An Investigation of the Cause of Initial Band Bending of a Cleaved Clean n-GaAs(110) Surface
- PDF / 437,754 Bytes
- 7 Pages / 420.48 x 639 pts Page_size
- 65 Downloads / 151 Views
Mat. Res. Soc. Symp. Proc. Vol. 54. '1986 Materials Research Society
342
EXPERIMENTAL As is well known, it is difficult to reproducibly evaporate ultra-low coverage metal by using a regular metal evaporator in which the quartz thickness monitor and the sample surface are roughly the same distance from the bead. The lowest metal coverage on a III-V semiconductor surface ever To remedy this reported is the 0.003 MLobtained by Ludeke et al. [121. deficiency and make our ultra-low coverage study more feasible, we have designed and built an ultra-low coverage multi-metal evaporator based on a geometric factor of 1/200 [151 (Fig. 1). In this design the ratio of the metal coverage on the sample to that on the monitor is esample -__-
6monitor
2 X-mCos as 2 co m ZcsC
1
I
(I )
200
When using a regular evaporator, it is generally feasible to obtain a reproducible metal coverage of 0.02 MLby calibrating the evaporation rate of the
metal evaporator for a few minutes and exposing the sample to the evaporator for a few seconds. But, with a geometric factor of 1/200, we are able to
Our calireproducibly evaporate a metal with a coverage as low as 10-4 ML. bration tests lead us to estimate that this evaporator gives an error of less than 30% for 10-4 ML. The accuracy of the metal coverage is improved at higher coverages. Naturally, in the ultra-low coverage regime, one is concerned with the contamination of the cleaved surface under study. To prevent the contamination of the clean and the consequent surface, the experiments were performed in an ultra-high vacuum chamber with the base pressure in the 10-11 torr range. The pressure was kept in the same range when the noble metals were Some samples were left in the chamber for several hours. Other evaporated. samples were exposed to a fake evaporator--a hot tungsten filament without metal evaporation--to determine the effect of heating during evaporation on the surface band bending. No band bending was observed during either of these tests. The data were collected in three photoemission spectroscopy (PES) experiments by using a He lamp and synchrotron radiation as the photon
sources. By following the Ga 3d core level and the valence band spectra (when using the He lamp) or by following the Ga 3d and the As 3d core level spectra (when using the synchrotron radiation) for different noble metal
2... .. :
.c
_a ....b -------------------------COOLING__ WATER
Fig. 1.
12 27 0 Zm=3"; .c= .7 ; Z =44.1"; as= .50; b is the metal bead to be evaporated; m is the thickness monitor; t is the shutter; s is the sample; c is the UHV chamber. By using a long tube, a geometric factor of 1/200 is obtained and the shadow of the thickness monitor and its holder is not cast on the surface of the sample.
S
343
coverage, the profile of the n-GaAs(ll0) surface band bending versus the noble metal coverage was recorded. Mo discrepancy of experimental results between different photon sources or different runs was detected. Other
experimental details can be found in References 5 and 13. RESULT
Data Loading...
