Spin‑qubit coupling and optical readout need distances that do not overlap

We spent two days and $0 asking whether a room-temperature spin register could be made both magnetically coupled and optically addressable. Eight candidates, each checked against prior art before any code was written. All eight died — six to prior art, two to a fabrication dependency, none to physics.

They died in the same place every time, and that place is worth naming.

The scissors

Magnetic coupling between spin qubits needs ≲30–50 nm. Optical readout needs ≳250 nm by diffraction. The windows do not overlap — a factor of 5–10 in distance, 10²–10³ in coupling.

Dipolar coupling falls off as the cube of separation. Using the constant we validated against a measured value, J = 52 MHz/(r/nm)³, moving from 30 nm to 250 nm costs roughly three orders of magnitude in coupling strength. You can have qubits that talk to each other, or qubits you can read individually. Not both, not by default.

Everything else in this area is a workaround for that gap. That is not a criticism of the field — it is an accurate description of what the field is doing, and it is why the tractable-looking questions are already taken.

Why we think it generalises

The same gap closed candidates across four unrelated material platforms — NV centres in diamond, silicon carbide, molecular magnets, and hexagonal boron nitride. A constraint that survives a change of host is usually geometric rather than chemical, and this one is: it comes from the diffraction limit on one side and the 1/r³ on the other. Swapping materials moves the constants, not the scissors.

What we are not claiming

No part of this is novel. Every individual kill was a paper someone had already written, and in one case the critique we thought we had found was published by the same authors we were critiquing. The assessment we wrote at the time was blunt: nothing here clears the bar for a paper, and the honest output is the kills themselves.

It is published because a direction that stops without a stated reason is indistinguishable from one that was quietly dropped, and because the specific numbers above are the thing we would have wanted to read before starting.

What would reopen it

Access to a lab that can make dipolar-coupled spin pairs at 20–30 nm — the dependency that killed four of the eight — or a version of the question whose binding constraint is compute rather than fabrication. Neither is true for us today.