Scandium (Sc)

In 1871 Dmitri Mendeleev published a description of an element nobody had ever seen. He called it eka-boron, placed it in a gap in his table, and specified what it would be like: an atomic weight near 44, an oxide of formula Eb₂O₃ with a density around 3.5, colourless salts, and — most usefully for anyone hunting it — an expectation that it would be found in minerals containing yttrium and erbium.

Lars Fredrik Nilson found it in 1879 at Uppsala, in exactly those minerals, while processing euxenite and gadolinite in pursuit of erbium. He obtained about two grams of the oxide and named the element for Scandinavia. He does not appear to have connected it to the prediction. His colleague Per Teodor Cleve did, within months, and wrote to Mendeleev to tell him.

The prediction that made the table a law

Mendeleev's arrangement was one of several competing schemes in the 1870s, and its gaps could easily have been read as evidence against it. What settled the argument was that three of the gaps were filled by elements whose properties matched his descriptions.

Gallium came first, in 1875, and Mendeleev famously wrote to Lecoq de Boisbaudran to tell him his measured density was wrong — which it was. Scandium was the second, in 1879. Germanium followed in 1886 and matched eka-silicon so closely that the argument was effectively over.

The scandium numbers hold up well. Its atomic weight is close to 45 against a predicted 44; its oxide is Sc₂O₃ with a density near 3.9 against a predicted 3.5; and the salts are indeed colourless, since the ion has no d electrons left to absorb visible light. The one thing Mendeleev did not anticipate is that his prediction would be verifiable at all in his lifetime.

Nobody produced the metal until 1937, and a pound of it at 99% purity did not exist until 1960.

Common, and nowhere in one place

Scandium's crustal abundance is around 22 parts per million. That is more than lead, vastly more than silver, gold or platinum, and comparable to cobalt. It is not, in any ordinary sense, a rare element.

It is nonetheless one of the hardest to buy, because it is almost never concentrated. The scandium ion has an awkward combination of charge and size: it is too large to substitute comfortably for aluminum and too small and highly charged to follow calcium, so it slots at trace levels into iron- and magnesium-bearing silicates and is spread thinly through ordinary rock. It also does not travel with the rare earths into the pegmatites where they accumulate, despite sitting at the head of that group.

Genuine scandium minerals exist and are extraordinarily scarce. Thortveitite, from Iveland in Norway, is the best known and has been mined in quantities better measured in kilograms than tonnes.

Practically all scandium therefore comes out as a by-product of processing something else: uranium tailings at Zhovti Vody in Ukraine, titanium dioxide pigment production, the red mud left by bauxite refining, and nickel-cobalt laterite operations — the Nyngan project in New South Wales being the first mine developed primarily for scandium rather than incidentally to it. World production has historically been a few tonnes a year, a quantity that would be an accounting error in any other metal market, and the price has been high and erratic in consequence.

Half a per cent that changes what you can weld

Add a few tenths of a per cent of scandium to aluminum and it precipitates as Al₃Sc — extremely fine particles whose crystal structure matches the surrounding aluminum lattice closely enough to sit coherently within it. Coherent precipitates are unusually effective at obstructing dislocations, and they are also remarkably good at pinning grain boundaries in place.

The strength increase is useful. The grain-boundary pinning is transformative, and it is the reason the alloy exists commercially at all.

High-strength aluminum alloys have a chronic problem with welding: the heat-affected zone beside the weld recrystallises, the carefully developed microstructure is lost, and the joint is substantially weaker than the parent metal. Designers work around this with rivets. Scandium suppresses recrystallisation, so an aluminum-scandium alloy can be welded and keep most of its strength across the joint.

Soviet aerospace exploited this first — welded aluminum-scandium components appear in MiG-29 airframes — and the Soviet Union had the advantage of scandium recovered from its uranium industry while Western manufacturers had essentially no supply. The alloys later reached bicycle frames, baseball bats and lacrosse shafts, where the marketing outran the metallurgy somewhat. The current serious interest is in additive manufacturing: Airbus developed Scalmalloy, an aluminum-magnesium- scandium powder designed for laser powder-bed fusion, where the extremely rapid solidification of the process suits an alloy whose whole benefit is fine grain structure.

Scandium is also, for a transition metal, remarkably light — closer to aluminum than to titanium — so alloying with it costs almost nothing in weight.

Two markets that do not need much of it

Solid oxide fuel cells run on a ceramic electrolyte that conducts oxide ions, conventionally zirconia stabilised with yttria. Substituting scandia raises the ionic conductivity substantially, which allows the cell to operate several hundred degrees cooler for the same performance — and operating temperature is the dominant constraint on how long the rest of the stack survives. Bloom Energy's units are the best-known commercial application.

The older use is lighting. Scandium iodide added to a mercury vapour discharge produces a spectrum unusually rich in emission lines spread across the visible range, which gives high efficiency together with a colour rendering good enough for film and television. Metal halide lamps of this type lit stadiums and film sets for decades before LEDs displaced them.

One stable isotope, and why that is the reason it is scarce

Scandium has exactly one stable isotope, scandium-45. Its neighbours have six and five respectively.

That is not a coincidence, and it explains the element's cosmic scarcity better than anything about Earth's geology does. Nuclei with an odd number of protons are systematically less stable than their even-numbered neighbours, because protons pair off within the nucleus and an unpaired one is less tightly bound. Odd-numbered elements therefore have fewer stable isotopes — often one, never more than two — and correspondingly fewer ways for stellar nucleosynthesis to end up making them.

Plotted against atomic number, the abundance of the elements zigzags for exactly this reason, and scandium sits in a conspicuous trough between two peaks. It is scarce in the solar system before it is ever scarce in a rock.

Scandium-46, with a half-life of about 84 days, is produced by neutron irradiation and used as a gamma-emitting tracer in oil refineries and pipelines to follow flow and locate blockages.