BIOTITE: A POOR CHOICE FOR HIGH PRECISION 40AR/39AR GEOCHRONOLOGY

A growing set of coexisting biotite and sanidine 40Ar/39Ar dates reveal that biotites are generally older than sanidine. This observation makes biotite a poor candidate for high precision geochronology studies of volcanic rocks and coupled with ubiquitous loss of 39Ar during irradiation suggests general caution for all biotite dates. In cases such as the ca. 28 Ma Fish Canyon Tuff (FCT) 40Ar/39Ar standard there are clearly inherited biotite grains that come from lithic fragments as single crystal analysis yield apparent ages as old as 33 Ma. In many other cases it is not obvious what causes age discordance between the mineral phases, but we consider that the sanidine ages are more reliable. We have undertaken a biotite/sanidine dating study of a mid-Tertiary volcanic section that has well-controlled stratigraphy. Two samples of each of four ignimbrites confirm previous observations that biotite consistently yields an older date compared to sanidine, however the discordance is not always significant at 2 sigma. For these 28 to 32 Ma ignimbrites, biotite dates range between 0.03 and 1.72 Ma (typically ~0.2 Ma) older than sanidine pairs. For the two samples from each unit, the biotite data are less reproducible compared to sanidine data and in one instance two biotite dates from the same ignimbrite differ by 1.5 Ma. In vacuo encapsulation experiments on some of the biotites confirm that a significant amount of 39Ar is lost via recoil during irradiation. Typically about 0.8% 39Ar is lost during irradiation with more altered samples recoiling out up to 5% of their total 39Ar. The amount of 39Ar loss is far more than predicted by simple surface ejection and the loss correlates to degree of biotite alteration. These experiments document that significant bias towards old dates will be encountered due to irradiation loss of 39Ar. U/Pb zircon data from the oldest tuff (Hells Mesa) are slightly complex, but indicate an age of 32.5±0.1 Ma that is about 0.3 Ma older than both biotite and sanidine (assuming FCT sanidine=28.02 Ma and 40K lambda total=5.543e-10/a). We are presently determining U/Pb zircon dates for the ignimbrites to constrain their maximum eruption ages, evaluate magma residence time and to provide additional data points for the intercalibration of the U/Pb and 40Ar/39Ar methods.