Laser-assisted atom probe tomography (APT) was used to measure the indium mole fraction x of c-plane, MOCVD-grown, GaN/In(x)Ga(1-x)N/GaN test structures and the results were compared with Rutherford backscattering analysis (RBS). Four sample types were examined with (RBS determined) x = 0.030, 0.034, 0.056, and 0.112. The respective In(x)Ga(1-x)N layer thicknesses were 330 nm, 327 nm, 360 nm, and 55 nm. APT data were collected at (fixed) laser pulse energy (PE) selected within the range of (2-1000) fJ. Sample temperatures were = 54 K. PE within (2-50) fJ yielded x values that agreed with RBS (within uncertainty) and were comparatively insensitive to region-of-interest (ROI) geometry and orientation. By contrast, approximate stoichiometry was only found in the GaN portions of the samples provided PE was within (5-20) fJ and the analyses were confined to cylindrical ROIs (of diameters =20 nm) that were coaxial with the specimen tips. m-plane oriented tips were derived from c-axis grown, core-shell, GaN/In(x)Ga(1-x)N nanorod heterostructures. Compositional analysis along [0 0 0 1] (transverse to the long axis of the tip), of these m-plane samples revealed a spatial asymmetry in charge-state ratio (CSR) and a corresponding asymmetry in the resultant tip shape along this direction; no asymmetry in CSR or tip shape was observed for analysis along [-1 2-1 0]. Simulations revealed that the electric field strength at the tip apex was dominated by the presence of a p-type inversion layer, which developed under typical tip-electrode bias conditions for the n-type doping levels considered. Finally, both c-plane and m-plane sample types showed depth-dependent variations in absolute ion counts that depended upon ROI placement.
Download