Search

Combinatorial problems arising in puzzles, origami, and (meta)material design have rare sets of solutions, which define complex and sharply delineated boundaries in configuration space. These boundaries are difficult to capture with conventional statistical and numerical methods. Here we show that convolutional neural networks can learn to recognize these boundaries for combinatorial mechanical metamaterials, down to finest detail, despite using heavily undersampled training sets, and can successfully generalize. This suggests that the network infers the underlying combinatorial rules from the sparse training set, opening up new possibilities for complex design of (meta)materials.

We determine the kinetic phase diagram for nucleation and growth of crystal phases in a suspension of charged colloids. Exploiting the seeding approach in extensive simulations, we calculate nucleation barrier heights for face-centered cubic (fcc) and body-centered cubic (bcc) phases for varying screening lengths and supersaturations. We determine for the entire metastable fluid region the crystal polymorph with the lowest nucleation barrier, and find a regime close to the triple point where metastable bcc can form due to a lower nucleation barrier, even though fcc is the stable phase. For higher supersaturation, we find that the difference in barrier heights decreases and we observe a mix of hexagonal close-packed, fcc, and bcc structures in the growth of crystalline seeds as well as in spontaneously formed crystals. Our kinetic phase diagram rationalizes the different crystallization mechanisms observed in previous work.

Using computer simulations, we investigate the phase behavior of hard-core spherocylinders with a length-to-diameter ratio L/σ=5 and coated by a soft deformable corona of length λ/σ=1.35. When quasi-two-dimensional layers are formed in smectic and solid phases at low temperatures, the competition between the two intrinsic length scales of the parallel aligned particles leads to the stabilization of different in-plane lattices of nonconventional symmetry, including low-density hexagonal, square, and high-density hexagonal crystals, as well as an intriguing dodecagonal quasicrystal. Our Letter opens up the opportunity to control the assembly of anisotropic nanoparticles into structures with preengineered symmetry-dependent physical properties.

Using simulations, we study the diffusion of rodlike guest particles in a smectic environment of rodlike host particles. We find that the dynamics of guest rods across smectic layers changes from a fast nematiclike diffusion to a slow hopping-type dynamics via an intermediate switching regime by varying the length of the guest rods with respect to the smectic layer spacing. We determine the optimal rod length that yields the fastest and the slowest diffusion in a lamellar environment. We show that this behavior can be rationalized by a complex 1D effective periodic potential exhibiting two energy barriers, resulting in a varying preferred mean position of the guest particle in the smectic layer. The interplay of these two barriers controls the dynamics of the guest particles yielding a slow, an intermediate, and a fast diffusion regime depending on the particle length.

We perform an extensive computational study on the phase behavior of hard banana-shaped particles, and show that biaxial, twist-bend, and splay-bend nematic phases are metastable with respect to a smectic phase for a system of hard bent spherocylinders. However, if the smectic phase is destabilized-either by polydispersity in the particle length or by curvature in the particle shape-stable biaxial, twist-bend, and splay-bend nematic phases are obtained. This provides a unified and consistent picture on the subtle role of particle shape on the phase behavior of bent rods.

We use computer simulations to study the existence and stability of a biaxial nematic N_{b} phase in systems of hard polyhedral cuboids, triangular prisms, and rhombic platelets, characterized by a long (L), medium (M), and short (S) particle axis. For all three shape families, we find stable N_{b} states provided the shape is not only close to the so-called dual shape with M=sqrt[LS] but also sufficiently anisotropic with L/S>9,11,14,23 for rhombi, (two types of) triangular prisms, and cuboids, respectively, corresponding to anisotropies not considered before. Surprisingly, a direct isotropic-N_{b} transition does not occur in these systems due to a destabilization of N_{b} by a smectic (for cuboids and prisms) or a columnar (for platelets) phase at small L/S or by an intervening uniaxial nematic phase at large L/S. Our results are confirmed by a density functional theory provided the third virial coefficient is included and a continuous rather than a discrete (Zwanzig) set of particle orientations is taken into account.

We develop a lattice-based Monte Carlo simulation method for charged mixtures capable of treating dielectric heterogeneities. Using this method, we study oil-water mixtures containing an antagonistic salt, with hydrophilic cations and hydrophobic anions. Our simulations reveal several phases with a spatially modulated solvent composition, in which the ions partition between water-rich and water-poor regions according to their affinity. In addition to the recently observed lamellar phase, we find tubular and droplet phases, reminiscent of those found in block copolymers and surfactant systems. Interestingly, these structures stem from ion-mediated interactions, which allows for tuning of the phase behavior via the concentrations, the ionic properties, and the temperature.

Using computer simulations, we investigate the phase behavior of a system of particles interacting with a remarkably simple repulsive square-shoulder pair potential and report the formation of a novel (and stable) pyrochlorelike crystal phase. The lattice structure of the pyrochlorelike phase formed in our simulations possesses two inherent length scales corresponding to the inter- and intratetrahedral neighbors. We show that it can be used to fabricate a photonic crystal which displays complete photonic band gaps in both the direct and inverted dielectric structures.

