Electrical Characterization of Transition Metal Silicide Nanostructures Using Variable Temperature Scanning Probe Microscopy

Abstract

Cobalt disilicide (CoSi2) islands have been formed on Si(111) and Si(100) through UHV deposition and annealing. Current-voltage (I-V) and temperature-dependent current-voltage (I-V-T) curves have been measured on the islands using conducting atomic force microscopy (c-AFM) with a doped diamond like carbon cantilever. Thermionic emission theory has been applied to the curves and the Schottky barrier heights, ΦB, and ideality factors, n, for each island have been calculated. Barrier heights and ideality factors are evaluated as functions of temperature, island area, and each other. While all islands were prepared in UHV conditions, one set was removed from UHV and measurements were performed in ambient conditions while the other set remained in UHV. The islands measured in ambient conditions were known as "air-exposed samples" due to the fact that the surface was assumed to be passivated upon exposure to atmospheric conditions. The islands measured in UHV were known as "clean samples" because the surface was not passivated. Air-exposed samples were CoSi2 islands on Si(111) and exhibited a negative linear correlation between the barrier height and the ideality factor. Measured values of ΦB on the air-exposed samples approached reported bulk values. Measurements from CoSi2 islands on clean Si(111) and Si(100) revealed no correlation between ΦB and n. Furthermore, it was observed that the measured barrier heights of CoSi2 islands on clean Si surfaces are ˜0.2 — 0.3 eV less than the barrier heights measured from CoSi2 islands on air-exposed surfaces. This negative shift in the clean surface barrier heights was attributed to Fermi level pinning by the non-passivated silicon surface states. Additionally, a slight trend toward lower barrier height as a function of decreasing island area was detected in all samples. This trend is attributed to increased hole injection and generation-recombination in the smaller islands, but it may also be due to effects caused by increased spreading resistance as the island size decreases. Non-linearity in activation energy plots, as well as correlations between decreasing barrier height and decreasing island area-to-island periphery ratio, are attributed to generation-recombination. These measurements indicate that the Schottky barrier height decreases and ideality factor increases with decreasing temperature, even if there is no direct linear correlation between ΦB and n. These temperature-dependent relationships are attributed primarily to hole injection and generation-recombination, with barrier height inhomogeneity as a minor effect.

Titanium silicide (TiSi2) islands have been formed by UHV deposition of titanium on atomically flat Si(100) and Si(111). Scanning tunneling microscopy (STM), scanning tunneling spectroscopy (STS), and a variant of current imaging tunneling spectroscopy (CITS) have been used to characterize single electron tunneling (SET) through the islands. SET is observed to occur in the islands and is evaluated based on the predictions of the orthodox model. The observation of SET suggests that the Schottky barrier could be effective in future SET-based electronic devices. SET was not observed as often as expected, however, suggesting that there is a mechanism limiting SET. Possible mechanisms for SET limiting are evaluated and it is concluded that SET is limited due to a combination of Schottky barrier lowering, a low resistance substrate, and Fermi level pinning by the non-passivated surface states of the silicon. These factors make SET in TiSi2 islands on silicon potentially too variable to be used in future devices unless the SET-limiting mechanism is resolved.