Currently, plant proteins are fractionated to ingredients with high purities, but an often ignored point is the impact of the extraction and fractionation process on protein functionality. To allow a fair and effective comparison, it is key to understand the changes in protein's aggregated state occurring in the extracted ingredients during processing. We review conventional and upcoming plant protein extraction and fractionation processes (on pulses and oilseeds) and focus on how the processing history influences the macroscopic functional properties of the proteins. To establish this link, we dive into seed morphology and give an overview of the plant seed composition. In addition, we explain the essence of each process step and how it impacts the protein's aggregated state. The latter is linked to the macroscopic functionality (foaming, emulsification, and gelation). We identified three major protein structure-changing steps in the conventional protein extraction process: defatting, alkaline extraction, and isoelectric point precipitation. These steps lead to large, insoluble aggregated structures, which strongly impacts the protein macroscopic functionality. Milder extraction methods reduce these alterations, but a potential consequence is the presence of non-proteinaceous components, which could give challenges in sensory and nutritional aspects and affect the techno-functional properties of the ingredient. The take-home-message is that we need to consider the process-induced change of the protein aggregated structures, which are likely to dominate the functionality over the protein's molecular parameters.

Complex interfaces stabilized by proteins, polymers or nanoparticles, have a much richer dynamics than those stabilized by simple surfactants. By subjecting fluid-fluid interfaces to step extension-compression deformations, we show that in general these complex interfaces have dynamic heterogeneity in their relaxation response that is well described by a Kohlrausch-Williams-Watts function, with stretch exponent β between 0.4–0.6 for extension, and 0.6–1.0 for compression. The difference in β between expansion and compression points to an asymmetry in the dynamics. Using atomic force microscopy and simulations we prove that the dynamic heterogeneity is intimately related to interfacial structural heterogeneity and show that the dominant mode for stretched exponential relaxation is momentum transfer between bulk and interface, a mechanism which has so far largely been ignored in experimental surface rheology. We describe how its rate constant can be determined using molecular dynamics simulations. These interfaces clearly behave like disordered viscoelastic solids and need to be described substantially different from the 2d homogeneous viscoelastic fluids typically formed by simple surfactants.