3 physicists asked what happens if you cut a photon in half

Pull a mirror out of the way while a single photon is halfway through bouncing off it, and you don’t get half a photon going one direction and half going the other. You get a rainbow.

That’s the claim from a trio of Norwegian physicists in a new paper (there’s an arxiv.org link), and I’ll admit the answer landed further from my intuition than I expected.

Start with what a photon is, because the whole result hangs on it. A photon is a single particle of light, and under normal conditions you can’t divide it. But it’s also not a particle in the tidy sense. It doesn’t sit at one spot. It’s an extended thing, smeared out in space and time.

Why the world isn’t a permanent acid trip

Splitting and merging photons isn’t part of daily life, and there’s a reason you should be glad. If shining one color of light through glass routinely split or combined photons, every surface would throw off colors that were never in the source. The universe would look like the most vivid legal LSD trip you can picture. It doesn’t. Hence, no LSD.

Photons do split and combine, but only under specific conditions. The medium the light travels through has to respond to the light itself. Do that and one color can fan out into many.

Physicists call the everyday version linear. Splitting and combining photons is nonlinear, and to force nonlinearity you usually need a very sensitive medium or a very intense source, like a laser.

Yanking a mirror away mid-reflection doesn’t look nonlinear at first glance. Sit with it for a minute and it clearly is. What hides that is the way we’re trained to picture single photons hitting mirrors.

The half-silvered mirror trap

Take a partially reflective mirror. Send one photon at it and the photon either passes through or reflects. We say it enters a superposition of both paths, with the odds set by how reflective the glass is. Put a detector on each path and the moment one clicks, the superposition collapses and the other path vanishes.

Both detectors never click at once. You never record half a photon each way.

Now the tempting move. Take a fully reflective mirror and rip it away partway through the reflection. By the same logic, the photon should sit in a superposition of reflected and transmitted, the odds set by when you pulled the mirror relative to the photon’s “size.” Try to measure it, the superposition collapses, one detector clicks.

That is not what happens. Once I stopped and actually worked through it, the naive picture was obviously wrong. Seeing why takes a little extra baggage.

Why a sharp edge costs you a rainbow

Time and frequency are two sides of the same coin. Play a note on a piano and there’s a time-domain picture, a steady pressure wobble that runs for a while. That note maps to a single frequency at a single amplitude. Messier sounds, chords or staccato, carry messier structure in time and need more frequencies stacked together, each with its own amplitude.

This holds for every time-varying signal, and then some. Growing up on the farm, I listened to AM radio on a tube set that was old even then, and the music kept getting chopped by a clicking noise. That was our electric fence zapping errant grass, a misbehaving sheep or a horny bull.

Each short sharp jolt of current threw off a short electromagnetic pulse, and an angry bull. A very brief pulse in time smears across a very broad slice of spectrum, including, to my irritation, the AM band. The shorter an event, the more frequency it takes to build. Flip it around and a single tone that never changes needs almost no spectrum at all.

Photons at mirrors obey the same rule. While the photon reflects, the field varies smoothly, turning the incoming wave into the reflected one. The transmitted wave doesn’t exist, so its amplitude sits at a contented zero.

Then you yank the mirror. The reflected wave’s amplitude drops to zero and the transmitted wave jumps up from zero. Two hard edges, and hard edges demand far more bandwidth than the original photon carried. The cut-off photon is still in a superposition of reflected and transmitted, but now it also has a sharp edge, and that edge takes a spread of photons at different frequencies.

Cut a photon in half and you make a rainbow. As far as I can tell, those new photons are themselves in a superposition of reflected and transmitted. But because there can be many of them, you could measure transmitted and reflected light at the same time. That’s the part the half-silvered mirror never allowed.

The experiment nobody’s run yet

Pulling this off in a lab will be hard. You’d need a source that fires single photons on demand with a very narrow spectral bandwidth, which stretches each one out in time so the extra photons from the cut are visible. Then you’d need to flip the mirror at exactly the right moment.

A bathroom mirror won’t do. The authors calculate the switch from reflective to transmitting has to happen in about 10 femtoseconds. A femtosecond is 10-15 s, and that’s too fast to physically move anything. Some materials, semiconductors among them, can be flipped from reflective to transmitting in 30 to 100 fs using ultrafast laser pulses as the trigger. The catch is the trigger itself. That laser pulse is loud, and filtering it out so you can see the faint photons made by cutting the long single photon will be a real problem.

We already have a hint this works. The same kind of mirrors get used to shorten ultrashort pulses, which means the reflected pulses come out carrying more frequencies than they went in with, so new photons must be getting made. We just haven’t caught it for single photons yet.

Give it about a year. I’d bet we will.

Physical Review Letters, 2026, DOI: 10.1103/94pm-hp34