Superheavy Elements

Everything from rutherfordium onwards has been made rather than found. There is no ore, no extraction, no supply chain and, for most of these fifteen elements, no sample that ever contained more than a handful of atoms at once. They are grouped here not because they share a column — they occupy five different ones — but because they share a set of problems that no other part of the table has: they must be built one atom at a time, identified indirectly, and studied before they decay.

The first five of them, elements 104 to 108, are also counted among the transition metals, since they begin a 6d series and sit under hafnium through osmium. That page treats them as the tail of the d-block and hands the rest here, which is the right division of labour: what is genuinely known about these atoms is nuclear, not chemical, and reading them as ordinary transition metals would credit them with a body of chemistry that has never been measured. The ten heaviest, 109 to 118, have no assigned family colour on this site at all, because assigning one would be a prediction dressed as an observation.

Building a nucleus that does not want to exist

Two nuclei have to be pushed together hard enough to overcome their mutual electrostatic repulsion and then, crucially, not so hard that the resulting object flies apart. The window between those two conditions is narrow, and it narrows further with every proton added.

Two strategies split the field for thirty years. Cold fusion, developed at GSI in Darmstadt, fires a medium-weight beam — chromium, iron, nickel, zinc — at a lead or bismuth target. The compound nucleus forms relatively cool and sheds only one neutron, which improves survival odds, and this route produced elements 107 to 112. Its problem is that cross-sections collapse as the beam gets heavier: by element 113 the production rate had fallen to something like one atom every few months of continuous running.

RIKEN in Japan pushed the cold-fusion route to its limit anyway, firing zinc-70 at bismuth-209 for nine years and recording three atoms of element 113 — one in 2004, one in 2005 and a third in 2012 whose decay chain was clean enough to settle the case. Dubna had a competing claim from a different reaction, but RIKEN's chain terminated in well-characterised nuclides and theirs did not, and the joint working party awarded priority to RIKEN. Nihonium became the first element named in Asia.

Hot fusion with a calcium-48 beam took over from there. Calcium-48 is a doubly magic, neutron-rich nucleus that happens to be just stable enough to use as a beam material, and firing it at actinide targets — plutonium, americium, curium, berkelium, californium — produced elements 114 through 118 at Dubna, mostly in collaboration with Livermore and Oak Ridge, who supplied targets no one else could make. The berkelium target used for tennessine had to be produced at Oak Ridge, shipped to Russia and used quickly, because berkelium-249 has a half-life of 330 days and the experiment was racing its own target's decay.

Nobody sees the new atom. It recoils out of the target, is separated from the beam by a gas-filled magnetic separator, and is implanted in a silicon detector. What is recorded is the implantation followed by a chain of alpha decays, each with a characteristic energy and timing, and the identification rests on that chain terminating in a nuclide already known. When the chain ends in unknown territory instead, the claim is much weaker — which is precisely what the priority disputes were about.

Why the half-lives fall off, and why they might not keep falling

Adding protons adds electrostatic repulsion that grows as the square of the charge, while the strong force that holds the nucleus together is short-ranged and grows only linearly with the number of nucleons. Past bismuth, alpha decay becomes energetically favourable for every nuclide; past about element 100, spontaneous fission joins it, and the liquid-drop picture of the nucleus predicts that a nucleus of around 104 protons should fission essentially instantly.

It does not. Rutherfordium isotopes survive for seconds to minutes, which is many orders of magnitude longer than the liquid drop allows. The explanation is shell structure: just as electrons fill shells with closures at the noble gases, protons and neutrons fill nuclear shells, and a nucleus near a shell closure gains extra binding that raises the barrier against fission. Every superheavy element that exists at all exists because of that shell correction.

