The influence of geometry on the local and global packing of particles is important to many fundamental and applied research themes, such as the structure and stability of liquids, crystals and glasses. Here we show by experiments and simulations that a binary mixture of hard-sphere-like nanoparticles crystallizing into a MgZn₂ Laves phase in bulk spontaneously forms icosahedral clusters in slowly drying droplets. Using advanced electron tomography, we are able to obtain the real-space coordinates of all the spheres in the icosahedral clusters of up to about 10,000 particles. The local structure of 70–80% of the particles became similar to that of the MgCu₂ Laves phase. These observations are important for photonic applications. In addition, we observed in simulations that the icosahedral clusters nucleated away from the spherical boundary, which is distinctly different from that of the single species clusters. Our findings open the way for particle-level studies of nucleation and growth of icosahedral clusters, and of binary crystallization.

Assembling binary mixtures of nanoparticles into crystals, gives rise to collective properties depending on the crystal structure and the individual properties of both species. However, quantitative 3D real-space analysis of binary colloidal crystals with a thickness of more than 10 layers of particles has rarely been performed. Here we demonstrate that an excess of one species in the binary nanoparticle mixture suppresses the formation of icosahedral order in the self-assembly in droplets, allowing the study of bulk-like binary crystal structures with a spherical morphology also called supraparticles. As example of the approach, we show single-particle level analysis of over 50 layers of Laves phase binary crystals of hard-sphere-like nanoparticles using electron tomography. We observe a crystalline lattice composed of a random mixture of the Laves phases. The number ratio of the binary species in the crystal lattice matches that of a perfect Laves crystal. Our methodology can be applied to study the structure of a broad range of binary crystals, giving insights into the structure formation mechanisms and structure-property relations of nanomaterials.