Assessment of strategies for measuring chloride-36 in natural samples

The cosmogenic isotope 36Cl (half-life of 3.01 x 105 years) has been applied to problems in hydrogeology and surface exposure dating for a number of years. However its measurement in natural samples is complicated by its low abundance with respect to the stable isotopes of chlorine (36Cl/Cl = 10-10 to 10-16), and the presence of the isobar 36S (0.02% of S). Conventional Accelerator Mass Spectrometry (AMS) techniques for measuring 36Cl require tandem accelerators of 5 MV or larger. Only these machines can accelerate the 36Sn+ ions to high enough energies for them to be separated in gas-filled magnets, or in dE/dx analysis in the final detector. The present work investigates the possibility of 36Cl analysis on smaller, more economical machines. It is demonstrated that 36Ar3+ and 36S3+ ions, generated from argon and SF6 gas within the 2 MV tandem accelerator at the University of Toronto, will not preclude 36Cl measurements. A gas ion source was developed in which Cl- ions are created when a negative ion beam transfers its charge to the gas ICl, and the Cl- ions are then extracted. The efficiency of Cl- generation was about 20% for a Cu- beam. Because of the instability of sulphur iodide (S2I2) at room temperature, and the resonant charge transfer reaction to ICl, this gas source should discriminate against 36S. It was also shown that in the collision of S - and Cl- ions with NO2 gas at about one milliTorr pressure, the S- was attenuated by about two orders of magnitude relative to the Cl-. Finally, the 25 MV Holifield accelerator at Oak Ridge, Tennessee was used to measure 36Cl with the completely stripped 36Cl 17+ ion. Because the completely stripped sulphur species has a charge of 16+, it was possible to separate this isobar from 36Cl in a magnet. These measurements detected 36Cl in seawater for the first time, and showed, unexpectedly, that the 36Cl/Cl ratio varies in the world ocean.