Extrapolating the shell closures beyond the last known one gives the island of stability — a region of unusually long-lived nuclides expected near a magic proton number and a neutron number of 184. Where exactly is contested. The older macroscopic-microscopic calculations put the proton closure at 114; more recent self-consistent mean-field models favour 120 or 126, and different choices of nuclear force give different answers. What all of them agree on is that the island lies at higher neutron numbers than anything reachable with current beams and targets. The heaviest isotopes made so far sit on the shoreline, not the island: half-lives around flerovium and copernicium are longer than the trend predicts, but they are measured in seconds, not the years the more optimistic predictions once suggested.

The reason that gap is so hard to close is arithmetic, and it works against the experimenter twice. A fused nucleus can be no more neutron-rich than the two objects fused to make it, and the most neutron-rich projectile anybody has found usable is calcium-48, with 28 neutrons to 20 protons. Fired at the most neutron-heavy actinide target that can be fabricated, it makes something that then has to throw off several more neutrons to carry away its excitation energy before it can survive at all — and what survives lands short of the predicted closure. The second problem is worse. The beams that have to replace calcium-48 in order to get past element 118 — titanium-50, chromium-54 — carry fewer neutrons per proton than calcium-48 does, so every step up in proton number moves the product further from the neutron count that would make it live longer. Reaching higher and reaching the island are not the same journey, and past a point they are opposite ones.

Chemistry with a sample size of one

A chemical experiment on an element that exists as one atom for four seconds sounds like a contradiction. It is not, provided the question is posed statistically rather than descriptively.

The method is gas-phase chromatography. A newly made atom is swept along a channel whose walls are gold or silicon dioxide and whose temperature falls steadily along its length, and where it sticks is recorded by the alpha decay it emits when it gets there. Repeat that with enough atoms and the distribution of deposition temperatures gives an adsorption enthalpy — a real thermodynamic quantity — which can be compared with the same measurement made on the lighter members of the column.

The results are the whole scientific point of this end of the table, because relativistic effects on the electrons should be strongest here. A 7s electron in element 112 travels fast enough for its relativistic mass gain to matter; the orbital tightens and the electrons in it become harder to engage in bonding. The prediction was that copernicium might behave more like a noble gas than like mercury, and that flerovium might be volatile rather than lead-like. Experiments at both Dubna and GSI found copernicium and flerovium to be volatile and weakly interacting with gold, but the adsorption enthalpies have been argued over, sample sizes are in the single digits, and different groups have drawn different conclusions from data sets of a few atoms. This is the frontier where the periodic law itself is under test, and the evidence is thin by design rather than by neglect.

Why this end of the chart is a settlement rather than a measurement

Everywhere else in the table, a discovery was not seriously in doubt and only the label was ever argued over. Here the two questions collapse into one. A claim is a handful of decay chains; deciding whose chains were convincing decides who gets to name the element; so for these fifteen tiles the printed table is recording the outcome of an adjudication, not the result of an observation.

That adjudication has both a history and a rulebook, and each is a subject in its own right. The three-decade quarrel between Berkeley and Dubna over elements 104 to 109 — two countries teaching from irreconcilable tables, with the names unfixed until 1997 — belongs with the elements that nearly had other names. The procedure that came out of that quarrel, from the placeholder an element carries before it has a real name to the suffix rule that left tennessine no alternative ending, belongs with how elements get their names.

Two consequences are worth carrying into the rest of this page. Priority in this range turns on the quality of a decay chain and nothing else, which is exactly how the rival claims to element 113 above were separated. And a name arrives years behind the atom, so the word on a tile and the experiment that earned it frequently belong to different decades.

Where the table stops

Element 118 completes the seventh row, and getting further is not a matter of persistence. Calcium-48 on californium-249 gives 118; the next step needs a target of einsteinium or fermium, which cannot be made in the quantities a target requires, so heavier beams — titanium-50, chromium-54 — must be used instead, at cross-sections perhaps a hundred times smaller. Facilities at Dubna and RIKEN are running those experiments now, on timescales measured in years per atom. Element 119 would open an eighth row and, on the simplest orbital-filling arguments, eventually require g orbitals, which no element has ever used. Whether the resulting objects would be periodic in any useful sense is exactly the question the single-atom experiments are already probing.