Accelerator radiocarbon dating of art, textiles, and artifacts

A. J. T. Jull.

Research Scientist at the NSF Arizona AMS Facility and Department of Geosciences at the University of Arizona, in Tucson, Ariz.

This article is reproduced from Nuclear News, June 19998, and is based on a paper presented at the ANS Winter Meeting, held November 16-20, 1997, in Albuquerquete N.M. c 1998 American Nuclear Socitey. All rights Reserved.


Accelerator mass spectrometry allows present-day scientists to look into the past by radiocarbon dating of relics such as cloth, artwork, and ancient writings. Accelerator mass spectrometry (AMS) is a technique for direct measurement of the concentration of radioisotopes. Its primary use is for radiocarbon dating of small samples of carbon, although many measurements have also been made on the longer-lived radionuclides such as 26Al, 10Be, 36Cl, and 129I, which have applications to geology and marine studies.

AMS has become an accurate and precise method for dating many types of materials - including such interesting items as the Shroud of Turin and the Dead Sea Scrolls, which will be discussed later—where only a small sample can be spared. A radiocarbon measurement can be obtained on a sample of ~0.5 mg of carbon, and measured to ±40 years in uncalibrated radiocarbon age in a measurement time of 30–40 minutes on each sample.

About one carbon nucleus in a trillion contains two extra neutrons, giving a mass of 14. This carbon-14 is radioactive and decays with a half-life of 5730 years. For historical reasons, uncalibrated radiocarbon measurements are often referred to a half-life of 5568 years. However, this inconsistency is corrected during calibration [the reason for using the (Willard F.) Libby half-life of 5568 years instead of the correct one of 5730 years has to do with the finding in about 1962 that the true half-life was 5730±30 years. This creates an error in the "raw" age of about 2 percent. Since nearly all applications where the precise age is needed require calibration, this difference is removed in the calibration process].


Carbon-14 is produced in the upper atmosphere by nuclear reactions induced by cosmic rays on nitrogen (see Fig. 1). Nearly all the carbon in the atmosphere is present as carbon dioxide (CO2). The CO2 in the atmosphere maintains an equilibrium with the biosphere and the oceans. Because plants absorb carbon from the atmosphere during photosynthesis, and as animals eat plants, the animals will also contain the same level of 14C as the plants and the atmosphere. When a plant or animal dies, it ceases to take up 14C, and thus no longer maintains an equilibrium level of 14C. The amount of 14C in the carbon from this material will then decay. As with any radioactive decay, the number of 14C atoms decaying in a given time is proportional only to the number of 14C atoms present. A radiocarbon age can be calculated by comparing the amount of 14C in a sample with that in "modern" material, defined as 1950 AD. We can equally well use a different standard if we know its relation to "modern," or 1950 AD. Radiocarbon ages are then quoted as "years before present" (BP). The formula used for this calculation is:

Radiocarbon age (years BP) = -t • ln(14C in sample/14C in "modern.")

We can calculate the radiocarbon age from the Libby (Willard F. Libby) mean life of 14C (t ), 8033 years, the natural logarithm (ln) of the ratio of 14C in the sample to 14C in 1950 AD (pre-bomb) material. For practical reasons, which are discussed later, the value of "modern" is defined by reference to two primary standards of known radiocarbon content. These two standards were measured by many different laboratories to determine the value of the standards relative to "modern." Because the production rate of 14C is not a constant, we need to make corrections for this effect, as discussed in the following sections.

The first attempt to use radiocarbon for dating was the work of Libby and his co-workers, 50 years ago, using counting of the decays of the radioactive isotope. In the 1950s, gas-counting methods were perfected, and later, liquid scintillation counting has also been used, as we will discuss later. Large sample sizes were needed for both counting methods, which limited their usefulness in such applications as studies of artwork, where only small samples could be taken. Accurate dating also had to wait for a good calibration of the radiocarbon time-scale in the 1960s, using an absolute chronology based on tree rings. The radiocarbon time-scale has now been calibrated with tree rings to more than 10000 years before present, and beyond that using a coral chronology (Stuiver, et al., 1993).

