Contact Dependence of Carrier Injection in Carbon Nanotubes: An Ab Initio Study

We combine ab initio density functional theory with transport calculations to provide a microscopic basis for distinguishing between good and poor metal contacts to nanotubes. Comparing Ti and Pd as examples of different contact metals, we trace back the observed superiority of Pd to the nature of the metal-nanotube hybridization. Based on large scale Landauer transport calculations, we suggest that the optimum metal-nanotube contact combines a weak hybridization with a large contact length between the metal and the nanotube.

Figure 1

Charge density redistribution $\Delta \rho (\mathbf{r})={\rho }_{\mathrm{Me}/\mathrm{C}}(\mathbf{r})-{\rho }_{\mathrm{Me}}(\mathbf{r})-{\rho }_{\mathrm{C}}(\mathbf{r})$ in (a) Pd and (b) Ti monolayers interacting with a graphene layer, indicating regions of charge depletion and excess with respect to the superposition of isolated layers. (c) Schematic double-layer geometry in top view, with the cutting plane used in (a) and (b) indicated by the dash-dotted line.

Figure 2

(a) Electronic band structure $E(k)$ of a Pd monolayer interacting with a graphene layer. (b) Details of $E(k)$ in (a) near the Fermi level, defined as $E_{\mathrm{F}}=0$.

Figure 3

(a) Schematic geometry of the $(6,6)$ nanotube in contact with metal leads, used in the calculation of the contact reflection coefficient $\rho$. (b) Contact reflection coefficient $\rho$ as a function of the nanotube-metal coupling $\Delta$ and the contact length $L_{\mathrm{c}}$. (c) Cuts through the contour plot (b) at selected values of $\Delta$, showing $\rho$ as a function of $L_{\mathrm{c}}$. The effective contact length $L_{\mathrm{c}}^{\mathrm{eff}}$ is emphasized by a heavy solid line in (b) and by data points in (c).