Tuesday, May 15, 2007

Uranium Thorium dating of stars

Here at the lounge, we tend to follow the lemming pack in believing that when it comes to U decay chain dating, smaller is better. Smaller blanks, smaller spots, miniaturization is the name of the game. So imagine my surprise when I read that somebody had done astrogeochronology on an entire star!

The paper, and some related lecture notes are available on the web, so I had a look, and this is my best attempt at an explanation:

Anna Frebel and colleagues are interested in primitive stars- stars which have few metals (metal=anything heavier than He) in them. This is also useful because there is an Fe interference on the best U peak. There are also C-N molecular interferences. However, in their star survey they found a star with low Fe, low C, and R-process enrichment. (The R process is the nucleosynthetic process that produces U).

What they do is to calculate the expected initial U and Th production based on one of the other stable R process elements they observed (Os, Ir, Eu). They then assume the deficit between measured and expected Th and U is due to radioactive decay at the standard rate.

They get a mean age of about 13.2 Ga, +/- around 2.5 Ga, depending on how the errors are crunched. For the Th, the main source of error was the model of the initial concentration; for the U the main error was in the measurements.

Obviously, there is still work to be done. It would be easy for sub-permil U/Pb hotshots to rubbish this work, but considering what the first U/Pb dates from the 1920's were like, I reckon this is a pretty good number. And it should only get better.
The different results from different reference elements means that either they need to model R vs S abundances of the reference elements, or that the theoretical R values are wrong for the supernova(e) that generated the heavy elements in this star. But as an analysis, this looks like a good enough result to put the modelers and theoreticians on the back foot.

Hat tip: Bad Astronomer.

No comments: