Scrutinising thick films of AlInN
To produce GaN-based lasers that are latticematched, there needs to be a switch from the conventional choice of cladding material, AlGaN, to the alloy AlInN.
Unfortunately, very little is known about the characteristics of AlInN layers with a thickness comparable to that of the cladding "“ but researchers from Japan are addressing this omission by produced these structures on sapphire, before scrutinising them with a range of techniques, including X-ray diffraction, cross-sectional transmission electron microscopy and spectroscopic ellipsometry.
Prior to this recent work by a collaboration between Nagoya Institute of Technology and Meijo University, there have been very few reports of AlInN films greater than 100 nm thick "“ and for those thicker films, no details have been provided.
To study thick AlInN films, the team from Japan took a 2-inch GaN-on-sapphire template, loaded it into an MOCVD chamber, and deposited Al0.17In0.83N layers with thicknesses of 40 nm, 90 nm and 300 nm, using conventional precursors, a susceptor temperature of 830 ËšC, and a pressure of 13.3 kPa.
Films of AlInN were deposited at 0.6 μm/hr, a growth rate that is markedly lower than that used for GaN. "However, this is not necessarily slow, considering conventional AlInN growth," explains corresponding author Makoto Miyoshi.
X-ray diffraction scans produced a strong peak for AlInN and periodical fringe peaks. Taken together, this indicates that AlIN is a single-phase crystal, probably forming an abrupt interface with GaN. The X-ray spectra had a high degree of correlation with a corresponding simulation, which confirmed the thickness and the composition of the AlInN alloys.
To determine crystal quality, the researchers turned to X-ray rocking curves, atomic force microscopy and cross-sectional, dark-field transmission electron microscopy along two different crystal directions. These techniques revealed that the AlInN films have: a mosaicity that is strongly dependent on the underlying GaN; surfaces that feature flat areas and pits, which increase in size with the thickness of the film; and that these pits, originating from threading dislocations in the underlying GaN, are pure-edge and edge/screwmixed dislocations.
The root-means-square roughness of the surface of the 300 nm-thick AlInN film is 1.82 nm. According to Miyoshi and co-workers, this value is not as small as it is for atomically flat surfaces, but it is still "rather good", considering the influence of the pits.
Crosssectional, dark-field transmission electron microscopy images, acquired along two different crystal directions, reveal that the pits seen on the surface originate from threading dislocations in the underlying GaN.
The team plans to carry out more experiments with AlInN films. Their work will include understanding the mechanism of AlInN growth, growing free-standing GaN substrates with fewer threading dislocations, growing 500 nm-thick AlInN films, and controlling the impurity doping and conductivity of AlInN films.
Miyoshi and co-workers, this value is not as small as it is for atomically flat surfaces, but it is still "rather good", considering the influence of the pits.
However, the team believes that these AlInN films are not good enough for making lasers. "We need to decrease the surface pits," says Moiyoshi, who adds that the thickness of the film will have to be increased to around 500 nm.
