Gold (Au)

Almost every metal is grey. Iron, silver, aluminum, nickel, tin and platinum all reflect visible light more or less evenly, which is why they look like variations on the same colour. Gold does not, and copper does not, and for a long time nobody could say why. The answer turned out to require Einstein, which is not a sentence you expect to write about a substance Bronze Age smiths were already beating into foil.

The colour is a relativistic effect

In a heavy atom the innermost electrons are moving fast enough — a significant fraction of the speed of light — that their mass increases measurably. The consequence propagates outward: the 6s orbital contracts, the 5d orbitals expand slightly to compensate, and the energy gap between them narrows. In silver, one row up, the equivalent gap sits in the ultraviolet, so every visible wavelength bounces off and the metal looks white. In gold the gap has fallen to about 2.4 electronvolts, which lands squarely in the blue end of the visible spectrum. Blue light is absorbed. What comes back to the eye is everything else, and everything else is yellow.

This is not an ornamental curiosity. The same contracted 6s orbital is why gold is so reluctant to give an electron away, and why its Pauling electronegativity of 2.54 is the highest of any metal — higher than several nonmetals.

Gold that behaves like a halogen

Push that far enough and gold stops acting like a metal at all. Its affinity for an extra electron is greater than sulfur's, and in the compound cesium auride the gold sits as a genuine Au⁻ anion, with cesium as the cation. The result is not a shiny alloy: it is a transparent yellow-brown semiconducting salt that dissolves in liquid ammonia. Group 11 is supposed to contain metals. Gold, at the bottom of it, can be talked into forming the negative ion of an ionic compound.

The everyday face of that same reluctance is corrosion resistance. Gold is untouched by oxygen, by water, by every individual mineral acid. It yields to aqua regia — nitric acid supplying the oxidation, hydrochloric supplying chloride ions that pull the gold into a stable tetrachloroaurate complex — and, industrially far more importantly, to dilute cyanide solution in the presence of air. The MacArthur–Forrest cyanide process, patented in 1887, is what made ore containing a few grams per tonne worth digging, and it arrived within a year of the discovery that mattered most.

Two billion years in one basin

In 1886 an itinerant prospector found gold-bearing conglomerate on a farm outside what became Johannesburg. The Witwatersrand Basin has since yielded somewhere near two-fifths of all the gold humans have ever recovered — no other deposit on Earth comes close to that share.

The rock is Archean, roughly 2.9 billion years old, and geologists have argued about its origin for a century: a fossil placer, where rivers concentrated grains eroded from older rocks, or a hydrothermal system that deposited gold in place. The evidence now favours a modified placer, and one detail keeps the debate interesting — carbon seams in the reef, possibly the remains of microbial mats, are among the richest gold horizons in the basin. Chasing those reefs downdip has produced the deepest human excavations anywhere, working faces some four kilometres below the surface where the rock arrives at around 60 °C.

Where the atoms were made

Elements up to iron are built by fusion inside stars. Gold is not; it requires rapid neutron capture, the r-process, in an environment of extraordinary neutron density. For decades the default assumption was a particular class of supernova. Then in August 2017 the merger of two neutron stars was detected in gravitational waves and followed optically as a kilonova, and the light curve carried the signature of freshly synthesised heavy elements — several Earth masses of them from one event.

That did not settle it. Mergers are rare and, crucially, slow to happen after the parent stars form, which makes it hard to explain the heavy elements already present in very old, very metal-poor stars. Magnetorotational supernovae and magnetar flares are both under active consideration as earlier contributors. The honest position is that neutron-star mergers are demonstrated gold factories, and are probably not the only ones.

A cube twenty-two metres on a side

Cumulative production through the end of 2025 is estimated by the World Gold Council at close to 220,000 tonnes. Melt all of it together and the result is a cube about 22 metres on an edge — a single mid-rise building's worth of metal, spread across every wedding ring, bar, reserve vault and circuit board on the planet, and slightly under one troy ounce per living person.

Roughly 44% of that stock is jewellery. Most of the rest is investment metal and central-bank reserves. Industrial use is a small slice by mass and an indispensable one by function: gold does not grow an oxide layer, so a gold-plated contact still conducts after twenty years in a damp connector, and gold bonding wire a few tens of micrometres across still joins the die to the package inside a great many chips.

One hundred nanometres on a telescope mirror

Two extremes show what the metal can do physically. It is the most malleable of all metals: a single gram can be beaten out to about a square metre of leaf roughly 100 nanometres thick, thin enough to transmit a greenish light. And it reflects infrared radiation at around 98–99%, better than aluminum or silver in that band, which is why the eighteen beryllium segments of the James Webb Space Telescope carry a gold coating of similar thickness. The total gold on all eighteen mirrors comes to about 48 grams.

The isotope, and the transmutation that works

Gold is mononuclidic — every gold atom found in nature is gold-197, and with no mixture to average over, its atomic weight is pinned down more tightly than that of most elements. Radioactive gold-198, with a half-life of 2.7 days, was used in early brachytherapy as implanted seeds.

The alchemical ambition has been achieved, twice over, and it is a terrible business. In 1980 a Berkeley team led by Glenn Seaborg stripped protons from bismuth nuclei in a heavy-ion accelerator and made a few thousand atoms of gold. Neutron irradiation of mercury-196 in a reactor also yields gold. Both routes work; both cost overwhelmingly more than the gold is worth, which is the most reliable thing anyone has ever established about turning base metal into precious.