Issue 1 | Physical Review Research, published 19 October 2026

Building large “cat states” of light on demand, with feedback instead of luck

Our understanding

Slide 1: Light in two states at once, every time? Status: proposal, a theoretical recipe for large cat states of light on demand, nothing built yet. Schematic: an orange and a white light wave drawn on top of each other with opposite phase.Slide 2: A cat state: two light waves at once. Schematic: wave A in orange and wave B in white overlap with opposite phase, labelled both at once.Slide 3: Today, big ones come by luck. Schematic: twelve attempts in two rows, eleven crossed out in red and labelled thrown away, one blue box with the two waves labelled kept.Slide 4: The proposal: a qubit steers, feedback corrects. A qubit between two mirrors with an orange light wave; orange light leaves to the right. A blue loop runs from the qubit to a measure box, then to a correct box, and back into the outgoing light.Slide 5: In their analysis, the simpler setup comes out ahead. Schematic bars without values for two criteria, average fidelity and non-Gaussian features: in both, the blue bar for this proposal is longer and marked higher than the grey bar for the three-mode scheme.Slide 6: Still a proposal. Open: how much survives real cavity loss and imperfect detectors. Schematic interference fringes: a grey dotted pattern before and an orange pattern shifted slightly sideways after a tiny push, the test the authors suggest. Link to anyon.news.
Slide 1 of 6

A cat state of light is a light wave that is in two very different states at once, named after Schrödinger’s famous cat. Such states are useful for sensing and for protecting quantum information. Today, large ones are mostly made by luck: you try many times and keep the rare success. The authors propose a recipe that works every time. A single qubit in a mirror box shapes the light, and a measurement is fed back to correct it. It is a proposal on paper.

The details

What they did

A superposition of two large, distinct light fields, often called an optical Schrödinger-cat state (our inference: the abstract says “coherent-state superposition”), is a resource for quantum sensing, quantum repeaters and bosonic error-correcting codes, but making large ones usually relies on probabilistic schemes. The authors propose a deterministic protocol: a qubit coupled dispersively to a light mode, as available today in cavity quantum electrodynamics, plus electro-optical feedforward. In their analysis the single-mode protocol beats an advanced three-mode scheme based on Gaussian operations and photon-number-resolving detectors, both in average fidelity and in quantum non-Gaussian features. They also propose the sensitivity of interference fringes to weak displacements as a conclusive witness of quantum interference.

Why it matters

Deterministic sources of large non-Gaussian states are one of the bottlenecks of optical quantum technology; a recipe built from existing cavity-QED ingredients is attractive. This is a theoretical proposal.

What is still open

What fidelities survive realistic cavity loss and detector inefficiency in an experiment? (our inference)

Get next Monday’s issue.

Every Monday, the week’s most important quantum papers, explained in five minutes. Free.

More from issue 1