(Color online) (a) STM image of the surface acquired with a Si tip. (b) A contact footprint formed on a -type by a Si tip. The contact was made at 2 V and , i.e., at in conductance. Before making this footprint, the tip was used for the conductance measurements and subject to touching the surface several times. Therefore, the crater roughly indicates the average initial size of the NWs formed between the Si tip and the surface.
(Color online) Examples of transient conductance trace observed in the high-conductance region . Two traces in each panel represent the conductance recorded during the tip retraction and the reapproach, respectively. In many cases, the conductance simply jumps down and up, as shown in (a), but occasionally exhibits a couple of plateaus, mostly in the subquantum range as shown in (b) and (c). The plateaus are nearly reproducible in the retraction and the reapproach traces.
(Color online) The and the conductance histograms at in the high-conductance region . The histograms are plotted against to cover the wide distribution of the junction conductance. A few peaklike features are commonly observed in two histograms.
An example of transient conductance trace observed on the junction and in the low-conductance region . This trace shows a well-defined plateau in the range.
(Color online) Conductance histograms in the low-conductance region obtained under different sample biases. Histograms (a), (b) and (c), (d) are for the and the junctions, respectively. The vertical axis in each panel is appropriately scaled so that we can directly compare the intensity of each plot. The histogram in (a) under positive biases shows appreciable intensity at low biases.
(Color online) The curves in the low-conductance region obtained under different sample biases. As in Fig. 6, the plots (a), (b) and (c), (d) are for the and the junctions, respectively. Note that each plot has a different scale for , for clearly displaying the behavior of . Below , all curves show a nearly linear decrease with increasing , whereas they exhibit more moderate and nonlinear dependence at higher conductances. The latter behavior is particularly evident for the curves in (a).
(Color online) The attenuation length of the conductance deduced from the linear part of the curves in Fig. 6 is plotted as a function of the sample bias. The filled circles and squares represent the data for the and the junctions, respectively. Despite some scatter, the magnitude of the attenuation length is around 0.5 nm and shows no systematic dependence on the bias and the junction type.
(Color online) The quadratic fit (solid lines) to the nonlinear part of the curves in Fig. 6(a).
Article metrics loading...
Full text loading...