Ar-40/Ar-39 dating of Ordovician K-bentonites in Laurentia and Baltoscandia

Several Ordovician K-bentonites occurring widely in eastern North America a nd western Europe were dated using the Ar-40/Ar-39 technique to test previo usly suggested inter-continental correlations. The three thickest and most widespread bentonites - Deicke, Millbrig (North America) and Kinnekulle (Sw eden and Denmark) - were examined. Single-grain analyses of phenocrystic bi otites yield 100% concordant plateau ages of 449.8 +/- 2.3 Ma (2 sigma inte rnal error) (Deicke), 448.0 +/- 2.0 Ma (Millbrig) and 454.8 +/- 2.0 Ma (Kin nekulle in Sweden). The altered biotites from the Denmark Kinnekulle sample show discordant age spectra with highly variable plateau ages being most l ikely caused by recoil redistribution of Ar isotopes between biotite and in terlayer secondary alteration products during neutron irradiation. Such rec oil artifacts are likely evident in previously published Ar-40/Ar-39 data f or these units, causing bias towards spuriously old ages. In such cases whe re intragrain alteration is evident and discordant age spectra are common, geologically meaningless plateau ages may be obtained which are less accura te than integrated ages, hence a 100% concordance criterion for acceptance is useful. The inconsistent Ar-40/Ar-39 ages and geochemical contrasts betw een the North American and Swedish bentonites imply distinct volcanic episo des, and preclude a specific correlation between the Millbrig and Kinnekull e K-bentonites. The precise Ar-40/Ar-39 ages indicate a minimum duration of 7 +/- 3 Myr for the Diplograptus multidens graptolite biozone. In comparis on with previous U/Pb data, the results are useful for quantifying the bias (ca. 1%) between current calibrations of the Ar-40/Ar-39 and U/Pb geochron ometric systems. This bias is most likely due to errors related to K-40 dec ay constants and neutron fluence monitors used in Ar-40/Ar-39 dating. (C) 2 001 Elsevier Science B.V. All rights reserved.