A clock built into unstable atoms
Radiometric dating works because certain isotopes are unstable — their nucleus spontaneously transforms into a different, more stable element at a fixed statistical rate, unaffected by temperature, pressure or chemical bonding. That rate is described by a half-life: the time for exactly half of a starting quantity to decay. Because decay is a memoryless random process at the level of individual nuclei, the population-level decay is a clean exponential, and measuring the ratio of remaining parent isotope to accumulated daughter isotope tells you, unambiguously, how much time has passed since the clock was "reset".
N(t) = N0 * (1/2)^(t / t_half) rearranged for age: t = t_half * log2( N0 / N(t) ) carbon-14: t_half ≈ 5,730 years (organic material, up to ~50,000 yr) uranium-238: t_half ≈ 4.47 billion yr (zircon crystals, whole-Earth timescale) potassium-40: t_half ≈ 1.25 billion yr (volcanic rock, bracketing fossil layers)
Radiocarbon: dating what was once alive
Carbon-14 is continuously produced in the upper atmosphere by cosmic rays striking nitrogen, and living organisms constantly exchange carbon with the atmosphere through photosynthesis and respiration, keeping their internal ¹⁴C ratio matched to the atmosphere's. The moment an organism dies, that exchange stops and the ¹⁴C it already contains begins decaying with a 5,730-year half-life at a fixed rate, while the far more abundant, stable ¹²C stays constant — so the shrinking ¹⁴C/¹²C ratio is a direct clock. Because the half-life is short, radiocarbon dating is useful only up to roughly 50,000 years before the remaining ¹⁴C becomes too sparse to measure reliably, which makes it the standard method for archaeology and recent-Holocene climate work but useless for anything geological.
Why atmospheric ¹⁴C is not perfectly constant — calibration
Cosmic ray flux, and therefore atmospheric ¹⁴C production, has varied over time due to changes in solar activity and Earth's magnetic field strength, so a raw radiocarbon age is not automatically a calendar age. Scientists correct for this using calibration curves (the internationally maintained IntCal series) built by radiocarbon-dating material of independently known age — principally tree rings dated by dendrochronology, and marine and speleothem records — producing a lookup that converts a raw ¹⁴C age plus its uncertainty into a calibrated calendar-year probability distribution.
Long timescales: bracketing rock, not the fossil itself
Fossils themselves almost never contain a datable radioactive parent-daughter pair in useful concentration, so paleontologists instead date volcanic ash or lava layers above and below a fossil-bearing rock unit, using isotope systems with much longer half-lives suited to millions or billions of years: potassium-argon (and its more precise sibling argon-argon) for volcanic minerals, and uranium-lead for zircon crystals that crystallise in magma and lock in essentially zero initial lead. Because these methods date the volcanic rock, not the fossil, the fossil's true age is bracketed between the ages of the layer just below it and just above it.
Sources of error, and why cross-checking matters
Every method assumes a closed system since formation — no parent or daughter isotope added or removed except by decay — and geologists check this assumption using concordant results from two independent isotope systems on the same sample (for example, uranium-lead measured two different ways within the same zircon crystal), or by cross-referencing radiocarbon with dendrochronology as described above. Discordant results are themselves informative: they usually flag metamorphic reheating, weathering, or contamination that reset part of the clock.
What the simulation shows
This simulation animates a population of unstable parent atoms decaying stochastically into daughter atoms, letting you watch the exponential curve emerge from individually random decay events and read off an estimated age from the remaining parent fraction — the same logic, at vastly different timescales, behind both carbon dating a bone and uranium-lead dating a zircon crystal billions of years old.
Frequently asked questions
Why can't carbon-14 dating be used on rocks or very old fossils?
Carbon-14 only exists in organisms that were exchanging carbon with the atmosphere while alive, so it cannot date rock directly. Its 5,730-year half-life also means that after about 50,000 years so little ¹⁴C remains that the ratio can no longer be measured reliably, well short of dinosaur or Earth-formation timescales.
How do scientists date dinosaur fossils if radiocarbon does not reach that far back?
By dating volcanic ash or lava layers directly above and below the fossil-bearing rock using long-half-life systems like potassium-argon/argon-argon or uranium-lead, which bracket the fossil's age between two independently dated layers rather than dating the fossil itself.
What does 'half-life' actually mean for a single atom?
It is a statistical statement about a population, not a prediction for an individual atom — any single unstable nucleus has a fixed probability of decaying in any given time interval regardless of its age, but with enough atoms that constant probability produces the smooth, predictable exponential decay curve the dating methods rely on.
Try it live
Everything above runs in your browser — open Isotope Dating and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.
▶ Open Isotope Dating simulation