One the biggest challenges for quantum computers is the incredibly short time that qubits can retain information. But a new qubit from Princeton University lasts 15 times longer than industry standard versions in a major step towards large-scale, fault-tolerant quantum systems.
A major bottleneck for quantum computing is decoherence—the rate at which qubits lose stored quantum information to the environment. The faster this happens, the less time the computer has to perform operations and the more errors are introduced to the calculations.
While companies and researchers are developing error-correction schemes to mitigate this problem, qubits with greater stability could be a more robust solution. Trapped-ion and neutral-atom qubits can have coherence times on the order of seconds, but the superconducting qubits used by companies like Google and IBM remain below the 100-microsecond threshold.
These so-called “transmon” qubits have other advantages such as faster operation speeds, but their short shelf life remains a major disadvantage. Now a team from Princeton has designed novel transmon qubits with coherence times of up to 1.6 milliseconds—15 times longer than those used in industry and three times longer than the best lab experiment.
“This advance brings quantum computing out of the realm of merely possible and into the realm of practical,” Princeton’s Andrew Houck, who co-led the research, said in a press release. “Now we can begin to make progress much more quickly.”
The team’s new approach, detailed in a paper in Nature, tackles a long-standing problem in the design of transmon qubits. Tiny surface defects in the metal used to make them, typically aluminium, can absorb energy as it travels through the circuit, resulting in errors in the underlying computations.
The new qubit instead uses the metal tantalum, which has far fewer of these defects. The researchers had already experimented with this material as far back as 2021, but earlier versions were built on top of a layer of sapphire. The researchers realized the sapphire was also leading to significant energy loss and so replaced it with a layer of silicon, which is commercially available at extremely high purity.
Creating a clean enough interface between the two materials to maintain superconductivity is challenging, but the team solved the problem with a new fabrication process. And because silicon is the computing industry’s material of choice, the new qubits should be easier to mass-produce than earlier versions.
To prove out the new process, the researchers built a fully functioning quantum chip with six of the new qubits. Crucially, the new design is similar enough to the qubits used by companies like Google and IBM that it could easily slot into existing processors to boost performance, the researchers say.
This could chip away at the main barrier preventing existing quantum computers from solving larger problems—the fact that short coherence times mean qubits are overwhelmed by errors before they can do any useful calculations.
The process of getting the design from the lab bench to the chip foundry is likely to be long and complicated though, so it’s unclear if companies will switch to this new qubit architecture any time soon. Still, the research has made dramatic progress on one of the biggest challenges holding back superconducting quantum computers.
