Random deposition of conducting nanoparticles on a flat two dimensional (2D) substrate leads to the formation of a conducting path at the percolation threshold. In sufficiently small systems significant finite size effects are expected. However, in the 2D square systems that are usually studied, the random deposition means that the main effect of small system sizes is that stochastic fluctuations become increasingly large.
We have performed experiments and simulations on rectangular 2D nanoparticle films with nanoscale overall dimensions. The sample geometry is chosen to limit stochastic fluctuations in the film’s properties. In the experiments bismuth nanoparticles with mean diameters in the range 20-60nm are deposited between contacts with separations down to 300nm. At small contact separations there is a significant shift in the percolation threshold (pc) and the conducting
path formed close to pc resembles a nanowire. Percolation theory describes the experimental onset of conduction well: there is good agreement between predicted and measured values of the power law exponent for the correlation length.
Atomic clusters can be produced in a size range (100nm to 0.5nm) that bridges the gap between the limits of current lithographic fabrication technologies for integrated circuits and the atomic/molecular regime. The work presented here aims to combine established top-down device processing with bottom-up engineered cluster assembly. Conducting cluster deposition and standard optical fabrication techniques have been used to produce wires on a textured (V-grooved) substrate. The lengths of the wires (ranging from 2μm to 1mm) are defined simply by the separation of NiCr/Au contacts. The deposited nanoparticles range in size from 20-100nm and in principle define the width of the nanowire. In-situ conductance measurement allows precise control of the deposition process and the onset of conduction in the wire is readily monitored as a function of deposition time. The effectiveness of the surface templating technique is demonstrated by SEM and AFM imaging carried out after deposition. The surface coverage is seen to vary from <20% on the unpatterned (normal-to-beam) surface (which is required to be non-conducting) to >100% at the apexes of the V-grooves used to promote growth of the wire. Self assembly of the nanoparticles leads to completion of a wire between the pre-formed contacts with no possibility of a parasitic conduction path. Wires formed through this technique currently have minimum widths of ~1μm but straightforward extensions of the technique should soon allow nanowire formation.