Designing miniature wires for small-scale electronics

The last few decades have seen an amazing miniaturization of silicon microelectronics, to the extent that modern transistors are approaching quantum limits. Further miniaturization requires novel materials with a well-defined atomic structure that allows information to be processed and stored with a small and uniform number of charge carriers. Nanostructured materials have attracted the attention of researchers because they promise both improved processing and energy efficiency.

Carbon nanotubes (CNTs) were discovered in 1991 and are a possible alternative to silicon. Their ability to transport electrons in one dimension has excited immense interest, but other electronic properties limit their use. First, the electronic properties depend on the CNT chirality — the direction along which a graphene sheet is rolled into the nanotube — but no chirality-specific synthesis is available. Second, intertube interaction is relatively strong so CNTs bundle together, which changes the overall electronic properties. Third, current flow through CNT-electrode contacts is poor.¹ We have shown that inorganic nanowires (INWs) made from novel alternatives to carbon do not suffer from the limitations of CNTs.

Earlier calculations had indicated the high stability and 1D metallic nature of molybdenum-sulphide (Mo₆S₆) and molybdenum-chalcohalide (Mo₆S_6−xI_x) INWs. Recent experimental advances in the synthesis of such wires motivated our present theoretical study of their electronic-transport properties.² We find that Mo₆S_6−xI_x nanowires are effective conductors and less prone to bundling than CNTs. Our calculations also show an easy flow of electrons through the contact between a gold (Au) electrode and a Mo₆S₆ nanowire. In addition, Mo₆S₆ nanowires have novel electronic properties: when twisted they switch between metallic and semiconductor properties, thus making them suitable for use as a nanoswitches or potentiometers.

Figure 1. (a) Structure of a molybdenum-chalcohalide Mo₆S_{6- x}I_x (x=2) nanowire (NW): side view and cross section. (b) Structure of Mo₆S₄I₂nanowires arranged hexagonally in a plane normal to the wire axes. (c) Deviation from the equilibrium binding energy (E₀) as a function of the relative unit-cell size, a/a_eq. (d) Contour plot of the nanowire binding energy in this lattice (eV/unit cell). CNT: Carbon nanotube.

We employed density-functional-based calculations extended with a Green's function formalism to predict the properties of INWs. The structure of a Mo₆S_6−xI_x nanowire, shown in Figure 1(a), consists of a Mo backbone decorated by sulfur and iodine ligands. Iodine content can take values in the range x=0…6. The most stable isomer is for x=2 and bundles, as shown in Figure 1(b). The longitudinal deformation energies indicate a remarkable axial stiffness comparable to a CNT of similar diameter, as shown in Figure 1(c).³ All wires are metallic, regardless of the iodine arrangement, and their density of electronic states is similar to that of conducting CNTs. As a consequence, the transport properties around the Fermi level of a Mo₆S₄I₂ wire and any armchair CNT are very similar.³ The highly anisotropic interaction between the wires in their bundle³ gives a considerably smaller effective mutual interaction — see Figure 1(d) — which is in contrast to the strong intertube attraction in CNT ropes.⁴ It is easier to separate a free-standing INW than a free-standing CNT from their bundles.

Figure 2. (a) Schematic view of the geometry of a Mo₆S₆nanowire in contact with the gold (Au) electrode. (b) Transmission coefficient of the free periodic nanowire without gold contacts (black) and for the Au –(Mo₆S₆)₂₀–Au system (red). E_F: Fermi energy.

Next, we considered a pointlike contact between a Mo₆S₆nanowire and gold electrodes.⁵ The investigated geometry pictured in Figure 2(a) corresponds closely to the preferred structures of molybdenum-sulfide clusters on an Au(111) surface.^6,7 The transmission of an Au – (Mo₆S₆)₂₀ – Au system around the Fermi level, as shown in Figure 2(b), is close to the ideal transmission per conducting channel. With CNTs, such good contacts require a segment of five or more nanometers of CNT embedded into the electrode.¹ The Au – Mo₆S₆ contact is electron-transparent due to a unique division of labor: sulfur attaches the nanowire strongly to the gold electrode while an unperturbed current flows through the direct, extended Au – Mo channels.⁵

Figure 3. (a) Geometry of twisted Mo₆S₆nanowire. The electrodes are considered as semi-infinite, ideal, straight nanowires. (b) Transmission as a function of twisting angle (in units of conductance quantum, 2e²/h).

As a metallic system, the molybdenum-sulfide nanowire would not serve as a logic element in electronic devices. However, twisting the Mo₆S₆ wire as pictured in Figure 3(a) introduces a metal-semiconductor transition, and the band gap increases linearly and monotonically with torsion angle: see Figure 3(b).⁸ This is in clear contrast to twisted CNTs whose conductance oscillates under increasing torsion.⁹ The metal-semiconductor transition in the wire is due to a loss of symmetry, which causes an avoided level crossing around the Fermi energy. Since the gap opens linearly with the increase in torsion angle, the wire could be employed as a nanoswitch or potentiometer.⁸

In conclusion, the molybdenum-chalcohalide and molybdenum-sulfide nanowires exhibit many exciting properties, some resembling closely those of CNTs and others providing distinct advantages. They show exciting promise as unique building blocks for next-generation electronic devices. Future work will focus on the integration of these promising structures into nanoelectronic circuits.