For semi/non-polar GaN grown on sapphire, we always found that at different azimuth angle, the on-axis XRD rocking curve FWHM is very different. Some paper mentioned that this kind of anisotropic property is related to the larger mosaic tilt, reduced coherent length, or dislocation and stacking fault. But there is still not a clear conclusion about this question. Does anybody know how to analyse this phenomenon? how to distinguish the contribution from different factors?