The practical use of accelerator mass spectrometry was shown in 1977 by two groups simultaneously at McMasversity and at the universities of Toronto and Rochester (N.Y.) (Nelson et al., 1977; Bennett et al., 1977). The great advantage of using AMS is that we can measure the isotope ratio of 14C to stable carbon directly. The number of applications of AMS today is large, and so we will focus on a general overview of some interesting applications that will give some flavor for the variety of uses of the method.

Libby’s measurements on 14C in the 1940s were done by counting the decays of the 14C, using samples of several grams of carbon-black powder (see Anderson et al., 1946). Subsequent developments made this method obsolete, and more accurate methods using gas-proportional counters and liquid-scintillation counters were developed. These methods relied on the observation of a decay of the radioactive carbon atoms. When a 14C atom decays, it emits a beta particle, which can be counted in a gas by the electrical pulse it generates. In a liquid scintillation counter, the beta particle excites the emission of light from a complex organic molecule or "scintillant." Because only about 13.5 decays per minute occur in one gram of modern carbon, it was necessary to use fairly large samples of several grams of carbon.

It was recognized that direct measurement of the number of 14C atoms in the sample would greatly enhance the sensitivity, and several unsuccessful attempts were made in this direction using conventional mass spectrometry. In 1977, as already mentioned, two papers (Nelson et al., 1977 and Bennett et al., 1977) were published simultaneously in Science, reporting on the development of such a method, which added a particle accelerator into the mass spectrometer to produce an accelerator mass spectrometer.

This technique has allowed the measurement of radiocarbon in samples of much less than a milligram, or more than a thousand times less material than is needed for the older counting methods. This has led to a great increase in the use of 14C dating in applications to artwork, where conservation of the work requires removal of the smallest sample possible.

By the end of 1997, some two dozen AMS laboratories were in operation around the world, with more in the planning stages.

Calibration of time-scale

If the amount of 14C produced in the atmosphere were always the same, then we could calculate a "radiocarbon age" using the equation we have discussed directly as an estimate of sample age. Unfortunately, things are more complex. This was recognized soon after Libby published his first Curve of Knowns (Arnold and Libby, 1949). The cosmic rays striking the upper atmosphere fluctuate in intensity with time by a small amount due to changes in the magnetic fields of the sun and the earth. Fluctuations in the amount of carbon dioxide in the atmosphere can also affect the concentration of 14C in the CO2.

fig2 Some of these effects are illustrated in Fig. 2, which shows the increasingly large difference between radiocarbon and true age from 7000 to 15000 years BP. This deviation is much smaller less than 7000 years ago. More recently, we have learned that short-term changes in 14C in the atmosphere can be signals of climatic changes.

Because of the effects, we need to calibrate the radiocarbon age against something of known age. We can use tree rings, since they have annual growth bands and can be counted for the last 4500 years continuously. Beyond that, we can correlate the overlap of older pieces of wood to get a continuous chronology for over 11000 years.

At some periods of time, there is a smooth dependence of 14C on the known age. In others, due to fluctuations in the 14C in the atmosphere at the time the wood grew, we will get fluctuations in 14C also. This will have the effect of broadening the calibrated age range.

In these cases, the smallest possible error in the original measurement is advantageous, but may not reduce the final calibrated age range much. An example that results in a wide age range is a raw radiocarbon age of 1200±70 years BP, which will give a calibrated age range with 95 percent confidence of 660–990 AD. If we can reduce the errors to ±40 years BP, however, we will still obtain a calibrated age range of 687–975 AD. In an optimal time period, such as most of the 15th century, the calibrated age range may even be smaller than the uncalibrated age errors.

