ABSTRACT
Although studies of the optical properties of single crystalline III-nitrides have been conducted during the last four decades, knowledge in the areas of defects and impurities is still being developed. –is is related to the quality of the materials, which still lags far behind what has been demonstrated for the lower bandgap III-V compounds. –e gradual development of growth procedures for III-nitrides has been focused on thin £lms grown on foreign substrates for the development of inexpensive device structures. On the other hand, the materials suitable for accurate studies of optical spectra related to dopants and most other defects are strain-free bulk samples or layers grown homoepitaxially on such bulk substrates [1]. –e availability of such samples has so far been very limited for GaN and AlN, while for InN, only heteroepitaxial samples grown on foreign substrates such as sapphire exist to date. Bulk material of GaN and AlN now exists with a density of structural defects (dislocations) <106 cm−2 in GaN and even lower in AlN. Such material is suitable for optical studies of defects, provided the density of impurities and other point defects are also low (a spectroscopic line width of <1 meV is desirable). –e concentration of residual impurities in nominally undoped bulk GaN (mainly donors like O and Si) is at best about 1016 cm−3 for GaN, and considerably higher (>1018 cm−3) for bulk AlN. For samples grown on foreign substrates, the linewidth in photoluminescence (PL) spectra is typically several meV due to the inhomogeneous strain. For InN, this problem continues to exist, and in addition, InN samples so far are degenerate n-type that cause an additional broadening of the PL spectra. Alloys of III-nitrides are now only available as heteroepitaxial layers, and such samples su¢er an additional broadening due to the alloy composition ¬uctuation potentials.
