M. Járó - L. Költő szerk.: Archaeometrical research in Hungary (Budapest, 1988)

Dating - BENKŐ Lázár: Thermoluminescence dating of Hungarian archaeological sites (potteries, hearths, calcite)

From a measurement of the TL emitted by the quartz sample (natural TL, NTL), and by a proper laboratory calibration giving known artificial laboratory doses (artificial TL, ATL), one can estimate the archaeologically accumulated radiation dose. By comparing this latter with the annual dose (dose-rate) received by the quartz grains, the absolute age can be given by the equation Total dose since kiln-firing Age (in years) = Annual dose In this simple age equation, many complicating factors are involved. Some of them, which are relevant to our investigations, are discussed in the following sections where the experimental techniques and results will be given along with the estimation of error limits. It will be seen that a number of inherent uncertainties imply a severe limitation in improving the accuracy of the TL age. The overall error is between ±5 and ±10% of the age. However, TL is considered as a useful technique in archaeology, being potentially the unique tool to date a great variety of inorganic materials even beyond 50 000 years — which is approximately the upper limit for radiocarbon dating. As far as findings of fired clay are concerned, in the present practice TL dating should be reserved for sites on which there is insufficient material for radiocarbon, or an ambiguity exists in the radiocarbon results. Experimental techniques and procedures To apply the above age equation, reliable and well-reproducible NTL, ATL and dose-rate measurements are required. During the first period of our investigation (appr. ten years ago), TL measurements were performed with a Harshaw 2000A,B TL analyser [2, 3]. As far as the main features of this apparatus, in comparison with the standard model, are concerned, the max. temperature was extended up to 600 °C, with the possibility of integration between any two predetermined temperatures. The optical filters can easily be interchanged. In addition to this, TL measurements required a higher flexibility in the heating programme, too. This has been achieved by developing an appropriate circuitry for preheating. From 1985 on, the Daybreak photon counting system has been used. This apparatus is designed specifically for TL research in natural materials. It is modular, expandable and compatible with a wide range of system components. The temperature controller has an extremely stable, reproducible digital ramp for heating rates 0-25 °C, automatic repeat ramp for the background glow curve, a low-power, low-volume glow oven for fast cooling and evacuation, and a pile-up compensating ratemeter that extends the single photon counting dynamic range to beyond ten million counts per second. The system is packaged in two parts: an electronic enclosure and the glow oven assembly with PMT—housing (the PM tube is an EMI 9635QA selected for very low dark count). The glow curves are recorded by a Canberra multichannel analyser and then evaluated numerically. As the potteries of prehistoric origin in the Carpathian basin are extremely poor in quartz grains, a special sample preparation technique was developed to get a higher yield in grains. The occurrence of the grains in the appropriate size interval was found to be as low as a part of 10"

4 of the whole material or less. In our procedure the sherds were crushed to fragments of appr. 0.5 cm, divided into small portions and subjected to repeated attacks in HCl and HF by making use of an ultrasonic bath. The clay matrix was successively eliminated by rinsing in distilled water. The complete decomposition of the fragments was followed by continuous etching of the residual grains in HF for an hour.