How Long Does Light Bounce Inside a Mirrored Box?
How Long Does Light Bounce Inside a Mirrored Box?
This one started as pure curiosity, nothing more practical than that. Looking into it, I found plenty of other people had already chased the same oddly specific question. Strangely comforting, that. Finding out the rabbit hole already had company.
The question itself is simple enough. Take a box, mirror every inch of the inside, fire a single pulse of light through a gap, and imagine sealing that gap at the speed of light itself. What happens to the light after that?
Does it bounce between the walls forever? Does it fade out quietly? Does it end up as nothing but a trace of heat? Or does the premise itself, closing a box "at the speed of light," fall apart before the light ever gets a chance to answer?
I wasn't looking for anything useful when I started. I wanted to know what happens to light when it runs out of anywhere left to go.
A one-meter mirrored box sounds like a simple thought experiment, right up until you ask how good "mirrored" really is. The short answer: nothing built so far keeps light forever, though the best laboratory mirrors come remarkably close. What follows walks through why, using NIST's own measurements of how quickly trapped light actually disappears.
The Question Hides a Physical Trap
Here's the part the question quietly skips over. Nobody can close a lid "at the speed of light" with a hand, a switch, or a hinge. Light would already be gone before any mechanical part finished moving even a millimeter.
Light crosses a one-meter path in about 3.3 nanoseconds, tens of millions of times shorter than a blink. A 2023 review of blink physiology puts the full blink movement at roughly 250 to 450 milliseconds, though the part that actually blocks vision, the eyelid's occlusion window, is closer to 40 to 200 milliseconds. A blinking eyelid, next to that, might as well be standing still.
Light trapped inside a real mirrored box does not bounce forever. Even mirrors reflecting 99.9 percent of the light lose enough with every bounce that most of it disappears within a few millionths of a second, though the best laboratory cavities, built with mirrors reflecting 99.99 percent or more, have kept a measurable trace alive for tens of microseconds.
None of that happens the way the question implies, though, shooting light in and then slamming a door shut. Light is too fast for that. The honest version is closer to building a box that starts almost completely sealed, with a gap thin enough for one pulse to slip through, and closing that gap optically the instant the light arrives.
Simple-sounding physics questions have a habit of hiding a trap in the premise itself. Even the long-running dispute over who truly deserves credit for discovering the universe's expansion turns out to be far messier than the tidy textbook version suggests, once you start pulling at it.
Knowing that real mirrors leak is one thing. Watching exactly how fast that leak adds up, inside a box this small, is where the number gets uncomfortable.
The First Few Microseconds Inside the Box
In the first instant, the pulse crosses the box and reaches the far wall in about 3.3 nanoseconds. For the next few reflections, assuming the mirrors are smooth and the beam went in well aimed, the light travels close to a clean, narrow line, wall to wall.
That doesn't last long. A cube's six flat walls, each parallel only to the one directly facing it, have no curvature anywhere to pull a drifting beam back toward center, so diffraction slowly broadens it over many trips across. Add microscopic surface flaws and any tiny misalignment, and what began as a tight beam turns messier with every pass.
Most of the damage, though, comes from simple accumulation. Say the walls reflect 99.9 percent of the light on each hit, an excellent mirror by ordinary standards. That still leaves a tenth of one percent lost on every single bounce, and in a one-meter box, bounces come roughly every 3.3 nanoseconds, closer together still if the beam entered at an angle.
Run that loss forward, counting the mirrors alone, and about 63 percent of the light's original energy is gone in roughly 3.3 microseconds, tens of thousands of times shorter than a single blink.
That tenth of a percent is the mirror's entire loss budget, though, and it doesn't cover what the box itself adds. Light that spreads past the edge of a finite wall is simply gone. A slight misalignment walks the beam sideways instead of returning it. In air rather than vacuum, the gas absorbs and scatters a little more on its own.
So 3.3 microseconds is the optimistic figure, not the pessimistic one. Push the mirrors and the vacuum toward what real labs have actually built and the decay stretches out by more than an order of magnitude, still thousands of times shorter than the blink a person would use, without thinking, to describe something as instant.
All of that describes energy disappearing, not light slowing down, and the distinction is worth sitting with. The instinct to picture light gradually grinding to a halt, like a ball losing momentum on grass, is almost exactly wrong.
What the Billiard-Ball Intuition Gets Right, and Wrong
The billiard-ball picture isn't a bad instinct. Something really is draining away with every bounce. It's just aimed at the wrong quantity. What shrinks is the amount of light left, not its speed, and whatever survives a reflection keeps moving at exactly the pace it started with.
