The stone that lights up when you rub it
Someone I know stood in a dark room with two milky quartz pebbles and rubbed them hard against each other. They flashed. Yellow. She had called it sparks first, then stopped and corrected herself: not sparks. A flash. A glow, the whole pebble briefly bright.
That correction is the interesting part. Sparks are hot. This is not.
It has a name: triboluminescence. Light from mechanical action. Crush sugar, peel tape in the dark, rub two quartz pebbles, and you can get light with no heat at all.
What is happening, as far as anyone knows: when a crystal with no centre of symmetry is strained or ground along its own faces, charge separates. One face ends up positive, the other negative. When the charge recombines, it discharges through gas at the contact and excites it. The emission is nitrogen's: the N2 spectrum, measured by Zink and coauthors in 1976. So the light you see is not only the stone. Part of it is the gas at the surface, lit up.
The gas has to be there for that route. Take it away and the charge has nothing to jump through. There is a second route to light in quartz, and I get to it further down.
This is a cousin of piezoelectricity, not the same thing. Quartz is piezoelectric: squeeze it and charge appears across the crystal. That is how a quartz watch keeps time. Triboluminescence is the messier relative, where the crystal fractures or grinds and the separated charge jumps a gap instead of running down a wire.
It is not fully understood. That is the honest state of the question, and I would rather say so than dress up a guess.
Why quartz does it: it has no centre of symmetry, and it does not conduct well. Those two things together let charge build until it jumps. You will see it said that about half of all crystalline materials show it in some form. That number is inherited, not measured. It traces back to a 2013 paper about fracture models (Chandra, Chandra and Jha, J. Lumin. 135:139), not to a census, and the reviews that repeat it cannot agree on it: one says nearly 50% of inorganics and 30% of organic solids, another lists five different percentages in a single sentence. The only real counts I can find are small and targeted. Chandra and Zink surveyed 45 metal sulfates and found 36 of them lit up: 80% of one family, not half of everything. So read the big number as estimated, not observed.
One correction, because I wrote the first version wrong. No centre of symmetry is the usual route, not the gate. A reader pushed back with a centrosymmetric crystal that flashes anyway: hexakis(antipyrine-O)terbium(III) triiodide, structure confirmed centrosymmetric at 160 K and brilliantly triboluminescent (Clegg, Bourhill and Sage, Acta Cryst E 58 (2002) m159). Doped fluorites do it too, and their lattice is centrosymmetric as well. The likeliest reason is local asymmetry: an impurity or a defect breaks the symmetry right where the charge needs breaking. That fits the wider state of the question, in their words rather than mine: triboluminescence has several possible origins, and no single explanation covers all of it.
How anyone knows the gas is involved, the part that convinced me: change the gas and the colour changes. Run the rub in a low-pressure chamber, or swap the air for argon, and the nitrogen blue drops out while a different emission takes over. If the light came only from the stone, the atmosphere would not matter. It does.
Humidity is a separate, in-air effect. Water vapour adsorbs on the surface and leaks the charge away before it can build. Dry air flashes, damp air often will not. That is an air-phase knob: it tells you about the air, and only about the air.
The test, if you want to run it on your own stone:
Two pebbles of translucent milky quartz. Not one. Rub them together. Bone dry. Damp air conducts the charge away before it can flash. That is why the test says dry, and why a humid day can kill it. A dark room, and five minutes for your eyes to adjust. The flash is faint. Look at it with daylight eyes and you will see nothing and call it a failure. Rub hard and fast. Slow grinding gives nothing. Watch the contact point, not the whole stone.
No flash is a normal result, not a broken stone. Humidity, opacity, and how a particular pebble fractured all matter.
One more thing, because it separates this from a spark: the flash is reported underwater too, where no air spark can form. The demo is Dietrich Zawischa's, a physicist at Leibniz University Hannover, and it is one demo: a science radio piece repeats it and credits him. His wording is the whole source: no sparks, and under water the effect is the same. I have found no published measurement of it, and I am not going to write that it flashes steadily when nobody has timed it. What the tank settles is the spark. It does not settle the gas.
And gas is not the only way quartz makes light. Chapman and Walton (J. Appl. Phys. 54, 5961, 1983) fitted the flash from a crack tip in quartz to a blackbody at roughly 2800 K, a thermal emission from a zone about a nanometre wide. That one needs no gas at all, and it would work underwater just as well. So the tank rules out a spark and leaves me with two candidates that both survive: gas discharging at the contact, or heat at the crack tip.
So the gas has two possible places to come from: a thin film still clinging to the stone, or gas dissolved in the water itself, released at the crack when the crystal fails. Humidity cannot separate those two. Underwater there is no air humidity to vary. They separate on a clock.
That gives the test I want to run: rub the pebbles under water and time it. If the flash fades over ten to sixty seconds, that is a finite pocket being spent: a film washing off, or air trapped in the rough contact draining away. If it holds, something is still feeding gas to the contact, and the water is the source.
The sharper version: degas the water. Boil it, cool it under a sealed film so no air gets back in, and rub again. Zawischa runs the same trick on sugar: degas it in vacuum with no nitrogen around and the triboluminescence dies. If the quartz weakens in degassed water, the gas is coming out of the water. Then bubble argon through the water instead of air and the nitrogen blue should drop.
One hedge I keep until those runs are done: I have not checked whether the underwater light is still mostly nitrogen or picks up the crystal's own thermal emission, the blackbody Chapman and Walton fitted. In air, the gas-swap runs say the light is mostly the air. Underwater, I will only say that once the degassed and argon runs agree.
I have never seen anyone try this on a river cobble. It gets done with sugar and tape as a party trick, and it gets skipped on the stone in your pocket. That seems backwards to me. The cobble came out of a river with a story; this is one cheap way to hear part of it.
If you try it, tell me what you saw. Colour, where it flashed, whether it worked at all. And if it failed, that is data too. The first read is free at the desk.