INTERCALIBRATION OF STANDARDS, ABSOLUTE AGES AND UNCERTAINTIES IN 40AR39AR DATING

Abstract

The 40Ar39Ar dating method depends on accurate intercalibration between samples, neutron fluence monitors, and primary 40Ar40K (or other external) standards. The 40Ar39Ar age equation may be expressed in terms of intercalibration factors that are simple functions of the relative ages of standards, or equivalently are equal to the ratio of radiogenic to nucleogenic K-derived argon (40Ar*39ArK) values for one standard or unknown relative to another. Intercalibration factors for McClure Mountain hornblende (MMhb-1), GHC-305 biotite, GA-1550 biotite, Taylor Creek sanidine (TCs) and Alder Creek sanidine (ACs), relative to Fish Canyon sanidine (FCs), were derived from 797 analyses involving 11 separate irradiations with well-constrained neutron fluence variations. Values of the intercalibration factors are RMMhb-1FCs = 21.4876 +/- 0.0079; RGA-1550FCs = 3.5957 +/- 0.0038; RTCsFCs = 1.0112 +/- 0.0010; RACsFCs = 0.04229 +/- 0.00006, based on the mean and standard error of the mean resulting from four or more spatially distinct co-irradiations of FCs with the other standards. Analysis of 35 grains of GHC-305 irradiated in a single irradiation yields RGHC-305FCs = 3.8367 +/- 0.0143. Results at these levels of precision essentially eliminate intercalibration as a significant source of error in 40Ar39Ar dating. Data for GA-1550 (76 analyses, 5 fluence values), TCs (54 analyses, 4 fluence values), FCs (380 analyses, 40 fluence values) and ACs (86 analyses, 11 fluence values) yield MSWD values showing that the between-grain dispersion of 40Ar*39ArK values is consistent with analytical errors alone, whereas MMhb-1 (167 analyses, 4 irradiations) and GHC-305 (34 analyses, 1 fluence value) are heterogeneous and therefore unsuitable as standards for small sample analysis. New K measurements by isotope dilution for two primary standards, GA-1550 biotite (8 analyses averaging 7.626 +/- 0.016 wt%) and intralaboratory standard GHC-305 (10 analyses averaging 7.570 +/- 0.011 wt%), yield values slightly lower and more consistent than previous data obtained by flame photometry, with resulting 40Ar40K ages of 98.79 +/- 0.96 Ma and 105.6 +/- 0.3 Ma for GA-1550 and GHC-305, respectively. Combining these data with the intercalibration approach described herein and using GA-1550 as the primary standard (1.343 x 10-9 mol/g of 40Ar*; [McDougall, I., Roksandic, Z., 1974. Total fusion 40Ar39Ar ages using HIFAR reactor. J. Geol. Soc. Aust. 21, 81-89.]) yields ages of 523.1 +/- 4.6 Ma for MMhb-1, 105.2 +/- 1.1 Ma for GHC-305, 98.79 +/- 0.96 Ma for GA-1550, 28.34 +/- 0.28 Ma for TCs, 28.02 +/- 0.28 for FCs, and 1.194 +/- 0.012 Ma for ACs (errors are full external errors, including uncertainty in decay constants). Neglecting error in the decay constants, these ages and uncertainties are: 523.1 +/- 2.6 Ma for MMhb-1, 105.2 +/- 0.7 Ma for GHC-305, 98.79 +/- 0.54 for GA-1550, 28.34 +/- 0.16 Ma for TCs, 28.02 +/- 0.16 Ma for FCs, and 1.194 +/- 0.007 Ma for ACs. Using GHC-305 as the primary standard (1.428 +/- 0.004 x 10-9 mol/g of 40Ar*), ages are 525.1 +/- 2.3 Ma for MMhb-1, 105.6 +/- 0.3 Ma for GHC-305, 99.17 +/- 0.48 Ma for GA-1550, 28.46 +/- 0.15 Ma for TCs, 28.15 +/- 0.14 Ma for FCs, and 1.199 +/- 0.007 Ma for ACs, neglecting decay constant uncertainties. The approach described herein facilitates error propagation that allows for straightforward inclusion of uncertainties in the ages of primary standards and decay constants, without which comparison of 40Ar39Ar dates with data from independent geochronometers is invalid. Re-examination of 40K decay constants would be fruitful for improved accuracy.