(a) A tilted plan-view SEM image of hexagonally arranged nanocolumns, which have the cross-sectional shape of hexagon. Except a few relatively larger columns, the sizes and heights of those columns are quite uniform. (b) A cross-sectional SEM image showing the bottom of a nanocolumn. The bottom edges of the masks (100 nm in thickness) of slanted walls define the 250 nm hole diameter. The column cross-sectional size is around 300 nm.
Plan-view (a) and cross-sectional (b) SEM images of the overgrowth sample. The layers of overgrowth, nanocolumn, and GaN template are indicated with arrows and labeled by I, II, and III, respectively. A cross-sectional SEM image demonstrating the interface between the template and nanocolumn layers and the masks in between is shown in part (c). The air gaps between nanocolumns are partially filled during the overgrowth process. The original column walls are depicted by the vertical dashed lines.
AFM images of in size of the control sample (a) and the overgrowth sample (b). From the images, the pit density of and surface roughness of 0.834 nm in (a) and the pit density of and surface roughness of 0.411 nm in (b) can be obtained.
(a) Cross-sectional CL image of the overgrowth sample with an electron acceleration voltage of 15 kV. The SEM image taken at the same location as the CL image is shown in part (b). The nanocolumn layer has the higher CL emission efficiency when compared to the other two layers.
Plan-view CL images of 5000-time magnification with electron acceleration voltages of 6 kV (a) and 15 kV (b) at about the same location. The larger bright area of stronger emission in a shallower layer (a) can be seen.
CL images of 1000-time magnification with electron acceleration voltages of 3 kV (a) and 15 kV (b) at about the same location. The domain structure in the overgrown layer with the domain size in the range of several tens of microns can be seen. For reference, the plan-view SEM image of the same location is also shown in part (c). The arrows indicate the boundary of a domain.
(a) Depth-integrated XRD rocking curves in the (0002) plane of the three samples. The FWHMs of the control, nanocolumns, and overgrowth samples are 220, 303, and , respectively. As shown in part (b), the rocking curve of the nanocolumns sample can be decomposed into two components, including the broader one from the nanoclumns and the narrower one from the template beneath. Their FWHMs are 629 and , respectively.
Depth-dependent rocking curve FWHMs of the control and overgrowth samples based on the three-beam XRD measurement in the (01-13)/(0-11-2) plane. Near the top surface of the overgrowth sample, the crystal quality is significantly improved when compared to the GaN template.
Depth-dependent edge dislocation density, screw dislocation density, and lateral domain size of the overgrowth sample based on the two-beam XRD measurement. The crystal quality represented by those values near the overgrowth surface is significantly better than that of the control sample, which is described by the values of , , and 824 nm for the skew dislocation density, edge dislocation density, and lateral domain size, respectively.
Temperature-dependent integrated PL intensities of the nanocolumns, control, and overgrowth samples. The numbers indicate the ratios of the integrated intensities at 10 K over those at room temperature for the three samples. Among the three samples, the nanocolumns sample has the highest emission efficiency.
Normalized PL spectra of the three samples at 10 K (a) and room temperature (b). The PL spectral features of the nanocolumn sample are significantly redshifted from those of the control and overgrowth samples.
Normalized Raman-shift spectra of the three samples at room temperature. The lower and higher Raman spectral features of the nanocolumn sample correspond to the signals from the nanocolumns and the template beneath, respectively.
Comparisons of AFM and XRD measurement data between the overgrowth and control samples.
Article metrics loading...
Full text loading...