The key parameter controlling the glass transition of colloidal suspensions is ϕ, the fraction of the sample volume occupied by the particles. Unfortunately, changing ϕ by varying an external parameter, e.g. temperature T as in molecular glass formers, is not possible, unless one uses thermosensitive colloidal particles, like the popular poly(Nisopropylacrylamide) (PNiPAM) microgels. These however have several drawbacks, including high deformability, osmotic deswelling and interpenetration, which complicate their use as a model system to study the colloidal glass transition. Here, we propose a new system consisting of a colloidal suspension of non-deformable spherical silica nanoparticles, in which PNiPAM hydrogel spheres of ∼ 100 − 200 µm size are suspended. These non-colloidal 'mesogels' allow for controlling the sample volume effectively available to the silica nanoparticles and hence their ϕ, thanks to the T-induced change in mesogels volume. Using optical microscopy, we first show that the mesogels retain their ability to change size with T when suspended in Ludox suspensions, similarly as in water. We then show that their size is independent of the sample thermal history, such that a well-defined, reversible relationship between T and ϕ may be established. Finally, we use space-resolved dynamic light scattering to demonstrate that, upon varying T , our system exhibits a broad range of dynamical behaviors across the glass transition and beyond, comparable with those exhibited by a series of distinct silica nanoparticle suspensions of various ϕ.

Wheat gluten includes two major proteins classes, gliadin (25–60 kg/mol) and glutenin polymers (100 to > 2,000 kg/mol) each comprising several polypeptides routinely identified by size-exclusion chromatography and electrophoresis. Gluten proteins are rich in glutamine (30%) and contain several repeated sequences, linking them to the wide class of intrinsically disordered protein (IDP). Here we showed that an ethanol/water (EtOH/W, 50/50, v/v) extract of an industrial gluten, comprising 1/3 of glutenin polymers and 2/3 of gliadin, underwent liquid-liquid phase separation (LLPS) below 14 °C, leading to two coexisting phases, respectively rich and poor in protein. As the quenching depth increased, proteins of lower and lower molecular weight joined the rich phase, akin to what would have been obtained for a polydisperse polymer sample. Within the rich phase the mass ratio of glutenin over gliadin decreased from 2.5 to 0.5 as the temperature dropped from 14 °C to −0.8 °C. Concomitantly the concentration in glutenin polymers increased up to 143 ± 6 g/L (at 9 °C) and then stopped to evolve, suggesting that the binodal line intersected the gelation line below this temperature. Applying the Flory-Huggins (FH) lattice model for each gluten protein classes, we demonstrated that their partitioning in the coexisting phases followed a same temperature dependency. However, some gliadin species joined the rich phase above their critical temperature. Here, specific interactions with the glutenin polymers through weak forces were exemplified. The study demonstrated the relevance of the Flory-Huggins (FH) lattice model in predicting phase behavior even when applied to complex protein mixtures.

Both an experimental and a theoretical investigation of the fracture propagation mechanisms acting at the process zone scale in glassy polymers are presented. The main aim is to establish a common modeling for different kinds of glassy polymers presenting either steady-state fracture propagation or stick-slip fracture propagation or both, depending on loading conditions and sample shape. On the experimental point of view, new insights are provided by in-situ AFM measurements of the viscoplastic strain fields acting within the micrometric process zone in a brittle epoxy resin, which highlight an extremely slow unexpected steadystate regime with finite plastic strains of about 30% around a blunt crack tip, and accompanied by propagating shear lips. On the theoretical point of view, we apply to glassy polymers some recently developed models for describing soft dissipative fracture that are pertinent with the observed finite strains. We propose a unified modeling of the fracture energy for both the steady-state and stick-slip fracture propagation based on the evaluation of the energy dissipation density at a characteristic strain rate induced in the process zone by the competition between the crack propagation velocity and the macroscopic sample loading rate.