soapbox

So, I’ve still got some measurements to take tomorrow to confirm Feng’s results from yesterday, but it looks like the Schottky gates used to form the quantum dot on the most recent chip sent by our collaborators are giving huge (>100 nA) leakage currents to the two-dimensional electron gas (2DEG) layer approximately 100 nm below the surface of the silicon-germanium (SiiGe) heterostructure. And yes, 100 nA is a huge current (approx. 16 trillion electrons per second) when you’re in the business of counting single electrons in real time.

Hopefully, this is just a local anamoly with this particular chip, and not a more systematic concern. We have some eveidence for this being the case, since they’ve gotten more promising results with the same sort of devices before. Still, though, this is a bit alarming, since the signals we’ll eventually want to detect are on the order of several or tens of picoamps. Any experimentalist can tell you that signal-to-noise ratios less than unity are bad.

Anyways, the materials science remark in the title is mostly aimed at the specific defects occuring in SiGe heterostructures that causes this problem. Apparently, point defects, such as vacancies in the crystal lattice, propogate upward through the heterostructure as additional layers are epitaxially grown, resulting in large pyramidal pits in the surface, some of which cut straight through to the 2DEG layer underneath. Trying to put a metallic Schottky gate on top of one of these pits leads to instant failure of the Schottky barrier (as there’s no semiconductor between the 2DEG and the metal) and results in the sorts of large leakage currents Feng and I have been measuring. This is a well-known problem and has historically been difficult to fix or work around.

As it stands with our work, the collaboration has opted for the most common solution: make more to account for the lower yield of usuable samples caused by this problem. At one point, we investigated using a thin insulating layer of alumina (a.k.a. aluminum oxide or sapphire, if single crystal) before evaporating metal for the Schottky gates, but I think this led to some other problems. We’ll probably keep going about it the way we have been, but, for any of y’all with brilliant suggestions, the floor is now open.

this is a placeholder post for Leslie’s post (which she said she’d write either before or during colloquium).