Rheology Control in Reduced Fat Oil in Water Emulsions

In reduced-fat oil-in-water emulsions, the transition from a close-packed droplet network to a dilute droplet dispersion is governed by the dispersed-phase volume fraction, droplet size distribution, and continuous-phase viscoelasticity. A full-fat mayonnaise at 70–80 wt% oil typically behaves as a viscoplastic material with an apparent viscosity at 50 s⁻¹ of 5–20 Pa·s and a critical strain below 10% in oscillatory amplitude sweeps; reducing the oil phase to 3–30 wt% without continuous-phase structuring collapses the apparent viscosity to 0.05–0.50 Pa·s and eliminates the yield stress necessary for suspension stability. The resulting emulsion separates by creaming because the oil droplets are free to rise through the low-viscosity aqueous phase, and the absence of droplet packing prevents the formation of a weak elastic network. Rheological characterization should therefore be performed under standardized rotational and oscillatory conditions: ISO 3219-2:2021 for rotational viscometry, ASTM D2196-20 for non-Newtonian viscosity, DIN 53019-1 for coaxial cylinder geometries, and ISO 13320:2020 for droplet size distribution by laser diffraction. A cone-and-plate fixture of 50 mm diameter and 1° cone angle, with Peltier temperature control at 25 °C, is adequate for low-fat dressings below 40 wt% solids; for spoonable formulations with yield stresses above 10 Pa, a roughened parallel plate or vane geometry prevents wall slip. The structuring task is not simply to increase viscosity; it is to recreate the yielding behaviour, thixotropic recovery, and small-strain elasticity that fat-filled droplet networks provide, while maintaining acceptable shear thinning during pumping and spooning. This requires independent control of the continuous-phase polymer network and the droplet interface, because a polymer that thickens the aqueous phase can simultaneously destabilize the emulsion through depletion flocculation, bridge droplets, or compete with emulsifier adsorption.

What Mechanisms Govern Yield Stress Development in Reduced-Fat Dressings Under High-Pressure Homogenization?

Three mechanisms dominate yield stress development in reduced-fat dressings below 30 wt% oil: depletion flocculation by non-adsorbing hydrocolloids, particle gelation by colloidal microcrystalline cellulose, and thermally induced protein or starch aggregation. High-pressure homogenization at 150/40 bar first-stage/second-stage settings reduces droplet Sauter mean diameter to 1–5 µm and increases interfacial area, but the small droplet size alone does not generate continuous-phase yield stress. When xanthan gum is incorporated at 0.05–0.40 wt% before homogenization, the rod-like polysaccharide is excluded from the gap between approaching droplets, producing depletion flocculation with an osmotic pressure differential that weakly aggregates the emulsion. The resulting yield stress, measured as the shear stress at 0.1 s⁻¹ after a defined pre-shear, generally falls between 0.5 Pa and 20 Pa in model dressings, while the Herschel-Bulkley flow behaviour index n ranges from 0.12 to 0.50. The high-shear homogenization step may transiently disrupt the flocculated network; apparent viscosity at 50 s⁻¹ often declines by 15–40% from the premix value, but thixotropic recovery after 120 s rest returns to 80–95% of the original yield stress. This behaviour is accessible to a three-interval test at 0.1 s⁻¹, 100 s⁻¹, and 0.1 s⁻¹ on a controlled-stress or controlled-rate rheometer. Formulators should record the shear stress at the yield point, the storage modulus G′ in the linear viscoelastic region at 1 Hz, and the phase angle δ; phase angles below 45° indicate elastic network dominance. Inline process control uses a pressure transducer upstream of the homogenizer second stage; a pressure drop above 0.5–1.0 bar across the feed line frequently indicates excessive depletion flocculation, which may lead to pump cavitation and an unstable feed. This condition is particularly acute in low-fat formulations using xanthan at concentrations above 0.25 wt% combined with salt above 1.5 wt%; the electrolyte screens electrostatic repulsion and strengthens depletion attraction, requiring the addition of xanthan downstream of the homogenizer or after a starch pre-gelatinization step.

Acidified pourable dressings with 3–10 wt% oil present a separate processing problem because the low dispersed-phase volume means that the emulsion is only weakly structured, and the primary thickening agent must remain functional at pH 3.0–3.8 and after pasteurization at 70–85 °C. Direct substitution with native starch is limited by heat, acid, and shear; acid hydrolysis and granule rupture reduce viscosity during storage. Modified waxy maize starches such as acetylated distarch adipate and hydroxypropyl distarch phosphate provide process tolerance, but their rheology depends on gelatinization intensity and the order of addition relative to homogenization. When starch is gelatinized in the aqueous phase at 70–85 °C for 10–20 min before oil addition, the swollen granules create a continuous yield-stress fluid that can be later homogenized at 60–100 bar first stage; however, excessive pressure after gelatinization can rupture granules and cause a viscosity drop of 30–60% at 50 s⁻¹. If the starch is not fully gelatinized, droplet coalescence occurs because the continuous phase is too mobile; if the starch is overprocessed, the emulsion thins and serum separation increases. A composite approach uses xanthan gum at 0.05–0.15 wt% to provide shear thinning and starch at 2.0–4.0 wt% to provide body; the two dry powders must be pre-blended with sugar or salt to prevent local hydration lumps. In batch pasteurization with a scraped surface heat exchanger, the product should reach 75–85 °C in the centre and hold for 30–60 s; the heat exchanger rotor speed should not exceed 300 rpm to limit shear-induced granule fragmentation. On production lines, the most frequent failure is post-pasteurization homogenization of swollen starch, where lobe pumps and high-shear mixers disturb the granule network and produce a grainy, low-viscosity product. Published data for this specific configuration is limited; validation on the actual line is required because batch pasteurization and continuous heat exchange create different shear histories.