From experimental studies, it is well known that colloidal particles suspended in a near-critical binary solvent exhibit interesting aggregation phenomena, often associated with colloidal phase transitions and assumed to be driven by long-ranged solvent-mediated (SM) interactions (critical Casimir forces), set by the (diverging) correlation length of the solvent. We present the first simulation and theoretical study of an explicit model of a ternary mixture that mimics this situation. Both the effective SM pair interactions and the full ternary phase diagram are determined for Brownian disks suspended in an explicit two-dimensional supercritical binary liquid mixture. Gas-liquid and fluid-solid transitions are observed in a region that extends well away from criticality of the solvent reservoir. We discuss to what extent an effective pair-potential description can account for the phase behavior we observe. Our study provides a fresh perspective on how proximity to the critical point of the solvent reservoir might influence colloidal self-assembly.

In order to investigate the coupling between the degrees of alignment of elongated particles in binary nematic dispersions, surfactant stabilized single-wall carbon nanotubes (CNTs) have been added to nematic suspensions of colloidal rodlike viruses in aqueous solution. We have independently measured the orientational order parameter of both components of the guest-host system by means of polarized Raman spectroscopy and by optical birefringence, respectively. Our system allows us therefore to probe the regime where the guest particles (CNTs) are shorter and thinner than the fd virus host particles. We show that the degree of order of the CNTs is systematically smaller than that of the fd virus particles for the whole nematic range. These measurements are in good agreement with predictions of an Onsager-type second-viral theory, which explicitly includes the flexibility of the virus particles, and the polydispersity of the CNTs.

We present a new numerical scheme to study systems of nonconvex, irregular, and punctured particles in an efficient manner. We employ this method to analyze regular packings of odd-shaped bodies, both from a nanoparticle and from a computational geometry perspective. Besides determining close-packed structures for 17 irregular shapes, we confirm several conjectures for the packings of a large set of 142 convex polyhedra and extend upon these. We also prove that we have obtained the densest packing for both rhombicuboctahedra and rhombic enneacontrahedra and we have improved upon the packing of enneagons and truncated tetrahedra.

Using simulations we identify three dynamic regimes in supersaturated isotropic fluid states of short hard rods: (i) for moderate supersaturations, we observe nucleation of multilayered crystalline clusters; (ii) at higher supersaturation, we find nucleation of small crystallites which arrange into long-lived locally favored structures that get kinetically arrested; and (iii) at even higher supersaturation, the dynamic arrest is due to the conventional cage-trapping glass transition. For longer rods we find that the formation of the (stable) smectic phase out of a supersaturated isotropic state is strongly suppressed by an isotropic-nematic spinodal instability that causes huge spinodal-like orientation fluctuations with nematic clusters diverging in size. Our results show that glassy dynamics and spinodal instabilities set kinetic limits to nucleation in highly supersaturated hard-rod fluids.

Using simulations of hard rods in smectic-A states, we find stringlike clusters of up to 10 interlayer rods exhibiting dynamic cooperativity. We also find non-gaussian diffusion and heterogeneous dynamics due to the equilibrium periodic smectic density profiles, which give rise to permanent barriers for layer-to-layer diffusion. This relaxation behavior is surprisingly similar to that of nonequilibrium supercooled liquids, although there the particles are trapped in transient (instead of permanent) cages.

We present an efficient and robust method based on Monte Carlo simulations for predicting crystal structures at finite temperature. We apply this method, which is surprisingly easy to implement, to a variety of systems, demonstrating its effectiveness for hard, attractive, and anisotropic interactions, binary mixtures, semi-long-range soft interactions, and truly long-range interactions where the truly long-range interactions are treated using Ewald sums. In the case of binary hard-sphere mixtures, star polymers, and binary Lennard-Jones mixtures, the crystal structures predicted by this algorithm are consistent with literature, providing confidence in the method. Finally, we predict new crystal structures for hard asymmetric dumbbell particles, bowl-like particles and hard oblate cylinders and present the phase diagram for the oblate cylinders based on full free energy calculations.

We study the isotropic-to-crystal transformation in a mixture of colloidal hard rods and nonabsorbing polymer using computer simulations. We determine the height of the nucleation barrier and find that the critical cluster consists of a single crystalline layer growing laterally for all polymer fugacities considered. At lower supersaturation, the free energy of a single hexagonally packed layer increases monotonically with size, while the nucleation barrier of a second crystalline layer is extremely high. Hence, the nucleation of multilayer crystals is never observed. Multilayer crystals form only in the spinodal decomposition regime, either where, in an intermediate stage, single crystalline membranes coalesce into multilayer clusters or where, at higher polymer fugacity, smaller clusters of rods stack on top of each other to form long filaments. Eventually, these transient structures evolve into a thermodynamically stable bulk crystal phase.

