Attempts to understand potential deficiencies in chemical procedures for AMS

A major research focus of the DREsden Accelerator Mass Spectrometry (DREAMS) facility is sample preparation: in-house and at cooperating laboratories. Besides routine applications, developments are mainly driven by the users' demands such as new radionuclides (e.g. short-lived Be-7), radionuclides from new matrices or in larger sample amounts and at lower concentrations. Within the chemical preparation of BeO for Be-7,Be-10, Al2O3 for Al-26, AgCl for Cl-36, and Fe2O3 for Fe-60 and actinide AMS targets, we investigated and improved mostly particular steps: For oxide preparation aging for several hours and two-times rinsing of hydroxides is recommended, especially for Be(OH)(2) and Al(OH)(3), to prevent losses in the final preparation steps. Rinsing of freshly precipitated hydroxides can yield to losses as high as 31%. For Al(OH)(3) rinsing with H2O (pH5) instead of pH8-9 is recommended for further reducing redissolving. For Fe-60 and coprecipitated actinides (the latter tested by U6+ and Er3+), rinsing of overnight-aged Fe(OH)(3), yield to 2.6-3.5% losses. The depletion of the isobar Li-7 for (BeO)-Be-(7) is easily gained by hydroxide precipitation and rinsing. The quality of BeO and Al2O3 AMS targets can be monitored and improved (e.g. by using larger ion exchanger columns) by using the so-called quality factor, which is the current of the stable nuclide (Be-6 or Al-27), normalised to the current of the standard of the same batch. Finally, there is a high potential for ion exchange as a pre-enrichment tool for Cl for large ice samples containing low Cl-nat concentrations and low absolute amounts of Cl-nat.