Lutetium (Lu)

Lutetium closed the rare earths. It was the last of the naturally occurring lanthanides to be separated, it sits at the end of the series, and its discovery in 1907 finished a project that had begun with a black stone from a Swedish quarry a hundred and twenty years earlier — a history covered on the terbium page.

It closed the series three times over, because three people did it independently within months of each other.

The three chemists who all had it in 1907

Georges Urbain in Paris took Marignac's ytterbia apart and reported two elements from it to the Paris Academy of Sciences on 14 November 1907, naming the heavier one after Lutetia, the Roman name for the city he worked in. Carl Auer von Welsbach in Vienna presented the same split to the Imperial Academy on 19 December, with detailed spectroscopic evidence, calling his cassiopeium. Each accused the other of having been influenced by the other's unpublished work.

Charles James, at the University of New Hampshire, had done it too. He had arguably the largest and purest quantity of the new element of any of the three, obtained by the fractional crystallisation methods he had already used to extreme lengths on thulium. On learning that Urbain had announced first, James did not publish his claim at all.

Priority went to Urbain in 1909 on the basis of the publication date, and lutetium is his name. Cassiopeium stayed in general use in German-language chemistry into the 1950s, so for four decades the same element had two names depending on which language you were reading. The spelling was standardised from lutecium to lutetium in 1949.

James got nothing at the time and a posthumous award named after him from the American Chemical Society. He is the clearest case in the rare earths of a chemist losing a discovery not to better work but to a slower decision to publish.

Does lutetium belong under yttrium?

Every printed periodic table places lanthanum below yttrium in group 3, with lutetium tucked away at the far end of the detached lanthanide row. There is a serious argument that this is backwards.

Lutetium's 4f subshell is completely full, so its outermost occupied orbital is a 5d — the configuration of a d-block element, and precisely what group 3 should contain. Lanthanum, by contrast, has no 4f electrons at all, which makes it an odd choice to head a series defined by f-orbital filling. Physical trends support the same reading: run melting point, density, ionisation energy and ionic radius down scandium and yttrium and the next value fits lutetium better than it fits lanthanum.

An IUPAC project convened to settle group 3 membership issued a provisional recommendation in 2021 favouring scandium, yttrium, lutetium and lawrencium. Textbooks have not moved, partly because doing so would require redrawing a table that most readers learned in a particular shape. It is one of the few genuinely unresolved structural questions about the periodic table, and lutetium is at the centre of it.

Lutetium is also the endpoint of the lanthanide contraction, which makes it the hardest, densest and highest-melting member of the series — properties that come from the fourteen f electrons shielding the nucleus poorly and letting it pull everything in tightly.

Two drugs that changed what lutetium is for

Until recently lutetium was an element with an interesting position and almost no market. That changed with lutetium-177.

Lutetium-177 has a half-life of 6.6 days and emits beta particles that travel a couple of millimetres in tissue — far enough to kill a small cluster of cells around wherever the atom lodges, short enough not to reach much beyond it. It also emits weak gamma rays at 113 and 208 keV, which a gamma camera can image.

That last detail is what makes it valuable. The same injection that delivers the treatment can be photographed, so clinicians can confirm the drug went where it was supposed to and calculate the dose the tumour actually received. Treatment and imaging in one molecule.

Two drugs built on it are now in wide use. Lutetium-177 dotatate, licensed in 2018, targets somatostatin receptors on neuroendocrine tumours. Lutetium-177 vipivotide tetraxetan, licensed in 2022, targets prostate-specific membrane antigen in metastatic prostate cancer that has stopped responding to hormonal treatment, and demonstrated an overall survival benefit in the trial that supported its approval. Demand for the second outran manufacturing capacity almost immediately, and supply of the isotope — produced without inactive lutetium mixed in by irradiating ytterbium-176 and separating the product chemically — became the constraint on how many patients could be treated.

The heavier alternative for the same targets is actinium-225, which substitutes alpha particles for lutetium's betas and is far scarcer.

A detector that is its own radiation source

Lutetium's other modern role is inside the machine rather than inside the patient. Positron emission tomography relies on catching pairs of 511 keV photons, which requires a scintillator that is dense, high in atomic number, bright, and above all fast — because measuring the tiny difference in arrival time between the two photons localises the annihilation along the line between the detectors.

Lutetium oxyorthosilicate, and its yttrium-substituted variant, meet all four conditions better than anything else available, with a decay time in the tens of nanoseconds. They are the standard detector material in modern PET scanners, and they are what makes time-of-flight PET possible.

They also come with a permanent radioactive background. About 2.6% of natural lutetium is lutetium-176, a beta emitter with a half-life of 37.6 billion years, and there is no practical way to remove it from the crystal. Every PET detector therefore counts a steady trickle of decays from its own scintillator, whether or not a patient is present.

For most clinical imaging this is a negligible nuisance. For very low count-rate applications it is a genuine floor on sensitivity, and it is one reason some detector designs avoid lutetium entirely. But manufacturers have also turned it into an advantage: the lutetium-176 spectrum is known precisely and is always present, so the scanner has a built-in reference against which to check each detector's gain and energy calibration daily, with no external source to handle, store or account for.

Zircon's memory

The same slow decay drives one of geology's most informative isotope systems. Lutetium-176 decays to hafnium-176, and the two elements behave very differently when a magma crystallises.

Zircon takes up hafnium readily and rejects lutetium almost entirely, so a zircon crystal locks in the hafnium isotope ratio of the melt it grew from and then stops changing. Because the same crystal can be dated precisely by uranium-lead, a single grain yields both when it formed and what the isotopic character of its source was — whether the magma came from freshly melted mantle or from remelting of much older crust.

Applied to the oldest zircons on Earth, some of them over four billion years old and surviving as detrital grains long after their parent rocks were destroyed, the lutetium-hafnium system is one of the few direct windows onto the composition of the earliest continental crust.

Lutetium remains the most expensive of the lanthanides by weight. It is present in the smallest quantity in almost every ore, and it is the last element off the end of every separation train, which means the cost of extracting it includes the cost of getting everything else out first.