We study self-diffusion in complex fluids within dynamic density functional theory and explicitly account for the coupling to the fluctuating background. Applying the formalism to nematic and smectic liquid crystals, we find the temporary cages formed by neighboring particles to compete with permanent barriers in nonuniform systems, resulting in non-Gaussian diffusive motion that in different directions becomes correlated. Qualitative agreement with recent experiments demonstrates the importance of explicitly dealing with time-dependent self-consistent molecular fields.

We report a numerical study of the rate of crystal nucleation in a binary suspension of oppositely charged colloids. Two different crystal structures compete in the thermodynamic conditions under study. We find that the crystal phase that nucleates is metastable and, more surprisingly, its nucleation free-energy barrier is not the lowest one. This implies that, during nucleation, there is insufficient time for subcritical nuclei to relax to their lowest free-energy structure. Such behavior is in direct contradiction with the common assumption that the phase that crystallizes most readily is the one with the lowest free-energy barrier for nucleation. The phenomenon that we describe should be relevant for crystallization experiments where competing solid structures are not connected by an easy transformation.

Using confocal microscopy, we investigate the structure of binary mixtures of colloidal hard spheres with size ratio q=0.61. As a function of the packing fraction of the two particle species, we observe a marked change of the dominant wavelength in the pair-correlation function. This behavior is in excellent agreement with a recently predicted structural crossover in such mixtures. In addition, the repercussions of structural crossover on the real-space structure of a binary fluid are analyzed. We suggest a relation between crossover and the lateral extension of networks containing only equally-sized particles that are connected by nearest-neighbor bonds. This is supported by Monte Carlo simulations which are performed at different packing fractions and size ratios.

We study the kinetic pathways for the isotropic-to-nematic transition in a fluid of colloidal hard rods. In order to follow the formation of the nematic phase, we develop a new cluster criterion that distinguishes nematic clusters from the isotropic phase. Applying this criterion in Monte Carlo simulations, we find spinodal decomposition as well as nucleation and growth depending on the supersaturation. We determine the height of the nucleation barrier and we study the shape and structure of the cluster. More specifically, we find ellipsoidal nematic clusters with an aspect ratio of about 1.7 and a homogeneous nematic director field. Our results are consistent with theoretical predictions on the shape and director field of nematic tactoids. Classical nucleation theory gives reasonable predictions for the height of the nucleation barrier and the critical nucleus size.

Phase diagrams of hard and soft spheres with a fixed dipole moment are determined by calculating the Helmholtz free energy using simulations. The pair potential is given by a dipole-dipole interaction plus a hard-core and a repulsive Yukawa potential for soft spheres. Our system models colloids in an external electric or magnetic field, with hard spheres corresponding to uncharged and soft spheres to charged colloids. The phase diagram of dipolar hard spheres shows fluid, face-centered-cubic (fcc), hexagonal-close-packed (hcp), and body-centered-tetragonal (bct) phases. The phase diagram of dipolar soft spheres exhibits, in addition to the above mentioned phases, a body-centered-orthorhombic (bco) phase, and it agrees well with the experimental phase diagram [Nature (London) 421, 513 (2003)]. Our results show that bulk hcp, bct, and bco crystals can be realized experimentally by applying an external field.

Extensive simulations of a hard sphere fluid confined between two planar hard walls show the onset of crystalline layers at the walls at about 98.3% of bulk crystallization density rho(f) independent of the wall separations L(z), and is, hence, a single wall phenomenon. As the bulk density far from the wall rho(b) increases, the thickness of the crystalline film appears to increase logarithmically, with (rho(f)-rho(b)) indicating complete wetting by the hard sphere crystal of the wall-fluid interface. Increasing rho(b) further, we observe a jump in the adsorption which depends on L(z) and corresponds to capillary freezing. The formation of crystalline layers below bulk crystallization, the logarithmic growth of the crystalline film, its independence of L(z), and its clear distinction from capillary freezing lend strong evidence for complete wetting by the hard sphere crystal at the wall-fluid interface.

Density functional theory and computer simulation are used to investigate sedimentation equilibria of colloid-polymer mixtures within the Asakura-Oosawa-Vrij model of hard sphere colloids and ideal polymers. When the ratio of buoyant masses of the two species is comparable to the ratio of differences in density of the coexisting bulk (colloid) gas and liquid phases, a stable "floating liquid" phase is found, i.e., a thin layer of liquid sandwiched between upper and lower gas phases. The full phase diagram of the mixture under gravity shows coexistence of this floating liquid phase with a single gas phase or a phase involving liquid-gas equilibrium; the phase coexistence lines meet at a triple point. This scenario remains valid for general asymmetric binary mixtures undergoing bulk phase separation.

We develop an efficient simulation scheme to study a model suspension of equally sized colloidal hard spheres and nonadsorbing ideal polymer coils, both in bulk and adsorbed against a planar hard wall. The many-body character of the polymer-mediated effective interactions between the colloids yields a bulk phase diagram and adsorption phenomena that differ substantially from those found for pairwise simple fluids; e.g., we find an anomalously large bulk liquid regime and, far from the bulk triple point, three layering transitions in the partial wetting regime prior to a transition to complete wetting by colloidal liquid.