The occurrence of natural radioactive carbon in the atmosphere provides a unique opportunity to date organic materials as old as roughly 60,000 years. Unlike most isotopic dating methods, the conventional carbon-14 dating technique is not based on counting daughter isotopes. It relies instead on the progressive decay or disappearance of the radioactive parent with time.

The discovery of natural carbon-14 by American chemist Willard Libby of the United States began with his recognition that a process that had produced radiocarbon in the laboratory was also going on in Earth’s upper atmosphere—namely, the bombardment of nitrogen by free neutrons. Newly created carbon-14 atoms were presumed to react with atmospheric oxygen to form carbon dioxide (CO₂) molecules. Radioactive carbon thus was visualized as gaining entrance wherever atmospheric carbon dioxide enters—into land plants by photosynthesis, into animals that feed on the plants, into marine and fresh waters as a dissolved component, and from there into aquatic plants and animals. In short, all parts of the carbon cycle were seen to be invaded by the isotope carbon-14.

Invasion is probably not the proper word for a component that Libby calculated should be present only to the extent of about one atom in a trillion stable carbon atoms. So low is such a carbon-14 level that no one had detected natural carbon-14 until Libby, guided by his own predictions, set out specifically to measure it. His success initiated a series of measurements designed to answer two questions: Is the concentration of carbon-14 uniform throughout the plant and animal kingdoms? And, if so, has today’s uniform level prevailed throughout the recent past?

After showing the essential uniformity of carbon-14 in living material, Libby sought to answer the second question by measuring the radiocarbon level in organic samples dated historically—materials as old as 5,000 years from sources such as Egyptian tombs. With correction for radioactive decay during the intervening years, such old samples hopefully would show the same starting carbon-14 level as exists today. This was just what Libby’s measurements indicated. His conclusion was that over the past 5,000 years the carbon-14 level in living materials has remained constant within the 5 percent precision of measurement. A dating method was thus available, subject only to confirmation by actual application to specific chronologic problems.

Since Libby’s foundational studies, tens of thousands of carbon-14 measurements of natural materials have been made. Expressed as a fraction of the contemporary level, they have been mathematically converted to ages through equation 5 above. Archaeology has been the chief beneficiary of radioactive-carbon dating, but late glacial and postglacial chronological studies in geology have also been aided greatly.

Improvements in measurement accuracy and the ever-mounting experience in applying carbon-14 dating have provided superior and more voluminous data with which to better answer Libby’s original questions. It is now clear that carbon-14 is not homogeneously distributed among today’s plants and animals. The occasional exceptions all involve nonatmospheric contributions of carbon-14-depleted carbon dioxide to organic synthesis. Specifically, volcanic carbon dioxide is known to depress the carbon-14 level of nearby vegetation, and dissolved limestone carbonate occasionally has a similar effect on freshwater mollusks, as does upwelling of deep ocean water on marine mollusks. In every case, the living material affected gives the appearance of built-in age.

In addition to spatial variations of the carbon-14 level, the question of temporal variation has received much study. A 2 to 3 percent depression of the atmospheric radioactive-carbon level since 1900 was noted soon after Libby’s pioneering work, almost certainly the result of the dumping of huge volumes of carbon-14-free carbon dioxide into the air through smokestacks. Of more recent date was the overcompensating effect of man-made carbon-14 injected into the atmosphere during nuclear bomb testing. The result was a rise in the atmospheric carbon-14 level by more than 50 percent. Fortunately, neither effect has been significant in the case of older samples submitted for carbon-14 dating. The ultimate cause of carbon-14 variations with time is generally attributed to temporal fluctuations in the cosmic rays that bombard the upper atmosphere and create terrestrial carbon-14. Whenever the number of cosmic rays in the atmosphere is low, the rate of carbon-14 production is correspondingly low, resulting in a decrease of the radioisotope in the carbon-exchange reservoir described above. Studies have revealed that the atmospheric radiocarbon level prior to 1000 bce deviates measurably from the contemporary level. In the year 6200 bce it was about 8 percent above what it is today. In the context of carbon-14 dating, this departure from the present-day level means that samples with a true age of 8,200 years would be dated by radiocarbon as 7,500 years old.

The problems stemming from temporal variations can be overcome to a large degree by the use of calibration curves in which the carbon-14 content of the sample being dated is plotted against that of objects of known age. In this way, the deviations can be compensated for and the carbon-14 age of the sample converted to a much more precise date. Calibration curves have been constructed using dendrochronological data (tree-ring measurements of bristlecone pines as old as 8,200 years); periglacial varve, or annual lake sediment, data (see above); and, in archaeological research, certain materials of historically established ages. It is clear that carbon-14 dates lack the accuracy that traditional historians would like to have. There may come a time when all radiocarbon ages rest on firmer knowledge of the sample’s original carbon-14 level than is now available. Until then, the inherent error from this uncertainty must be recognized.

A final problem of importance in carbon-14 dating is the matter of sample contamination. If a sample of buried wood is impregnated with modern rootlets or a piece of porous bone has recent calcium carbonate precipitated in its pores, failure to remove the contamination will result in a carbon-14 age between that of the sample and that of its contaminant. Consequently, numerous techniques for contaminant removal have been developed. Among them are the removal of humic acids from charcoal and the isolation of cellulose from wood and collagen from bone. Today contamination as a source of error in samples younger than 25,000 years is relatively rare. Beyond that age, however, the fraction of contaminant needed to have measurable effect is quite small, and, therefore, undetected or unremoved contamination may occasionally be of significance.

A major breakthrough in carbon-14 dating occurred with the introduction of the accelerator mass spectrometer. This instrument is highly sensitive and allows precise ages on as little as 1 milligram (0.001 gram [0.00004 ounce]) of carbon, where the older method might require as much as 25 grams (0.9 ounce) for ancient material. The increased sensitivity results from the fact that all of the carbon atoms of mass 14 can be counted in a mass spectrometer. By contrast, if carbon-14 is to be measured by its radioactivity, only those few atoms decaying during the measurement period are recorded. By using the accelerator mass spectrometer, possible interference from nitrogen-14 is avoided, since it does not form negative ion beams, and interfering molecules are destroyed by stripping electrons away by operating at several million volts.

The development of the accelerator mass spectrometer has provided new opportunities to explore other rare isotopes produced by the bombardment of Earth and meteorites by high-energy cosmic rays. Many of these isotopes have short half-lives and hence can be used to date events that happened in the past few thousand to a few million years. In one case, the time of exposure, like the removal of rock by a landslide, can be dated by the presence of the rare beryllium-10 (¹⁰Be) isotope formed in the newly exposed surface of a terrestrial object or meteoroidal fragment by cosmic-ray bombardment. Other applications include dating groundwater with chlorine-36 (³⁶Cl), dating marine sediments with beryllium-11 (¹¹Be) and aluminum-26 (²⁶Al), and dating glacial ice with krypton-81 (⁸¹Kr). In general, the application of such techniques is limited by the enormous cost of the equipment required.