Low-temperature reactions: Tunnelling in space.

active sites for the growth of the cylindrical branches. These nanocrystals were grown by adding a molecularly dissolved PFDMS solution into a poor solvent for PFDMS, and then aging the resulting precipitate. As the surfaces of the nanocrystals need to be available as seeds for growing cylindrical micelles, surfactants could not be used to regulate the formation and size of the nanocrystals. This was done instead by optimizing the growth time after the addition of the PFDMS solution into the poor solvent. Once the nanocrystals were prepared, PFDMS-bearing diblock copolymers of an appropriate composition were added to the solution, and crystallization-driven epitaxial growth of cylindrical micelles occurred gradually from the surfaces of the PFDMS nanocrystals (Fig. 1a → b). The length of the micellar cylinders was tuned by arresting the epitaxial growth at specific times, and also by adjusting the mass ratio between the added copolymer (PFDMSb-PI with corona-forming PI blocks, shown in blue in Fig. 1) and the PFDMS homopolymer core. More impressively, it was also possible to sequentially add a second or third copolymer, such as PFDMS-b-PMVS (with PMVS denoting poly(methylvinyl silane) and shown in red in Fig. 1), after the initial diblock copolymer (PFDMS-b-PI) had been consumed. This yielded cylindrical block co-micelles as individual radiating arms (Fig. 1b → c). Because the nucleated terminus of each cylinder was blocked, these micellar cylinders only grew outwards from the seed crystal. Furthermore, it was also possible to detach these arms from the nanocrystal core and obtain them as separate entities: this was achieved by crosslinking the coronas of the block co-micellar cylinders and subsequently dissolving the nanocrystalline PFDMS core (Fig. 1c → d). This yielded individual cylindrical block co-micelles of well-defined lengths, which were not centrosymmetric as their composition was determined by the earlier steps. This work demonstrates the potential of epitaxial growth, or ‘living’ assembly, for creating unprecedented micellar architectures, and points to new research opportunities and breakthroughs. For example, the individual cylinders produced may further self-assemble in blockselective solvents to yield hierarchical micrometre-sized micelles in which molecular structure and ordering are controlled at both the nanometre and micrometre scales. Also, careful control over the size of the initial PFDMS seed particles may have intriguing implications. Increasing the size of the particles to the micrometre scale should allow the production of particles that closely resemble those found on lotus leaves. One can further imagine replacing some of the coronal blocks with lowsurface-tension fluorinated polymers6. The resultant fluorinated arm-bearing microparticles would presumably be highly oiland water-repellent, and may provide promising candidates for creating selfcleaning surfaces. ❐