Computation of shear viscosity of colloidal suspensions by SRD-MD.

The behaviour of sheared colloidal suspensions with full hydrodynamic interactions (HIs) is numerically studied. To this end, we use the hybrid stochastic rotation dynamics-molecular dynamics (SRD-MD) method. The shear viscosity of colloidal suspensions is computed for different volume fractions, both for dilute and concentrated cases. We verify that HIs help in the collisions and the streaming of colloidal particles, thereby increasing the overall shear viscosity of the suspension. Our results show a good agreement with known experimental, theoretical, and numerical studies. This work demonstrates the ability of SRD-MD to successfully simulate transport coefficients that require correct modelling of HIs.

Study of the B1-B2 transition in colloidal clusters.

The possible mechanisms for the B1 (NaCl-type) to B2 (CsCl-type) transition in crystalline colloidal clusters of equally sized particles are studied by means of two computational techniques: metadynamics and nudged elastic band calculations. The system is modelled by a screened Coulomb potential. Different interaction ranges are considered. The transition from a perfect NaCl cubic cluster to a full CsCl cluster is forced by metadynamics, revealing a transition path with intermediate metastable configurations in which planes are shifted one by one. The presence of metastable configurations in the transition path, corresponding to a certain number of NaCl planes turned into CsCl, has clear analogies with the known Hyde and O'Keeffe mechanism for ionic crystals, with some important differences due to finite-size effects. These comprise the fact that the transition starts by shifting a surface plane by means of a row-by-row mechanism that has no analog in bulk crystals. The energy barriers between the local minima in the transition path are calculated, showing that the barriers strongly depend on the screening length, in such a way that the B1 metastable phase can have very long lifetimes when the interaction is sufficiently long-ranged.

Aggregation in colloidal suspensions: evaluation of the role of hydrodynamic interactions by means of numerical simulations.

Numerical simulations constitute a precious tool for understanding the role of key parameters influencing the colloidal arrangement in suspensions, which is crucial for many applications. The present paper investigates numerically the role of hydrodynamic interactions on the aggregation processes in colloidal suspensions. Three simulation techniques are used: Brownian dynamics without hydrodynamic interactions, Brownian dynamics including some of the hydrodynamic interactions, using the Yamakawa-Rotne-Prager tensor, and stochastic rotation dynamics coupled with molecular dynamics. A system of monodisperse colloids strongly interacting through a generalized Lennard-Jones potential is studied for a colloid volume fraction ranging from 2.5 to 20%. Interestingly, effects of the hydrodynamic interactions are shown in the details of the aggregation processes. It is observed that the hydrodynamic interactions slow down the aggregation kinetics in the initial nucleation stage, while they speed up the next cluster coalescence stage. It is shown that the latter is due to an enhanced cluster diffusion in the simulations including hydrodynamic interactions. The higher the colloid volume fraction, the more pronounced the effects on the aggregation kinetics. It is also observed that hydrodynamic interactions slow down the reorganization kinetics. It turns out that the Brownian dynamics technique using the Yamakawa-Rotne-Prager tensor tends to overestimate the effects on cluster diffusion and cluster reorganization, even if it can be a method of choice for very dilute suspensions.

Kinetically driven ordered phase formation in binary colloidal crystals.

The aggregation of binary colloids of the same size and balanced charges is studied by Brownian dynamics simulations for dilute suspensions. It is shown that, under appropriate conditions, the formation of colloidal crystals is dominated by kinetic effects leading to the growth of well-ordered crystallites of the sodium-chloride (NaCl) bulk phase. These crystallites form with very high probability even when the cesium-chloride (CsCl) phase is more stable thermodynamically. Global optimization searches show that this result is not related to the most favorable structures of small clusters, which are either amorphous or of the CsCl structure. The formation of the NaCl phase is related to the specific kinetics of the crystallization process, which takes place by a two-step mechanism. In this mechanism, dense fluid aggregates form at first and then crystallization follows. It is shown that the type of short-range order in these dense fluid aggregates determines which phase is finally formed in the crystallites. The role of hydrodynamic effects in the aggregation process is analyzed by stochastic rotation dynamics - molecular dynamics simulations, and we find that these effects do not play a major role in the formation of the crystallites.

Tracer diffusion in colloidal suspensions under dilute and crowded conditions with hydrodynamic interactions.

We consider tracer diffusion in colloidal suspensions under solid loading conditions, where hydrodynamic interactions play an important role. To this end, we carry out computer simulations based on the hybrid stochastic rotation dynamics-molecular dynamics (SRD-MD) technique. Many details of the simulation method are discussed in detail. In particular, our choices for the SRD-MD parameters and for the different scales are adapted to simulating colloidal suspensions under realistic conditions. Our simulation data are compared with published theoretical, experimental and numerical results and compared to Brownian dynamics simulation data. We demonstrate that our SRD-MD simulations reproduce many features of the hydrodynamics in colloidal fluids under finite loading. In particular, finite-size effects and the diffusive behavior of colloids for a range of volume fractions of the suspension show that hydrodynamic interactions are correctly included within the SRD-MD technique.

Aggregation kinetics and gel formation in modestly concentrated suspensions of oppositely charged model ceramic colloids: a numerical study.

