Radiometric dating of rhyolites by conventional K/Ar method: methodology

Rhyolite magmatism represents a substantial part of calc-alkaline volcanic rocks in continental margin and/or back-arc setting and frequently associates with epithermal mineralization. Their precise and reliable dating is important for understanding the magmatic evolution, as well as the related metallogenetic processes. Nowadays, the higher precision is reached especially by applying ⁴⁰Ar/³⁹Ar and/or single grain U/Pb dating. However, using a proper methodology, even with the conventional K/Ar age determination, we can reach highly reliable results with precision of 3 % relative for single datings (at 1σ resp. 68 % confidence level). Improved methodology has been applied to rhyolites of the Jastrabá Fm. and related epithermal mineralization in the Kremnické vrchy mountain range in central Slovakia. Obviously, dated samples were carefully selected, knowing their geological setting and processes affecting the rocks. Generally, dating of whole-rock samples can not provide reliable results. Selection of target phases based on careful petrographic investigation, including a use of BSE images and electron microprobe analyses of K-bearing phases, represents an important step. Phases affected by epigenetic alteration should not be dated. Target phases in rhyolites of the Jastrabá Fm. in order of decreasing K-content are: hydrothermal K-feldspar (adularia, 12,6–13,7% K), magmatic K-feldspar (sanidine, 8,4–10,7% K), subsolidus K-feldspar (7,9–9,6% K), biotite (7,1–8,1% K), spherulitic groundmass (3,9–7,0% K), kfs-groundmass (3,6–6,1% K), glass (3,3–4,6% K), amphibole (0,53–0,72% K), plagioclase (0,3–1,2% K). Separation of selected target phases is carried out using heavy liquid, electromagnetic separator, shaking and handpicking (for final cleaning) on grain-size fractions 0.63–0.4 mm, 0.4–0.25 mm and 0.25–0.125 mm, 300–500 g each. As a rule, all datable phases are collected. Potassium is determined by flame photometry with a Na buffer and Li internal standard with relative analytical error 2%. Argon is extracted from the samples by RF fusion in Mo crucibles, in previously baked stainless steel vacuum system. Pure ³⁸Ar spike is added from gas pipette system and the evolved gases are cleaned using Ti and SAES getters and liquid nitrogen traps, respectively. The purified Ar is transported directly into the mass spectrometer and Ar isotope ratio is measured in the static mode, using a 15 cm radius magnetic sector type mass spectrometer built in Debrecen. The relative analytical error of ³⁸Ar spike is 2%, the relative analytical errors of ⁴⁰Ar/³⁸Ar and ³⁶Ar/³⁸Ar isotope ratios determination are 1%. Age of the sample is calculated using the decay constants suggested by Steiger and Jäger (1977) and isotopic composition of natural potassium ³⁹K - 93.2581%, ⁴⁰K - 0.01167%, ⁴¹K - 6.7302%, assuming that the rock or mineral has been a closed system for K and ⁴⁰Ar_rad concentrations. Analytical error is given at 68% confidence level (1σ). Usually it varies around 3 % relative. Despite the perfect sample preparation and analytical work some results are dubious owing to natural reasons and should be eliminated from further consideration. There are several ways to check reliability of results: (1) Kconcentration should correspond to the dated phase; (2) percentage of ⁴⁰Ar_rad should be high enough; (3) consistent results on different fractions/phases; (4) consistent results on samples of the same unit (statistical testing); (5) testing by the isochrone method – identification of phases with excess ⁴⁰Ar_rad (xenocrysts and/or plagiclase and amphibole phenocrysts) and ⁴⁰Ar_rad loss (glass). One has to be always aware of statistical aspects, especially confidence interval of a single datum. Appropriate statistical methods should be used in evaluation of multiple data. With multiple data for geological units differences in age smaller than error of a single datum can be recognized – one of the possibilities is a graph of normal distribution densities. If results do not fit with geological relationships something must be wrong – either radiometric dating or more likely our geological assumptions. Our results on rhyolites and related mineralizations of Kremnické vrchy are in other presentation of Lexa and Pécskay.
   Acknowledgements: Research was supported by the Hungary-Slovakia S&T SK 27/06 grant (APVV grant SK-MAD-01106), VEGA grant 2/0171/08, grant 1506 of the Slovak Ministry of Environment and OTKA grant (Hungarian National Scientific Foundation) no. K68153.