like material by a network of a gelator, without relying on saturated fat crystals. Gelators include ethylcellulose, which forms a polymer network when dissolved in hot oil and cooled; waxes such as rice bran, candelilla and sunflower waxes, which form crystal networks at a few per cent; monoglycerides; mixtures of phytosterols and oryzanol, which form tubules; and indirect methods in which an emulsion or foam stabilised by protein is dried and sheared in oil. Oleogels are studied as replacements for saturated and trans fats in spreads, bakery products, processed meats and ice cream, and as carriers for lipophilic nutrients. Their challenges are mouthfeel, oil binding, oxidation, cost and regulatory approval of some gelators.
Hydrogels are water-swollen networks of hydrophilic polymers, including the protein and polysaccharide gels already described, and engineered particles of micrometre size made by extrusion, emulsification or coacervation. Microgel particles of whey protein, alginate or pectin are used to entrap probiotics, flavours and bioactives, protect them during processing and digestion, and release them at the right place; whey protein microparticles of about 1 µm also serve as fat replacers because they feel creamy. Emulsion gels, where droplets stabilised by protein are embedded in or themselves form a gel, combine the properties of both and are models for yoghurt and cheese.
Nanoemulsions are emulsions with very small droplets, commonly defined as having mean radii below about 100 nm, though definitions vary. Because their droplets are much smaller than the wavelength of light, they can be transparent or slightly turbid; because Brownian motion dominates gravity at that size, they hardly cream; and their large surface area can increase the bioavailability of lipophilic compounds such as vitamin D, vitamin E, carotenoids, curcumin and omega-3 fatty acids.
Nanoemulsions are made by high-energy methods, mainly high-pressure valve homogenisers operated at high pressure with several passes, microfluidisers, which collide two jets in an interaction chamber, and ultrasonic devices; or by low-energy methods that exploit the spontaneous formation of fine droplets when composition or temperature changes, such as spontaneous emulsification and the phase inversion temperature method with non-ionic surfactants. Low-energy methods usually need high surfactant concentrations, which can be undesirable in food.
Nanoemulsions are kinetically, not thermodynamically, stable, which distinguishes them from microemulsions, which form spontaneously and are thermodynamically stable within a defined range of composition. Their main instability is Ostwald ripening, especially when the oil phase has some water solubility, as with flavour oils; it is inhibited by mixing in a poorly water-soluble oil such as a long-chain triglyceride. Questions remain about the digestion, absorption and safety of very small particles, so regulators in several jurisdictions evaluate nanomaterials in food specifically.
A student uses this subject to see dairy products as structures rather than recipes. When learning cream separation, ice cream manufacture, yoghurt making or cheese, the student can explain each step in colloid terms: why milk is homogenised hot, why cream is whipped cold, why yoghurt milk is heated hard, why calcium chloride is added to cheese milk. In examinations the subject gives questions on Stokes' law, HLB blends, overrun, the chymosin cleavage site and gelation pH, and on how microscopy and light scattering work.
A product or process technologist with five years' experience uses this knowledge to solve recurring problems: cocoa settling in chocolate milk, serum separation in a new acidified drink, an ice cream that melts too fast or feels icy, cream that will not whip after a supplier changed, cheese curd that is too soft, or yoghurt that weeps whey. They choose and dose stabilisers and emulsifiers, set homogenising pressures and temperatures, and use particle-size and microscopy results to confirm causes before changing a formulation.
A senior technical director uses colloid science to steer product strategy and troubleshooting across plants. They judge whether a cheaper stabiliser or emulsifier blend will hold up across seasons and markets, approve process changes that alter homogenisation or heat treatment, and decide when an instability in a long-life product justifies a recall or a reformulation. They also negotiate ingredient specifications, interpret supplier data critically and guide teams developing reduced-fat, high-protein and plant-dairy hybrid products where colloidal behaviour is less familiar.
A senior researcher uses this subject as the foundation for structural and mechanistic research: measuring interfacial rheology of milk proteins, modelling aggregation and gelation kinetics, designing emulsion gels and oleogels with targeted texture and nutrient release, and studying how structure governs digestion. They combine microscopy, scattering and rheology to build quantitative structure-function models, review the safety questions around nanostructured foods, and train students to connect molecular interactions with the behaviour consumers notice. They also referee claims for new stabilisers and structuring agents, asking for evidence of mechanism as well as performance.
This manual gives mechanisms and representative values; globule and micelle sizes, gelation pH, homogenising pressures, stabiliser doses and overruns vary with milk, equipment and product and must be set by trial. It does not derive DLVO theory, Mie theory or light-scattering mathematics in full, nor treat interfacial rheology methods in detail. It does not list which emulsifiers and stabilisers are permitted in each product in each jurisdiction; these must be read in the Codex General Standard for Food Additives and national law. Ice cream and cheese technology are covered in their own subjects. Research on oleogels and nanoemulsions is moving quickly, and regulatory positions on nanomaterials in food differ between jurisdictions.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.