Interlaboratory study of the ion source memory effect in 36 Cl accelerator mass spectrometry Stefan Pavetich a,⇑ , Shavkat Akhmadaliev a , Maurice Arnold b , Georges Aumaître b , Didier Bourlès b , Josef Buchriegler a,c , Robin Golser c , Karim Keddadouche b , Martin Martschini c , Silke Merchel a , Georg Rugel a , Peter Steier c a Helmholtz-Zentrum Dresden-Rossendorf, Bautzner Landstraße 400, 01314 Dresden, Germany b Aix-Marseille Université, CEREGE CNRS-IRD, F-13545 Aix-en-Provence, France c University of Vienna, Faculty of Physics, VERA Laboratory, Währingerstraße 17, 1090 Vienna, Austria article info Article history: Received 18 December 2013 Received in revised form 24 February 2014 Accepted 25 February 2014 Keywords: Accelerator mass spectrometry (AMS) Volatile elements Chlorine Negative ion sputter sources Long-term memory effect Interlaboratory comparison abstract Understanding and minimization of contaminations in the ion source due to cross-contamination and long-term memory effect is one of the key issues for accurate accelerator mass spectrometry (AMS) measurements of volatile elements. The focus of this work is on the investigation of the long-term memory effect for the volatile element chlorine, and the minimization of this effect in the ion source of the Dresden accelerator mass spectrometry facility (DREAMS). For this purpose, one of the two original HVE ion sources at the DREAMS facility was modified, allowing the use of larger sample holders having individual target apertures. Additionally, a more open geometry was used to improve the vacuum level. To evaluate this improvement in comparison to other up-to-date ion sources, an interlaboratory compar- ison had been initiated. The long-term memory effect of the four Cs sputter ion sources at DREAMS (two sources: original and modified), ASTER (Accélérateur pour les Sciences de la Terre, Environnement, Risques) and VERA (Vienna Environmental Research Accelerator) had been investigated by measuring samples of natural 35 Cl/ 37 Cl-ratio and samples highly-enriched in 35 Cl ( 35 Cl/ 37 Cl  999). Besides investi- gating and comparing the individual levels of long-term memory, recovery time constants could be calculated. The tests show that all four sources suffer from long-term memory, but the modified DREAMS ion source showed the lowest level of contamination. The recovery times of the four ion sources were widely spread between 61 and 1390 s, where the modified DREAMS ion source with values between 156 and 262 s showed the fastest recovery in 80% of the measurements. Ó 2014 Published by Elsevier B.V. 1. Introduction Accelerator mass spectrometry (AMS) is an ultrasensitive method for the measurement of isotopic ratios. Usually, ratios of radioactive nuclei to their stable isotopes are measured [1–3]. Due to the direct detection and counting of the radioisotopes, AMS does not depend primarily on radioactive decay parameters such as branching ratios and half-lives. The suppression of interfer- ing molecular isobars enables, depending on the measured isotope, the measurement of isotopic ratios down to 10 16 –10 12 which cannot be achieved by any other analytical technique. Therefore, AMS covers a wide spectrum of applications which range from dat- ing organic material to geomorphology and astrophysical studies [1–3]. Chlorine-36 (half-life = (3.013 ± 0.015)  10 5 a [4]) is a cosmo- genically and anthropogenically produced radionuclide. It is commonly used for groundwater dating (e.g. [5,6]), surface rock exposure dating (e.g. [7]), bomb peak studies (e.g. [8]) and extra- terrestrial material and cosmic ray variations studies (e.g. [9]). Due to the necessity of reaching the charge state 5+ to insure an unambiguous characterization and detection of 36 Cl, AMS measurements of 36 Cl were usually performed with large accelera- tors capable of producing terminal voltages of at least 5 MV that in addition provide the necessary energy for separating 36 Cl from the naturally superabundant 36 S. Recently, Martschini et al. [10,11] showed that measurements of 36 Cl with sufficient sulfur suppres- sion are also possible at lower energies. Besides the suppression of sulfur, a well working ion source is crucial to gain high-quality data in 36 Cl-AMS. AMS ion sources are usually based on the Middleton concept of sputtering negative ions with Cs [12,13]. Negative ions are emitted from the sample http://dx.doi.org/10.1016/j.nimb.2014.02.130 0168-583X/Ó 2014 Published by Elsevier B.V. ⇑ Corresponding author. Tel.: +49 351 260 3685. E-mail address: s.pavetich@hzdr.de (S. Pavetich). Nuclear Instruments and Methods in Physics Research B 329 (2014) 22–29 Contents lists available at ScienceDirect Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb