Period 7

The seventh row is the only one that human beings finished. Six of its thirty-two members were present on Earth before anyone went looking; the other twenty-six were assembled, one nuclear reaction at a time, over seventy-six years. It is also the only row in which not a single element has a stable isotope, and the only one whose later members were named before anyone could say with confidence what they were chemically like.

The six are francium, radium, actinium, thorium, protactinium and uranium. Two of them occur in quantity and the remaining four are transient products of their decay, which is why francium — at the foot of an entirely ordinary column — is among the scarcest naturally occurring elements there is. Crustal abundances, the decay chains and the 5f chemistry all belong to the actinides; what follows is about the row.

Thirty-two elements and four different subjects

Thirty-two is as wide as a row gets, and period 6 matches it — but nothing else about the two rows' internal divisions is comparable. Reading period 7 straight across means changing subject three times, and the four stretches are studied by largely separate communities using largely different equipment.

• Francium and radium complete the 7s subshell, and both make more sense as the last entries in their columns than as the first entries in this row. Francium's oddities are the group 1 trend taken to its end and radium's are group 2's; the alkali metals and alkaline earth metals are where those trends are argued through.
• Actinium to lawrencium are the fifteen members lifted out of the row and printed underneath the table. This is the only stretch with a substantial experimental chemistry behind it, and the actinides page is where that chemistry sits.
• Rutherfordium to copernicium are the row's d-block. Each has been probed by a small number of experiments and no more, and what those experiments can and cannot establish is treated with the superheavy elements.
• Nihonium to oganesson fill 7p and are the least characterised elements in existence. The questions they raise about the periodic law itself are taken up further down this page.

One feature of the row is purely a matter of naming, and it belongs to the row rather than to any one square in it. Both of the table's two names taken from a person who was alive to hear it sit here: seaborgium, approved in 1997 over a rule written expressly to stop it, and oganesson, approved in 2016 without anyone raising the question again. Period 7 is the row that was named while the people who made it were still in the room.

Twenty-six that were not here

The boundary was crossed in 1940 at Berkeley, when Edwin McMillan and Philip Abelson bombarded uranium with neutrons and identified the element beyond it. Naming it neptunium, after the planet beyond Uranus, made the point explicitly: the table had been extended past the edge of what nature supplied.

Everything after that came from a laboratory, and the pace tells its own story. The 1940s and 1950s added elements at a rate of roughly one a year; the 1960s through the 1990s added them at one a decade, amid competing claims from Berkeley, Dubna and Darmstadt; the last four took until 2016. The techniques by which the heaviest were built, and the long argument over who built which, belong with the superheavy elements.

The day the table had no gaps

On 28 November 2016 IUPAC approved the names nihonium, moscovium, tennessine and oganesson for elements 113, 115, 117 and 118. That completed the seventh row, and with it the periodic table had, for the first time since Mendeleev drew it in 1869, no empty positions anywhere.

It is worth pausing on how unusual that state of affairs is. The table's entire nineteenth-century reputation rested on its gaps: Mendeleev's willingness to leave a space and predict what would fill it was what turned a classification into a theory. A hundred and forty-seven years later there were none left to leave. Any further element does not fill a gap; it starts a new row.

Twenty orders of magnitude of half-life

The single trend that runs cleanly across period 7 is not a chemical one. It is nuclear stability, and it falls off a cliff.

Thorium-232 has a half-life of about fourteen billion years and uranium-238 about four and a half billion — long enough that both are still mined rather than manufactured. Plutonium-244 manages eighty million years, curium-247 sixteen million. By the end of the actinide stretch the figures are in minutes: mendelevium, nobelium and lawrencium are counted in tens of minutes at best. The 6d elements that follow last seconds to hours depending on the isotope, and the last few members of the row are measured in milliseconds. Oganesson's known isotope decays in under a millisecond.

That is roughly twenty orders of magnitude across a single row, which no other property of any row approaches. The cause is a competition between two forces with different reach. The strong nuclear force binds only immediate neighbours, so its contribution grows roughly in proportion to the number of nucleons. Electrostatic repulsion acts between every pair of protons at any distance inside the nucleus, so it grows roughly as the square of the atomic number. Somewhere the second overtakes the first, and past bismuth it has: every nuclide in this row is unstable, and the margin gets worse with each element added.

The practical consequence is that period 7's population is not a fixed set of substances but a set of processes. Uranium can be stockpiled; oganesson cannot be stored at all, and the total quantity ever produced amounts to a handful of atoms.

Does the row still obey the table?

The deepest question this row raises is whether it is a period in the same sense the others are.

Chemical periodicity exists because electrons occupy shells and because the outermost shell determines behaviour. Both premises weaken at high nuclear charge. Relativistic contraction of the s orbitals, and the spin-orbit splitting of p orbitals into distinguishable j = 1/2 and j = 3/2 components, grow steeply with Z, and by the end of this row they are large enough to reorder the picture rather than merely perturb it.

Oganesson is the clearest test case. It occupies the noble gas column, and a noble gas is noble because its outermost p subshell is closed and uniform. In element 118 the 7p subshell is calculated to be split by several electronvolts into two components that behave almost as separate shells. The same calculations predict that oganesson has a positive electron affinity — it should bind an extra electron, which no other noble gas does — and that its electron density is smeared to the point where the familiar shell structure of an atom is barely discernible. Several groups have predicted it would be a solid at room temperature, not a gas.

Flerovium sits under lead and has been predicted to behave more like a volatile, weakly interacting species than like a heavy metal. Copernicium under mercury has attracted similar predictions. Whether those predictions survive is being checked one nucleus at a time, and slowly.

None of this means the periodic law fails. It means the law was always a statement about electronic structure, and electronic structure at Z near 120 is not the structure that Mendeleev's chemistry was generalising from. A column tells you where an element sits; in this row it has stopped being a reliable guide to how it behaves, and the honest tables — including this one — decline to give the last ten members a family colour for exactly that reason.

Where an eighth row would go

Extending the table past 118 is not simply a question of trying harder, and it may not be possible at all.

On the simplest filling arguments, element 119 begins an eighth row with an 8s electron, and somewhere in that row 5g orbitals would be occupied for the first time — a subshell with eighteen places that would make the row fifty elements long. Pekka Pyykkö's calculated extension of the table, published in 2011, predicts that the filling order in that region abandons the neat n + l pattern entirely, with 8s, 5g, 6f, 7d and 8p levels interleaving in an order that depends on the details of the relativistic calculation.

There is also a limit somewhere. A crude argument attributed to Feynman puts it near Z = 137, where a point nucleus would require the innermost electron to exceed the speed of light; more careful treatments that give the nucleus a finite size push the boundary out to somewhere between 150 and 175, with different models giving different answers. Nobody is close to testing it. The practical limit arrives much sooner, set by the vanishing probability that two nuclei will fuse and survive, and by the fact that no target material heavier than californium can be made in the quantity a target needs.

Period 7 may therefore turn out to be the last complete row of the periodic table for a very long time — finished in 2016, and unlikely to be followed.