Praseodymium (Pr)

For forty-four years the periodic table contained an element that did not exist. Carl Gustaf Mosander announced didymium in 1841, having separated it from the cerium fraction he was picking apart in Stockholm. He named it from Greek didymos, twin, because it appeared to be lanthanum's inseparable companion — always present, never quite distinguishable.

Didymium was accepted. It appeared in textbooks and in Mendeleev's table, it had an atomic weight, salts and a characteristic absorption spectrum, and manufacturers sold didymium compounds. The first crack came in 1879, when Marc Delafontaine and Paul-Émile Lecoq de Boisbaudran pulled samarium out of it, which showed didymium was at least contaminated. Nobody suspected that what remained was two more elements in almost equal parts.

A hundred crystallisations, and two colours

Carl Auer von Welsbach settled it in Vienna in June 1885. Working with the double salt ammonium didymium nitrate, he crystallised it, took the fractions, recrystallised each of them, and repeated the operation more than a hundred times. Slowly the material sorted itself into two groups: one whose salts were green, and one whose salts were pink.

He named the green fraction praseodymium, from prasios, leek-green, plus the old name — the green twin. The pink one became neodymium, the new twin. It took several more years for the chemical community to accept the split, because the evidence was a colour separation achieved after a hundred tedious operations that nobody wanted to repeat.

Why rare-earth colours are so pure

The colours that let Auer see what he was doing come from transitions between 4f orbitals, and those transitions behave differently from the ones that colour transition-metal compounds.

A copper or cobalt ion gets its colour from d electrons in the outermost occupied shell, exposed to whatever surrounds them. Change the ligands and the colour shifts — copper sulfate is blue, copper chloride solutions are green. The 4f electrons of a lanthanide sit beneath filled 5s and 5p shells and barely feel their surroundings at all. Their absorption bands are consequently very narrow, almost line-like, and they land at essentially the same wavelength whatever compound the ion is in.

That is why praseodymium's green and neodymium's pink-violet are so distinctive, so consistent between salts, and so useful as identification. It is also the basis of every rare-earth laser and phosphor: a sharp, well-defined energy level that does not smear out when you put the ion in a crystal.

Didymium outlived its own disproof

The odd afterlife of this story is that didymium is still sold, still by that name, and still as a mixture — because for one purpose the mixture is exactly what is wanted.

A glassblower working at a torch faces an intense yellow-orange glare from sodium in the hot glass, concentrated almost entirely at the sodium D lines around 589 nanometres. Didymium glass absorbs a narrow band right there. The flare disappears, the glassworker can see the shape of the work through the flame, and the rest of the visible spectrum passes through almost unaltered — a filter that removes one colour rather than dimming everything.

Most of the sodium-line absorption is neodymium's contribution; praseodymium extends the protection further into the infrared and shifts the overall balance. Nobody separates them for this application, because there is no reason to. Lampworkers' safety spectacles, some welding filters and certain photographic filters are all specified as "didymium", which means that the name of a disproved element from 1841 remains a live product designation in a catalogue today.

Sold as NdPr, because separating them is not worth it

The same logic dominates the modern market. Praseodymium and neodymium sit adjacent in the series, are the hardest pair in the light rare earths to separate cleanly, and both work in permanent magnets: Pr₂Fe₁₄B has a slightly higher anisotropy field than the neodymium compound and a slightly lower saturation magnetisation, and substituting some praseodymium for neodymium changes magnet performance very little.

So the industry does not bother. Mixed NdPr oxide and NdPr metal — typically about three-quarters neodymium — is quoted, traded and shipped as a single commodity, and a large share of the world's magnets are made from it. This is a rare case of the chemistry that frustrated nineteenth-century analysts being simply accepted rather than overcome, on the grounds that the separation costs more than it is worth.

Where praseodymium is used alone, it is usually for colour. Praseodymium-doped zirconium silicate is the standard yellow stain for ceramic tiles and sanitaryware, chosen for its stability in a glaze firing. Praseodymium also goes into aircraft-grade magnesium alloys, into the carbon electrodes once used for studio and projection arc lighting, and into fluoride glass fibre amplifiers operating at 1.3 micrometres — a technology that worked but lost the telecommunications market to erbium at 1.55 micrometres, where fibre loss is lower.

Two thousandths of a degree above absolute zero

Praseodymium has exactly one natural isotope, praseodymium-141, which is unusual for a lanthanide and means its atomic weight is known with great precision. That single nucleus has a specific physical use at the extreme cold end of experimental physics.

Adiabatic demagnetisation cools by aligning magnetic moments in a strong field, removing the heat that produces, then switching the field off so the moments randomise and absorb energy from their surroundings. Using electron moments gets to millikelvin. Going lower requires using nuclear moments, which are thousands of times weaker and correspondingly harder to align — copper stages need enormous magnets and long precooling.

Praseodymium-141 in the intermetallic PrNi₅ offers a shortcut. The praseodymium ion's 4f electrons have no net moment in the crystal field, but they polarise in an applied field and amplify the magnetic field seen by the nucleus by a factor of ten or more. The nucleus therefore behaves as if it had a far larger moment than it does, a phenomenon called hyperfine enhancement, and can be polarised with a modest magnet at a modest starting temperature. Nuclear demagnetisation stages built from PrNi₅ reach a fraction of a millikelvin, and they are the standard first stage in the cryostats used to study superfluid helium-3 and other phenomena that only appear when almost all thermal motion has been taken away.