Characterization of Thick 4H-SiC Hot-Wall CVD Layers
- PDF / 1,033,982 Bytes
- 6 Pages / 417.6 x 639 pts Page_size
- 15 Downloads / 174 Views
premier semiconductor material for fabrication of high power and power microwave electronic devices. In addition, SiC devices can operate at elevated temperatures allowing them to be used
at either high ambient temperature or with reduced cooling requirements in nominal ambients. In recent years substrate and CVD developments have made great progress, thereby allowing for the development of SiC-based high power high frequency devices such as Metal-Semiconductor Field Effect Transistors (MESFETs)[ I]. Other high power devices such as diodes, MOSFETs and GTOs have also been demonstrated [2,3,4]. While current levels of uniformity of layer thickness and doping are sufficient for production use, the current wafer diameters are not. Thus work needs to continue on layer uniformity for increasing wafer sizes as they become available to be ready for production when it becomes commercially viable.
EXPERIMENTAL SETUP The wafers were grown in a horizontal hot-wall CVD reactor. The key component of the reactor is the graphite susceptor which is similar in construction to those described in Ref. 5 and 6. The susceptor is heated inductively and is designed to obtain good heating uniformity over a large area. The temperature is measured by a pyrometer focused on a position at the leading edge of the susceptor. The susceptor is also tapered to compensate for gas reactant depletion which may be quite severe in hot-wall systems. The precursors, silane and propane, are diluted in a high flow of purified hydrogen. Growth temperatures in the range of 1500-1700'C were used. Nitrogen was used as the n-type dopant and trimethyl aluminum (TMA) was used as the p-type dopant. The thicknesses and thickness uniformities were measured by observing the cleaved edge of the sample in a scanning electron microscope (SEM). The difference in doping between the highly doped substrate and the epitaxial layer gives sufficient contrast for determining the 167 Mat. Res. Soc. Symp. Proc. Vol. 572 ©1999 Materials Research Society
thickness. The doping and doping uniformities were measured by a mercury probe capacitancevoltage (C-V) profiler. RESULTS AND DISCUSSION Epitaxial Layer Uniformities Epitaxial layers were deposited on prototype 75 mm diameter wafers, cleaved and then measured along the flow direction for layer thickness by cross-section scanning electron microscopy (SEM). The results are shown in Figure 1. The uniformity is calculated as the standard deviation divided by the mean. The "epi-crown" (the raised ridge along the wafer
---
25 -
"20 U 1 15 UQ)
Average: 27.2 jim .... Uniformity: 1.21%/o
circumference) was not included in the calculations.
Additional wafers were grown and characterized by C-V measurements to determine the doping uniformity. For a doping level of 7.5x 1015 cm-3 n-type the resulting uniformity
.... ..--
50 10 0
was 7%, also calculated as standard deviation
divided by the mean. Similar growth runs on
10
..........-- ....-.. -------....
20
30
40
50
60
70
Distance from upstream edge (mm)
75 mm wafers forp
Data Loading...
