Clustering of oil droplets in foods : Implications on rheology, tribology and sensory perception

Many foods with a high palatability include significant quantities of fat. Fat influences next to the sensory perception and palatability also the microstructure of foods. It is thus challenging to reduce the fat content in many food items without negative impact on texture or perception. Fats and oils are mostly found as droplets in both solid and liquid matrices. Those droplets can be distributed homogeneously or inhomogeneously throughout a system. Previous research has shown that inhomogeneity (clustering) in the distribution of fat, in model systems, changes the fat related sensory perception and changes the mechanical properties. However, so far, the processes to obtain an inhomogeneous (clustered) fat distribution neither allow control over the process nor incorporation into foods. Therefore, this research aimed to understand the influence of fat phase inhomogeneity through controlled clustering of oil droplets on physical-chemical properties (rheology, tribology, cluster morphology), in-mouth behaviour and sensory perception of liquid o/w emulsions, emulsion gels and emulsion-filled gels. The findings of this research can help to control the microstructure of foods to design the physicochemical and sensory properties of foods. Further, the insights might help in reducing the fat content of foods without negative impact on mechanical properties and sensory perception.  Key conclusions/findings Oil droplet cluster size and interaction strength (within oil droplet clusters) can be controlled in liquid o/w emulsions, and semi-solid emulsion filled gels. Two methodologies provide good control over the size and interaction strength: hetero-aggregation and clustering with cross-linkers. For hetero-aggregation, the cluster size is controlled by the volume ratio of charged droplets and the strength by the charge of the emulsifier. For chemical clustering, the extent of clustering depends on cross-linker content. With regards to oil droplet inhomogeneity in gels, this process was controlled for overall strongly bound clusters (chemically and strong electrostatic interactions). An alternative approach is the incorporation of oil-rich or oil depleted filler particles into a gel matrix, introducing an inhomogeneity on a larger length scale. Inhomogeneity strongly affects rheological/mechanical properties and sensory perception of fat-related attributes. Clustering of oil droplets in o/w emulsions significantly increases viscosity due to entrapment of water in the oil droplet cluster leading to an increase in effective volume fraction and improved lubrication, allowing to enhance the perception of fat-related attributes. Emulsions with inhomogeneous oil droplet distribution were shown to be perceived as more thick and creamy. In gels, inhomogeneity mainly affects the modulus but can make gels also more brittle. If strong cluster interactions are present, graininess perception is enhanced. The effects on sensory perception in gels are less distinct. When taking increases in gel stiffness and brittleness into consideration, clustering in gels can be used to reduce the overall fat content.  Tribological characterisations of fat containing foods correlate to their sensory properties but depend on the presence of saliva When taking interactions between emulsions and saliva into account, tribological parameters correlate stronger with fat-related perception, then without saliva. We developed a new tribological setup to measure the friction properties of intact foods and boli samples. Material properties of emulsion gels can be directed by designing droplet-droplet interactions The rheological properties at small and large deformations of emulsion gels can be designed to vary between soft glassy and gel-like materials, by increasing the interaction strength between oil droplets. Overall, we conclude that by changing the oil droplet distribution in foods and the droplet-droplet interactions, the microstructure, physicochemical and sensory properties can be changed controllably