In the period of about 1700-1950 AD, the number of rapid fluctuations on 14C content due to solar activity, and also due to the addition of a lot of "dead" carbon dioxide to the atmosphere by the burning of fossil fuels, makes precise calibrated ages in this region impossible. Sometimes, some time periods can be excluded, but in general the entire range is quoted as the calibrated age. This time period has sometimes been dubbed the "Stradivarius gap" to illustrate the limitations of radiocarbon dating to age determination of some types of artwork.

fig3 After 1950, an additional source of 14C has been added to this already complex picture. Because of contamination of the atmosphere by above ground nuclear weapons tests between 1950 and 1963, periods after 1950 AD are characterized by higher than "modern" levels of 14C (Levin and Kromer, 1997). Figure 3 shows the 14C content of the post-1950 atmosphere. This actual amount of 14C can be used to " date" an object to a specific time period in the last 30 years. This spike reached its peak in 1963–64 at about twice the natural level. Since the Atmospheric Test Ban Treaty, this value has declined, due to mixing with the oceans to about 110 percent "modern" in 1997.

Accelerator Mass Spectrometry


Radiocarbon dating using AMS differs from the decay-counting methods in that the amount of 14C in the sample is measured directly, rather than after waiting for the individual radioactive decay events to occur. This makes the technique 1000 to 10 000 times more sensitive than decay counting. This sensitivity is achieved by accelerating sample atoms as ions to high energies using a particle accelerator, and using nuclear particle detection techniques. A photograph of the AMS facility at the University of Arizona is shown in Fig. 4.

The AMS method has generally improved since its inception, so that external precision of about ±0.5 percent in 14C content, or ±40 years in uncalibrated radiocarbon age are possible on 1-mg-sized samples. Samples as small as 100 µg have been successfully dated to about ±150 years BP. Experimental studies on even smaller samples are under way at several laboratories. When multiple targets are used, we can reduce the error to about ±0.25 percent or ±20 years in radiocarbon age.


Figure 5 shows a diagram of the Arizona AMS system. Some other laboratories use different equipment, but the basic principles are the same. The machine was originally designed by General Ionex Corporation, of Newburyport, Mass., although frequent improvements have been made by our laboratory. The same basic accelerator design is in use at the universities of Oxford, Toronto, and Nagoya and the Centre des Faible Radioactivités at Gir-sur-Yvettee, France, for AMS radiocarbon measurements. In 1992, we installed a new high-intensity sputter source from National Electrostatics Corporation (NEC), of Middleton, Wis.

The system consists of the following basic components and sequence of events:

  1. The ion source generates negative carbon ions (about 35 ºA C-) by Cs sputtering from a graphite target. At the University of Arizona, we measured each target for ~5 minutes, and the complete target wheel of 32 targets is rotated a total of six times, giving a total analysis time on one sample of about 30 minutes. Usually, we run one wheel of 32 targets per day.

  2. The injection magnet performs the initial separation of the negative ions by mass. At this point, molecular ions such as hydrides of carbon (CH-) are also present. N- is unstable, so an important possible interference is removed. Masses 14 and 13 are alternately injected into the accelerator.

  3. The accelerator generates a high voltage of about 2 million volts, and accelerates the C- ions toward the central part of the machine, which is at high voltage and is usually called the "terminal."

  4. The terminal houses a gas canal known as the "stripper." The 2-million-volt C- ions enter the canal and interact with a gas. Because they are moving so fast, they lose several electrons from their electron cloud, and as a result become positively charged. Any molecules, such as CH-, are destroyed in this process. The positively charged ions are accelerated away from the positively charged terminal, to the exit of the accelerator.

  5. The ions exit the accelerator, and are then separated by energy and charge, using an electrostatic deflector. This device deflects a beam of ions using an electrostatic field, and a narrow defining exit slit. The C3+ ions are selected.

  6. The C3+ ions are now passed through a set of magnets. If mass 13 is injected, the 13C beam stops in a metal cup, and the current is measured. If mass 14 is injected, the 14C passes through a second magnet, and then hits an energy-sensitive solid-state detector. This detector has the property that it produces a pulse proportional in height to the energy of the ion, for every ion hitting the detector. The number and the energy of the ions are separated by computer, and the 14C can be distinguished from any other ions which are counted. The count rate for a modern sample is around 100 counts per second.

