British Museum Natural Radiocarbon Measurements VI

Dates listed below are based on measurements made up to May 1968, and cover a period during which the technique of gas proportional counting using CO2 was gradually replaced by liquid scintillation counting using benzene. The gas counting measurements were carried out by the method and techniques previously described (Barker and Mackey, 1968) the only modifications being the replacement of some old electronic units by more stable solid-state equipment; proportional counting results are indicated in the text by (P) at the end of the relevant sample descriptions. Liquid scintillation counting, which is now the preferred method in this laboratory, is carried out using a Packard Tri-Carb liquid scintillation spectrometer model 3315/AES fitted with selected low-noise quartz-faced photomultipliers. Normally 3 ml of benzene is prepared from each sample. This is dissolved in 12 ml of scintillation grade toluene containing 5 gm/liter of scintillator (PPO) and the solution is measured in a standard low-potassium glass vial at a temperature of 0°C. Photomultiplier E.H.T., amplifier, and channel width settings are optimized for C14, and measurements are carried out at ca. 65% efficiency of detection for C14 to eliminate interference from any tritium which may be present in the benzene. Under these circumstances the background is approx. 8.6 cpm and the modern (95% Aox) is approx. 24.0 cpm. Samples are counted in groups of 3 to 5 together with background and modern reference samples and are measured for at least one week, the instrument being set to cycle at 100 min intervals. In this period, the counts accumulated are such that the background is always measured to a statistical accuracy of better than 1% and most other samples to a higher accuracy than this. Background and modern counts used in the calculation of each result are only those relevant to the period of measurement of that particular sample. Statistical analysis of groups of replicate measurements made under these conditions over a very long period of time has demonstrated the excellent long-term stability of the equipment and indicates that the technique is quite capable of achieving results of very high statistical accuracy when required.