Using Neutron Activation Analysis to Establish the Provenance of Pottery

030

During the last 25 years, professional archaeologists have become accustomed to relying on a wide variety of experts, both on the dig and afterward. Today no excavation would go into the field without an architect and photographer, or hesitate to call in a numismatist or bone specialist to study the collected material. A geologist, a paleobotanist and an anthropologist are other frequent members of an expedition staff. Scientific techniques like carbon 14 and thermoluminescence are often used for dating. And now the nuclear physicist is helping solve the problem of where pottery was manufactured.

The provenance of pottery has always been of concern to archaeologists, for provenance may help to answer important questions of cultural influence, patterns of trade and, maybe, even trace migrations.

Elaborate pottery classifications based on form and decoration enable the archaeologist to determine chronological, cultural and even regional differences. (For example, the Iron Age pottery of the Northern Kingdom of Israel shows distinct differences from the pottery of the contemporaneous Southern Kingdom of Judah.) However, deciding what is an import and what is a good imitation is still a hard knot to untie.

Neutron activation analysis is the most recent technique for untying it.

Understanding this new technique requires an understanding of the nature of pottery. Pottery is made of clay to which other material, such as crushed rock or powdered straw, has been added to decrease shrinkage and increase strength (temper). The shaped object is then heated until certain chemical changes occur. (These chemical changes prevent the vessel from turning into mud every time it is washed.)

The clay itself is formed from fine rock particles which result from the weathering of larger rocks. Trace elements from the “parent” rocks thus become a part of the clay. (A trace element is an element which is found in minute quantities that range from unmeasurable to approximately 5%.) Obviously clays from different parent rocks contain different trace elements in different percentages, depending on the geological processes by which the rocks were formed; and this is the basic datum on which neutron activation analysis operates.

Because the usual methods of chemical analysis cannot detect the presence of trace elements, nuclear physics has entered the breach. Each trace element can be identified and measured because it has at least one radioactive isotope which decays (that is, returns to a non-radioactive state) at a specific rate by giving off nuclear particles of specific energy. The rate of decay is called the “half-life”—the time required for the isotope to decay 50%.

Some elements have too short a half-life to be useful in this kind of analysis. For example, aluminum has a half-life of 2.3 minutes. Other elements such as bromine volatilize and disappear during the firing of the vessel. However, scientists have now settled on about twenty elements which can be used for trace element identification. As to these elements, the technique is so sensitive that it can detect one part in a million.

To obtain the clay samples, the surface of the vessel is first scraped clean to remove possible contamination from paint, slip or weathering. Then, between 50 and 100 milligrams of clay are taken either by a 031sapphire scraper or a drill bit. A scraper rather than a drill bit is used on thin pottery. A drill bit is used to take the sample from the edge of a potsherd, so if the vessel is restored no sampling marks can be seen.

The preparation of the sample differs according to the laboratory. Lawrence Radiation Laboratory mixes 100 mg. of clay with a cellulose binder and compacts the mixture into a pill using a hand operated hydraulic press. Each pill is then wrapped in pure aluminum foil and placed in a cylindrical capsule. Brookhaven National Laboratory selects a 40 mg. sample which is then placed in a fused quartz capsule which in turn is placed in a glass container within a cylindrical capsule (a total of three capsules).

The samples are then placed in a nuclear reactor and irradiated with neutrons. The chemical elements in the clay are thus made radioactive, so that their energy levels can be measured as they decay and gradually return to non-radioactive state.

After the samples are taken from the nuclear reactor they are placed in the “counter”—the colloquial name for a germanium gamma ray spectrometer—which “counts” or measures the energy levels given off by various radioactive trace elements, thus allowing each element to be identified and quantified.

Every group of clay samples which is analyzed includes reference material; that is, samples of local clay obtained either from local clay beds or from pottery of unquestioned local manufacture. These reference samples are used for comparison with the samples of unknown provenance. (In addition, each laboratory currently performing neutron activation analysis—the University of California at Berkeley, Brookhaven National Laboratory in Upton, New York, and the Hebrew University in Jerusalem—is building up a “library” of known clay samples which also serve as reference samples for comparative purposes.)

