Xps Analysis of the Sapphire Surface as a Function of High Temperature Vacuum Annealing
- PDF / 277,667 Bytes
- 4 Pages / 420.48 x 639 pts Page_size
- 37 Downloads / 176 Views
XPS ANALYSIS OF THE SAPPHIRE SURFACE AS A FUNCTION OF HIGH TEMPERATURE VACUUM ANNEALING E.D. Richmond Naval Research Laboratory, Code 6816, Washington D.C. 20375-5000
ABSTRACT: For the first time the (1102) surface of sapphire has been investigated by X-ray photoelectron spectroscopy to ascertain chemical changes resulting from annealing in vacuum at 1300" C and 14500 C. As received substrates had a substantial surface C contaminant. For substrates that were chemically cleaned before inserting them into the MBE system no trace of carbon is detected. A residual flourine contaminant results from the cleaning procedure and is desorbed by the vacuum annealing. Spectra of annealed substrates are compared to the unannealed chemically cleaned substrates. The annealed substrates exhibit 0.4 to 0.5 eV shift to higher binding energy of the Al peak and a 0.3 eV shift to higher binding energy of the 0 peak. In addition, a 2% depletion of oxygen from the surface occurs. INTRODUCTION: For the past few years NRL has been investigating the effect of high temperature annealing of the (17102) sapphire substrates annealed at 9000 C, 11000 C, 13000 C, and 14500 C prior to deposition of a silicon epilayer, because of the importance of silicon on sapphire (SOS) for 1 radiation-hard microelectronics. The films were investigated for strain properties and their growth 2 kinetics . In a comparison with chemical vapor deposited (CVD) SOS, the Molecular Beam Epitaxy (MBE) SOS contained an order of magnitude less microtwin volume, less strain, better crystalline 3 quality, higher mobilities, and lower interface states" One potential source of the differences between CVD SOS and MBE SOS may be due to the chemical changes of the surface due to the high temperature vacuum annealing. EXPERIMENT: The 3" (1102) sapphire wafers were chemically cleaned using the procedure shown in Table 1. 4 They were then inserted immediately into the NRL Si MBE/Surface Analytical System . The substrates were analyzed in the analysis chamber which consists of a VG ESCALAB MkIl with a dual anode Mg/Al source. The sources are operated at 300W and 600W for the Mg and Al sources respectively. For each source, a widescan was taken of each sample with a pass energy of 50 eV, and then high resolution scans at steps of 0.1 eV and a pass energy of 10eV were taken of the photoelectron peaks Al(2p), Al(2s), O(ls), C(ls), F(ls). Also the Auger spectra for the transitions O(KVV) and AI(KLL) were obtained. C1 slits were used throughout. The angle between the sample normal and the detector is 15" . RESULTS AND DISCUSSIONS Figure 1 shows a widescan of a chemically cleaned but unannealed sample. The only peaks which are discernible are the AI(2p), Al(2s), O(2s), O(ls), and F(ls) photoelectron peaks and the O(KVV) and F(KVV) Auger peaks. The F(KVV) peak and the O(2s) peak are not labelled and occur around 610 eV and 35 eV respectively. The C(ls) peak at about 295 eV and the C(KVV) Auger peak around 1000 eV are obviously absent. This is confirmed by high resolution spectra, which, in
Data Loading...
