Inflow, Outflow, Yields, and Stellar Population Mixing in Chemical Evolution Models

Brett H. Andrews, David H. Weinberg, Ralph Schönrich, Jennifer A. Johnson

Published 2016-04-28Version 1

Chemical evolution models are powerful tools for interpreting stellar abundance surveys and understanding galaxy evolution. However, their predictions depend heavily on the treatment of inflow, outflow, star formation efficiency (SFE), the IMF, the SNIa delay time distribution, stellar yields, and mixing of stellar populations. Using flexCE, a new, flexible one-zone chemical evolution code, we investigate the effects of individual parameters and the trade-offs between them. Two of the most important parameters are the SFE and outflow mass-loading parameter, which shift the knee in [O/Fe]-[Fe/H] and the equilibrium abundances, respectively. One-zone models with simple star formation histories follow narrow tracks in [O/Fe]-[Fe/H] that do not match the observed bimodality in this plane. A mix of one-zone models with variations in their inflow timescales and outflow mass-loading parameters, as motivated by the inside-out galaxy formation scenario with radial mixing, reproduces the high- and low-alpha sequences better than a single model with two infall epochs. We present [X/Fe]-[Fe/H] tracks for 20 elements assuming three different SN yield models and find some significant discrepancies with observations, especially for elements with strongly metallicity-dependent yields. Analyzing the high dimensional abundance space probed by surveys like APOGEE, GALAH, and Gaia-ESO will require more advanced statistical techniques. We applied one such technique, principal component abundance analysis, to the simulations and data to reveal the main correlations amongst abundances and quantify their contributions to variation in abundance space. PC1 and PC2 of the stellar population mixing scenario are dominated by alpha-elements and elements with metallicity-dependent yields, respectively, and they collectively explain 99% of the variance. flexCE is available at https://github.com/bretthandrews/flexCE.