Evaluation of Gap States in Hydrogen-Terminated Silicon Surfaces and Ultrathin SiO 2 /Si Interfaces by using Photoelectr
- PDF / 408,274 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 68 Downloads / 206 Views
Mat. Res. Soc. Symp. Proc. Vol. 500 © 1998 Materials Research Society
by 4.5%HF and 40%NH 4 F treatment, respectively, followed by pure water rinse for 3sec to remove residual fluorine atoms from the surface. Hydrogen bonding features on the surfaces so prepared were monitored by Fourier-transform infrared attenuated total reflection (FT-IR-ATR) and ultraviolet (Hell) excited photoelectron spectroscopy (UPS). The thermal oxidation was carried out in a furnace at a temperature of 1000°C in dry 02. The energy distributions of filled gap states for these samples were evaluated from total photoelectron yield spectra which were measured as a function of incident photon energy in the region from 4 to 6eV, with an energy resolution of -20meV, and normalized by the incident photon flux. RESULTS AND DISCUSSION The typical photoelectron yield spectrum for monohydride-terminated p+ Si(l 11) with a boron concentration of -5xl0
18
3
cm- is shown in Fig. 1. The spectrum shows a signal dynamic range of
about eight orders of magnitude. Obviously, an extremely low yield for photons with energies below 4.8eV is obtained. The surface Fermi-level position determined by a Kelvin probe method is located at 4.7eV below the vacuum level, namely shifted toward midgap by at least 0.25eV from the bulk Fermi-level. This Fermi-energy shift is mainly attributed to hydrogen-induced passivation of acceptors in the near surface region during the wet-chemical process [5,6]. In fact, after 5min annealing at 350'C the surface Fermi-level shifts close to the bulk position without any change in not only the yield spectrum but also surface hydrogen termination as confirmed by UPS measurements [5]. In addition to this, for both
as-prepared and annealed samples, no
p+-Si(111)
CD.
photovoltage under illumination conditions for the photoemissions was detected within a detection limit of -lmV, and the surface concentrations for ionic contaminants such as Na+ were proved by X-ray photoelectron spectroscopy to be at least an order of magnitude lower than needed to explain the Fermi-level position for as-prepared sample. The yield between the Fermi-level positions before and after annealing at 350'C indicates that a few gap states still exist on the surface. The yield spectra
l
from differently-doped samples, which were
0
l0
Z
5()
NA: 5x1018 cm
m 105
S
0 _j
8 8
Lu
5: Z 103
0
0
s
A
"" j 101 0
prepared with the same cleaning procedure, were also measured and compared as represented in
EoCA d,
a
,0
1-1.
0.
4.0
Fig.2. A fairly big difference in the yield among the samples is observed. However, this is not so surprising because the yield spectrum reflects the integrated energy distribution of filled states. In order to evaluate the band bending in the surface, or the tailing of the valence band edge for heavily-doped samples, the observed yield
5.6 5.2 4.8 4.4 PHOTON ENERGY (eV)
Fig. 1 Total photoelectron yield spectrum for as-prepared or annealed Si(1 11) with a boron 3 concentration of -5x1018 cm of which the surface was passi
Data Loading...