Aggregation kinetics and gel formation in aqueous suspensions that undergo heteroaggregation are studied by means of Brownian dynamics simulations. The simulated system, described in a previous paper [M. A. Piechowiak, A. Videcoq, F. Rossignol, C. Pagnoux, C. Carrion, M. Cerbelaud, R. Ferrando, Langmuir, 2010, 26(15), 12540-12547.], is constituted of two kinds of synthesized, almost equally sized colloids: silica particles that are negatively charged and alumina-coated silica particles that are positively charged. The interactions between colloids are modeled by the DLVO potential. Several compositions are analyzed, from silica-rich to alumina-rich cases. The particle volume fraction φ is varied in the range 6-12%. The study of the aggregation kinetics allows us to clarify the effect of those variations on the clustering process. Gelation is analyzed by detection of spanning clusters in each x-, y-, z-direction of the cubic simulation box. Percolating networks start to be observed from φ = 7%, a low value of the volume fraction close to the solid volume fraction experimentally measured in sediments of those suspensions.

Oppositely charged model ceramic colloids: numerical predictions and experimental observations by confocal laser scanning microscopy.

Fluorescent silica and alumina-like spherical particles with almost equal sizes are synthesized. Dilute aqueous suspensions are prepared with various ratios of those colloidal particles that exhibit opposite surface charges. These suspensions undergo heteroaggregation for a wide range of compositions. The structure of the formed aggregates is analyzed by means of confocal microscopy. The experimental results are compared to those of Brownian dynamics simulations in which the interactions between colloids are modeled by the DLVO potential. Good agreement between experiments and simulations is obtained.

Simulations of heteroaggregation in a suspension of alumina and silica particles: effect of dilution.

The influence of dilution on the aggregation process of suspensions composed of two kinds of oxide particles (alumina positively charged particles d(1)=400 nm and silica negatively charged particles d(2)=250 nm) has been studied by computer simulations. Two kinds of simulations have been performed: Brownian dynamics simulations to study the aggregation process and its kinetics and global minimization searches to find the most stable configurations of aggregates. We show that the rate of dilution has a strong influence on the structure and on the shape of aggregates in Brownian dynamics simulations. By confronting these aggregates with the stable aggregates found by global minimization, we demonstrate that they are metastable and their shape is explained by the competition between the kinetics of aggregate coalescence and the kinetics of aggregate reorganization into more stable configurations.

Simulation of the heteroagglomeration between highly size-asymmetric ceramic particles.

Aggregation phenomena of dilute suspensions composed of two kinds of oxide particles (alumina d(1)=400 nm and silica d(2)=25 nm) have been studied by computer simulations. These particles are oppositely charged and so are prone to heteroagglomerate. The interaction between particles has been modeled by the DLVO potential. Two kinds of simulations have been performed: Brownian dynamics simulations to study the aggregation kinetics and global minimization searches that permit the examination of the most stable configurations of agglomerates. We demonstrate that aggregation should occur also for quite large fractions of added silica (even when 200 silica particles are adsorbed on each alumina particle) and that aggregates are likely to present chainlike shapes. Both findings are in agreement with experiments.

Heteroaggregation between Al2O3 submicrometer particles and SiO2 nanoparticles: experiment and simulation.

The aggregation process of a two-component dilute system (3 vol %), made of alumina submicrometer particles and silica nanoparticles, is studied by Brownian dynamics simulations. Alumina and silica particles have very different sizes (diameters of 400 and 25 nm, respectively). The particle-particle interaction potential is of the DLVO form. The parameters of the potential are extracted from the experiments. The simulations show that the experimentally observed aggregation phenomena between alumina particles are due to the silica-alumina attraction that induces an effective driving force for alumina-alumina aggregation. The experimental data for silica adsorption on alumina are very well reproduced.

Agglomeration of alumina submicronparticles by silica nanoparticles: application to processing spheres by colloidal route.

In aqueous media, heterocoagulation between submicronic alumina (400 nm) and nanometric silica (25 nm) leads to the adsorption of silica on the alumina surface. By controlling the coverage rate of alumina particles, this adsorption destabilizes the suspension that leads to a very porous network of agglomerated particles. This work shows that the structure is all the more open as the density of charge carried by the two oxides is high and the ionic strength in the suspension low. From such a flocculated suspension, a new colloidal process to fabricate ceramic spheres is proposed which is based on a size increase of agglomerates. Under a controlled rotation of the vessel, electrostatic attraction between the surface charges of opposite polarity induces a size increase of agglomerates until the formation of spheres occurs. It has been shown that the mechanism of growth is poisoned by species adsorbed such as ions. Nevertheless, this new process proves very promising because it leads to a narrow size distribution of spheres by colloidal way, which can be subsequently consolidated by sintering, with a smooth surface.

Scaling and universality of self-organized patterns on unstable vicinal surfaces.

We propose a unified treatment of the step bunching instability during epitaxial growth. The scaling properties of the self-organized surface pattern are shown to depend on a single parameter, the leading power in the expansion of the biased diffusion current in powers of the local surface slope. We demonstrate the existence of universality classes for the self-organized patterning appearing in models and experiments.