Report on TMS Workshop on "Doping, Dopants and Carrier Dynamics in Wide Bandgap Semiconductors" in Copper Mountain Resort, Colorado , April 2-6, 2000

Report on TMS Workshop on "Doping, Dopants and Carrier Dynamics in Wide Bandgap Semiconductors" in Copper Mountain Resort, Colorado, April 2-6, 2000 .

This meeting was sponsored by Office of Naval Research (ONR) and TMS, and organized by F Schubert, Boston University. The practical arrangements were as usual handled in an excellent way by Allwood Inc. There were about 50 participants, the weather was good, and everybody returned back home with an embarrassingly strong tan.

Acceptor doping remains an important topic, and was discussed in a number of contributions. Be is believed to be the shallowest acceptor in GaN, and may readily be introduced in MBE growth. (It is more hazardous in MOCVD growth). P-type doping at 2x10 ¹⁷ cm ^-3 was reported with Be doping, with a mobility of 200 cm ²/Vs at 300 K.

Co-doping was discussed extensively, it now seems that this technique is becoming appreciated in the III-N community. The theory of co-doping to improve the p-type conductivity was reviewed, with applications suggested besides GaN also for AlN and ZnO. Good p-type material was suggested for C-O co-doped AlN (2C per O). For ZnO good p-type behavior is expected for N acceptors co-doped with Al or Ga (e g 2 Al per N). The first experimental results on p-type ZnO seem to verify these ideas.

Experimental demonstration of a much improved p-type conductivity for Mg-doped GaN with O co-doping in MOVPE was given. These results seem to confirm the theoretical predictions that both the hole activation and the hole mobility are drastically improved by co-doping at levels up to the 10 ²⁰ cm ^-3 range. It was also claimed by some authors that the co-doping would increase the solubility of Mg in GaN. A rumour was around that the latest high power laser diodes announced by Nichia were made using co-doping in the p-region.

Preliminary reports were given on a similar role of H in co-doping, i e a substantial improvement of both Be and Mg p-doping was observed in the presence of atomic H in MBE growth.

Not much progress was made on the identification of optical spectra for donors and acceptors in GaN. To get a definite identification of defect-related spectra much better material quality is needed, very little can be done with hetero-epitaxial layers, due to the dislocation-induced broadening of the spectral linewidth. On topic discussed was the 3.27 eV PL spectrum in GaN, which is indeed very common in GaN. No final conclusion was given, those working with Mg doping reassured this spectrum was the Mg acceptor signature. The same spectrum was observed with C or Si doping, however, or even in undoped samples grown in an equipment which has never seen Mg! More work is needed here...

Theoretical models of the limitations of doping (n-type and p-type) for different wide bandgap materials was reviewed, in terms of the position of the Fermi level stabilization energy E _FS. The ease of doping the material n- or p-type is then determined by the position of E _FS vs the band edges for each material. In general strong n-doping requires a large electron affinity χ, while good p-doping requires a small work function φ. For a recent literature reference see e g Phys Rev Lett 84, 1232 (2000).

Studies of donor dopants in GaN and AlGaN under high hydrostatic pressure were reported. It was previously found that O shows a metastable behavior, i e a deep state at high Al compositions in AlGaN. It was shown that the shallow Si donor has in fact a similar behavior, in contrast to most theoretical predictions. These donors reveal the deep level at high pressures in GaN.

Polarization doping has been found to be very important for III-N structures. A report was given of polarization doping of holes in AlGaN/GaN heterostructures. A 2D hole gas is expected to exist in suitably gated structures, at densities 10 ¹³ cm ^-2 or higher. This is of great interest for improvement of the base spreading resistance in AlGaN/GaN HBT structures. The induced base transistor (IBT) concept was discussed, based on n-GaN/AlGaN/GaN/n-InAlGaN, where a thin GaN QW base has been inserted. Such a structure was claimed to be a strong competitor to a regular HBT structure.

An interesting approach to improved p-type conductance is the p-type superlattice doping utilizing the strong polarization fields in AlGaN/GaN systems. Previously only the improvements in the lateral p conduction in the structure were reported. New EBIC experiments measuring both the vertical and the lateral minority carrier diffusion lengths L _D were reported. A ratio of 2.6 between the lateral and vertical L _D values was found. This ratio may be reduced by grading the superlattice structure. The data are relevant for the base conductance of HBTs.