In‐situ TEM study of the formation of the smallest possible fullerenes on metal surfaces

Abstract

The nucleation and growth of carbon on catalytically active metals is one of the most important techniques to produce a large variety of graphenic nanomaterials. The most prominent species that have been grown on metals by vapor or solid phase deposition are graphene and carbon nanotubes. Small closed cages such as fullerenes have hardly been observed to nucleate catalytically on metal surfaces. The formation of fullerene molecules, in particular under realistic growth conditions is difficult to observe at high spatial resolution. The smallest possible fullerene is C 20 which is a dodecahedron composed of 12 pentagons. This molecule with exteme curvature is in a hybridization state between sp 2 and sp 3 . Until now, only few studies reported the synthesis of C 20 due to its instability and high reactivity. Isolated cages of C 20 haven't been observed to date by electron microscopy. Here we report the nucleation and growth of spherical carbon cages, some of them corresponding to the smallest possible fullerenes starting with approximately the size of C 20 , on metal surfaces [1]. The experiments were carried out in‐situ in a transmission electron microscope (TEM) by using a heating stage. The samples were prepared on few‐layer graphene suspended on standard Cu grids for electron microscopy. Different transition metal layers (Co, Fe, Ru) with thickness of 5 nm were deposited by cathodic sputtering onto the graphene layers. After an initial heating and cooling cycle of the samples, small carbon cages appeared on the graphene layers around the periphery of metallic nanoparticles. Fig.1 shows schematically how the experimental procedure was carried out. A series of examples for the observed structures is shown in Fig. 2. The contrast of these circular features closely resembles the appearance of C 60 in TEM images. However, the diameter of most of the observed cages ranges between 0.35 and 0.4 nm which is clearly smaller than the diameter of C 60 (0.7nm). No isolated cages were observed; the small cages always appeared as aggregates and in many cases as an ordered arrangement, in particular when the cages were encapsulated by a graphenic shell. The prerequisite for the nucleation of the cages was an uncovered metal surface. The cages persist after cooling to room temperature. In order to identify the elemental composition and the bonding states of the observed cages, electron energy‐loss spectra with a monochromated electron beam were taken at energy resolution of 0.2 eV. To relate the observed contrast in the TEM images to fullerene‐like clusters, image simulations were carried out by using the EMS (Electron Microscope Simulator) simulation program. Polymerized and unpolymerized C 20 were simulated. Fig. 3 shows the calculated appearance of the aggregate of three C 20 molecules on a monolayer of graphene. The observations are in accordance with the simulated images of polymerized C 20 molecules. The nucleation of the cages occurs by the dissolution of carbon in the metal at high temperature and the diffusion through the bulk, followed by the segregation on the surface upon cooling. Since the C 20 cages are less stable than larger fullerenes, their formation should be driven by kinetics under non‐equilibrium conditions. Due to their large curvature and inherent reactivity, the cages tend to polymerize.