Resolvable miscalibration of the 40Ar/39Ar geochronometer

U/Pb and 40Ar/39Ar isotopic dating techniques are the most widely applied geochronometers, both capable of 0.1% internal precision. A robust intercalibration between the two isotopic systems is fundamental for reconstructing short term processes and events in geologic time. However, whereas the U decay constants are known precisely (to ca 0.1%), the currently used 40K decay constant (5.543×10^{-10}/yr, (1)) is associated with an unstated uncertainty that is about an order of magnitude larger than the former, making high-resolution comparisons of ages from the two isotopic systems impossible. We present an indirect calibration by comparing radio-isotopic ages derived from both isotopic systems of rapidly cooled volcanic rocks in order to minimize effects from protracted cooling history. Eleven data pairs of 206Pb/238U and conventional 40Ar/39Ar ages exhibit a bias between the two isotopic systems ranging from >-1.5% for young rocks to ca -0.5% for rocks as old as 2 Ga (possibly even smaller for rocks >2 Ga), with the 40Ar/39Ar ages being consistently younger. All Mesozoic and Paleozoic samples display a bias of about -1%. Most of this bias is probably the result of miscalibration of the electron capture decay constant of 404→ 40Ar (λ40Kec) by ca -1%, in combination with a miscalibration of smaller magnitude and opposite sense of the β- decay constant (λ40Kβ-) of 40K→ 40Ca. Bias greater than 1% for younger Cenozoic samples probably reflects pre-eruptive zircon saturation (magma residence time) whose effects become proportionately negligible beyond ca. 200 Ma. Whereas the currently used decay constant for 40K (see above) is based on an arguably arbitrary selection from counting experiments associated with large and sometimes incomprehensible uncertainties (mostly from experiments conducted in the 1940s to 1960s) two recent recalibrations of λ40Ktotal using liquid scintillation counting techniques suggest precise and mutually consistent values of 5.553 ± 0.013×10^{-10}/yr (2) and 5.554 ± 0.018×10^{- 10}/yr (3), respectively. If these values are assumed correct, λ40Kβ- would require a change of +0.3% if combined with the more robust -1% change for λ40Kec, which is in agreement in sense with our observation of a decreasing bias for ages >1 Ga. Applying the statistical methods described in Kwon et al. (2002) to our larger data base in order to simultaneously evaluate λ40Ktotal and the age of the commonly used flux monitor Fish Canyon sanidine (FCs) yields 5.530 ± 0.038×10^{-10}/yr for the former, which agrees with the constants given by above but is associated with a relatively large uncertainty (0.7%), and 28.28 ± 0.06 Ma (0.2%) for the latter. The recalculated value for FCs is in agreement in sense and magnitude with that of an astronomically calibrated age of 28.21 ± 0.06 Ma (4). 1. R. H. Steiger, E. Jaeger, Earth and Planetary Science Letters 36, 359 (1977). 2. K. Kossert, E. Guenther, Applied Radiation and Isotopes 60, 459 (2004). 3. A. Grau Malonda, A. Grau Carles, Applied Radiation and Isotopes 56, 153 (2002). 4. K. F. Kuiper, J. R. Wijbrans, F. J. Hilgen, Terra Nova 17, 385 (Aug, 2005).