GaN and Related Materials for Device Applications
- PDF / 3,345,970 Bytes
- 5 Pages / 576 x 777.6 pts Page_size
- 72 Downloads / 309 Views
still working.2 This was probably the first indication of potentially excellent reliability in a material where defect propagation is much more difficult than in II-VI compounds. The GaN materials and device efforts at AT&T Bell Laboratories, RCA, and other industrial laboratories were mostly terminated because of management impatience and the seemingly intractable growth issues just described. From the mid-1970s to the late 1980s, only a few groups, notably that of Isamu Akasaki in Japan, continued the push for improved materials properties. In this period, the few GaN papers at conferences were usually relegated to Friday afternoon sessions when most attendees had already left. A key breakthrough was reported in 1989 by the Akasaki group3 who noticed that highly resistive (compensated) Mgdoped GaN showed p-type conductivity after exposure to the electron beam in a scanning electron microscope. This allowed fabrication of electroluminescent diodes using p-n junctions for the first time. Subsequently Nakamura et al.4 found that GaN(Mg) grown by metalorganic chemical vapor deposition (MOCVD), which was invariably highly resistive, could also be made conducting by simple thermal annealing under a N2 ambient at around 700°C. This process was reversible when NH 3 was used as the annealing ambient, implicating hydrogen as the passivating species. The unintentional passivation by atomic hydrogen of p-type dopants during growth and cooldown after MOCVD or metalorganic-molecular-beam-epitaxy (MOMBE) growth is a common phenomenon in III-V semiconductors and has been identified as the cause of high (and variable) contact resistances and time-dependent current gains in GaAs/
AlGaAs heterojunction bipolar transistors.5 Hydrides such as NH 3 , AsH3, or PH 3 can still crack sufficiently on the cooling substrate after growth is completed to provide a relatively high concentration of rapidly diffusing atomic hydrogen. This hydrogen forms a neutral dopant-hydrogen complex with acceptor species, passivating or removing their electrical activity. The process is reversed by one of two different methods—either thermal annealing to dissociate the complex and cause the hydrogen to leave the crystal or at least bond to another hydrogen atom to form a stable molecule, or injection of minority carriers (electrons) that dissociate the complexes even at room temperature through localized electronic excitation. The phenomenon of residual hydrogen incorporation in MOCVD and MOMBE growth of p-GaN can be eliminated with MBE in which solid Ga and plasmadissociated N2 are used as the source chemicals. To this point, there has been more emphasis placed on MOCVD growth of GaN because of the larger experience base and the fact that this later technique was the first to produce device-quality material. In this issue, the University of Illinois group headed by Hadis Morkog reviews the advances made in MBE growth of GaN and related materials in recent times. Of particular interest is the ability of the lower growth temperature techniques such as MBE and
Data Loading...
