Fabrication of superconducting qubits

Hello there!

I might guess that you already know a lot about quantum computations and qubits. But if some of you have never been in a cleanroom and want to see, how superconducting qubits are actually fabricated, I have made several photos inside a cleanroom in the Neel Institute – CNRS in Grenoble and will share it with some basic comments.

Disclaimer 1: If you are already an expert in nanofabrication, you will not get much from this article. It is written mostly for those who are just curious about the topic.

Disclaimer 2: The precise recipes are different in different cleanrooms. In general, though, the algorithm will be similar. I will show one rather more simple which we use in the Neel Institute – CNRS in Grenoble.

Disclaimer 3: All the pictures inside the cleanroom were made on a properly cleaned device. Do not try to repeat it without permission.

So, let us start...

Our goal is in fact building aluminum structures with some of the parts less than 1 um on a silicon substrate. The critical element of these structures is called ‘Josephson junction’ and in our case it must be a sandwich of two aluminum plates with an aluminum oxide between them. For details see the picture below (See details here: https://en.wikipedia.org/wiki/Josephson_effect)

To reach this goal we will apply a Double angle evaporation principle: a mask used as a stencil will be put on top of a silicon wafer (a thin polished slice of silicon) and than a metal coating will be applied to the surface at two different angles.

Wearing special cleanroom suits we are entering the clean room and here is what we see there...

The light is actually yellow. Putting filters on all windows and lamps is needed to avoid a premature exposure of photoresist (the same idea as using red lamps for developing pictures in a darkroom).

I have brought my 2-inch silicon wafer in a special container. Next we are going to etch it with an oxygen plasma in order to clean its surface. To do so we put the wafer inside a Reactive Ion Etching machine, pump it and than switch it on.

After this procedure is finished and we get a wafer with a cleaned surface, we can start preparing the mask. We need to make a double layer mask to be able to use the double angle evaporation making the second layer offset from the first one. So, the plan is to make the first layer of an electron resist (a polymer which degrades after interaction with an electron beam) about 700 nm thick. See more: https://en.wikipedia.org/wiki/Electron-beam_lithography). And than make the second layer of a resist about 250 nm – the latter being more resistant to impact of an electron beam.

First we prebake the wafer on a special heating plate during 5 min (t=180°C).

Than we put the wafer on a surface of a centrifuge and add a drop of a resist on top of it, close the lid and rotate the wafer with a speed of 4000 rotations per minute during 30 seconds. As a result of this procedure the resist spread in an even layer of desirable thickness. We bake it again to make the polymer harder and after repeat the procedure to fabricate the second layer of a resist.

To control the precise thickness of both layers we use an interferometer.

After both resists are ready, we load the wafer into a lithographer which will 'draw' the mask with an electron-beam. As a result during the deposition stage the places which were irradiated will easily unstick in the solvent, but the remaining resist will stay. Another tricky part is this one: since the top layer of a resist is more sustainable than the lower one, we can destroy the lower layer without disturbing the top one. And after the deposition we will get bridges on the top layer resist over hollows in the lower layer.

Oh, it's a queue here! We should wait.

The lithographer is located inside another room which is even cleaner than the rest of this place.

Only specially trained people can work with this equipment. Now a staff member loads out the wafer in one of the chunks. After we will make the lithography using a Remote Desktop connection and take back our wafer during the next unload in 24 hours.

...

Next day we have come and taken the wafer with an exposed resist. Though we can not see nothing new with our own eyes, let us have a closer look after the depositioning stage.

(picture will be later)

We go to the chemical room to put the wafer with an exposed resist in a special solvent (methyl isobutyl ketone diluted in isopropanol 1:3) and than we crash it in the isopropanol to create the mask.

Now we have our mask! And we can see it.

We have just checked the mask quality with an optical microscope and are satisfied by the results. Now it is time to evaporate the metal. To do so we load the wafer in a double angle evaporator 'PLASSYS' and pump it for 2 days to reach a really high level of vacuum.

After the two days have passed we run the script which makes the aluminum evaporation by using electrons. First, we deposit 20 nm of aluminum on the wafer tilted by the angle +35°. Than we let oxygen in the chamber to form an oxide on the surface of the first aluminum layer. After we make another deposition with tilting -35°.

We retrieve the wafer with all the surface covered by aluminum. And now we need "to tear off the stencil" - meaning to make a lift-off.

We go back to the chemical room and submerge the wafer into the NMP (N-Methyl-2-pyrrolidone) for several hours. If some parts of the metal stick too hard, we remove them in an ultrasonic bath. After all stages being done we rinse the wafer with acetone, ethanol and isopropanol. Dry it by blowing with pure nitrogen – it is important do not let any liquid to stay on the surface.

Finally, the wafer is ready. We have aluminum structures on the silicon surface with Josephson junctions.

To see the junctions better we will use SEM – the Scanning electron microscope.

P.S. I would like to express my gratitude to Thibault CHARPENTIER, who has helped me a lot and taught me to work in the clean room, and to Karthik BHARADWAJ for making SEM pictures for me, when I haven't yet passed the SEM-training.

P.P.S. Sorry for grammatical mistakes in this text, if there are any. I will proofread it later, now it is time to go home. Thank you for your attention and feel free to comment.