Archaeometry 26, zyxwvutsrqp 1 zyxwvutsrqp (1984), 15-20, Printed in Great Britain zyxwvu RESEARCH NOTES AND APPLICATION REPORTS RADIOCARBON MEASUREMENT BY ACCELERATOR MASS SPECTROMETRY: AN EARLY SELECTION OF DATES R. GILLESPIE, J. A. J. GOWLETT, E. T. HALL and R. E. M. HEDGES Oxford Radiocarbon Accelerator zyxwvut Unit, Research Laboratory for Archaeology and the History of Art, Oxford University, 6 Keble Road, Oxford OX1 3QJ, England INTRODUCTION Early work and progress with the Oxford accelerator mass spectrometer (AMS) has been des- cribed elsewhere (Hedges 1981, Gillespie, Hedges and White 1983). The purpose of this note is to demonstrate the variety of applications made in the first few months of its operation as a reliable dating system. The system operating at present has changed little from the rather general technical des- cription given in the references mentioned above. Briefly, samples containing zyx - 5 mg of carbon are first pre-treated chemically. In the particular case of bone, amino acids are isolated and purified from the insoluble protein matrix (Gillespie, Hedges and Wand, n.d.). The purified sample is oxidized to C 0 2 , and converted to C2Hz. A graphite target is prepared by pyrolysing the acetylene with a tantalum wire heated to - 2000 "C. This produces a highly compact deposit about 4mm long and 0.75 mm diameter. Such a target is capable of consistently pro- ducing large stable beams (usually 10-15pAC-) in the ion source (Hedges, Wand and White 1980, Wand, Gillespie and Hedges, n.d.). It is our usual practice to prepare 2 targets from each initial organic sample. Standards are prepared similarly, by cracking C2H2 made from NBS (new) oxalic acid. A number of known-age targets are also prepared, which can be used to test various parts of the overall process. These include the direct cracking of benzene types A-F supplied by the C-14 laboratory at Harwell, combustion of dendrochronologically dated wood, or the total chemical processing of previously dated large samples of bone. In general, 20 targets are loaded into the ion source, typically comprising 12 unknowns, 4 oxalic acid standards, and a selection of known-age and background samples. Measurements are made of the I4C count, and the beam currents of "C-, 12C3+, 13C3+. I4C is identified in a detector after the ions have been accelerated to an energy of 8 MeV, by means of a tandem electrostatic accelerator. Non-14C ions are removed at 3 stages: a 90" magnet injecting beam into the accelerator: a 90" analyser magnet which selects 14C3+ following the stripping of C- to Cn' within the accelerator terminal, and a Wien (Ex B) filter. The particle flux from the oxalic standard entering the detector is 99-99.6% 14C. I2C3+ and I3C3+ are both measured in the focal plane of the analyser magnet. The position of the I3C3+ beam is sensed, and this information is incorporated in a slow feed back loop to the terminal voltage, keeping the latter stable to zyxwvu 2 100 V. 12C3+ and 13C3+are measured only when the inflector magnet is set to inject beams of mass 12 or 13. This is achieved by changing the energy of the extracted -ve ion beam from the ion source over the distance between the entry and exit focal planes of the magnet. A pulsed voltage in 2 steps is applied to an insulated section of the beam line and magnet box. Typically, "C- is injected for 1.5 msec and 13C- for 6 msec. The 15