Electrical Characterization of Ultra-Thin Oxides Grown on Silicon Surfaces Cleaned in Ultra-High Vacuum
- PDF / 408,385 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 4 Downloads / 206 Views
High-precision
current source Gauge ...........
.
Fib ............
VVoltmeter Oxygen inlet ..........
-1.59
mm
I'M-\ ............................... H
UHV chamber
i 2
Fig. 1. Schematic drawing of the experiment. needles in a row, the separation between each needle being 1.59 mm. The tips were cleaned in acetone and propanol prior to mounting in the UHV chamber; no in-situ cleaning was performed. The probe was brought down on the cleaned samples using a micrometer screw. In contact, the load on each needle was about 100 grams. A current I (typically 10 YA) was passed through the sample via the outer two needles using a programmable current source. The voltage V across the inner two was then measured with a digital voltmeter. The setup is shown in Fig. 1. The sheet resistivity of the sample is obtained from V where c is a geometric correction factor. The probe was centered on the sample, giving c = 3.8 [10]. The sample was exposed to 4000 L of molecular oxygen at room temperature. The resistivity was measured during the oxidation. The oxygen was supplied to the chamber through a leak valve while the exposure was monitored by an ion-gauge situated 1.5 m away from the sample. The gauge was mounted in such a way that electrons and light from the gauge did not reach the sample. The presence of the gauge did not affect the measured resistivity. The gauge was calibrated to another ion-gauge close to the sample (which was turned off during the experiment), allowing the exposure at the sample position to be determined. We estimate the exposure accuracy obtained in this way to be ±20%. The qualitative features of these measurements for nominally identical samples are well reproduced. At least two samples of each of the four types described above were measured. The resistivity was measured at both polarities of the current, taking the average of the two. Performing this average cancels contributions from spurious effects, such as thermovoltages. The changes in surface conductance AG was then obtained from the averaged resistivities as AG i/p - I/po, where po is the sheet resistivity of the clean sample. BACKGROUND Surface states can affect the surface conductance through two mechanisms; the states themselves can be conducting, and the states can change the conductivity in the space charge layer which extends around 1-10 pm below the surface. For surface states to form conducting channels, they need to be metallic and to have a reasonable dispersion. Even so, the scattering in such channels is very strong, in particular at the steps between terraces on the surface. Thus when measurements are performed on scales much larger than the average terrace width (100 nm or less), the contribution from such channels is expected to be very small. In addition, there is no evidence that oxygen on silicon induces such surface states in the band gap at the Fermi level.
294
10
10
-
C
A8
T--- 4
_ _
I
B
2
7 0 2, 0
o 2-------
0
--- .-.--.
--
---
-2
•---------..........
.........
-4
0
(100)2xl
0.2
-6
(111)7x7
0
Data Loading...
