Sample

For STM, the sample must be electrically conductive. In a lab setting, this would typically be Highly Oriented Pyrolytic Graphite (HOPG). HOPG is extremely flat — which prevents the tip from crashing during lateral movement — and regular enough to serve as calibration reference, using the known diameter of carbon atoms. However, as you might expect, it is not a common item and therefore tends to be expensive.
Vítor Barbosa reported resolving carbon atoms using carbon transfer paper, the kind used in handcraft work for copying drawings onto fabric. This is surprising, since it is not designed to be that precise and contains contaminants like clay to improve its structural bonding. Pencil leads use a similar mixture and are actually cheaper — and famously useful for making graphene.
I decided to make something a bit more interesting: heating sugar until only the carbon remains (example here). The result looks more regular, is pure black, and is even cheaper if you have sugar and a kitchen stove.
It also makes for some Breaking Bad references. I will introduce it on video with “[HOPG] is expensive, so I decided to cook some crystal. It should be around 99.1% pure.”
Another experiment I’d like to try is levitating the sample using an array of opposing magnets (like this).