Electrical Contacts

Making electrical contacts to carbon nanotubes is in principle straightforward. A small amount of raw SWNT material, which usually consists of highly entangled ropes, is first ultrasonically dispersed in some organic solvent, typically dichloroethane, and then spun onto an oxidized silicon wafer on which a large array of electrodes have been fabricated with conventional electron-beam lithography, metal evaporation and lift-off [4]. An Atomic Force Microscope (AFM) is then used to determine whether there is an individual SWNT or a small rope (by apparent height) bridging two or more electrodes. Alternatively, the electrodes are aligned and fabricated on top of a nano-tube that has been identified by AFM using predefined alignment markers on the substrate [5]. However, this latter procedure often results in nano-tubes being cut between the electrodes due to electron beam damage during the lithography process. Figure 1 is a schematic of a typical three-terminal device geometry for transport measurement. The AFM image shows an individual SWNT molecule lying across two metal electrodes, which are used as source and drain, respectively. A variety of metals have been used for electrical leads including gold and platinum. A typical electrode width is on the order of 100 nm and the source-drain spacing varies between ^100 nm and |xm. In most experiments, another nearby electrode or a doped silicon substrate underneath the SiO2 is used as a gate to electrostatically modulate the carrier density of the nanotube under study.

Room temperature transport characteristics fall into two distinct types. The first type of nanotubes shows no or weak gate voltage dependence of the linear-response conductance. These nanotubes are identified as the metallic type. A strong gate dependence is observed for the second type of devices (as we will show below), which indicates that these tubes are semiconducting.

For metallic nanotubes, the measured two-terminal resistance is often dominated by the contact resistance between the nanotubes and the metal electrodes. In early transport measurements, the nanotubes typically formed a tunnel barrier of high resistance of ~ 1 MO with the electrodes. However,



Fig. 1. Typical device geometry for electrical transport measurement

metallic nanotubes have an ideal intrinsic two-terminal resistance of only h/4e2 or 6.5 k^ since there are only two propagating subbands crossing at the Fermi energy. Thus in these measurements, the bias voltage dropped almost entirely across the contacts, and tunneling phenomena dominated the transport. The high contact resistance is likely due to a combination of extrinsic factors such as granularity of the contacts and contamination at the interface, and intrinsic ones as have been considered theoretically by several authors [6,7,8]. In order to observe the intrinsic transport properties of metallic nanotubes and certain transport phenomena, low-resistance electrical contacts to the nanotubes must be used. Soh et al. [9] have made such contacts by evaporating metal on top of the nanotubes grown directly on a silicon chip with a chemical vapor deposition method. The measured two-terminal resistance was as low as ~ 10 k^. Similar reduction in contact resistance was achieved by using planarized and briefly annealed gold electrodes [10]. In a third approach, a laser pulse was used to weld the two ends of suspended nanotubes into gold electrodes [11].

0 0

Post a comment