  7. The ratio of 14C to the 13C current is compared to that for the standard samples. Appropriate corrections are made for the small natural variations of 13C/14C ratio, and for the background. The radiocarbon age can now be calculated. Finally, the radiocarbon age is calibrated using the curves we have already discussed.

Sample preparation

In order to do a measurement on a real sample, which may be quite dirty or contaminated, it must be cleaned. Up to now, we have assumed that a sample has already been removed and converted into a form from which the 14C age can be determined.

For our "age" to have meaning, we must know that a sample was removed from a representative piece of the material in question, and that all contaminants that might affect the age have been removed. The sampling of an object such as a textile is relatively straightforward. A sample should be removed from a part of the object that is clearly not a later addition, and it should be as free of visible contaminants—such as hairs, soot, plant fragments, etc.—as possible. Contaminants can have a considerable effect on ages of older materials, but for less than about 1000 years of age, the amount of contaminants required to produce a significant age effect are large. For example, a 10 percent contamination of an 800-year-old sample with modern material would produce an 80-year shift in age.

Most radiocarbon laboratories adopt a minimum "standard" pretreatment, consisting of soaking the sample sequentially in dilute hydrochloric acid, distilled water, dilute sodium hydroxide, distilled water, acid again, and then distilled water until the washing water is neutral. The acid step removes carbonates, such as from wind-blown dust, hard water, or soil, and the base step removes many soluble organic materials, such as fatty acids. For many pieces of artwork, additional chemical cleaning steps are often employed, such as successive solvent extractions using a range of organic solvents. In the case of silk samples (discussed later), we use a series of solvents employing hexane, ethanol, and methanol, followed by a final rinse in distilled water. In other cases, e.g., removal of protein-based glues from canvas, soaking the canvas sample in hydrochloric acid followed by 1 percent sodium hydroxide at 50 °C for one to two days is often required.

At the end of the chemical cleaning, the samples are dried and inspected. A sample is then ready for conversion to a form in which it can be placed in the accelerator mass spectrometer for analysis. A few milligrams of the cleaned sample is placed in a glass tube along with some cleaned copper oxide. The tube is attached to a vacuum system and evacuated. The sample is heated to about 900 °C to produce CO2. Other compounds produced are mainly water and nitrogen oxides. The pressure of cleaned and dried CO2 gas is measured in a known volume (to determine its size), and from that, the carbon content of the sample. Yields of 40–45 percent carbon are typical from most organic materials. The CO2 is then transferred to a line for graphite production. The graphite is produced by conversion of CO2 to carbon monoxide over hot zinc, and this gas is then converted to graphite over hot iron, at about 600 °C. The graphite powder (~0.5mg) samples are packed into target holders and mounted in a 32-position target wheel, which is inserted into the accelerator ion source. Eight standard graphite targets, made from oxalic acid standards I and II produced by the U.S. National Institute of Standards and Technology, are run in each wheel. Results are calculated by comparison of the 14C content of the samples to these two standards and normalization to "modern" 1950 AD. A correction is made for the small blank introduced during the chemistry, of about 0.3±0.1 percent "modern" carbon. After blank subtraction, the maximum determinable age is 49 000 years BP.

At our laboratory, we quote the uncalibrated radiocarbon ages with an error equivalent to plus-or-minus one standard deviation, or one sigma. This error range is the 68 percent confidence interval for the estimate of the age that is quoted. Two standard deviations define a 95 percent confidence interval. Usually, when a calibrated age is quoted, both one- and two-sigma ranges are reported.

As has already been discussed, the error range of the calibrated age may be significantly different from that of the error in the uncalibrated "radiocarbon age," which is quoted in years before present. For radiocarbon ages, errors of about ±50 years on recent material are typical for accelerator measurements. Higher precision is usually achieved by longer counting times, or, preferably, measurements on multiple targets of the same material. The best errors quoted using high-precision AMS are about ±20 years (see later).

