A limiting factor in conventional 4D STEM technology is the speed at which diffraction images can be acquired at each pixel in the STEM image. The speed is restricted by the frame rate of the conventional CCD cameras (30 fps for 512 x 512 images) and the speed by which 4D STEM datasets can be written and saved on disk. Such slow speeds limit the amount of data researchers can collect in a given timeframe and are not practicable for beam sensitive specimens with limited allowable dose before they are damaged or structurally transformed.
Materials and Methods
This experiment utilized cross-section specimens isolating individual ZnO nanowires (NW). As previous publications show, the growth direction of these nanowires on a GaN substrate can be guided as a function of substrate Ga-ion bombardment, which results in different local surface boundaries to partition NW heterojunctions into different segments and orientations.
Previously, HRTEM measurements determined tolerance of crawling nanowires to strain and lattice mismatch. However, it has been difficult to obtain conclusive quantitative strain measurements due to the beam-sensitive nature of the specimen and lack of a statistically adequate number of diffraction measurements from the substrate-NW interface region.
Employing a STEMx system with the OneView IS camera, two (256 x 85 x 1024 x 1024) datasets were obtained (each in less than two minutes) from NWs grown on bombarded and non-bombarded substrates. The OneView camera is the quickest and highest performance fiber coupled CMOS camera that takes 16 MP images at up to 25 fps. Using the STEMx system, the speed of the scanning probe is harmonized to the frame rate of the camera to collect 4D STEM diffraction datasets at full speed (up to 300 fps at 512 x 512 pixels).
In addition to the visual comparison of averaged diffraction patterns at diverse interface regions, 2D strain maps were calculated along orthogonal axes within the diffraction pattern and with reference to the unstrained GaN substrate. Using the STEMx Analytical Suite, the relative in-plane, out-of-plane, shear and rotation at the interface can be rapidly judged for large datasets (each >40 GB).
Overall, existing strain mapping results show the character and degree of lateral strain (ε_uu) more relaxed in directions parallel to the nanowire long axis through the ion bombarded regions (Figure 1) compared to the localized intense lateral strain points perpendicular to the same axis (ε_uu in Figure 2). Otherwise, the relative degree of compressive strain (ε_vv) was observed to be about the same in either case.
Figure 1. HAADF (top) of nanowire cross-section. Expanded view of CBED diffraction map taken over 256 x 85 pixel area corresponding to dotted outline and extracted using STEMx software shown below. Localized points of strain in the lateral (ε_uu) crystalline dimension and compressive strain (ε_vv) are shown below.
Figure 2. HAADF (top) of nanowire cross-section. Expanded view of CBED diffraction map taken over 256 x 85 pixel area corresponding to dotted outline and extracted using STEMx software shown below. Localized points of strain in the lateral (ε_uu) crystalline dimension and compressive strain (ε_vv) are shown below.
Further data collection is necessary to supplement these conclusions. Current experiments hint at new paths to explore understanding and controlling lateral epitaxy of nanocrystals. The implications of these results can be wide-ranging; touching on areas of future fabrication of nanowire-based devices. For instance, in combining top-down and bottom-up fabrication techniques or incorporating II-V and III-nitride semiconductors into industrial processes.
By using the STEMx system, the fast frame rate OneView camera was employed to drive the STEM probe on beam-sensitive ZnO NWs, grown on GaN substrates. High-speed 4D STEM data acquisition pointedly reduced the specimen damage previously encountered during data collection (saving time and money). The information obtained with the OneView camera offers high-resolution data and is capable of direct observation of single systems over large length-scales. Intelligent modeling of this data allows a visualization of nanoscale effects in an overall microsystem. Such results provide insights into the mechanisms of nano epitaxial growth and foundational information to include these systems into forthcoming technologies.
This information has been sourced, reviewed and adapted from materials provided by Gatan, Inc.
For more information on this source, please visit Gatan, Inc.