Cornell's Krypton Fix: Tantalum Heat Drops From 400°C to 200°C

Four hundred degrees Celsius — that's the wall that's kept the world's best qubit material out of ordinary chip factories. Today on Quickly Quantum: a Cornell team says they've knocked that wall down to two hundred, using something you'd probably associate with headlights, not quantum computers — krypton gas. Does that actually get tantalum qubits into a real semiconductor fab, or is two hundred degrees still too hot? Before that, in the headlines: a Cambridge theory paper says 'magic' isn't the whole story for quantum advantage, Caltech physicists watched decades-old physics play out live on a quantum simulator, a three-billion-dollar SPAC merger takes its next regulatory step, and two senators want the power grid quantum-hardened. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Thursday, August 20, 2026. Let's get into it.

Here's a theory paper that quietly shrinks the space of things quantum computers can claim credit for — which, done right, is a good thing. Quantum algorithms run on what's called magic states, special starting setups that act as computational fuel; without them, a quantum computer is no better than your laptop. The old assumption was that if a state has magic, it's useful. A new study out of Cambridge's Cavendish Laboratory, published in Physical Review Letters, says that's not quite right. Lead author J.J. Thio and the team point to something called Kirkwood-Dirac negativity — a probability that's allowed to go negative, a concept tracing back to physicist Paul Dirac. When that value stays positive throughout a computation, a classical computer can simulate the whole thing on a laptop. When it goes negative, classical simulation gets exponentially harder — that's where a real quantum edge might live. As the Cavendish Laboratory's own account, @DeptofPhysics, put it on X, 'magic alone may not be enough. The key may lie in negativity: Kirkwood-Dirac negativity.' Physicist David Arvidsson-Shukur, from Cambridge's Hitachi Laboratory, put it more bluntly: 'magic is necessary but not sufficient to unlock quantum computers' full power.' No new hardware here — just a sharper ruler for what counts as genuinely quantum.

Thirty-five strontium atoms in a chain, held by laser tweezers, just confirmed something physicists have believed on paper for four decades. Caltech researchers, working with theorists at Paris-Saclay and the Technical University of Munich, used a neutral-atom quantum simulator — basically a stripped-down quantum computer built for one job — to directly measure the energy levels predicted by conformal field theory, the math that describes how wildly different materials behave identically at a phase transition. Think water boiling, or a magnet losing its pull: professor Jason Alicea calls it universality — the messy microscopic details wash out and only a few essential features survive. The team tuned their atom chain to the tipping point between two quantum phases and read off the energy 'rungs' predicted by the Ising and tricritical Ising models — the first time anyone's measured those rungs experimentally rather than just calculating them. Co-lead author Xiangkai Sun says the tools borrow directly from quantum computing hardware: 'we are at the point where we can use them to do fundamental physics research.' There's no commercial angle here — this is basic science using quantum hardware as a very precise microscope, and the team's already eyeing two-dimensional systems where classical computers may not be able to keep up at all.

Three billion dollars is the valuation attached to a company that, by its own account, is still pre-revenue. EigenQ, which builds post-quantum cryptography and quantum-safe infrastructure for government and defense customers, and its SPAC partner Silicon Valley Acquisition Corp, just confidentially filed a draft S-4 registration with the SEC — the next formal step toward taking EigenQ public. The pro forma enterprise value stays around three billion, backed by roughly two hundred fifteen million dollars sitting in the SPAC's trust account, with a target close in the fourth quarter and a planned Nasdaq listing under the ticker EIGQ. EigenQ's pitch rests on timing: federal agencies are racing to migrate off vulnerable encryption before quantum computers can break it — the so-called harvest-now-decrypt-later threat, where someone steals encrypted data today and waits to crack it open later. That's a real deadline problem for government buyers. What it isn't, yet, is a company generating revenue, and SPAC valuations in this sector have swung wildly before. A filed S-4 is a procedural milestone, not a done deal — there's real distance between paperwork and actually ringing the Nasdaq bell in the fourth quarter.

Diraq just planted its first American flag, opening a research lab in Chicago's Illinois Quantum and Microelectronics Park — the same hub we've noted this month as it tries to become a national quantum manufacturing center. Diraq builds silicon spin qubits, a different bet than the superconducting chips from Google and IBM: tiny electron spins trapped in silicon that, in theory, could ride existing chip-manufacturing infrastructure straight to millions of qubits on a single chip. The new lab, plugged into the park's 'On-Ramp' program at innovation hub mHUB, gives Diraq cryogenic refrigeration and measurement gear to complement its Sydney headquarters, and CEO Andrew Dzurak says the company plans to grow the Chicago team over the next year. The public target: a commercially useful system with many thousands of physical qubits by 2029. That's a roadmap claim, not a demonstrated result — a lab opening is infrastructure, not a technical milestone. But it does add another serious silicon player to a Chicago quantum cluster that's drawn both big names and, as we've noted, some local pushback over what residents are actually getting out of the deal.

