Besides the facilities for electrically characterizing the nano-antennas under ambient conditions we can also do it in the vacuum environment of a SEM. An additional benefit is, that one can visually inspect if samples change due to dc fields. In the following I will show the setup and explain how it can be used to study electromigration on the nanoscale.
For the experiments we are using the Zeiss GeminiSEM 450 which we upgraded with Kleindiek MM3A-EM micromanipulators and a Keithley 2604B sourcemeter. The manipulators can be move by hand for a rough alignment but also electrically with nanometer precision. In order to be able to contact the electrodes we design a holder with a large cutout. For mounting everything the chamber of the SEM must be opened.
When everything is built in, the electrodes of individual antennas can be contacted, a voltage be applied and the current be measured. Being mounted within the vacuum of a SEM has the advantage that effects from the air or surface water layers can be excluded. Furthermore, when ramping up the voltage one can directly see at which point the antennas “change” .
Electromigration is the effect that within a conductor the material migrates when a current is running over longer times. This is cause by collisions of the moving electrons with the much heavier atoms and is mainly dependent on the temperature, material composition and the intensity of the current. In the semiconductor industry this is a very unwanted effect, but in nano-technology research it is quite suitable effect for creating nanometer gaps.
To achieve that, a small constriction is prefabricated in a wire such that the resistance and, hence, the temperature will be highest there. By very carefully ramping up and down the voltages over many cycles, the atoms can slowly be migrated out of this region until a single-atom bridge or nanometer-sized gap appears.
We tried to do the same with our nano-antennas and asked an experienced gap migrator for help. Unfortunately, it did not work on his setup as the connectors always broke. Hence, I implemented a program for conducting electromigration within our SEM to have a vacuum environment and a close look at every step. The program worked quite nicely but the nano-world was quite strange: Instead of material moving out of the constricted region it actually went in!*
*This is most likely due to the single crystallinity of our structures. Usually, people deal with polycrystalline material which consists of many grains. When a currents flows, the electrons move from grain to grain and scatter slightly at the boundaries, this means they lose energy and the material heats up. This is not the case for our structures and the surface scattering on nanometer-sized constrains does not seem to make this up.