Related to the question of quoted measurement precision is the question of accuracy. This refers to the ability to obtain the correct value for a sample of known age, within the measurement errors. Repeated measurements of the same sample should scatter around the accepted "true" value, with a distribution characterized by the standard deviation of the measurements. Comparisons of results among different laboratories is another check of accuracy. The best intercomparisons of this kind are done on samples where the different laboratories do not know the expected age beforehand. Intercomparison checks are performed every few years among radiocarbon laboratories using different techniques.

Examples of applications

Shroud of Turin dating

fig6 The Shroud of Turin is a textile which is of great general interest. It has a very unusual image of what appears to be a crucified man, and has been considered by many to be possibly the burial cloth of Jesus Christ (Fig. 6). This cloth will be exhibited in Turin this year (1998), for the first time in 20 years. Its recorded history dates back to the 14th century in France, with an uncertain record prior to that time.

Because of the controversy surrounding this textile, radiocarbon dating was needed to establish its age. Before the advent of AMS dating, however, an unacceptably large amount of cloth would have had to have been removed from the Shroud for dating. After many years of discussions, samples of the Shroud were finally removed in April 1988 by scientists from Turin and the British Museum. Samples were removed from a corner of the main cloth, which had previously been sampled in 1973. Each of the three laboratories involved in the measurements (at the universities of Arizona and Oxford, and at the Swiss Federal Institute of Technology, Zurich) received approximately 2 cm2 of cloth (about 50 mg), and three control samples. fig7 A complete discussion of these results is reported in the journal Nature, by Damon et al. (1989). The Shroud measurements show a good example of the effects of the calibration curve. The mean value of the three laboratories is 691±31 years BP. If we plot this value on the calibration curve (Fig. 7), we can see that the calibrated age range is smaller than ±30 years, at one sigma (1273–1288 AD), but at 2 sigma there are three intercepts on the curve due to a bump at about 1380 AD. At 2 sigma, the calibrated age range is then 1262–1384 AD. This age range, however, is not much larger than the 2-sigma range of the uncalibrated age of ±60 years. As these fluctuations in the calibration are quite common, reducing the error in calibrated age is not always possible. At a specific point in the calibration, such as at about 1280 AD, a steep slope in the curve allows very specific age assignment. Using the curve in Fig. 7, we could calculate the errors for 3 sigma, which defines a confidence interval of 99.5 percent. Although scientists have no doubt about the age of the Shroud, this result is still a matter of much discussion by those who do not accept these scientific results. The interested reader can look at the Web site for some other opinions.

Dead Sea Scrolls

The Dead Sea Scrolls is the name given to a large number of parchment and papyrus manuscripts found in the area of Qumran and elsewhere in the Judean Desert. The Qumran scrolls are generally considered to be the work of a Jewish religious group called the Essenes, who lived in this region up until 68 AD, when the area was destroyed by the Romans. The conventional wisdom is that the Essenes first came into existence in the early second century BC (see, for example, Shanks, 1992). The scrolls consist of three distinct types of documents.The first are complete books of the Bible in ancient Hebrew. The second are interpretations (or "pesher") of biblical texts by the writers, which often show quite extreme or apocalyptic interpretations. There are some other literary scrolls related to the ideas of the writers. Third, there are rules of the community and business documents of the time. Finally, there are two different scrolls, one called the Temple scroll, giving detailed plans for the temple and religious rules, and another, the Copper scroll, giving the location of some buried treasure.

The chronology of the Dead Sea Scrolls is important, and their ages were originally estimated using the technique of paleography, that is, the age assigned to a particular style of writing. Paleography assigns most of the scrolls to 150–30 BC, with a few dating to earlier times (because of an archaic script used). Obviously, a real question is the validity of the paleographic age assignment.

