Homo-Epitaxial and Selective Area Growth of 4H and 6H Silicon Carbide Using a Resistively Heated Vertical Reactor
- PDF / 1,232,732 Bytes
- 6 Pages / 417.6 x 639 pts Page_size
- 66 Downloads / 181 Views
173 Mat. Res. Soc. Symp. Proc. Vol. 572 ©1999 Materials Research Society
Pyrom : I
Figurel: Diagram showing the construction of the above-mentioned resistively heated CVD reactor. (1) Resistively heated filament (2) wafer carrier or susceptor and (3) graphite cup. It consists of a double walled, water-cooled vertical chamber, made out of stainless steel. It has three view ports one of which is used for pyrometric temperature measurement. It is also equipped with a RHEED gun for in-situ RHEED measurements to characterize the grown layers immediately after growth. The susceptor is resistively heated with a graphite filament with power consumption at 1700'C of 8kW. The challenge in the development of the resistive heating system has been the decrease of the filament lifetime due to hydrogen etching. Hydrogen is present in the reactor as a carrier gas and decomposition product of the reactant gases (Sil-4 and C3H 8 ). Hydrogen etching is a function of the filament temperature and the partial pressure of hydrogen in the region of the filament. Studies at EMCORE indicate that filament failure is likely after a 5% reduction in the filament cross sectional area. In order to increase the filament lifetime, hydrogen should be eliminated from the filament region and/or the filament temperature minimized. Our current design involves the use of a thin (20mil thick) graphite susceptor which drops into an open cup resulting in a 20mil to 40mil air gap between the filament and the susceptor. This configuration is used because in minimizes the temperature offset between the filament and susceptor. The base of the cup is purged with argon as shown in figure 1 to reduce the partial pressure of hydrogen in the vicinity of the filament. If the hydrogen can be eliminated the filament lifetime will be determined by carbon evaporation. For our geometry the predicted filament lifetime is over 1000 hours at a filament temperature of 2000°C. Because we have not been able to completely exclude hydrogen from the filament the current filament life is still determined by hydrogen etching. We have been able to achieve filament life times of 15 to 20 hours by this method. In our bid to achieve higher filament lifetimes (>100 hours), we are developing a growth process, which uses argon as the shroud, thus reducing the volume of hydrogen significantly. We are also designing a new cup and susceptor, which will lock in place to reduce the flow of hydrogen into the cup and allowing a higher flow of argon for purging.
174
EXPERIMENTS Homoepitaxial Growth / Reactor Calibration For the growth of a- SiC, we used (0001) SiC substrates with the polished growth surface tilted at angles between 30 and 80 from the basal plane. This tilt angle ensures that "step-flow" homoepitaxial growth occurs and that good morphology is obtained [3],[4],[5],[6]. In order to grow high quality a- SiC epitaxial layers and also to calibrate our CVD reactor, we have been working on optimization of the growth parameters including growth temperature, chamber pressure, reactant
Data Loading...
