Nickel Cluster Growth on Defect Sites of Graphene: A Computational Study

Abstract

The first extraction of graphene in 2004 led to a wide range of experimental and theoretical studies aimed at better understanding and exploiting the unique properties of this novel two-dimensional material. Among the many potential applications, which have been suggested, are uses of graphene as a substrate in high-performance catalysis and as a component in circuit-board technology. In particular, graphene s high surface area and conductivity have motivated proposals to use it as a substrate for growing and/or anchoring metal nanoparticles in high-performance catalysts and other electrochemical devices. However, the activity of such carbon-supported metal catalysts is strongly dependent on the dispersion and stability of the metal clusters on the support (i.e. the ability of the substrate to stabilize metal clusters of various sizes on its surface). Thus, vacancy defects are expected to play a vital role in making graphene suitable for these applications by supplying highly active binding sites for adsorbing and stabilizing metal clusters. Indeed, finite populations of single and double vacancy defects are thermodynamically stable in graphene, and have been studied extensively. Density functional theory (DFT) calculations revealed that vacancy defects resulting from the removal of up to five C atoms reconstruct to form non-hexagonal rings (models are shown in the Supporting Information: Figures S1.b–f). Even larger holes have been observed in electron microscopy experiments. Defects may also play a critical role in using graphene components for circuit fabrication. For example, taking advantage of the Dirac fermions in graphene requires opening up its band gap to convert it from a conductor into a semiconductor. This conversion can be achieved by doping graphene with either B or N atoms; however, another possibility for accomplishing this could be the adsorption of small metal clusters on the surface. Because the adsorption of such clusters can be used to tune additional magnetic and transport properties of the substrate, it might also provide a technique for controlling an additional set of electromagnetic properties. The catalytic nature of Ni is well established, and Ni nanoparticles are commonly used to catalyze the synthesis of carbon nanostructures. Owing to the strong affinity between Ni and C, the incorporation of Ni atoms into carbon nanostructures, grown using Ni catalysts, has been observed. Ushiro et al. reported that X-ray adsorption measurements detect Ni impurities in carbon nanostructures following nickel-catalyzed synthesis, which even treatment with acid is not able to remove. Moreover, Banhart et al. identified Ni impurities wrapped in onion-like graphenic particles by using electron microscopy. The work of Rinaldi et al. is even more supportive. Combining results from DFT calculations and high-resolution transmission electron microcopy measurements (HR-TEM) utilizing several in situ characterization techniques, they concluded that Ni atoms form very stable Ni– C compounds during nickel-catalyzed carbon nanotube (CNT) growth, which are incorporated into the final products. They also found unexpectedly strong adsorption of the Ni clusters on the CNT supports. However, despite the potential advantages of using Ni nanoparticles adsorbed on graphene, their catalytic and electromagnetic properties (with the exception of single and two Ni atoms adsorbates) remain mostly unexplored. Based on these findings, it would be expected that just as Ni nanoparticles might be used to tailor critical properties of defective graphene sheets, a graphene substrate might be used to modify the catalytic properties of nickel nanoparticles as well. To elucidate this potential interplay we employ DFT to study the adsorption of Nin nanoclusters on defective graphene (details in the Supporting Information). As substrate models we select graphene sheets with vacancy defects, resulting from the removal of x atoms (with x 5; see Figure S1 in the Supporting Information). To model the adsorbed Ni nanoparticles, we successively grew Nin clusters with n 10 and focused on the lowest energy adsorption configuration of each Nin cluster on each of these six graphene substrates (with and without vacancy defects). The binding energies (referenced against single Ni atoms and the graphene substrate) for the lowest energy configuration are summarized in Figure 1. The binding energies can be explained by three types of bond contributions. The first type of binding is between Ni atoms. As the cluster size increases the ratio of bulk to surface atoms increases so that the binding energy will asymptotically [*] Dr. W. Gao, Dr. J. E. Mueller, Dr. J. Anton, Prof. Dr. T. Jacob Institut f r Elektrochemie, Universit t Ulm 89081 Ulm (Germany) E-mail: timo.jacob@uni-ulm.de