| Graphene, a graphite monolayer presenting novel exciting properties, has attracted much attention recently in the scientific community as well as in the high-technology industry. In electronics, nanoribbons – narrow strips of graphene which happen to be semiconducting– could possibly allow further miniaturization of electronic devices such as transistors because of their atomic thickness. On the other hand, once making devices, the problem of metallic contacts, which can have critical impact at the nanoscopic scale, cannot be evaded. For example, metal induced gap states may short-circuit very short devices. With this in mind, the interaction of gold, palladium and titanium contacts with finite size graphene nanoribbons has been studied using ab initio density functional theory calculations.;This theoretical approach made it possible to study separately and then conjugate four important aspects of the metal-ribbon interaction: bonding, charge transfer, electrostatics and metal induced gap states. Another goal of this project was to study size effects related to the ribbons' dimensions and to estimate the minimal channel length necessary for a device to operate as expected without the unwanted effect of induced gap states. Aside from the high precision achieved, these calculations stand out from earlier studies because they take into account finite size effects which often prevail in small ribbons.;Using this model for the metal-nanoribbon junction, it was shown that, as for two-dimensional graphene, the bonding between a ribbon and a metal can be of two types depending on the electronic configuration of the metal. In the first case, physisorption, weak bonding resulting in a large separation distance between ribbon and electrode, is illustrated by the gold contact. On the other hand, titanium, because of its high density of states at the Fermi level, binds more strongly with graphene nanoribbons. This chemisorption is characterized by strong hybridization between the metal and the ribbon's orbitals. This leads to the apparition of intense evanescent gap states in the ribbon. As for palladium, it represents an intermediate case showing some but not as much hybridization.;For all three metals, right under the contact, we observe a net decrease of electron density in the ribbon in favour of the contact. Effectively, any kind of bonding is generally associated with charge transfer necessary to balance the work function difference. As expected, a metal with a large work function such as gold and palladium tends to attract electrons. However, this behaviour is surprising from titane whose work function is much smaller. This can only be explained by considering the large hybridization of the ribbon's orbitals. Moreover, the charge transfer is not restricted to the area under the contact as an evanescent charge can also be observed in the channel. This charge results from the occupation of metal induced gap states and its sign depends on the position of the charge neutrality point. The calculations presented don't take into account the effect of temperature. Nevertheless, the position of the Fermi level relative to the ribbon's bands suggests that at non-zero temperature, gold and palladium would p dope the graphene nanoribbons while titanium would have the opposite effect.;The charge redistribution in the junction generates a molecular-sized dipole which is responsible for the slowly varying three-dimensional potential profile, a particularity of low dimensional systems. A potential barrier thus appears at the interface and controls band curvature in the device, but it can also constitute an obstacle to efficient charge injection. The intimate relation between the titanium contact and the ribbons seems to maximize screening and reduce considerably the height of the barrier, which could turn out to be beneficial.;Finally, metal induced gap states slowly decaying in the channel have been observed directly in graphene nanoribbons for the first time. Their characteristic evanescent shape is clearly distinguishable. The extent of the gap states was measured to be on the order of a nanometer. It depends mostly on the ribbon's bandgap which varies both with its length and width. The bandgap, which can easily be measured experimentally, could thus be a useful tool to predict the penetration of gap states, a must-have information in order to limit the undesired effect of gap states since it sets for a device the minimal channel length (∼1-2 nm).;In summary, these theoretical results could be used as a guideline in the conception of electronic devices made of graphene nanoribbons in which doping and potential profile need to be optimized while making sure that metal induced gap states do not compromise normal operation. |
