< Back to previous page

Project

Finite Size Effects in Platinum-group Metal Nanostructures

Conductivity of nanostructures differ significantly from the bulk. One such manifestation is the resistivity scaling where the resistivity of metallic nanostructures increases drastically when the characteristic dimensions are reduced to the order of the carrier mean free path and is due to the additional scattering contributions from the interfaces and the grain boundaries. Though this phenomenon has been known for several decades, there is still no consensus over the relative importance of these scattering mechanisms in thin films and nanowires. Technologically, the increased resistivity has important implications: metallic nanowires are used as interconnect structures in microelectronic circuits and increased resistivity has led to problems such as higher latency, increased power consumption, more noise, and degraded reliability. Platinum-group metals have been proposed as promising candidates to replace the choice of metal, copper, however little is known about the scattering mechanisms in scaled structures of these metals.

This work aims to improve the current understanding of the scattering mechanisms in thin films (<30 nm) and nanowires (<100 nm2 in cross-section area) of the platinum-group metals. The scattering mechanisms in thin platinum-group metals films are studied and compared to copper through semi-classical resistivity modeling using the Mayadas-Shatzkes (MS) model and complemented with extensive microstructural characterization. An adaptation of the MS model for thin films was proposed that can model the resistivity of nanowires. MS models for wires and thin films describe the experiments with a consistent set of parameters.  For thin films, grain boundary scattering was established as the dominant scattering mechanism contributing to resistivity and assumes more importance for metals with higher melting points. However, resistivity is most sensitive to grain boundary scattering only when grain size becomes of the order of the mean free path but has little impact once the grain size becomes greater than a factor of 4-5 of the mean free path. The choice of cladding on films is found to significantly alter the surface scattering behavior. A novel scheme was proposed which can be used to fabricate nanowires with sub-100 nm2 in cross-section area. On examining the transport properties of Ru nanowires (quasi-one-dimensional metallic structures), grain boundary scattering was the dominant contributor to resistivity, except for very small cross-sectional areas where surface scattering contributes significantly due to a higher ratio of surface-to-volume. It is further understood that the resistivity was predominantly governed by the cross-sectional area of the nanowires and variations in aspect ratio (for a fixed cross-sectional area) have little impact. Furthermore, it was found that the mean free path is the single most critical parameter in determining the sensitivity of a metal to resistivity scaling. This results in a crossover where the resistivity of the platinum-group metals, though higher in bulk, become lower than that of copper at smaller dimensions. Finally, wafer-level reliability tests on ruthenium nanowires demonstrate the potential for robust reliability performance in interconnect structures.

Date:1 Oct 2013 →  2 Oct 2018
Keywords:Interconnects
Disciplines:Classical physics, Elementary particle and high energy physics, Other physical sciences, Condensed matter physics and nanophysics
Project type:PhD project