Over on X, a technical thread got quantum engineers talking this week, breaking down a new arXiv preprint that claims something quantum simulation hasn't managed before: running honest nonlinear physics on real hardware. The account @FreeDeathTV walked through the paper — arXiv 2608.13041 — arguing that the team kept the nonlinearity intact, rather than using the usual workaround of blowing the equations up into a giant linear approximation, and ran the viscous and inviscid Burgers' equation, a classic testbed for fluid turbulence, through a hybrid quantum-classical loop, paired with a noise-mitigation scheme that let the circuits run deeper than hardware normally tolerates. @FreeDeathTV called it a breakthrough and framed it as a step toward what they called PQC — though the paper itself is about fluid dynamics, not cryptography, so take that framing with a grain of salt. Here's the caveat: this is a preprint, not yet peer reviewed, we haven't independently confirmed it beyond this thread, and by @FreeDeathTV's own account, the demos are small and short-lived, with modest engagement on the post itself. If nonlinear PDEs really are becoming tractable on today's noisy hardware, that's a bigger deal than another qubit-count headline — but that's exactly the kind of claim that needs replication before I'd bank on it.

Two senators, one bipartisan bill, and a power grid that's still running hardware from decades before anyone worried about quantum hacking. Chris Coons and Mike Rounds introduced the Quantum Grid Utility Assurance and Resilient Defense Act — mercifully shortened to Quantum-GUARD — which would direct federal agencies to assess grid vulnerabilities and help utilities migrate to the post-quantum cryptography standards NIST finalized back in 2024. Coons frames the urgency plainly: quantum computing brings new economic opportunity, he says, but also tremendous cybersecurity risk, and the bill also tasks the Federal Energy Regulatory Commission with evaluating the reliability risks quantum computing poses to the grid. But here's the part that actually makes this hard, and it's not the algorithms — a quantum cybersecurity executive at Qtonic Quantum, posting as @QtonicQuantum on X, laid out the real obstacle: protective relays, RTUs, and SCADA equipment run for decades, much of it fielded before post-quantum standards existed, and 'you do not swap the cryptography on a relay the way you update a laptop.' This is a bill introduction, not passed law — the harder work, funding and timelines for actually retrofitting substation hardware, still has to get worked out in committee.

Colorado got a ribbon-cutting this week, as Infleqtion officially opened its new global headquarters, the Colorado Quantum Innovation Center in Louisville, with Governor Jared Polis on hand. Infleqtion's own account posted on X that the center anchors what they're calling 'America's Quantum Peak,' built around Infleqtion's neutral-atom technology — individual atoms trapped and controlled with lasers, the same basic approach behind some of the field's biggest recent qubit-count milestones. The building itself is a photo-op, and I'll say that plainly. But the number buried in the announcement that actually matters is a planned 2027 field test of quantum gravity gradiometry — sensing tiny variations in gravity to map what's underground, aimed at subsurface mineral mapping. That's the commitment worth holding Infleqtion to, not the ribbon.

Our main story today: the four-hundred-degree wall, and whether a Cornell team just found the crack in it. Here's the setup. Superconducting qubits — the same basic technology behind Google's and IBM's quantum chips — need to be built from materials that lose almost no energy to electrical resistance. For the last few years, the industry's favorite material for that job has been tantalum. It conducts better than the niobium it's replacing, and it forms a stable oxide layer on its surface that cuts down on a specific kind of energy loss called dielectric loss — think of it as a metal that doesn't accumulate the same kind of microscopic grime that steals a qubit's coherence, the length of time it stays usefully quantum before decohering into noise. The problem is how you actually put tantalum onto a chip. The standard method is called sputtering: you fire ionized gas at a solid chunk of tantalum, knock atoms loose, and let them land on a silicon wafer to form a thin film. Do that with argon, the usual gas, and you need to heat the silicon substrate above four hundred degrees Celsius to get the tantalum to form the right crystal structure — what's called the alpha phase, the high-conductivity, qubit-friendly version of the metal. And four hundred degrees is a genuinely awkward number, because it sits right at, or over, the thermal ceiling that standard commercial chip factories are built around. Go much hotter than that during what's called the Back-End-of-Line stage of fabrication, and you risk cooking the control circuitry already laid down underneath. Worse, at those temperatures, tantalum and silicon start to intermix at the boundary, forming a messy layer that directly degrades how long a qubit holds its quantum state. So a team at Cornell, led by assistant professor Valla Fatemi, tried something almost absurdly simple: swap the gas. Instead of argon, they sputtered with krypton — heavier, and it turns out that matters a lot. Because krypton atoms carry more mass than argon, the ions slam into the tantalum target with more kinetic momentum, and that extra force is enough to stabilize the same high-conductivity alpha-phase structure at a much lower temperature. How much lower? Four hundred degrees down to two hundred — that's not a marginal tweak, that's cutting the required heat in half, and it moves tantalum from right on the edge of what a fab can tolerate to comfortably inside the window most commercial Back-End-of-Line tool lines are actually built for. The team, publishing in Nature Materials, didn't just hit the lower temperature — the resulting qubits held up. Transmon qubits, the standard workhorse qubit design, built from krypton-sputtered tantalum films with compact twenty-micron capacitor gaps, reached internal quality factors — a measure of how little energy a resonator loses per cycle, essentially a coherence scorecard — up to sixteen point nine million. That's a genuinely strong number for a film grown at a temperature this low. So the physics checks out. The question the field's been circling — Google, IBM, and Rigetti all lean on tantalum-based designs — is whether two hundred degrees is actually the number that gets superconducting qubits built on a real commercial fab line, or whether it's still too hot for the tooling that would need to do it at volume. That's where the deeper read starts.

