Cornell Team Cuts Tantalum Qubit Fabrication Temperature in Half With Krypton Gas

[Caitlin_69]

Researchers at Cornell led by Assistant Professor Valla Fatemi have found a way to deposit high performance tantalum films for superconducting qubits at just 200 degrees Celsius instead of the roughly 400 degrees that has been required until now. The trick was swapping the usual argon process gas for krypton during magnetron sputtering, which promotes the correct alpha phase crystal structure at that much lower temperature.

Tantalum has become one of the favorite materials for superconducting qubits because tantalum based devices tend to hold their quantum state longer than most alternatives, but there has been a real manufacturability problem underneath that promise. Most commercial semiconductor fabrication lines run with a hard temperature ceiling near 400 degrees Celsius to avoid damaging existing circuitry, which left almost no margin between what tantalum needed to form correctly and what an actual factory could tolerate without breaking other components.

Cutting the required temperature in half changes that math significantly. It opens a much wider process window that could let commercial foundries integrate high performance tantalum films into existing automated tool lines without damaging the CMOS control circuitry or interconnect layers sitting underneath, rather than requiring an entirely custom facility built around tantalum's old thermal demands.

The films produced this way reportedly achieved internal quality factors up to 16.9 million, described by the researchers as being right at the world leading edge for this kind of material, and the work built directly on the same group's earlier characterization techniques developed for niobium based qubits using a similar sputtering approach.

This is exactly the kind of unglamorous materials science that rarely makes headlines but quietly determines whether quantum computing ever actually scales into something manufacturable at commercial volume rather than staying a boutique lab curiosity forever

A new ingredient for quantum computing: krypton gas | Cornell Chroniclenews.cornell.edu

[RainyDayFund76]

The manufacturability angle here is genuinely the underrated part of the entire quantum computing story right now. Everyone gets excited about qubit counts and error rates in headlines, but if you cannot actually make the material at commercial scale using existing factory tooling none of those flashy numbers matter much at all in practice.

[Ann]

Just swapping one noble gas for a slightly heavier one in an existing sputtering process to solve a problem this significant feels almost too elegant to be true, but that is often exactly how the best materials science breakthroughs actually work out in practice.

[GhostRider89]

Curious what the actual cost difference is between argon and krypton at whatever industrial scale a real foundry would eventually need.
Krypton is genuinely rarer and more expensive than argon as a raw material, so I hope that tradeoff was accounted for properly in their broader manufacturability argument here.

[Nomad]

Building directly on their own earlier niobium sputtering work is a nice example of how these quantum materials labs tend to compound their own institutional knowledge over time rather than starting fresh with every single new material they decide to investigate next.

[HighKey15]

A 200 degree ceiling opening up compatibility with existing CMOS back end of line processes is honestly the single most important detail buried in this whole announcement for anyone actually thinking seriously about scale someday. That is the difference between a boutique lab process and something an actual chip fab could eventually run on existing equipment

[CollapseState75]

The internal quality factor number is impressive on paper but I would genuinely want to see this reproduced independently by a different lab entirely before getting too excited about it. Quantum materials results have a rough history of not always replicating cleanly outside the original lab that first reported them

[Matticus]

Funding coming partly from the Department of War's Microelectronics Commons Program is a pretty clear signal of how seriously the national security side of quantum computing is now being taken at the federal level, separate entirely from the private commercial race that gets most of the actual public attention