Improving the accuracy of aluminium assay in purified quartz for in situ cosmogenic exposure dating

Precise and accurate measurement of the natural aluminium content in purified quartz powder, [Al], extracted from surface bedrock and/or sediment samples for in situ cosmogenic nuclide dating is critical for reliable 26Al exposure ages and 26Al/10Be ratios for burial dating. Numerous articles have cited issues regarding reliability of 26Al ages paired to 10Be ages, not only due to poor statistical error, but also due to uncertainty in [Al]. Loss of Al during aliquot preparation for ICPOES from bulk HF solutions of dissolved quartz or inaccurate assay via ICP (or AA) will result in reduced 26Al concentrations. In turn this will, depending on magnitude of the loss with respect to typical age errors, lead to lower-than-expected 26Al ages when compared to corresponding 10Be ages. If the loss is excessive, this may result in depressed 26Al/10Be ratios and thus a false burial signal. Over the past 4 years we have performed repetitive Al assays from a parent HF solution prepared from ~25 g of a low-iron glass sand (NIST-165a; ~320 ppm [Al]) using identical methods as for all quartz samples. The solution was used as an in-house laboratory standard for [Al] and shows a limiting 1σ of ±5% (n ~50) via ICP-OES at a commercial laboratory. The spread in paired NIST-165a duplicates measured per batch is better than ±1%. To determine the accuracy of our procedures, we prepared sets of solutions (1–10 ppm Al) from 3 different quartz powders using the standard addition method. Initial results indicate a 3-7% offset between [Al] based on the standard addition method and that based on the conventional calibration curve. Copyright (c) 2011 AMS12