The Zurich AMS laboratory had performed some AMS dating of scrool in 1991 (see Bonani et al., 1992), and we at the University of Arizona performed a new set of measurements in 1994. All the samples we studied were removed from the actual scrolls in April 1994 at the Rockefeller and Israel Museums in Jerusalem, apart from some material samples collected later in 1994. Small areas of parchment of about 0.25 cm2 were sampled. These samples were taken to our laboratory in Tucson for cleaning and AMS measurements.

fig8 The calibrated ages of the scrolls showed that they were made from about 150 BC to about 80 AD. The material used was sheepskin parchment and sometimes papyrus. Part of the apparently wide range is due to the shape of the tree-ring calibration curve in the 1st and 2nd centuries BC. Our results showed remarkable agreement with the paleographic dates. Figure 8 shows the radiocarbon ages of scrolls plotted on a "calibration curve" from tree rings, where the "age" of the samples on the x-axis is the paleographic age.

The Isaiah scroll (see Fig. 9) is an excellent example of a complete book of the Bible, which is on display at the Shrine of the Book at the Israel Museum. This document gave a radiocarbon age of 2141±32 years BP. We used this date as a test against some earlier studies by Bonani et al. (1992), who obtained an age of 2128±38 years BP. We can combine these two measurements to obtain a mean of 2136±24 years BP. Calibrating these dates gives a split range of 352–309 BC (13 percent probability) and 210–103 BC (86 percent probability). This sample’s expected age based on its style is 150-125 BC.

fig9 Another document, a commentary on the Book of Habbakuk (one of the "pesher" mentioned earlier), contains allusions to a prophetic figure called the "Teacher of Righteousness." Some writers have speculated that this literary figure can be associated with Christ. Our radiocarbon age of 2054±22 years BP almost certainly assigns this document to the pre-Christian period, with a calibrated age at 120–5 BC (95 percent confidence). This also demonstrates the achievable precision of such work. The Dead Sea Scroll measurements were done with precision in radiocarbon age of as little as ±20 years, which are some of the best measurements ever made by AMS 14C dating. Our results are reported in two papers (Jull et al., 1995) for those interested in more detail.

Asian Textiles

fig10 Another application of AMS to the dating of artistic works has been to a large number of Asian textiles, particularly silks. Most of the samples we have studied originated either from museums or from art dealers. Many of the materials appear on the market from time to time, sometimes from unidentified sources. It has become critical to buyers of such textiles that they know the age of the material. Dating silk can be problematical, as cleaning the material is very important. We have adopted a series of sequential solvent extractions in addition to the standard acid-base-acid pretreatment for silk samples. Some of the silks dated are very well preserved and quite beautiful in appearance.


Figure 10 shows a 16th century Indian textile that we dated on behalf of a commercial client. This textile, showing elephants, was dated to 1440–1650 AD (95 percent confidence). Two other examples are also shown. Figure 11 shows a blue embroidered robe collar with birds in flight.This material was expected to be Yuan dynasty (1279–1368 AD). Our 95 percent confidence calibrated age for this sample was 1328–1454 AD. We have also applied our methods with great success to other textiles, rugs, and carpets.

Wooden materials

fig12 Dating of wooden materials such as statues, furniture, tools, and other implements can also be carried out using AMS methods. An example of an earlier measurement by our laboratory (Fig. 12) is shown as an example. In the case of wooden artifacts, the age of the wood may add to the apparent age of the object. For example, if someone were to make a statue from a tree which had 100 years of tree rings, the age of the older wood might give an apparent age older than the actual time of fabrication. This problem can be minimised by taking several samples from opposite ends of the wood.


Not all samples turn out to be the "right" age. A big problem in the art world is the number of forgeries that appear for sale. fig13 This creates a problem, and independent and nonsubjective means of dating the materials is important. The small-sample capabilities of AMS radiocarbon dating mean that we can remove a few square millimetres of wood, canvas, silk, or other material for dating. In the case of paintings, dating a sample of the canvas provides a way of getting good information on the age of the artwork. Naturally, since this can greatly affect the value of the object, we need to be sure we have a good sample for dating. As a result, we get phone calls from all manner of sources: museums, art galleries, and private individuals. One example is shown in Fig. 13. This is a painting by a Cuban artist, Rene Portacarerro, which bears a date of 1945 AD. However, the canvas from this painting contains a 30 percent excess of bomb 14C. As we have already mentioned, this constrains the age to one or two possible years. In this case, the canvas dates it to 1962 or 1977 AD. The source of the sample often gives us some hints of the possible problems even before analysis. It is rare that a museum is surprised with our results. We have not as yet seen a valuable painting that came from someone’s attic or was left by their grandfather to be a genuine article.

