Electron transport in metallic dot arrays: Effect of a broad dispersion in the tunnel junction dimensions

Abstract

We have shown that for a one-dimensional multijunction array with a broad junction length distribution, we can increase the mean threshold voltage Vth for a given maximal tunnel resistance without increasing the scatter of Vth. For two-dimensional arrays we can increase strongly the output of devices which do not behave as open circuits, without any loss on the scatter on Vth, and still increase the latter with respect to a single-island device. The experimental background which justified the calculation will be described, as our model needs the experimental distribution of the island dimensions and junction lengths. Such a result may either show a way to increase the threshold voltages for a given technology, or allow for larger and therefore more controllable dimensions, or even extend the choice of materials.

Origin of yield problems of single electron devices based on evaporated granular films

“…7 One reason for this change of approach is the fabrication yield problem of granular structures. 1D systems of nanometer size clusters have been achieved, 6 but Coulomb blockade has not yet been observed in these systems. The situation is more optimistic for two-dimensional ͑2D͒ structures, but low yield poses a problem here as well.…”

“…Both these types of structures can be fabricated experimentally. 6 As for the 1D systems, a series of snapshots at different surface coverages of the growth process are shown in Figs. 1͑a͒-1͑e͒.…”

“…This corresponds well to experimental findings. 6 Therefore, while the fabrication of 1D granular systems remains an impressive achievement in technological terms, there is doubt that such systems will be feasible for making single electron devices.…”

“…In practice this means that the islands must be separated by less than 2 nm. 6 We have performed numerous simulations of the growth process and we have found that the conditions for Coulomb blockade devices are relatively unlikely to occur. The yield is about 10%.…”

“…More recently, however, interest has turned from exploration to application, and from stochastic granular structures to lithographically defined devices. This includes a combination of both, such as lithographically defined one-dimensional ͑1D͒ granular films 6 and the postprocessing of nanometer size clusters with high accuracy. 7 One reason for this change of approach is the fabrication yield problem of granular structures.…”