Growth Monitoring Using RHEED - Growth of AlN and GaN on c-Plane Sapphire

9.2 Growth Monitoring Using RHEED - Growth of AlN and GaN on c-Plane Sapphire

Figure 22 shows a typical growth history of GaN on c-plane sapphire using an AlN buffer layer documented with RHEED images, showing the 112(bar)0 and 11(bar)00 crystal directions. The first row of images shows c-plane sapphire "as is" after outgasing at about 500ºC for several hours (usually done overnight). After outgasing the NH_3 flux is turned on (5x10^-6 torr, jet or leak valve) and the substrate temperature ramped to about 900°C. After about 30 min at that temperature the next row of images were taken, showing the surface after nitradation. As can be seen, the AlN RHEED pattern starts to come in believed to be due to about a monolayer of AlN formed on top of the sapphire substrate.

Next the Al source shutter is opened (0.3 ML/s AlAs) at a substrate temperature of still 900°C. The AlN usually starts out to grow rough (3-D) as shown in the figure, but typically within a few minutes the growth becomes more and more 2-D as shown in the next row of images (after 30 min). AlN is grown until no more improvement in surface quality is noticed, which is usually between 10-30 minutes (about 200-500 Å).

Next the Ga source shutter is opened (7 ML/s GaAs) while the AlN shutter is still open, which is closed after about 1 min. This change "on the fly" is done to prevent contamination (especially oxygen) to build up on the AlN surface as well as to provide a somewhat gradual transition from AlN to GaN growth. Then the substrate temperature is ramped slowly to about 20° above the balance point to be in the good part of the hysteresis of the dips, which is explained in the next chapter. This corresponds to a substrate temperature of 820°C for a Ga flux of 6.8 ML/s (GaAs).

Sometimes the GaN starts out growing somewhat 3-D as shown in the figure (after 5 min.), and it usually helps to grow for a few minutes at relatively high substrate temperatures. After a couple of minutes the typical weak 2x2 reconstructed GaN pattern starts to come in as shown. The streaks in the GaN RHEED pattern typically become sharper over time, and sometimes a dramatic sharpening is seen as shown in the figure (here after about 30 minutes). Another set of images is provided showing the pattern 5° off axis.

Figure 22 (below): 112(bar)0 and 11(bar)00 RHEED pattern of GaN on c-plane sapphire with AlN buffer layer

Sapphire c-plane substrate "as is" after outgasing