So where does the missing tenth of a percent actually go? Real mirrors do three unglamorous things. They absorb a sliver of the light as heat, let a whisper of it pass straight through the coating, and scatter a trace sideways off microscopic roughness on the surface. None of that shows up in a single reflection. It shows up fast once the reflections number in the thousands, which in a box this small takes almost no time at all.
There's also a quieter assumption worth naming: that a box full of trapped light would visibly glow if you peeked inside. It wouldn't, and not for the reason most people expect. Seeing into the box requires an eye, a camera, or a sensor, and every one of those is itself a small hole the light can escape through.
Watching the box, in other words, is never free. It opens its own small door. Researchers who measure cavity losses use that door on purpose. They read the faint light escaping through a mirror or a detector port rather than trying to look directly inside. In one 2016 NIST resonator study, the decay of that leaking light became the tool for measuring something else entirely: tiny birefringence effects in the mirrors themselves.
That still leaves the other half of the question open. How good can a real mirror get, once someone builds one on purpose rather than buying it at a hardware store? The number exists. It's just not attached to a box shaped anything like this one.
What Scientists Have Actually Measured
I couldn't find a published account of anyone building a literal cube, firing a single visible pulse into it, and clocking how long the light lasted. That specific experiment doesn't seem to be out there. What does exist is something more useful: real precision cavities built to answer nearly the same question under carefully controlled conditions.
The finding across those devices is consistent. Even genuinely excellent mirrors buy light microseconds, not eternity, and the more carefully the cavity is engineered, the further that number stretches without ever reaching forever.
The NIST resonator study used two mirrors rated at 99.99 percent reflectivity, spaced 1.55 meters apart and sealed against a turbomolecular pump, aimed at mid-infrared light with a wavelength of 4.53 micrometers. Measuring mirror birefringence meant first pinning down how long that cavity holds light, and fitting the decay signal gave a photon lifetime of about 46.9 microseconds.
That is the same quantity as the 3.3-microsecond figure above, the time it takes for roughly 63 percent of the light to vanish, so the two are directly comparable. The cavity beats the simple estimate by about a factor of fourteen, on the strength of better mirrors, greater length, and total round-trip losses low enough to land close to what the mirrors alone would predict.
It's the same instinct behind checking whether plants can sense Earth's magnetic field: an effect too faint to notice directly, chased down by building an instrument sensitive enough to catch it anyway. In 1997, a separate NIST team modeled an evanescent-wave cavity ring-down design meant to push the technique past gases and onto condensed matter, films and liquids included. The proposed monolithic total-internal-reflection minicavity would let part of the circulating light reach material sitting just outside the reflecting surface, with calculated decay times on the order of a microsecond.
That geometry was a feasibility study run as a wave-optics model, not a built device. The evanescent-wave version of ring-down it belonged to was already working in the same NIST lab around then, though, using a superpolished prism inside a conventional cavity to detect iodine at coverages below a single molecular layer.
Separately, Fraunhofer IOF uses this exact ring-down technique to characterize ultra-low-loss mirrors, with published measurements above 99.996 percent reflectance and a stated ability to resolve reflectance into the 99.999 percent range, across wavelengths from 355 to 1550 nanometers.
None of these devices is a cube. Every one is a carefully aligned cavity, engineered to keep light traveling the same narrow path instead of letting it spread the way a plain box would. That distinction is easy to miss, and it's doing more work than the reflectivity numbers usually get credit for. Which means those numbers were never going to settle, on their own, what a plain cube would do.
Would a Real Cube Even Come Close?
Lower the reflectivity and the effect is straightforward. A 99 percent mirror loses ten times more light per bounce than a 99.9 percent one, so the light disappears roughly ten times faster. Raise it toward 99.999 percent and survival time stretches out substantially instead, though a figure that high is normally quoted for one specific wavelength and angle, not for light in general.
Swapping ordinary air for a vacuum buys time too, since air molecules, water vapor, and dust all quietly absorb or scatter a portion of the light. At the microsecond scale, though, the mirrors' own losses usually outweigh whatever the air was doing.
I originally assumed reflectivity alone would explain most of the gap between a household cube and NIST's cavities. It doesn't. Geometry is a variable of its own, sitting alongside the coating rather than downstream of it, and the flat-walled case sits at the unfavorable end of it.
