Full text loading...
HAADF STEM image of a typical nanoribbon structure investigated in this work with a magnified view of the HEMT layers shown in the top inset. Bottom inset illustrates several HEMT nanoribbons in cross-section with metal contact and covered in passivating Al2O3. Indicated composition of the layers is based on the nominal growth values.
A 5.5 nm probe shown in (a) was used to produce NBED patterns along the  zone axis, such as the reference pattern seen in (b). The growth-direction strain profile obtained from NBED for the nanoribbon HEMT structure seen in (c) is presented in (d).
(a) HAADF STEM image of HEMT structure used to create GPA strain maps. The g-vectors (0002) and (011) are first selected from a FFT image (b). Raw phase and amplitude images are calculated from each vector and shown in (c)–(f). Strain maps are calculated for both the growth direction (g)  and the in-plane direction (h)  with the scale shown to the far right.
Growth-direction strain maps with spatial resolutions of (a) 5 nm and (b) 2 nm generated from the HAADF STEM image shown in Figure 2(a) with the strain scale indicated at the bottom. (c) Linescans are obtained from each image as indicated in (a) and integrated over the whole field of view.
A summary of strain profiles generated using GPA, NBED, and FEA simulations. The inset shows a scaled cross section of the HEMT structure. Comparing the two experimentally obtained strain profiles highlights the advantages of each technique used, namely: better spatial resolution of GPA and lower noise of NBED. Two curves obtained from FEA simulations demonstrate how the modified structure based on experimental measurements of the layers' compositions improve the fit of the strain profile with experimental data over that profile obtained from the nominal structure.
Article metrics loading...