The Detector That Goes Quiet

Open another way through, and one counter stops.

To see how this can happen, imagine a small experiment assembled from a source of light, mirrors, two balanced beam splitters, and two detectors. Each splitter divides an incident beam equally between its outputs. We will use an ideal version first, with perfect components, so that imperfections do not conceal the question.

The source sends single photons into the apparatus. A beam splitter gives the light two possible routes. Mirrors bring those routes together at a second beam splitter, beyond which sit the detectors. Each detection registers as one event: a click, or a number added to a count.

Call the detectors A and B.

Block one route. Some photons are absorbed by the obstruction. Those that reach the second beam splitter are equally likely to register at A or B. Over many repetitions, both counters rise at approximately the same rate.

Now remove the obstruction. Both routes are open. Adjust their relative timing with the care this experiment requires.

Detector A receives the photons.

Detector B goes quiet.

B has not been switched off. Put the obstruction back, and it clicks again. Nor have the missing detections been lost from the total. With both routes open, they appear at A.

The oddity is that a route capable of contributing detections at B, when available together with the other route, can help prevent detections there.

If you imagine each photon as a little bead choosing a road, while the other road has no part in what happens, this is difficult to explain. Opening an unused road should not cancel the arrival of the bead.

Waves can cancel. A crest can meet a trough. But the light arrives in individual detection events, even when the source is arranged to send photons one at a time.

The experiment requires us to keep both facts.

Real experiments have losses, background counts, and imperfect alignment. Their dark outputs are judged through measurements rather than declarations of absolute darkness. In 1986, Philippe Grangier, Gérard Roger, and Alain Aspect reported experiments that brought the relevant behaviors together: light with a strong single-photon character at a beam splitter, and interference when the alternatives were recombined.

The clean experiment imagined here is a way into those measured facts. It is not a story about a photon deciding to surprise someone.

What makes it strange is more demanding: a rule that works extraordinarily well requires a different account of possibility from the one we use for roads and beads.

We will learn that rule. By the end, the silent detector should feel less like an inexplicable trick. You will be able to say what makes it quiet, what can make it click again, and which questions that explanation leaves unanswered.

Along the way, the same theory will give an atom somewhere to stop falling, give matter its patterns, let particles cross barriers, and place a sharp limit on what two distant systems could achieve by carrying separate local instructions.

The surprising darkness belongs to the theory that helps explain why there is a visible, durable world at all.

Quantum mechanics is often introduced as permission for anything to happen. Its actual power comes from being exacting about what can happen, with what probability, under which conditions. Sometimes it allows something a familiar picture forbids. Sometimes it forbids something that picture freely allows.

The empty counter is evidence of a prohibition.

Our story begins earlier, when people were trying to account for a much more familiar kind of light: the glow of something hot.