That isn't a guess on my part, as it turns out. The same 1997 NIST paper notes almost in passing that a resonator built from plane-parallel mirrors sits right at the edge of stability, carrying large diffraction losses precisely because nothing in the geometry refocuses the beam. Adding a single curved surface is what pulls such a cavity back into the stable regime and cuts those losses substantially.
A cube is the flat-walled case with no curved surface anywhere. A small tilt sends the beam walking sideways instead of returning it, and the diffractive spread has nothing working against it. A real box has seams and finite wall edges as well, and a beam that keeps widening gives that light more ways to find them.
Then add a wavelength mismatch. Coatings built for 99.99-plus percent reflectivity are generally optimized for a specific, narrow band, not the mixed spectrum of ordinary light. The household cube starts losing ground on nearly every front the lab cavity was built to control.
So: a genuinely built version of the box in the question, flat mirrored walls, no beam-shaping, air or an imperfect vacuum inside, would almost certainly fall short of what a vacuum-sealed lab cavity manages. How far short depends on things the question never bothers to specify. How wide the beam is. What wavelength it carries. How large and how continuous the mirrored walls really are, and how precisely they were aligned.
The flat-walled shape is a handicap from the moment the box is sealed, then, though not a guarantee that the light vanishes in a handful of bounces. None of that makes the cube a failure, exactly. It just means the real value of trapped light was never going to live inside a box shaped like the one in the question. It lives in the shapes scientists built on purpose, and in what those turned out to be good for.
Where This Kind of Trapped Light Actually Works
Nobody uses a mirrored box to store light the way you'd store water in a jar. That specific use case doesn't exist, and probably never will. But the exact mechanism this piece has been circling shows up constantly in instruments built to detect things too faint to see any other way.
The main one is called cavity ring-down spectroscopy. A laser bounces back and forth between two highly reflective mirrors many times over, and because the light effectively travels a distance far longer than the physical device holding it, even a trace gas that absorbs almost none of a given color leaves a measurable fingerprint on how quickly the light dies out. The decay that ruins the thought experiment is the whole point of the instrument: it sniffs out air pollutants and trace gases too faint for a single pass of light to ever catch.
None of this involves light actually slowing down, for what it's worth. That's a different, deliberately engineered trick physicists have pulled off under very different lab conditions elsewhere, a stranger puzzle in its own right.
On the commercial side, that same reflectance-measuring technique now qualifies mirrors for specialized laser, metrology, and photonics systems. An ordinary household mirror never needs that kind of precision, just enough clarity across every visible color at once to throw back a decent reflection.
A mirror that's 99.999 percent reflective sounds close enough to perfect that the missing thousandth of a percent should not matter. Over many thousands of reflections, though, that fraction compounds, and light escaping past the mirror's edge adds its own share on top. Perfect and almost-perfect look identical from a distance. Inside a cavity, they are different experiments.
What It All Comes Down To
Push the conditions to their theoretical limit: flawless, infinitely broad mirrors, a true vacuum, a seam that never leaks, and light's energy really would persist forever. Nothing in physics rules out that limiting case. But every real mirror falls short of that ideal in the same three small ways: some of the light warms it almost imperceptibly, some slips straight through the backing, and some scatters off a surface no polish can make perfectly smooth. In any box anyone could actually build, the light weakens, and most of what it loses ends up as an amount of heat too small to ever notice.
Chase the ideal case all the way to its edge, though, and another question shows up uninvited. Light behaves like a wave, and also like a stream of particles, and a stream of particles bouncing off a wall over and over should push on that wall, however faintly. Does light bouncing inside a sealed box exert a real, measurable force on the mirrors holding it? Would that force, however small, ever be enough to nudge the box itself?
One question closes and another opens before the first has even finished landing. I'm apparently not the only one who finds that kind of loose end more interesting than an answer.
Sources & References
- NIST — Speed of Light, Fundamental Physical Constants
- Fleisher, Long, Liu & Hodges (NIST) — "Precision Interferometric Measurements of Mirror Birefringence in High-Finesse Optical Resonators," Physical Review A 93, 013833 (2016)
- Willett, Maenner & Mayo — "The Perceptual Consequences and Neurophysiology of Eye Blinks," Frontiers in Systems Neuroscience (2023)
- Pipino, Hudgens & Huie (NIST) — "Evanescent Wave Cavity Ring-Down Spectroscopy with a Total-Internal-Reflection Minicavity," Review of Scientific Instruments 68, 2978 (1997)
- Fraunhofer IOF — Cavity Ring-Down: Reflectance Characterization of Ultra Low Loss Mirrors
Comments
Post a Comment