How We Know the Age of the Earth

• Radioactive decay runs at a fixed, constant-probability rate — a clock that starts when a mineral crystallises.
• The isochron method removes the need to assume how much daughter isotope was there at the start, which is the objection people usually raise.
• No original Earth rock survives plate tectonics, so the age comes from meteorites that were never reprocessed.
• Clair Patterson got 4.55 billion years in 1956, and seventy years of independent measurement has refined it, not replaced it.
• Independent decay systems agree. That mutual consistency is the real evidence, far more than any single date.

The clock

Some atomic nuclei are unstable and spontaneously convert into a different element. The critical property is how they do it: each nucleus has a fixed probability of decaying in any given interval, and that probability does not depend on the nucleus's history, its neighbours, temperature, or pressure. An atom that has sat unchanged for a billion years is exactly as likely to decay in the next second as a freshly minted one. Radioactive decay has no memory.

That property produces exponential decay, and with it the half-life — the time for half of a large population of parent atoms to convert to daughter atoms. Half-lives vary enormously between isotopes, which turns out to be extremely useful.

Parent → daughter	Half-life	Useful range
Carbon-14 → Nitrogen-14	5,730 years	Up to ~50,000 years
Uranium-235 → Lead-207	704 million years	Millions to billions
Potassium-40 → Argon-40	1.25 billion years	100,000 years to billions
Uranium-238 → Lead-206	4.47 billion years	Millions to billions
Rubidium-87 → Strontium-87	48.8 billion years	Billions

The first row is the source of a persistent confusion. Carbon-14 is never used to date rocks. After about ten half-lives so little parent remains that measurement becomes meaningless, which caps radiocarbon dating around 50,000 years and confines it to once-living material. Dating a 4-billion-year-old rock with carbon-14 would be like timing a geological era with a stopwatch — the objection that "carbon dating is unreliable at those timescales" is correct and irrelevant, because nobody does it.

For deep time you want an isotope whose half-life is comparable to the age you are measuring. Uranium-238's 4.47 billion years is almost perfectly matched to the age of the solar system.

The obvious objection, and the actual answer

If you know the decay rate and you measure the parent and daughter amounts, you can solve for elapsed time. But that calculation contains an assumption: that there was no daughter isotope present when the clock started. If a mineral formed already containing lead, a naive calculation would report it as far older than it is.

This is a real problem, and geochronologists solved it decades ago. The solution is the isochron method, and it deserves to be better known because it converts the assumption into a measurement.

The trick is to analyse several different minerals that crystallised from the same melt at the same time. Then, for each one, plot two ratios against each other: the daughter isotope relative to a stable, non-radiogenic isotope of the same element, versus the parent relative to that same stable isotope.

At the moment of formation, all these minerals shared one initial daughter ratio — so on the graph they start on a horizontal line. As time passes, each mineral's parent converts to daughter in proportion to how much parent it began with. The points rotate upward, and crucially they stay on a straight line. The slope of that line grows with time.

The result is a system with two unknowns and enough data to solve for both:

• The slope gives the age.
• The y-intercept gives the initial daughter ratio — measured, not assumed.

And there is a built-in integrity check. If the sample was disturbed — reheated, weathered, contaminated, partially melted — the points stop falling on a straight line. A bad sample announces itself. This is why "how do you know the initial conditions?" is a good question with a satisfying answer rather than a fatal flaw.

Zircon crystals deserve a special mention here, because they are close to an ideal clock by chemical accident. When zircon (ZrSiO₄) crystallises it readily accepts uranium atoms into its lattice but strongly rejects lead. So a fresh zircon starts with essentially zero lead, and virtually all the lead found inside one later must have come from uranium decay. Zircons are also brutally durable — they survive erosion, transport, burial and metamorphism that destroys the rock around them. And because zircon contains both uranium-235 and uranium-238, decaying at different rates into different lead isotopes, each crystal carries two independent clocks that must agree.

Why meteorites

Now the twist. Earth is geologically alive. Plate tectonics continuously drags crust down into the mantle and manufactures new crust at ridges; erosion, sedimentation and mountain-building rework whatever is left. The planet has been resurfacing itself for its entire existence.

