UC Santa Cruz

Europium is a useful proxy in identifying r-process elements such as Thorium and Uranium that act as long-term sources of generated heat in the mantle of terrestrial planetsdue to their radioactivity and long half-lives. Comparing these abundances to a proxy for the bulk-silicate mantle, such as Magnesium, allows us to describe the geothermal evolution of terrestrial planets. We use stellar data from two independent teams, the first being Delgado-Mena et al. which has 566, out of a total of 1059 stars, having measured Europium values, and the second team being Battistini and Bensby, which has a total of 378 stars, each with measured Europium values. Between these two datasets, there exist only 68 shared stars. With these datasets, we perform an analysis that aims to identify significant star-to-star variation in Europium. Our analysis involves the decorrelation of [Eu/H] values via a metallicity parameter in the form of either [Fe/H] or [Mg/H], as well as Teff while comparing the [Eu/H] vs. Age trend before the decorrelation to the same trend after the decorrelation. For the Delgado-Mena dataset, we perform our analysis twice; first, the individual analysis of the thin disk stars and second, the analysis of the total population according to the galactic populations: thin disk, thick disk, high, and hαmr. The important takeaways from this study are noted as such: the peak-to-peak [Eu/H] values are subject to inflation due to the correlation seen with both metallicity and Teff parameters. Metallicity parameters impact this inflation greater than the Teff parameter. Teff trends from the separate datasets are of similar magnitude but of different signs, which is indicative of a different problem outside the scope of the purpose of this study. Our analysis of [Eu/H] values after compensating for trends in both Teff and a metallicity term results in a tight dispersion (standard deviation) with these values for the Delgado-Mena data ranging from 0.064 to 0.087 dex. Each of these values is smaller than the quoted average error of 0.1 dex for the raw [Eu/H] values. As for the Battistini & Bensby data, we see the values 0.088 and 0.092 dex, both of which are more consistent with the quoted error of 0.08 dex. Comparing these results to the quadratic sum of the average errors from both datasets with a value of 0.13 dex, tells us that Delgado-Mena is overestimating their errors, whereas, for Battistini & Bensby, the results are inconclusive in determining whether they are over or underestimating. We find evidence for true outliers in the data, with as many as ten stars that exceed 3σ difference for the [Fe/H] & Teff detrended Delgado-Mena data with an average residual [Eu/H] value of 0.17 dex, where we would expect about two outliers for the size of our dataset. Similarly, we find five stars that exceed the same 3σ difference for the [Mg/H] & Teff detrended B&B data with an average residual [Eu/H] value of 0.29 dex. For both datasets, the percentage of stars that exceed the 3σ difference is about 1.5% of the [Eu/H] positive sample size. Averaging the standard deviations from the total population models, we find a [Eu/H] variation value of 0.085 dex which gives us a peak-to-peak range of ±0.26. What we find is little evidence of significant intrinsic star-to-star variation of [Eu/H]; with peak-to-peak ranges significantly decreased by our analysis. The motivating literature suggests a range of ±0.5 dex, we find a range of ±0.26 dex, and with relatively few extreme outliers, we are led to believe that the true variation of [Eu/H] is less than the literature suggests.