Colloidal interactions get patchy and directional

Biology is the master of self-assembly, and proteins are the proof. Relying on interactions specified by the sequence of amino acids, proteins are able to precisely assemble into a multitude of intricate configurations, each with a unique functionality dictated by its three-dimensional structure. At an abstract level, one can think about the sophistication of protein folding as arising from three main principles governing the interactions between assembling components: valency, directionality, and specificity. Valency, or patchiness, refers to the number and spatial distribution of interacting domains (Fig. 1A). When attractive patches are present, the number of ways two interacting components can be positioned relative to one another will be limited. Directionality refers to the tendency for building blocks to interact with an orientational dependence (Fig. 1B). For a highly directional interaction, the number of ways two components can be rotated relative to one another will be limited. Finally, specificity refers to the presence of chemical complementarity that facilitates a selective attachment between building blocks (Fig. 1C). For a component with a highly specific attractive interaction, the number of unique complementary partners that can be recognized will be limited. Each of these traits can be thought of as being orthogonal axes representing the types of interactions that can occur between assembling components. As each type of interaction becomes more well defined, the information content of the building block increases, and the number of plausible final states is decreased. Thus, the final assembled structure possesses a greater degree of programmed structural sophistication (1). Although biological systems have mastered these tools, it has long been the goal of chemists and physicists to instill artificial systems with similarly complex interactions so as to self-assemble advanced materials from the bottom up with limited human intervention (2). To date, researchers have been able to strongly instantiate one of these traits into a single structure, but developing synthetic methods that allow for the simultaneous and well-defined expression of a combination of these features has been a considerable challenge. In their work presented in PNAS, Diaz A. et al. (3) describe a method for synthesizing micrometer-sized colloidal particles that exhibit both valency and directionality, representing a significant advance toward realizing complex self-assembling materials (Fig. 1D).

Fig. 1.

When the three major types of colloidal interaction become more well-defined, building blocks are imbued with more information and can form more sophisticated structures. Valency (A), directionality (B), and specificity (C) describe the positional, orientational, and chemical recognition degrees of freedom of an interaction, respectively. (D) Colloids synthesized by Diaz A. et al. (3) present faceted DNA patches with programmed symmetry, allowing for the creation of advanced self-assembled materials.

Nanometer- and micrometer-scale colloidal particles have drawn the attention of numerous researchers due to the extraordinary size-dependent properties they possess and their ability to assemble into larger superstructures with interesting collective properties. However, for most of the decades-long history of this field, hard, spherical particles have lacked the features described above, interacting isotropically and packing only into dense, crystalline arrangements in equilibrium (4). While important for answering fundamental questions in condensed matter physics and colloid chemistry, the lack of structural diversity available from such systems has limited the manifestation of certain properties. For example, photonic band gap materials can be assembled from colloidal particles but show the most promising properties only when arranged in nonclose-packed, open frameworks (5). The introduction of anisotropic building blocks has allowed for more advanced structures that begin to manifest interactions that are valent and directional. For example, nematic liquid crystals exhibit orientational order (i.e., have directional interactions) but lack positional order, and plastic crystals exhibit positional order (i.e., have valent interactions) but lack orientational order. Polyhedral nanocrystals uniformly coated in ligands can show both patchy and directional interactions, but methods to produce highly faceted particles of this sort at the micrometer scale represent a synthetic challenge. Importantly, valency or directionality may in some cases arise dynamically as a result of the assembly process (e.g., through the bundling of surface ligands that break the symmetry of the underly particle interaction) (6, 7). Alternatively, DNA origami frameworks can be synthesized that host sub–20-nm particles and allow for well-defined patch-based architectures with open structures (8). However, it remains unclear if and how these methods could be scaled up to control the assembly of the micrometer-sized particles that would be necessary to realize crystals with photonic properties at visible wavelengths. To create more complex architectures in a rationally designed manner, particles must be synthesized such that their interactions are more well-defined in terms of their valency, directionality, and/or specificity. Broadly applicable methods that achieve this goal remain limited (9–14).

Synthetic routes to create particles with valency rely on a reduction in the interaction symmetry, usually created in the form of distinct chemical patches (Fig. 1A). In one approach, small particles are assembled into clusters of between two and nine components, which then undergo some form of chemical modification (e.g., polymerization) that permanently links them together (15, 16). The final result is one contiguous, solid structure where each of linked particles acts as a distinct patch that directs interparticle interactions. A second method preassembles particles onto a substrate or into a lattice and utilizes the contact points between adjacent objects to guide the acquisition of ligands that spatially defines a patch with a distinct surface chemistry (16, 17). Regardless of the synthetic route, prior work on creating particles with valency has faced a similar set of challenges. First, it is often difficult to control the number and distribution of patches on the particle. In most cases, the products of any given synthetic approach contain a mixture of particles with differing valency that must undergo a subsequent purification step to be used. Additionally, most procedures to date produce patches with a single symmetry that cannot be easily modified without major changes to the protocol (16).

