Structure, dynamics and the role of particle size in bicontinuous Pickering emulsions
MetadataShow full item record
Bicontinuous Pickering emulsions (or bijels) are a relatively new class of novel soft material with many potential industrial applications, including microfluidics, tissue engineering and catalysis. They are typically formed by initiating the spinodal decomposition of a binary liquid mixture in the presence of neutrally-wetting colloidal particles. The particles attach at the liquid-liquid interface and arrest the phase separation by jamming when the concentration of particles approaches the 2D close-packing limit. Predicted by simulations in 2005 and realized in the laboratory in 2007, many aspects of the bijels complex behaviour and properties have remained unexplored. This thesis expands the knowledge of the bijels structural and dynamical properties, while focusing specifically on the role of particle size. The bijels porosity (average interfacial separation L) according to simulations can be controlled by varying the size r and volume fraction ϕ of particles in the system (L ∝ r/ϕ). The inverse scaling of L with ϕ has been verified for one size of particle, but to access smaller values of L (to allow the structure to be used for a wider range of industrial applications) the scaling with r must be tested. Chapter 3 presents the first systematic study of reducing particle size in bijels made with the liquid pair water/lutidine (W/L).We find that a five-fold reduction in r only requires moderate modification to preparation methods (concentrations of reactants during particle synthesis and increased particle sonication time) and in principle allows L values of between 1 & 10 μm to be accessed in the W/L system, where previously 10 μm was the limit. We demonstrate that this reduced lower bound of L can be translated into a lower bound for polymerized bijels also. Unfortunately, reducing particle size even further (in the same way) reveals a law of diminishing returns, as the uptake fraction of particles to the interface also reduces as we reduce particle size. Hence, to reduce lengthscale even further, a new bijel fabrication paradigm is required. Unexpectedly, we find that the temperature quench rate becomes less important for smaller particles (which constitutes a direct material synthesis advantage) and develop a new theoretical framework to take account of this observation. Large particles promote domain pinch-off during the coarsening (due to a larger driving force towards spontaneous curvature) resulting in bijel failure when slow rates are used because the time required to jam is greater than the time required for depercolation. To further probe the bijels structure as a function of particle size and quench rate, and to account for the success/failure scenarios which seem not to depend on L, in Chapter 4 we quantitatively characterize the morphology by measuring distributions of interfacial curvatures. By computing area-averaged quantities to make valid comparisons, we find that smaller particles and faster quench rates produce bijels with greater hyperbolic `open' character, aligning with our understanding of bijel formation gained from Chapter 3. We compare to simulated bijel data and an estimate of the hyperbolicity of the bare liquids undergoing spinodal decomposition, validating the results. In addition, we uncover a time-dependent `mutation' of the curvature distributions when large particles are used, but not when smaller particles or a different liquid pair is used. The mutation appears to correlate with the propensity of the interfacial particles to form a 'monogel', whereby the interfacial particles develop permanent bonds and remain as a 3D percolating network after the interface is removed, although the precise mechanism of the mutation is still to be verified. Following the results from Chapters 3 & 4 it is clear that there are potentially microscopic phenomena in the bijel which result in macroscopic aging and/or a determination of macroscopic structural properties. To investigate further, we use diffusing-wave spectroscopy (a form of light scattering) to probe the microscopic dynamics of the interfacial particles and/or the particle-laden liquid-liquid (L-L) interface. We find that bijel dynamics show two-step (fast/slow) decay behaviour, with the dynamics slowing as the system ages. The two-step decay is very similar to that observed in colloidal gels formed by diffusion-limited cluster aggregation (DLCA), with the initial (fast) decay due to thermally-activated modes of the gel network, and the later (slow) decay due to the relaxation of internal stresses induced by gel syneresis. For a bijel, the internal stresses could be due to syneresis, but could also be due to the jamming transition and/or the monogelation process and/or the forces acting on the L-L interface by the particle layer. In terms of the aging, if the system does not form a monogel, the correlation functions can be (almost) rescaled on to a master curve, indicating the property of universal aging. If the system does monogel, the functions cannot be superimposed, implicating the monogelation process as a potential cause for a different kind of aging in this system. Due to the interesting differences found when changing the size of the stabilizing particles in a bijel, in Chapter 6 we combine large and small particles (making `bimodal' bijels) and look for evidence of particle segregation by size, quantitatively estimate the ratio of particle uptake fractions and measure kinetics. Larger particles are found to adsorb to the interface in twice the quantity as smaller particles, and we find evidence to suggest the preference of larger particles for interfaces curved in only one direction, corroborating results from previous Chapters. Bimodal bijels take longer to jam than an equivalent monomodal (standard) bijel, which is backed up by simulations and highlights the increased ability of the bimodal particles to reorganise at the interface before arriving at the jammed metastable state. Finally, we also observe that the lengthscale of a bimodal bijel can heavily depend on the quench rate used during the preparation, suggesting that quench rate could be used (as well as particle size, volume fraction and contact angle) as a lengthscale control parameter. This thesis adds to the bijel literature, building on previous experimental studies and verifying/contradicting simulations. Particle size is shown to be a pivotal parameter for bijel formation in the W/L system, with particles of size r = 63 nm proving more versatile (markedly less sensitive to quench rate) than particles of size r ≈ 300 nm. However, even-smaller particles (of the same type) do not provide any additional advantage. We also show how the particle size can not only control bijel porosity (according to L ∝ r/ϕ as predicted by simulations) but can control bijel topology (smaller particles result in structures with greater hyperbolic character). By monitoring the bijel structure over time (topology and dynamics) we have shown that the bijel (in some cases) continues to age for at least c. 1 hr (topology) and in all cases c. 1 day (dynamics). For the first time experimentally, we have used a bimodal dispersion of particles to stabilize a W/L bijel and have uncovered a potentially useful new way to produce samples with different porosities from the same starting mixture, by changing the quench rate. The knowledge of the interplay between particle size and quench rate along with the effect on bijel topology will both assist in the scaling up of processes for industrial-level production and inform future strategies for tailoring the structure for specific applications. Future research should focus on several remaining open questions. The volume fraction of r = 63 nm particles in the W/L system should be increased towards 10% and sonication procedures improved to allow good redispersion to test the lower bound of L, which we expect to be around 1 μm. Also, a new W/L fabrication paradigm should be devised which uses sterically-stabilized particles, to continue the reduction of r towards the value used in simulations (5 nm) in order to test the fundamental physics of bijel formation, specifically what value of interfacial attachment energy is needed for long-term stability. Bijel dynamics can be further probed by using a technique which allows a variation in the probe lengthscale (e.g. differential dynamic microscopy, DDM), as well as developing a better theoretical model for (multiple) light scattering in a bijel system to arrive at the mechanisms responsible for the anomalous aging, and compare to the predictions of monogelation. Finally, higher magnification/resolution microscopy should be used to look for particle segregation on the liquid-liquid interface (as seen in simulations) and to identify in real-space the locations of the changes in Gaussian curvature over time as measured in Chapter 4.