Now, here's what makes this more than just a materials-science curiosity. Cornell's own write-up quotes team member Olszewski laying out exactly why four hundred degrees was such a dangerous number in the first place: 'There's this whole set of tooling and fabrication lines that don't really go above 400 Celsius, and they're built for that. And tantalum on silicon, when you use the old method, was right on the border of that... There was little margin to do things reliably. Using krypton brought that threshold down to 200 Celsius.' That's the whole ballgame in one quote — it wasn't that four hundred degrees was impossible, it's that it left engineers almost no margin for error on production lines that weren't designed to run that hot in the first place. The funding trail here is worth noting too. This work was backed by the Department of War's Microelectronics Commons Program and the Air Force Office of Scientific Research, alongside Cornell's own NanoScale Science and Technology Facility. That's federal money going directly at 'can we actually manufacture this stuff,' not just 'can we make it work once in a lab' — which tells you where the government thinks the real bottleneck in scaling quantum hardware sits right now: not algorithms, not qubit counts, but fabrication. This also isn't Fatemi's first swing at this problem. The lab's earlier work bombarded niobium — the material tantalum has been steadily displacing — with argon ions to push its performance, and that characterization and nanofabrication groundwork is what let the team move quickly once they had the idea to try a heavier sputtering gas on tantalum instead. Cornell's own account, @CornellNews, summed up the result on X: krypton gas slashes the deposition temperature of tantalum, producing thin films with higher electrical conductivity — 'key for quantum computing,' as they put it. So does two hundred degrees actually solve the fab problem? Here's where I want to be careful, because this is exactly the kind of result that's easy to oversell. The paper reports a materials and process improvement — a real one, backed by a strong quality-factor number and peer review in Nature Materials — not a new qubit count or a new error rate. Nobody's announced a chip built this way yet. No company — not Google, not IBM, not Rigetti, all of whom already lean on tantalum in their own qubit designs — has said they're adopting this specific krypton process. And 'two hundred degrees fits inside standard Back-End-of-Line thermal budgets' is true in general, but whether it clears every piece of legacy tooling at every fab that would actually need to build these chips at volume is still an open question the paper doesn't answer, because nobody's tried it at that scale yet. And that's really the whole skeptic case in one line: this proves the chemistry works, it doesn't yet prove the manufacturing works, and those are very different claims. A foundry line has hundreds of tools, and hitting two hundred degrees on one wafer in one deposition chamber at Cornell is not the same claim as running that recipe cleanly, batch after batch, on a commercial tool that also has to talk to everything else already sitting on the wafer. Here's my read: this is the good, unglamorous kind of quantum news — the kind that doesn't get a splashy qubit-count headline but quietly removes a real bottleneck between the lab and the factory floor. Materials and fabrication problems like this one are exactly what turns 'we did this once in a university cleanroom' into 'we can build ten thousand of these reliably.' Tantalum won its spot as the field's favorite superconducting material because it's simply better at holding quantum information than niobium was — less dielectric loss, a more stable oxide. But a better material that only a handful of university labs can deposit correctly isn't a scaling story, it's a curiosity. Cutting the deposition temperature in half is the difference between a process that lives in one cleanroom and one a commercial foundry might actually run. That's the direction this needs to keep moving in — from a single result in one lab toward someone, a fab, a foundry partner, one of the big superconducting players, actually building on it. Time for the Hype Check. I'm putting this one at a 7. The physics is real, it's peer-reviewed in Nature Materials, the quality-factor number is genuinely strong, and the temperature drop is exactly the kind of unglamorous fix the field needs. It loses points because it's a fabrication result, not a device or a company committing to use it — until someone announces a wafer built on a real commercial line with this process, the manufacturing story is still a promise, not proof.

If krypton sputtering really does clear a path onto commercial lines, the winners are the foundries and the superconducting programs — Google, IBM, Rigetti — who get a cheaper route to scale, and the losers are the qubit architectures whose whole pitch has been that theirs is easier to manufacture than superconducting will ever be. If you want that kind of analysis waiting for you every weekday, follow Quickly Quantum wherever you get your podcasts. This has been Quickly Quantum, an AI-voiced podcast, created and built by a real human using today's cutting-edge technology. Nothing you heard on this show is financial advice. I'm Brian Lampert, and I'll catch you all tomorrow — take care!