From Sensing to Computing: Organic Qubits

Long coherence. High fidelity. Fast gates.

¹

I catch The Quantum Dragon doing the darndest things, but this was a real headscratcher. You can see that he had blocked off all of his senses except for taste, because even fire-breathing red dragons drink coffee. And yet there he was, programming a quantum circuit while blind, deaf, anosmic, and hypoesthetic. Despite the risk to my personal safety, I had to ask why and tapped him gently on the shoulder.

Of course he screamed “WHAT?!?!?,” which I can take in stride at this point, before removing his headset and blindfold.

Quantum Sensing Origins

It turns out that he was inspired by “NVision Expands from Quantum Sensing to Quantum Computing to Accelerate Discovery and Validation of New Therapies.” NVision designs molecules that have chemical properties that push the complexity of a problem into the molecule’s design, as opposed to the surrounding hardware. Its POLARIS enhances signals, such as the nuclear spins in sugars, making them visible to MRI so that we can visualize sugar metabolism in cancers and determine if treatments are working or not.

In other words, NVision is generally recognized as a quantum sensing company. In case you’re wondering, it was named well before Nvidia became popular. Yes, I asked. The name NVision came from the company embracing NVC (nitrogen-vacancy centers, commonly known as NV centers) first, but it shifted after 2 years to organic molecules. Spun out of Ulm University in 2015, the company now has 140 employees representing 27 nationalities, and with 40% holding PhDs. The company has raised €110M in funding but, more importantly, is commercializing its MRI solutions.

So, it’s a real company is what I’m saying, not just another startup in the quantum computing space.

Enter Quantum Computing

And this leads us to why The Quantum Dragon was trying to do quantum computing without most of his senses. The story goes that after working on molecules for MRI, NVision found molecules that did not work well with MRI. Rather than simply discarding them, NVision found that the molecules could work well as long-lived, addressable, tunable qubits with intrinsic photonic interfaces. The polarized sugars for MRI, for example, have up to 5 minutes of spin lifetime at room temperature.

If you’re familiar with NVC qubits, harking back to NVision’s NVC origins, think that but using organic molecules instead of synthetic diamonds, swapping molecular lattices for NVC’s carbon lattices. Similar to NVC doping, you have an inert host with a photoactive carbene doped in. The tech has actually been in development for 4 years.

How It Works

The molecules embed a photonic integrated circuit (PIC) with a nanometers-thin organic film above the silicon electronics. The PIQC (“pixie”), as they call it, adds the “Q” for the quantum layer. This promises long coherence with high fidelity from the molecules and fast gates from the photons. A nuclear spin is the qubit, an electron spin is its photonic interface, and photons do the operations between the molecular qubits.

Because the photons are not used as qubits, are not memories, and are only used for entanglement, they can be lost. Compared to maybe 2% acceptable optical loss for photonic quantum computers, 70% optical loss is acceptable here. Attempted entanglement can be repeated until it is successful. This architecture allows the qubits to be kept in isolation, while also enabling connectivity.

Conclusion

Never tap a dragon on its shoulder. In retrospect, I’m lucky I escaped with my eyebrows.

Also, I mentioned nuclear spins as qubits. Like other spin qubits, the architecture promises hundreds of thousands, perhaps up to millions, of qubits on a single wafer. This would allow the full quantum computer to occupy one server rack. Furthermore, like other photonic interconnect architectures, distributed quantum computing should be native.

The key takeaway is that this really isn’t new. NVision has been applying its technology to quantum sensing for over a decade. It’s commercializing quantum sensing. And that makes this approach more interesting than novel approaches without such a backstory.

1. Image generated by Google’s language model AI.