Colloidal Microcrystalline Cellulose and the Shear History Required for Network Formation

Colloidal microcrystalline cellulose is a water-insoluble particle gel that requires a precise shear history to form a yield-stress network in reduced-fat emulsions. The raw material is a coprocessed blend of microcrystalline cellulose and sodium carboxymethylcellulose; the latter provides electrostatic and steric stabilization, while the former forms hydrogen-bonded aggregates. Dispersion at 0.15–1.00 wt% in a high-shear inline mixer with a tip speed of 10–24 m/s activates the network, but the processing window is narrow. Below 10 m/s tip speed, dispersion is incomplete and the product remains grainy with low yield stress; above 24 m/s tip speed, excessive viscous heating raises the product temperature above 60 °C, and the network may thin irreversibly. The preferred activation temperature is 20–40 °C; below 10 °C hydration is slow, and above 60 °C the cellulosic network can be disrupted by thermal motion and shear. Once activated, the dispersion displays a yield stress between 0.5 Pa and 12 Pa and a storage modulus G′ above G″ across 0.1–100 rad/s. The network is stable from pH 2.0 to 11.0, tolerates 1–10 wt% sodium chloride, and does not require heat. In a low-fat dressing process, microcrystalline cellulose is best activated in a separate aqueous phase before oil addition; if it is added after emulsification, it may coat the oil droplets and reduce the available network density. Recovery after pumping is rapid; typically 80% of original yield stress returns within 60–180 s after shear removal. The most common batch-to-batch variation originates from insufficient activation energy rather than from raw material lot differences; therefore, the dispersion should be checked by measuring shear stress at 10 s⁻¹ after a defined pre-shear, and by recording the mixer motor current or power draw per product mass. A universal power-input threshold cannot be transferred reliably from laboratory rotor-stator devices; process capability must be established on the production mixer used for the batch.

When Heat-Set Whey Protein Networks Replace Fat in Acidified Dressings, Thermal History Controls Batch-to-Batch Rheology

Heat-set whey protein concentrate at 0.5–3.0 wt% can generate a continuous gel network in acidified oil-in-water emulsions if the protein is thermally denatured above 70–85 °C and cooled under quiescent conditions. The gelation mechanism involves unfolding of β-lactoglobulin, exposure of hydrophobic domains, and formation of disulfide-stabilized aggregates that span the aqueous phase. In a low-fat dressing at pH 4.0–4.6, the protein is positively charged and may adsorb strongly at the oil droplet interface; this creates bridging flocculation if homogenization occurs after denaturation, or fills reinforcement if homogenization precedes denaturation. The resulting emulsions show a storage modulus G′ between 10 Pa and 1000 Pa, depending on protein concentration, heating rate, and salt. Heating in a scraped surface heat exchanger at 85 °C with a residence time of 30–90 s produces a weak gel that is pumpable but prone to shear-induced breakdown; subsequent passage through a positive displacement pump at speeds above 300 rpm may reduce yield stress by 50–80% without recovery. This imposes an operational boundary: high-pressure homogenization must be performed before heat treatment, and recirculation after gelation should be limited to low-shear lobe or progressive cavity pumps with tip speeds below 0.5 m/s. The use of whey protein concentrate in clean-label formulations must also account for lactose and mineral content; variation in calcium from 0.3 wt% to 1.0 wt% shifts gel onset temperature by 5–10 °C and alters the final yield stress. Laser diffraction according to ISO 13320:2020 should be used to verify droplet size remains below 10 µm before thermal gelation; if the droplet size exceeds this value, protein bridging and serum separation can occur during storage at 4 °C for 28 days.

On a production line for spoonable reduced-fat mayonnaise, the sequence of high-shear mixing, homogenization, heat treatment, and back-pressure-controlled cooling determines whether the yield-stress specifications are met. Inline rheometry or differential pressure measurement across the cooling tube can be used to infer the apparent viscosity at shear rates of 10–100 s⁻¹, but these values must be correlated with off-line measurements according to ISO 3219-2:2021 and ASTM D2196-20. A common processing conflict arises when the same formulation is run on two different homogenizer types: a radial diffuser homogenizer may generate lower droplet shear than a flat-valve high-pressure homogenizer at the same nominal pressure, resulting in larger oil droplets and lower viscosity in the final emulsion. The reason is that energy density per pass depends not only on the pressure setting but also on the valve geometry and the flow rate. Consequently, formulation transfer from pilot to production should be based on energy density in kJ/m³ and droplet size distribution rather than pressure alone. Droplet size distribution measured by ISO 13320:2020 should be recorded before and after heat treatment; changes in D[4,3] greater than 10% during heating indicate partial coalescence or protein bridging. For clean-label low-fat dressings using citrus fibre or potato protein, published data for specific low-fat emulsion applications is limited, and the operating window may be as narrow as ±5 °C for thermal gelation or ±0.05 wt% for salt-sensitive protein aggregation. In such cases, process capability must be confirmed by design-of-experiment batches that include hold time, pH, and shear variables; no single rheology modifier or mixing condition is universally applicable. The tolerance for viscosity variation at 50 s⁻¹ is often set at ±10–15% of target in full-fat products, but reduced-fat emulsions may require tighter bounds of ±5–10% because the yield-stress cliff-edge is steeper when the dominant network is a hydrocolloid gel rather than close-packed oil droplets. The final product specification should therefore include not only a Brookfield viscosity at 25 °C but also a yield-stress recovery ratio and a serum separation index after storage at 4 °C for 28 days. Published data for this specific configuration is limited; the tolerable shear history for each hydrocolloid network must be established on the production equipment that will deliver the emulsion.