Argon (Ar)

In the late 1880s Lord Rayleigh set out to measure the density of nitrogen accurately, which was the sort of unglamorous metrological task he did superbly. He prepared it two ways. Nitrogen extracted from air, by removing the oxygen, water and carbon dioxide, came out at 2.3102 grams per litre. Nitrogen made chemically from ammonia came out at 2.2990. The difference was about half a per cent, and it would not go away.

Half a per cent is far too large to be experimental error in work of that quality and far too small to suggest anything dramatic. Rayleigh spent two years trying to make it disappear, and in 1892 published a short letter in Nature describing the discrepancy and asking readers for suggestions.

Chasing a discrepancy into a new column

William Ramsay wrote back, and the two men then pursued the problem along different lines with an understanding that they would share what they found.

Rayleigh took the older approach: sparking air with excess oxygen so the nitrogen was converted to oxides and absorbed, leaving whatever would not react. Ramsay passed air repeatedly over heated magnesium, which combines with nitrogen to form the solid nitride. Both were left with a small residue that refused to combine with anything, amounting to roughly one per cent of the original air, and its spectrum showed lines matching no known element.

They announced argon in 1894 to considerable scepticism, because a new element that fitted nowhere in the periodic table was easier to disbelieve than to place. It fitted nowhere because an entire group was missing. Within five years that group had been filled in, and Ramsay had a hand in every one of them.

Rayleigh received the 1904 Nobel prize in physics and Ramsay the 1904 prize in chemistry, in the same ceremony, for two halves of the same discovery.

Cavendish saw it in 1785 and let it go

The residue was not new. Henry Cavendish, working a hundred and nine years earlier, had sparked air with oxygen over alkali precisely to determine whether all of what was then called phlogisticated air could be made to react. Almost all of it could. A small bubble persisted that he could not remove however long he ran the experiment, and he recorded it as being not more than about one hundred and twentieth of the whole.

He published the observation and drew no conclusion from it, and for over a century nobody picked it up. Rayleigh and Ramsay both cited him explicitly once they knew what he had been looking at. The proportion he measured was close to right.

Named for idleness, and then caught reacting

Argos means idle or inactive, and the name was chosen for a substance that would combine with nothing at all. It held for a hundred and six years.

In 2000 a group at the University of Helsinki led by Leonid Khriachtchev and Markku Räsänen prepared argon fluorohydride, HArF, by photolysing hydrogen fluoride trapped in a frozen argon matrix and warming it slightly, identifying the product by its infrared absorption bands. The compound is stable only below about 27 kelvin and decomposes on warming. It is nonetheless a real neutral compound of argon, and it separates argon from the two lighter noble gases, for which no such thing has ever been made.

Almost all of Earth's argon is nuclear waste

Argon makes up close to one per cent of the atmosphere — the third most abundant constituent of air, present in far greater quantity than carbon dioxide. Very little of it came with the planet.

In the Sun and in the interstellar medium, the dominant isotope of argon is argon-36, produced directly by stellar nucleosynthesis. In Earth's atmosphere, argon-36 is a trace component and argon-40 accounts for over 99.6%. That inversion has one cause: argon-40 is the decay product of potassium-40, which is abundant in crustal rock, and it has been accumulating and leaking out of the solid Earth for four and a half billion years. The argon in the air is, essentially in its entirety, radiogenic — a by-product of radioactive decay that happens to be inert enough to have survived in the atmosphere rather than reacting away.

This also explains a long-standing embarrassment in the periodic table. Argon's atomic weight is higher than potassium's, even though argon comes first. Ordering the elements strictly by mass, as Mendeleev originally did, puts argon among the alkali metals and potassium among the noble gases, which is chemically absurd — and Mendeleev simply inverted the pair on chemical grounds, without being able to justify it. The justification arrived with Henry Moseley's demonstration in 1913 that the organising quantity is nuclear charge rather than mass. Argon is heavy for its position precisely because so much of it is the four-decades-heavier decay product of an element that sits next door.

A clock that only starts when the rock freezes

The same decay makes argon one of geology's most important dating tools, and the reason it works so cleanly is that argon is a gas that fits into no mineral lattice.

When a mineral crystallises from a melt, any argon present escapes. The clock is set to zero by the crystallisation itself. Every argon-40 atom found inside that crystal afterwards must have been produced in place by the decay of potassium within it, so the ratio of the two gives the time elapsed since the rock solidified.

Potassium-argon dating rewrote human prehistory almost immediately. In 1961 the method was applied to volcanic tuff in Bed I at Olduvai Gorge, and returned an age of roughly 1.75 million years for the hominin material found there — far older than the prevailing estimates, which had been guesswork built on faunal comparison. The antiquity of the human lineage roughly doubled on the strength of one measurement.

The modern refinement, argon-argon dating, irradiates the sample first to convert some of its potassium-39 into argon-39. Both isotopes can then be measured on the same tiny sample in one instrument, removing the need for a separate potassium analysis on a separate aliquot. Better still, the sample is heated in steps and the ratio measured at each stage: a rock that has stayed closed gives a flat plateau of consistent ages, and one that has leaked argon or been reheated gives a disturbed profile that announces itself. The method carries its own internal check on whether the answer can be trusted.

Cheap because oxygen pays for the plant

Argon's industrial dominance is in shielding. Welding molten metal in open air lets oxygen and nitrogen dissolve into the weld pool and ruin the joint; a blanket of argon prevents it, and argon is preferred over cheaper alternatives because it is denser than air and stays where it is put. Every TIG weld and most MIG welding on aluminum and stainless steel is done under it.

The larger tonnage goes into steelmaking. Argon-oxygen decarburisation, introduced in the 1950s and 60s, blows a mixture of argon and oxygen through molten stainless steel to burn out carbon while the argon dilution keeps the expensive chromium from oxidising along with it. That process is the main reason stainless steel stopped being a premium material and became a commodity one.

Beyond metals, argon fills the gap in double and triple glazing, where it conducts heat about a third less readily than air and costs little; fills incandescent and fluorescent lamps; provides the atmosphere for growing semiconductor crystals; and preserves documents, including the American Charters of Freedom, in sealed cases where oxygen would slowly destroy the parchment and ink.

All of it comes from the air, as a side stream of the same distillation columns that make oxygen and nitrogen, which is why a gas that constitutes only a fraction of a per cent of the atmosphere is nevertheless inexpensive.

Being appreciably denser than air, argon accumulates in pits, sumps, tank bottoms and ship holds and displaces the oxygen there. It is odourless and gives no warning of its presence, and confined spaces where it has collected have been the cause of industrial asphyxiation deaths.