Next Generation Stellar Models with Rotation

Abstract

This thesis details a line of work, carried out over several years, that has seen the development of a next generation, self-consistently evolved, rotating stellar evolution model grid. The forthcoming chapters describe their application in studying various observed phenomena. Stellar rotation is the third fundamental parameter of stellar evolution, alongside mass and metallicity, that sets the evolutionary track of a star. Thus, accurately modeling stellar rotation is essential in accurately predicting stellar behavior, and interpreting stellar data. The model grid developed in this thesis is one of only several constructed thus far in order to incorporate, and study the effects of stellar rotation on evolution. In Chapter 2, we describe the initial grid that was created to benchmark stellar behavior, as modeled in the 1D stellar evolution code MESA using well-studied, nearby stellar populations. We applied our models to the Pleiades, the Hyades, and the Praesepe galactic open clusters to derive cluster ages and metalicities. We found, contrary to a contemporaneous study carried out with an alternative model set, that stellar rotation did not appear to greatly affect the derived cluster properties in comparison to previous studies that used non-rotating stellar models. Thus, our results highlighted the uncertainty present due to unconstrained physics in stellar models, primarily the assumptions made in modeling stellar rotation, and convective mixing. We also demonstrated that the dominant effect of stellar rotation appears to be gravity darkening in our MESA models, versus main sequence lifetime extension (due to rotation-enhanced mixing) in the alternative model set. These model discrepancies necessitate further studies aimed at constraining the effects of stellar rotation (in addition to other uncertain aspects of stellar evolution). In Chapter 3 we extended the application of our models to study the extended main sequence turn off phenomenon, observed in all open clusters younger than about 2 Gyr. Our results confirm previous findings that it appears stellar rotation distributions are responsible for this phenomenon, but can not fully rule out an age spread in our modeling. While an combination of an age spread and rotation rate distribution often is the most statistically favorable model in replicating observations on color-magnitude diagrams, age spreads also introduce features that are not observed. In addition, rotation rate distribution appear to be able to account for the majority of the extended main sequence turn off morphology, even in absence of an age spread. The primary means by which stellar rotation creates an extended main sequence turn off is by gravity darkening according to our models. Our models also manage to simultaneously replicate the observed split main sequence (a feature common to open clusters younger than about 400 Myr) in the cluster NGC 1866. This provides strong evidence for stellar rotation being a primary factor in explaining these stellar population phenomena. In Chapter 4 we describe improvements made to our models in modeling the rotation behavior of stellar masses $\leq\ 1.8\ \rm M_{\odot}$. In our previous studies, modeling stars in this mass range was not done well, meaning that our grid had been limited in not being able to model rotation effects, e.g., in clusters aged greater than about 1.7 Gyr, which is roughly when stars of mass $1.8\ \rm M_{\odot}$ enter the main sequence turn off in our modeling. The work presented in Chapter 4 implemented two options for modeling the angular momentum evolution of stars $\leq\ 1.8 \rm M_{\odot}$: those of \cite{matt.etal:2015} and \cite{garraffo.etal:2018}. These stars require special treatment because stars in this mass range begin to develop surface convection zones, and surface magnetic fields. These magnetic fields couple to outgoing stellar winds and gradually slow the star in a process known as magnetic braking. In this work, we tested and compared our newly implemented magnetic braking models, finding that both fail to completely reproduce observed rotation periods in open cluster, but do provide some successes as well. Our results are, broadly speaking, in line with other stellar evolution models that employ the same magnetic braking formalisms. Thus, both of the implemented braking models serve as incomplete, but promising means of modeling the rotation rate evolution of stars with convective envelopes. We also check for our models ability to replicate the observed lithium depletion of solar-like stars, and are not able to reproduce it, indicating additional, missing physical processes are likely required.