Quandela makes over 1,000 spin-photon interfaces on a pilot production line

[ShadowRider91]

French quantum company Quandela, working with the C2N nanoscience centre and several universities, has shown it can fabricate more than 1,000 spin-photon interfaces on a foundry compatible semiconductor production line. That might not sound thrilling, but it is a big step for the field. Most quantum hardware results so far have come from what researchers call hero devices, a handful of carefully made samples that work brilliantly but cannot be reproduced at scale

A spin-photon interface is a small semiconductor structure that links a stationary qubit, stored in the spin of an electron, with photons that can carry that information elsewhere. The idea is that spins act as memory while photons move data around a network or between processors. That hybrid approach is one of the more promising routes to building larger quantum computers

The performance numbers are solid across the batch, not just in one lucky device. The team achieved seven-partite entanglement, with one spin entangled with six photons, and a spin-photon entangling fidelity of 93.5 percent. Photon indistinguishability averaged 95 percent and reached 98 percent in the best devices. Extraction efficiency was 80 percent, and spin coherence stretched to 1.94 microseconds, about two orders of magnitude better than before

The researchers put it neatly, saying the challenge is shifting from proving the spin-photon interface works to optimising and integrating it. That is what maturity looks like in any technology. You stop asking whether something is possible and start asking how to make lots of it reliably. The work is currently a preprint on arXiv with collaborators from France, Singapore, Denmark, Germany and Italy

This comes a day after we discussed Pasqal stepping back from a French defence programme, so France is clearly having a busy week in quantum. Do you think manufacturing is now the real bottleneck for quantum computing, more than the physics? And is 93.5 percent fidelity good enough to build on?

Quandela Builds Thousands Of Spin-photon Interfaces For Hybrid Quantum Computingquantumzeitgeist.com

[Inference Scott]

Moving from hero devices to a production line is the most important shift in quantum hardware right now. Every other industry went through the same transition. Transistors were hand made curiosities before they were mass produced. Quantum is finally starting down that road

[Andrew_17]

A hundred times better coherence is a big number. Spins losing their information too quickly has been a major limit for these systems

[Sigma98]

France seems to be positioning itself as Europe's quantum hub. Pasqal, Quandela, Alice and Bob and several others are all based there

The state has invested heavily and it shows. The UK should take note

[BigDog26]

Ninety three and a half percent fidelity is decent but not enough for large scale computing on its own. Error correction needs components well above 99 percent to work efficiently

Still, this is across a whole batch, which matters. Improving an average is easier once you have a repeatable process. The jump from here should be quicker

[DodgyDepot]

Foundry compatible is the key phrase. If these can be made using processes similar to existing chip fabs, costs fall and production scales far more easily than with bespoke lab equipment. That is how you get from research to products. The economics change completely. Suddenly quantum hardware can follow a cost curve

[NeverQuitRoss81]

Preprint again, so worth waiting for peer review. The numbers look plausible though, and Quandela has a solid track record of delivering on its photonic work

[Shaun5]

Seven photons entangled with one spin is the kind of resource you need for measurement based quantum computing. Building those cluster states reliably is one of the main hurdles for photonic approaches. Doing it with a chip made on a production line is impressive. It will be interesting to see how far that number can go