AlN is an ideal substrate material for deep UV, near UV and even high-efficiency blue emitters due to its wide band gap of about 6.2eV at room temperature. Sapphire is still the most commonly used substrate for heteroepitaxial AlN for UV applications, given the UV absorbing qualities of silicon and the cost of bulk AlN, however there are many different approaches in the literature to high quality AlN films on sapphire. We have investigated two different approaches on sapphire with a miscut of 0.4° towards m-plane: 'direct epitaxy' where growth is started at high temperature after a surface treatment, and the uses of a lower temperature nucleation layer.
Defect densities were assessed by X-ray diffraction (XRD) full width half maximum (FWHM) of and reflections and surface morphology by atomic force microscopy (AFM). Despite literature reports of it being successful the 'direct epitaxy' approach resulted in either an
optically rough surface or very high defect densities ( FWHM = 483 arcsec, FWHM = 1065 arcsec), and we believe that this approach may be very sensitive to substrate preparation and miscut. After optimisation the nucleation approach reproducibly gives an acceptable material quality ( FWHM = 351 arcsec, FWHM = 761 arcsec). When grown at a high temperature of 1250°C these surfaces have large steps around 8 nm high, but when capped at 1110°C these bunches are largely eliminated.
(Note some graphical crystallographic directions are missing from this text representation)