7.2 Development of a Compact NH_3 Jet - Original N_2 Jet

The original jet is shown in Figure 5 and Figure 6, placed about 70cm from the substrate using the central source flange of the Gen II MBE system. The nozzle consists of a tungsten tube with a laser drilled hole on the side on the order of 10's of microns. The tube is closed off at one end, while the other end is connected to a gas line. The tube is resistively heated and was tested to temperatures of up to 2000°C. The region around the nozzle is water cooled to prevent damage. The skimmer is placed on the order of 3mm from the nozzle on a x,y,z-stage connected to bellows to allow the skimmer to be centered on the jet beam. Pumping between the nozzle and skimmer region was accomplished by a 100 l/s (N_2) turbo pump backed by a grease free diaphragm roughing pump. Both pumps are designed to resist corrosive gases, including NH_3.

Results with this jet turned out to be fairly disappointing in terms of GaN growth. Fluxes were only on the order of 10^-8 torr at the substrate, which, as determined later (using NH_3), corresponds to a GaN growth rate on the order of 10-100 Å/hour. Another problem at the time were substrate temperatures. In and Sb-In turned out to evaporate to fast to be used to mount samples to the Mo sample blocks as often done in conventional MBE. Cold spots developed quickly and eventually the sample fell off. The sample was typically only during the first hour or so hot enough for growth to be expected. This problem was later solved by using mechanical mounts as described in another chapter. Despite the problems mentioned above GaN was detected by Auger on c-plane sapphire substrates used. The RHEED pattern was usually not very conclusive if GaN growth took place or not.

This jet was later modified in an attempt to increase the flux. Since the distance to the substrate can not be changed due to physical restrictions of our system the only significant parameter to improve on is the background pressure P_b between nozzle and skimmer to decrease beam attenuation. This was achieved by replacing the turbo pump by a 1000 l/s cryo pump. Also, a second differential pumping stage was added between the skimmer and a collimator inserted 20 cm from the skimmer. The turbo and roughing pumps originally placed between the skimmer and nozzle are now used to pump the region between skimmer and collimator. Fluxes achieved after these modifications were in the 10^-7 torr range, but no further growth attempts were made to date due to the successes of a new compact NH_3 jet design discussed in the next section.

The original jet should not be viewed as a failure though. Although it seems to provide insufficient fluxes for GaN growth to be practical, its capability for high nozzle temperatures T_o of up to 2000 K makes it for example an ideal candidate for FeN growth, which has great potential for magnetic recording applications. As reported by C. T. Rettner and H. Stein (Ref. 24) the N_2 sticking coefficient on Fe (111) is markly enhanced by using an energetic beam of N_2 as shown in Figure 7, which was obtained using a substrate temperature of 520 K. They used a jet with a resistively heated tungsten tube and N_2 seeded in H_2 to obtain translational energies of up to 4.3 eV. As can be seen from the figure, the sticking coefficient is about 1x10^-6 at 0.09 eV and over 0.1 at 4.3 eV, which seems to be the maximum attainable sticking coefficient. It was also found that the sticking coefficient increases by lowering the substrate temperature from 600 K to 300 K by about 50%.

Our group is currently preparing to explore FeN growth on a FeAl/AlAs/GaAs superstructure.

Figure 5 (Ref. 22): Original Jet

Figure 6 (Ref. 23): Photograph of Original Jet

Figure 7 (Ref. 24): Sticking Coefficient S of N_2 on Fe(111)
at normal incidenceand substrate temperature of 520 K


Table of Contents