III. NEUTRON ACTIVATION OF POTTERY (Abstract)

I. Perlman and F. Asaro

Department of Chemistry and Lawrence Radiation Laboratory

University of California, Berkeley

Significant progress has been made in neutron activation analysis of
pottery as a result of a combination of factors: a powerful reactor; the

novel lithium-drifted (silicon and germanium) detectors, which combine high
resoiving power with high sensitivity; high-quality computation facilities;

and the accumulation of considerable experience and of a background of data.
It is now possible to determine over 40 different elements in the pottery
samples, although such a complete analysis is usually not made. Many of

these elements can be determined to 1 percent, even if present only in con-
centrations of one part per million. In other elements, precision is only

a few percent, and some can be estimated with even less precision. Because
of the nature of the activation process, neutron activation analysis act-
ually tends to be most sensitive for the elements normally present in low
concentration.

The fundamental composition of the clays is very similar, but the minor
and trace elements provide a characteristic fingerprint; thus, if enough
elements are determined, it should be possible to characterize clays of a
particular area uniquely.

The analysis of pottery is valuable because clay is one of the most
ancient materials and is universally associated with man. People settling
in a new area seem to have brought their pottery with them, but they soon

started to use the clay that was available locally, while continuing to make

ceramic ware in the style of their country of provenience. Therefore, analysis
together with comparison of style provides good evidence of migration and
transplanting of groups of people.

Neutron activation analysis yields a wealth of data, which, by their
very quantity, present difficulties and bring out intrinsic complexities.

The first steps in handling this information are to collect local pottery from
one spot, establish a standard for the local production, and group the data.

The dispersion of the individual groups is then established by statistical ana-
lysis. Any particular sample can be fitted into a particular group with a pro-
bability dependent on the dispersion of this group. An interesting example of
the acuity of the procedure was found in the analysis of the clay plugs used
for sealing certain predynastic Egyptian jars. The clay plugs were found to

agree very well in their composition with that of "Nile mud" pottery found in
the vicinity, and the body of the jars was quite similar to clay that was also
found locally but which represented a material apparently used for different
purposes.

-21-

22

An important question is how many (and which) elements are sufficient to

characterize a shard adequately.   Sometimes, but notusually, one or two elements
are sufficient. The clay found in a place in southern Israel, for instance, is
characterized by a high hafnium content (12 parts per million), whereas most

other clays contain only 2 to 3 parts per million. In an analysis of a group
of Cypriot pottery, one piece was found to contain this high concentration of
hafnium, and the rest of the composition was subsequently found to match the
clay from Israel.

Thus fAr, 1,400 pieces of Cypriot pottery have been analyzed to obtain

a background on this type of ceramic ware. To provide a reliable analysis for
an unknown Cypriot sample, an estimated 10,000 pieces will be needed.    Simi-
larly, an adequate study of the Mediterranean pottery would require about

100,000 pieces. The present rate of analysis is about 2,000 pieces per year.
The rate of analysis may be increased, but the difficulty of handling and,

particularly, of recalling the information needed for interpretation of results
increases with the rate of acquisition of data.

To achieve high accuracy, it is necessary to irradiate a calibrated

sample, as well as the unknown, for comparison. About 2 years was required

to develop a reliable, homogeneous sample of suitable quantity, one which had
a representative composition and filled all requirements for such a standard.
Analysis can be made on a very small sample, but usually a 100-mg sample is

taken by use of a sapphire drill. Even this quantity is small enough so that
it can usually be removed in an inconspicuous place, thus causing no visible
damage to valuable pieces.