Some atoms slowly change into other atoms at a steady, known speed, like sand falling through an hourglass. By measuring how much has changed, scientists can tell how old something is. Carbon-14 works for things that were once alive, up to about 50,000 years old. Other "hourglasses" work for rocks billions of years old. Tree rings and yearly lake layers can be counted like a calendar. When completely different clocks give the same answer, we can trust it.
Carbon-14 dating only works on once-living things up to about 50,000 years old, and it's calibrated against counted tree rings, lake layers and corals.1 For rocks, scientists use other atoms with much longer half-lives. The key is consilience: methods based on different physics keep agreeing with each other.2
"Carbon dating is unreliable, and scientists just assume their answers. They carbon-date dinosaurs and get millions of years, which is impossible."
How carbon-14 works
- Cosmic rays hit nitrogen high in the atmosphere and make a small amount of carbon-14.
- Plants take it in as they grow; animals take it in by eating plants.
- When the plant or animal dies, it stops taking in new carbon. The carbon-14 it has slowly turns back into nitrogen, losing half every 5,730 years.
- After about 50,000 years there's too little left to measure.1
Nobody carbon-dates dinosaurs or rocks. Carbon-14 is gone long before then. Dinosaur ages come from dating volcanic rock layers above and below the fossils with other methods.
Calibration: checking the clock
Early radiocarbon dating assumed the air always had the same amount of carbon-14. By 1969 scientists knew that wasn't quite true, so they built calibration curves: they radiocarbon-date things whose age is already known some other way, and correct the results.1
- Tree rings: one ring per year, counted and matched across overlapping trees.
- Corals: dated with uranium-thorium; a 2009 calibration curve reaches back about 50,000 years.
- Lake Suigetsu, Japan: a 70-meter core of yearly lake layers (varves), counted back 52,000 years and checked against 651 radiocarbon dates. Its curve closely matched the coral curve from completely different data.1
The corrections typically make dates 10–20% older than raw dates, not younger.1
Many clocks, many ranges
| Method | Parent → daughter | Half-life | Used for |
|---|---|---|---|
| Carbon-14 | ¹⁴C → ¹⁴N | 5,730 years | Once-living material up to ≈50,000 years |
| Potassium–argon | ⁴⁰K → ⁴⁰Ar | 1.25 billion years | Volcanic rocks and certain minerals |
| Uranium–lead | ²³⁸U → ²⁰⁶Pb · ²³⁵U → ²⁰⁷Pb | 4.47 billion · 704 million years | Zircon crystals: two clocks in one crystal |
| Rubidium–strontium | ⁸⁷Rb → ⁸⁷Sr | 48.8 billion years | Rocks older than ~50–100 million years |
| Samarium–neodymium | ¹⁴⁷Sm → ¹⁴³Nd | 106 billion years | Very old rocks and meteorites |
| Luminescence | (trapped electrons since burial) | — | ≈300 to over 100,000 years |
Table from geologist G. Brent Dalrymple and Science Meets Religion.12
When different clocks agree
Rubidium–strontium gives 4.60 ± 0.07 billion years. Samarium–neodymium, a completely different element pair with a different half-life, gives 4.56 ± 0.08 billion years. The ranges overlap.2
Uranium–lead gives 3.56 billion years for the rock's formation and 1.85 billion years for a later heating event. Potassium–argon ages from the same area record the same 1.85-billion-year heating.2
The age of the Earth, 4.54 billion years, comes the same way: meteorites, Moon rocks and Earth's oldest crystals, dated by different methods in different labs, all land in the same range.3 See The young-earth story.
Young-earth creationists argue that decay rates may have been much faster in the past, or that rocks may have started with daughter atoms already present. Scientists answer that (1) decay rates have been measured under extreme heat, pressure and chemistry with no meaningful change; (2) "isochron" methods detect starting daughter atoms; and (3) if decay had been sped up enough to compress billions of years into thousands, the heat released would have melted the Earth's crust, a problem young-earth researchers have themselves acknowledged.
Sources
- Science Meets Religion, How reliable is radiocarbon dating?: half-life ≈5,730 years; ≈50,000-year limit; calibration since 1969; 2009 coral curve; Lake Suigetsu (70 m, 52,000 years, 651 samples, 2012); luminescence range; calibrated dates 10–20% older. Also: Oxford Radiocarbon Accelerator Unit, Calibration.
- G. Brent Dalrymple, Talk.Origins Archive, Radiometric Dating: methods, half-lives and ranges; Juvinas meteorite; Morton Gneiss.
- USGS, Age of the Earth: 4.54 billion years; meteorites 4.53–4.58; zircons up to 4.3; Moon rocks 4.4–4.5.
- Marked for a second source: the "heat problem" acknowledged by young-earth researchers, and tests of decay-rate stability under extreme conditions.