Radiocarbon Dating

Radiocarbon dates organic material — anything that was once alive — over roughly the last 55,000 years. It is the most extensively cross-checked dating method in use, with an internationally maintained correction curve built from samples of independently known age.

How it works#

Cosmic rays striking the upper atmosphere produce neutrons, which collide with nitrogen-14 and convert it to carbon-14. That carbon-14 oxidizes to carbon dioxide and mixes through the atmosphere, so every plant takes it up during photosynthesis and every animal takes it up by eating.

While an organism is alive its carbon-14 stays in equilibrium with the atmosphere. At death, intake stops. Carbon-14 decays with a half-life of 5,730 years while stable carbon-12 does not, so the ratio between them falls at a known rate.

After about ten half-lives the remaining carbon-14 falls below what instruments can distinguish from background. This sets the practical limit at roughly 50,000 to 55,000 years. Older material reads as "no measurable carbon-14," which is not a date.

Modern measurement uses accelerator mass spectrometry, which counts individual carbon-14 atoms rather than waiting for decay events. This is why milligram samples are now sufficient where Libby's original method needed grams.

The assumption that was wrong#

As Willard Libby formulated the method in the late 1940s, it assumed atmospheric carbon-14 had been constant. It has not been.

A central assumption of the original method was false. The method was not abandoned; it was calibrated. The variation is now measured directly rather than assumed away.

How calibration works#

Carbon-14 is measured in material whose calendar age is known independently, producing a curve that translates raw radiocarbon years into calendar years.

The current standard is IntCal20, which combines these with published uncertainty bands. It has been revised as data accumulated — IntCal04, IntCal09, IntCal13, IntCal20 — and the revisions have been small.

The tree-ring and varve records are counted, not calculated. Neither depends on radioactive decay, so they are independent of the method they calibrate. See Dendrochronology.

Validation against known ages#

Three laboratories in Arizona, Oxford, and Zurich worked blind on coded samples for the Shroud test, and the result ran against the expectations of many involved.

Common objections#

"A living snail was dated to 27,000 years old"

This happened. Freshwater molluscs build shells partly from dissolved carbonate derived from ancient limestone, which contains no carbon-14. The organism incorporates dead carbon and reads old.

This is the reservoir effect, described in 1963, and it is the reason no laboratory dates freshwater shell without a reservoir correction, or dates it at all when terrestrial material is available.

The same applies to marine organisms, which read several hundred years old because deep ocean water has been out of atmospheric contact. There is a published, regionally mapped marine reservoir correction for this.

A method that specifies where it does not apply, gives the physical mechanism, and publishes the correction is not refuted by someone applying it there.

"Carbon-14 has been found in dinosaur bones and coal"

Trace carbon-14 does appear in samples that should contain none. Reported values are generally 0.1 to 0.5 percent modern carbon.

Four sources produce readings at that level: in-situ production from uranium and thorium decay in surrounding rock converting nitrogen-14 to carbon-14; bacterial contamination, which is substantial in porous bone; contamination during collection and preparation; and machine background, which sits around 0.05 pMC and is measured by laboratories using anthracite and calcite blanks.

The decisive point is the size of the signal. If dinosaur bone were 6,000 years old rather than 65 million, it would retain roughly half its original carbon-14 — an unmistakable signal, hundreds of times what is measured. What is observed sits at the detection floor, which is what contamination produces and not what a young sample would.

"Fresh lava was dated as thousands of years old"

That is potassium-argon dating, not radiocarbon. The two are frequently conflated.

Radiocarbon cannot date lava at all, because it dates organic material that took up atmospheric carbon. The Mount St. Helens case is covered under Radiometric Dating.

"The atmosphere's carbon-14 may have varied more than assumed"

It varied more than Libby assumed, which is why the calibration curve exists.

The curve measures the variation directly using material of independently known calendar age. Breaking it would require tree-ring counting, varve counting, and uranium-thorium dating to fail simultaneously and in the same direction, despite resting on unrelated physics.

A young-earth model requires more than variation. It requires atmospheric carbon-14 to have been low enough that samples appear tens of thousands of years old when they are only thousands — a difference far outside the range the calibration data shows.

"Contamination makes all results unreliable"

Contamination is a real and constant concern, which is why laboratories publish pretreatment protocols and background measurements.

Its direction is known. Modern carbon contamination makes samples read younger, not older, and the effect is largest on old samples. A 40,000-year-old sample contaminated with 1% modern carbon reads about 7,000 years too young.

Contamination therefore cannot make a young sample appear ancient, which is what a young-earth conclusion would require.

What the method cannot do#

• It cannot date rocks, minerals, or anything that did not take up atmospheric carbon.
• It cannot date beyond roughly 55,000 years.
• It cannot date dinosaur fossils, which are outside both constraints by a wide margin.
• It gives dates as "BP," before present, fixed by convention at 1950 because atmospheric testing after that date makes later material unusable as a reference.
• Results are reported with statistical uncertainty. A radiocarbon date without an error range should not be trusted.

What the evidence shows#

The method's original assumption of constant atmospheric carbon-14 was wrong, and the correction is the calibration curve — built from tree rings and varves, which are counted rather than calculated and do not depend on radioactive decay.

Calibrated dates match historically documented ages across three millennia. Independent records from unrelated physics converge on the same curve, and revisions across four decades of updates have been small.

The known failure modes — reservoir effects, contamination, in-situ production — are documented, quantified, and corrected for, and contamination works in the direction opposite to what a young-earth conclusion needs.