The consequence is unavoidable: every rock on Earth is younger than the Earth. The oldest reasonably undisputed intact rock, the Acasta Gneiss in Canada's Northwest Territories, is about 4.03 billion years old. Rocks from the Nuvvuagittuq greenstone belt in Quebec have been dated up to 4.28 billion years, though that figure is contested. The oldest terrestrial material of any kind is not a rock at all but a scatter of individual zircon grains from the Jack Hills in Western Australia, around 4.4 billion years old, surviving as inclusions in much younger sandstone — orphan crystals that outlived every rock they were ever part of.

Those Jack Hills zircons set a hard floor: the Earth is at least 4.4 billion years old. They cannot give the ceiling.

For that you need material that formed at the same time as the Earth and then never got reprocessed. Certain meteorites are exactly that. Iron meteorites and chondrites are fragments of small bodies that condensed from the same disc of gas and dust as the planets and then, being too small to sustain geological activity, simply sat unchanged in space for four and a half billion years. They are the solar system's original packaging.

This was the insight behind Clair Patterson's work in the 1950s. Patterson dated lead isotopes from the Canyon Diablo iron meteorite — the object that made Arizona's Meteor Crater — and combined them with terrestrial samples in a lead-lead isochron. Published in 1956, his result was 4.55 ± 0.07 billion years. The modern value of 4.54 ± 0.05 sits comfortably inside his error bars.

Patterson's project had a famous side effect. To measure lead at these precisions he had to build the first ultra-clean laboratory, because ordinary labs, ordinary air and ordinary reagents were saturated with lead contamination. Tracing that contamination to leaded petrol, he spent much of the rest of his career campaigning against it — a public-health victory that emerged, unexpectedly, from a geology problem.

Refinements since have pushed further back. The oldest solids known in the solar system are calcium-aluminium-rich inclusions in the Allende meteorite, dated by lead-lead to about 4.567 billion years. Those inclusions mark the first material to condense as the solar nebula cooled — effectively the solar system's zero mark. Modelling of planetary accretion, supported by hafnium-tungsten dating of core formation, puts Earth's assembly within a few tens of millions of years after that.

Why the number is trustworthy

The strength of the case is not any single measurement. It is that many independent methods, resting on different physics, converge.

Different decay systems agree. Uranium-lead, potassium-argon, rubidium-strontium and samarium-neodymium have half-lives spanning two orders of magnitude and involve different nuclear processes — alpha decay, beta decay, electron capture. Run several on the same sample and they return the same age. For a systematically wrong answer, every one of these unrelated systems would need to be wrong by precisely the amount that preserves their agreement.

Decay rates are stubborn. Experiments have subjected radioactive materials to extremes of temperature, pressure, magnetic field and chemical bonding. Rates barely move. The tiny exceptions are well understood, confined to decay modes that involve the atom's own electrons, and far too small to matter here.

Nature ran the experiment for us. At Oklo in Gabon, uranium ore reached critical concentration roughly 1.7 billion years ago and sustained natural nuclear fission reactions for hundreds of thousands of years. The isotopic products left behind are exquisitely sensitive to the values of fundamental physical constants at the time. They match today's values. Whatever else has changed in 1.7 billion years, the physics underpinning these clocks has not.

Independent sciences constrain it. Stellar astrophysics gives the Sun's age from its structure and fusion rate. The cooling curves of white dwarfs, the main-sequence turn-off ages of globular clusters, and cosmological measurements of the microwave background all place the universe at around 13.8 billion years. Every one of these would have to be independently wrong, in a mutually consistent direction, for 4.54 billion to be badly off.

What it replaced

It is worth remembering how recent this knowledge is. In 1650, Archbishop James Ussher famously derived a creation date of 4004 BC from biblical genealogy. In the nineteenth century Lord Kelvin applied real physics — modelling the Earth as a body cooling from a molten start — and got 20 to 400 million years. Geologists and Darwin both objected that this was far too short for the sedimentary record and for evolution, and could not say why Kelvin was wrong.

He was wrong for a reason nobody could have known: radioactivity had not been discovered. The Earth's interior generates its own heat through radioactive decay, so it has cooled far more slowly than Kelvin's model allowed. The same discovery that invalidated his calculation supplied the tool that replaced it. Within a decade of radioactivity's discovery, Ernest Rutherford and Bertram Boltwood were using it to date rocks, and geology had a clock at last.

Further reading: The USGS publishes an accessible overview of radiometric dating and the age of the Earth. Patterson's original paper is Age of meteorites and the Earth, Geochimica et Cosmochimica Acta (1956). Brent Dalrymple's The Age of the Earth (1991) remains the standard book-length treatment.