When the data comes from the counter, the amount of each trace element has been determined for each sample—both for the reference material and the samples of unknown provenance. The data can then be analyzed in two ways. The first method involves calculating the mean or average amount of each trace element in the samples of known local clay. Since there is always some variation among samples, the amount of this variation is measured by computing the standard deviation, which is a measure of the homogeneity of a particular trace element in a group of samples. A group of similar samples will have a small standard deviation, while a disparate group will have a large one. Each sample of unknown provenance is then compared, element by element, to the group of locally manufactured samples. If the count for an element in the sample of unknown provenance lies at two standard deviations or more from the mean of the locally manufactured group then the odds are 20 to 1 that the sample of unknown provenance does not belong to that group; that is, it does not come from the same claybed as the reference samples and is therefore not of local manufacture. Since this is done for every element, and the total odds on whether or not a sample is of local manufacture is the product (not the sum) of all the odds, it is easy to tell whether a sample is of local manufacture. Once a sample is determined to be not of local manufacture, it is then compared with other reference samples in the laboratory’s “library” to determine whether a likely provenance can be identified from them for the unknown sample.

The second method of analysis compares every sample to every other sample (including reference samples) for all elements. This will result in a number which sums up the amount of similarity between every possible pair of samples. The group of samples is then rearranged until every sample lies next to the sample most similar to itself. All samples from the same area of manufacture will form a group or “cluster” since all will have high indices of similarity to each other. Imports will not lie in the same cluster as local ware but may form another cluster if they were manufactured in the same place.

To illustrate the use of neutron activation analysis in attempting to solve a particular archaeological problem, we may consider the technique as recently applied to Philistine material from Ashdod. Where the Philistines came from has long troubled archaeologists. The most similar pottery to Philistine ware is Mycenean IIIC1 pottery, 032which is found both in the Aegean and Cyprus. The Bible identifies Ashdod as a Philistine stronghold and excavations there by Dr. Moshe Dothan revealed a Philistine stratum in which both Philistine and Mycenean IIIC1 sherds were found. Five Philistine sherds and five Mycenean sherds from Ashdod were analyzed (by the first method described above), along with Mycenean sherds from three sites on Cyprus. The clay in the Mycenean sherds from Ashdod matched the Philistine ware from Ashdod, but not the sherds from Cyprus. So the Mycenean pottery from Ashdod was made from local clay and was not an import. Had the Mycenean samples matched one of the Cypriote groups, we would have known where this pottery came from—pottery which the Philistines might have brought with them from their original homeland.

Since the Mycenean pottery from Ashdod was manufactured only from local clay, the question of the original homeland of the Philistines remains unanswered. But if further excavation produces Mycenean pottery which the Philistines brought with them when they arrived in Palestine, then neutron activation analysis may help solve the problem of Philistine origins.

Neutron activation has also been used on 16 of the coffins from Deir el-Balach, discussed elsewhere in this issue of the BAR. This analysis has established that these coffins are of local origin, rather than having been imported from Egypt. With this information, Dr. Trude Dothan can proceed with some confidence to search for a local installation where the coffins were manufactured. Perhaps even more important, we know that the highest level of style, technique and art among the Deir el-Balach coffins was of Canaanite manufacture.

(For further details, see I. Perlman and F. Asaro, “Pottery Analysis by Neutron Activation,” Archaeometry, Vol. 11, p. 21 (1969); F. Asaro, I. Perlman, and M. Dothan, “Mycenean IIIC1 Ware from Ashdod”, Archaeometry Vol. 13, p. 169 (1971); I. Perlman, F. Asaro, and Trude Dothan, “Provenance of the Deir el-Balah Coffins,” Israel Exploration Journal, Vol. 23, p. 147 (1973).

During the last 25 years, professional archaeologists have become accustomed to relying on a wide variety of experts, both on the dig and afterward. Today no excavation would go into the field without an architect and photographer, or hesitate to call in a numismatist or bone specialist to study the collected material. A geologist, a paleobotanist and an anthropologist are other frequent members of an expedition staff. Scientific techniques like carbon 14 and thermoluminescence are often used for dating. And now the nuclear physicist is helping solve the problem of where pottery was manufactured. The provenance of pottery has […]