A test for systematic errors in Ar-40/Ar-39 geochronology through comparison with U/Pb analysis of a 1.1-Ga rhyolite

Important sources of systematic error in Ar-40/Ar-39 dating arise from unce rtainties in the K-40 decay constants and K/Ar isotopic data for neutron fl uence monitors (standards). The activity data underlying the decay constant s used in geochronology since 1977 are more dispersed than acknowledged by previous geochronologically oriented summaries, and compilations of essenti ally the same data in nuclear physics and chemistry literature since 1973 h ave consistently produced lower estimates (and larger assigned uncertaintie s) of the constants for K-40 ---> Ar-40 and K-40 --> Ca-40 decay. Consideri ng also uncertainties in K-40/K, and the questionable existence of a gamma- less electron capture K-40 --> Ar-40 decay direct to ground state, the tota l K-40 decay constant is known to no better than +/-2% at the 2 sigma level . Ar-40*/K-40 ratios for individual standards are known to better than +/-2 % in some cases, but interlaboratory discrepancies of more than 2% in the A r-40/Ar-39 ages of secondary standards like the Fish Canyon sanidine (FCs) suggest larger uncertainties. The very precisely determined decay constants for U-238 and U-235, and the existence of quantitative internal U/Pb, reliability criteria, offer an alt ernative basis for evaluation of both the 40K decay constants and the ages of Ar-40/Ar-39 standards. High precision U/Pb (zircon) and Ar-40/Ar-39 (alk ali feldspar). data from the 1.1-Ga Palisade Rhyolite provide a highly sens itive basis for comparison. Ten U/Pb analyses on abraded single zircons as well as one analysis of six fragmented and HF leached crystals yield a Pb-2 07/Pb-206 age of 1097.6 +5.2/-5.4 Ma (95% confidence, including decay const ant errors). Ar-40/Ar-39 incremental CO2 laser heating of single alkali fel dspar grains yields nine undisturbed age spectra with error-weighted platea u ages (based on conventional decay constants and an age of 28.02 Ma for FC s) from 1086.5 +/- 4.8 Ma (R-FCs(F239) = 52.7740 +/- 0.3062) to 1090.4 +/- 4.3 Ma (R-FCs(F238) = 53.0281 +/- 0.2746) (2 sigma, including analytical er rors only), with MSWD = 0.95. The weighted mean R-FCs(F239) value of these plateau steps (52.9011 +/- 0.2324), including irradiation-related errors, i s inferred to reflect the eruption-age R-FCs(F239) value of the Palisade Rh yolite alkali feldspar. Reconciliation of the Ar-40/Ar-39 and U/Pb results suggests that either the age of the Ar-40/Ar-39 Standard is older, or the K-40 total decay constant is smaller, than values in current use by geochronologists. Comparison wit h constraints from an historic eruption indicates a total K-40 decay consta nt of 5.37 x 10(-10)/yr and an age of 28.05 Ma for FCs. Further application s of this approach will provide more robust solutions and allow estimation of uncertainties. Copyright (C) 1999 Elsevier Science Ltd.