Methods to imbue particles with directionality require a reduction in the rotational degrees of freedom of interacting structures. This is commonly achieved in a manner that is analogous to the well-known biological principle of “lock and key” binding (18). Particles must have some degree of shape complementarity such that certain orientations are favorable relative to others. A simple but effective manifestation of this effect is the interaction between two flat facets of attractive particles, being more complementary in shape than the interaction of two curved surfaces (10). More sophisticated versions of the lock and key principle rely on a greater degree of shape complementarity between interacting particles (Fig. 1B). For example, a pair of interacting nanostructures presenting complementary concave and convex features will interlock when assembling, restricting several orthogonal degrees of rotational freedom and forming a highly directional colloidal interaction (14).

In their work presented in PNAS, Diaz A. et al. describe a method for synthesizing micrometer-sized colloidal particles that exhibit both valency and directionality, representing a significant advance toward realizing complex self-assembling materials.

When interactions are driven by surface ligand-based chemical bonding, a polymer brush may form on a flat facet in which molecules adopt more elongated conformations to minimize steric repulsion, thereby providing a directional attraction perpendicular to the surface. Because the polymer brush height is more elongated on flat vs. curved surfaces, this effect is more pronounced for interacting particles that are faceted compared with those that are spherical (19). Prior work has shown that particles with strong directional interactions can assemble into novel structures (11–13, 20), but methods to synthesize distinct patches that also express directional attraction have yet to be achieved (9).

The innovation presented in the work by Diaz A. et al. (3) lies in a clever synthetic approach that allows for colloidal particles to be imbued with both valency and directionality in a well-defined and controlled manner. In their procedure, polymer droplets created via an oil-in-water emulsion are functionalized with DNA strands appended at one end with a hydrophobic group that allows them to be laterally mobile across the surface of the droplets. When two sets of droplets functionalized with complementary DNA are mixed together, an ordered superlattice forms as a result of the hybridization of numerous complementary strands. Although two particles initially meet at a single tangent point, the mobile DNA accumulates at the points of intersection in order to maximize the number of DNA “bonds” between particles. Thus, DNA “patches” are formed at symmetric points defined by the coordination environment in the assembled superlattice. Furthermore, if the DNA interactions are sufficiently strong, the intersection of neighboring polymer droplets deforms and flattens to create additional DNA connections, eventually reaching an equilibrium that balances favorable DNA hybridization against the unfavorable costs of deformation and reduced entropy of recruited mobile ligands. Photopolymerization renders the now-faceted structure of the droplet preserved as a solid particle with immobile DNA ligands trapped and concentrated at the interparticle junctions (Fig. 1D). Elevated temperatures allow complementary particles to be dehybridized and recovered as colloids with both valency and directionality.

Importantly, the methods developed by Diaz A. et al. (3) outline a vast design space for particle synthesis, allowing for tailorability in both valency and directionality. By increasing the strength of the interparticle DNA hybridization and/or by changing the size of complementary droplets, the size of the patches could be controlled. In addition, the number and position of patches could be tailored by changing the symmetry of the lattice into which the particles assembled. Lattices formed from a mixture of droplets with different sizes are sensitive to the relative stoichiometry and size ratio of the two components being assembled. Consequently, judicious choice of these parameters allows for the design of n-valent particles with patches at well-defined symmetric locations. Alternatively, droplets could be assembled into two-dimensional lattices one unit cell in thickness, allowing for particles to be isolated bearing patches on only one hemisphere. These lower-symmetry particles make the design of more sophisticated, nonclose-packed structures more feasible. Finally, a model that accounts for the energetic balance between patch formation and facet development showed excellent agreement with the experimental results, facilitating a rational synthetic strategy for colloids with a diversity of multivalent and directional interparticle attractions.

This synthetic approach highlights the possibility of future methods based on deformable colloids that will allow for even greater degrees of structural complexity. However, the third key principle governing information-rich self-assembly, specificity, remains a challenge to incorporate into any synthetic methodology (Fig. 1C). The methods of Diaz A. et al. (3) are promising because of the use of DNA strands at each patch. Because oligonucleotides exhibit sequence-specific hybridization interactions as a consequence of their Watson–Crick base pairing, numerous orthogonal attractive pairs can form in a single homogeneous mixture. Therefore, placing oligonucleotides with distinct sequences at prescribed patch locations would allow for the simultaneous expression of valency, directionality, and specificity. Such particles could form the basis for highly complex but rationally designed structures that form spontaneously under mild conditions. Additionally, while micrometer-sized particles have interesting optical and rheological properties that make them useful, they lack the enormous diversity of size-dependent properties observed in inorganic nanoscale systems. Methods to miniaturize the strategies outlined here to be applicable to nanometer-scale particles impart new challenges but promise to vastly expand the toolbox of programmable materials. Regardless, the work of Diaz A. et al. (3) puts the field one step closer to achieving the same structural sophistication of proteins but with artificial colloidal materials, providing perfect control over self-assembly and the arbitrary manipulation of matter at the smallest length scales.