After the nanowire sample was drop-cast onto the e-chip devices, as shown in the schematic in Figure 1, any nanowire that spans the 3 m holes in the support film was contacted to the Au electrodes of the e-chips using Pt lines deposited in the FIB-SEM. Pt has been shown to make good ohmic contact with the p-type doped side of the GaAs nanowire and the Au catalyst on top of the n-type doped section of the GaAs nanowire diode.13-15 Figure 2a) shows a low-magnification SEM image of a GaAs nanowire diode after deposition onto the echips device. The four Au contacts acre clearly shown at the four corners of the micrograph. Figure 2b) shows a magnified view of a region where a nanowire spans the hole in the support film. The deposited Pt contact lines can be seen in this SEM micrograph as the bright features connecting the two ends of a nanowire diode to the Au contacts of the device (not shown in Figure 2b). The I-V measurement for this nanowire diode is shown in Figure 2c), confirming that good electrical contacts to both sides of the diode is established. We note that the diode properties (ideality factor) for this specific nanowire device are very similar to those that have been measured in the SEM while still attached to the substrate. Notable differences are the increase in Ohmic resistance, potentially due to the Pt point contacts and connecting lines, as well as a possible conductive surface layer. Neither of these effects, however, will influence the EBIC data that will be discussed next.
Figure 3a) shows a low magnification high-angle annular dark field image, obtained in the aberration-corrected JEOL ARM200CF STEM. Here, the Pt contacts can be identified at the top and bottom of the nanowire diode, in addition to the Au catalyst at the top of the GaAs nanowire.
It is clearly visible in Figure 3a) that only one nanowire is connected by the Pt contact. An atomic- resolution HAADF image is shown in Figure 3b), revealing the atomic structure of the GaAs nanowire at the p-i-n junction (as identified in prior EBIC measurements). The GaAs nanowire sample was tilted towards the (211) orientation which was the closest major zone axis that could be reached by the holder tilt limited by the small gap in the ultra-high-resolution gap of the
ARM200CF’s objective lens pole-piece. It should be noted here that the GaAs columns, which are separated in this projection by 81 pm are not resolved. This is attributed to both the large sample thickness at the center of the nanowire and the high-current probe-size used for the STEM-EBIC measurements. The electron-probe size used for the measurements shown here is ~ 100 pm, but lower current settings can reach a probe diameter of better than 70 pm.16
Figure 3c) shows the STEM-EBIC signal from the nanowire diode shown in Figure 3a). The signal was acquired by placing the stationary electron probe at specific distances from the p-n junction, while simultaneously recording the current created by the electron beam. It is important to note here that reference data taken far away from the p-i-n junction do not lead to any current signal, and the resulting data is shown in the Supplemental Online Material. The data clearly shows that the EBIC signal, once the electron probe approaches the diode junction, is significantly larger than the noise of the measured current signal. This suggests that smaller electron probe currents, and thereby smaller electron probe sizes, can be used in future experiments to further optimize the spatial resolution of these measurements. Figure 3c) shows the measured STEM EBIC signal as a function of position across the p-n junction but from acquired from different distances with respect to the surface of the same nanowire. The STEM-EBIC signal from the center of the nanowire is significantly higher compared to the signal taken near the surface of the sample. The higher signal could be simply due to the larger sample thickness at the center of the nanowire or due to the presence of a passivation layer. As previously reported, there is a passivation layer on the surface of the nanowire formed either by an amorphous carbon or a native oxide layer (or a combination of both).7 The data taken near the surface of the sample contains a larger volume fraction of these surface layers, which results in a reduction of the STEM-EBIC signal.
For each position along the center of the nanowire as well as at the edge, a second set of STEM-EBIC data was acquired after an hour of electron beam exposure to assess the effects of the electron beam on the sample. We note that the STEM-EBIC signal is reduced in the second measurements, with a large reduction in the STEM-EBIC signal observed for the data taken from the region near the surface of the nanowire. Similar effects have been previously observed and were attributed to interstitial and vacancy defects created by the electron beam exposure.7 These point defects can act as deep-level traps within the band gap of GaAs, and increase the probability of e-h recombination through the Shockley-Read-Hall process.