Other applications

There are many other important applications of AMS radiocarbon dating. First, paleoclimatology, as a significant part of the effort at our laboratory, is devoted to studies related to global climate change. In any such study, accurate dating of the material is a prerequisite. Also, as we mentioned earlier, some changes in the 14C content of the atmosphere can be triggered by climatic changes caused by ocean circulation. Second, we have pioneered studies of in situ 14C produced directly in rock surfaces by cosmic ray spallation reactions. This "in situ 14C" allows us to estimate surface exposure ages in geomorphological samples.

Several important applications for this technique exist, which include dating of fault scarp motions for earthquake hazard estimates, determining the episodic nature of volcanic eruptions, and estimation of erosion rates, especially when combined with data from other radionuclides such as 10Be. Radiocarbon dating of earthquake faults where charcoal has been deposited is a way of estimating the recurrence time of earthquakes.

Another interesting example of an application of radiocarbon is to the forest fire history in Yellowstone National Park. We were able to show a strong correlation between fire frequency, climate, and solar cycles. Fires became much more frequent during periods identified from local and regional climate proxy records as warm and dry, and a major pulse of fire-related debris-flow activity (enhanced flows of sediment and debris, including burned trees, in the years following large forest fires due to increased erosion) occurred at the height of the Medieval Warm Period, ca. 1050-1200 AD.

An important tool

AMS radiocarbon dating has become an important tool in age verification for works of art. Given appropriate consideration of its limitations, very specific age ranges can be assigned in most cases. Some samples fall into age ranges that result in a large spread of calibrated age. Nevertheless, a result that was expected to be 10th century and turns out to be 1700–1950 AD can give as much information as distinguishing between samples in a region for which very precise age ranges can be quoted.


Anderson, E. C., W. F. Libby, S. Weinbourn, A. F. Reid, A. D. Kirshenbaum, and A. V. Grosse, Phys. Rev. 72 (1947): 931.

Arnold, J. R. and W. F. Libby, "Radiocarbon dates," Science 113 (1951): 111–120.

Bard, E., M. Arnold, R. G. Fairbanks, and B. Hamelin 230Th-234U and 14C ages obtained by mass spectrometry on corals," Radiocarbon 35, no. 1 (1993): 191–199.

Bennett, C. L., et al., Science 198 (1977): 508–509.

Bonani, G., S. Ivy, W. Wölfli, M. Broshi, I. Carmi, and J. Strugnell, "Radiocarbon dating of 14 Dead Sea Scrolls," Radiocarbon 34, no. 3 (1992): 843–849.

Damon, P. E., et al., "Radiocarbon dating of the Shroud of Turin," Nature 337 (1989): 611–615.

Jull, A. J. T., D. J. Donahue, M. Broshi, and E. Tov, "Radiocarbon dating of scrolls and linen fragments from the Judean desert," Radiocarbon 37, no. 1 (1995): 11–19. (Published in slightly different form also as ‘Atiqot [Journal of the Israeli Antiquities Authority], 28, 86–91.)

Nelson, D. E., et al., Science 198 (1977): 507–508.

Levin, I. and B. Kromer, "Twenty years of atmospheric 14CO2 observations at Schauinsland Station, Germany," Radiocarbon 39, no. 2 (1997): 205–218.

Shanks, H., Understanding the Dead Sea Scrolls (1992), Random House, New York.

Stuiver, M. M., A. Long, and R. S. Kra, "Calibration 1993," Radiocarbon 35 (1993).

Taylor, R. E., Radiocarbon dating: An archaeological perspective (1987), J. Wiley and Sons, New York.