Carbon-14 dating can be reliable but cannot be applied to anything older than 80,000 years at the maximum.
Here is some information from a geology textbook I have:
Carbon-14 dating was first developed in the 1940's for use in archeology. It is the only isotopic decay system that occurs naturally in sedimentary rocks and fossils, so it is an exception to the rule that sedimentary rocks cannot be dated directly. The half-life of the system is so short (5,730 years) that all the primordial carbon-14 decayed long ago. However, carbon-14 is continually produced in the atmosphere by the bombardment of nitrogen-14 by cosmogenic neutrons. The nitrogen-14 loses a proton, thereby becoming carbon-14. This atmospheric radioactive carbon is assimilated into the carbon cycles of plants and animals while they live. The decay clock starts at the moment the tissue is no longer living and stops exchanging carbon with the environment. The nuclear reaction is:
Neutron + Nitrogen-14 yields Carbon-14 + proton
And the corresponding decay reaction is
Carbon-14 yields Nitrogen-14 ＋;;;;;Beta particle ＋;;;;;energy
The age equation is
t = 19.035 * 10^3 * log(A/A naught) in years
where A is the measured activity of the sample in distintegrations per minute per gram of carbon (dpm/g), and A naught is the initial activity. The currently accepted value for A naught is 13.56 + or - 0.07 dpm/g.
Carbon-14 dating is based on two assumptions: first, that the rate of carbon-14 production in the upper atmosphere is constant; second, that the rate of assimilation of carbon-14 by organisms is rapid relative to its rate of decay. Both of these assumptions appear to be valid.
The procedure for measuring carbon-14 in a sample is quite different from that for any other isotopic system. First, the sample is burned in a vacuum chamber to produce carbon dioxide, part of which is composed of carbon-14. The carbon dioxide gas is run through a chamber from which most of the other atmospheric gases have been removed. It is then concentrated in a copper counting tube, where a Geiger counter measures the decay activity. During measurement, the sample must be shielded from atmospheric gamma rays and neutron bombardment. In many cases, only a few grams of material are needed, but in others several kilograms may be required. Even with sufficient material, it takes 12 to 24 hours of counting per sample to get a measurement. For precision, multiple samples are counted and recounted to screen out background noise and to get consistent measurements with small error estimates.
The main limitations of carbon-14 dating for geology is its short half-life. The practical upper age limit is about 40,000 years, although with extraordinary care and extremely sensitive equipment, marginal dates of 60,000 to 80,000 years have been produced. Another problem has been uncovered by analyzing the carbon in dated tree rings. These studies show that the relative amount of atmospheric carbon-14 has fluctuated tremendously in historic times, decreasing significantly in the last 100 years as a result of the burning of trees and fossil fuels. Nuclear explosions have also added slight amounts of atmospheric carbon-14. A correction must be added to the date to compensate for these effects.
There are other pitfalls as well. Marine shells are notorious for picking up carbon that has been dissolved in seawater from ancient carbonate rocks. This ancient carbon is radioactively "dead" and depresses the carbon-14 count, giving the impression that the shell is much older than it really is. This is why an occasional living shell gives a very ancient date. In general, however, carbon-14 is the primary dating tool for Quaternary geology and archeology. It is possible to date directly any organic material that was once living, such as shells, wood, peat, bones, baskets, cloth, paper, food materials, and even some pottery and iron objects. Carbon-14 dating proved critical in dating the ice ages as well; undoubtedly, new uses will continue to be found for it.
Materials Suitable for Dating by the Carbon-14 Method
Material: Charcoal and wood
Amount required: 25 grams
Comments: Usually reliable, except for finely divided charcoal, which may adsorb humic acids; removable by treatment with NaOH. Subject to "post-sample-growth error"; that is, the difference in time between growth of a tree and use of the wood by humans.
Material: grains, seeds, nutshells, grasses, twigs, cloth, paper, hide, burned bones
Amount required: 25 grams
Comments: Usually reliable. These materials are short-lived and have negligible post-sample-growth errors.
Material: Organic material mixed with soil
Amount required: 50-300 grams
Comments: Should contain at least 1% organic carbon in the form of visible pieces. Efforts should be made to remove as much soil as possible in the field.
Amount Required: 50-200 grams
Comments: Often reliable, but intrusive roots of modern plants must be removed. The coincidence of peat formation with the occupation of archaeological sites requires careful consideration.
Amount Required: 50 grams
Comments: Often well preserved and reliable. Interior of tusks is younger than the exterior. Some ivory tools may have been carved from old rather than contemporary material.
Amount Required: 300 grams
Comments: Heavily charred bones are reliable. Lightly charred bones are not because exchange with modern radiocarbon is possible.
Amount Required: 1,000 grams or more
Comments: Organic carbon in bones, called collagen, is reliable. However, the organic carbon content is low and decreases with age to less than 2 parts per hundred
Material: Shells (inorganic carbon)
Amount needed: 100 grams
Comments: The carbon in the calcite or aragonite of shells may exchange with radiocarbon in carbonate-bearing groundwater. Shell isotope carbon may be initially enriched in carbon-14 relative to wood due to fractionation. It may also be depleted in carbon-14 due to incorporation of "dead" carbon derived by weathering of old carbonate rocks. The reliability of shell dates is therefore questionable.
Material: Shell (organic carbon)
Amount required: Several Kilograms
Comments: Organic carbon is present in the form of conchiolin, which makes up 1 part per hundred to 2 parts per hundred of modern shells. Dates may be subject to systematic errors due to uncertainty of initial carbon-14 activity of this material.
Material: Lake marl and deep-sea or lake sediment
Amount Required: Variable
Comments: Such material are datable on the basis of the radiocarbon content of calcium carbonate. Special care must be taken to evaluate errors due to special local circumstances.
Material: Pottery and iron
Amount Required: 2 to 5 kilograms
Comments: Pottery sherds and metallic iron may contain radiocarbon that was incorporated at the time of manufacture. Reliable dates of such samples have been reported.