Observations are increasingly probing the relationship between star formation rates and local conditions in disk galaxies, but the mechanisms that set these rates are not well understood. In particular, while it is widely believed that energetic feedback self-regulates star formation, detailed models of this process have not yet been developed. H II regions and supernovae in giant molecular clouds (GMCs) can both trigger star formation and truncate it (by cloud destruction), and feedback-driven diffuse-interstellar medium (ISM) turbulence can both enhance and suppress the creation of new GMCs. Additional elements that control star formation include: gas self-gravity, which gathers the ISM into massive clouds; the stellar component's gravity, which confines the ISM vertically; heating and cooling processes, which separate the ISM into dense and diffuse thermal phases; sheared orbital rotation and Coriolis forces, which suppress converging flows; and magnetic fields, which mediate angular momentum exchange. Dr. Eve Ostriker (University of Maryland College Park) and collaborators will study how these processes interact to regulate large-scale star formation, identifying the dominant effects and developing quantitative predictions in terms of common observables. The research will involve a series of focused numerical simulations, separately targeting three regimes of star formation: galactic center, mid-disk, and outer disk. The approach to be taken is distinct from other recent numerical work in that the simulations will concentrate on local (1 to 1000 parsec) rather than global (30 to 30,000 parsec) scales, directly following processes that are often treated using sub-grid prescriptions in galaxy evolution models. Each simulation domain will be large enough to capture the important galactic environmental effects, but small enough so that the turbulence and density structure of the gas is well resolved by the numerical grid. Preliminary studies have shown that vertical resolution of the ISM disk is crucial to correctly estimating the surface density of star formation.
It is expected that this work will have broader impacts on workforce, education, and research infrastructure. They include (1) training and mentoring of students, including a PhD student working on modeling and analysis, and additional graduate students (co-)advised by Dr. Ostriker in related observational and scientific computation projects, (2) integration of scientific research results and computational modules in instructional curricula for undergraduate and graduate courses. (3) outreach activities for the public, including creation of visually-rich web pages and presentations at open house events, and (4) development, implementation, and testing of new tools for computational hydrodynamics/magneto-hydrodynamic simulation, to be shared with the community as part of an established code package.
All of the stars in our own Milky Way and other galaxies are made of material that was originally much less dense, known as the interstellar medium (ISM). In composition, the ISM is primarily hydrogen gas in either atomic or molecular form, but it also contains about 10 percent helium, trace amounts of other elements in gaseous form, and also small solid particles that astronomers call "dust." The ISM has a complex physical and thermal structure, and is highly dynamic. Most of the ISM consists of clouds of cold gas embedded in an intercloud medium of much warmer gas. The clouds and intercloud medium are in constant motion, with individual structures merging and shearing apart in response to pervasive multi-scale turbulence. For the atomic medium, the cold clouds have comparable pressure to that in the surrounding medium, but density 100 times larger. As a consequence, these clouds tend to fall toward the midplane of the galaxy, where they attract each other and can collect into very massive giant molecular clouds (GMCs). Star formation takes place when dense knots of gas within GMCs collapse gravitationally. Most of the stars that form have masses comparable to that of our Sun or lower, but some of them are much more massive, and these rare but powerful stars have a dramatic impact on their own natal GMCs and beyond, into the surrounding ISM. One of the most important effects of massive stars is to heat the whole ISM with ultraviolet (UV) radiation, which maintains the temperatures of both the cold clouds and the warm intercloud medium. Another -- perhaps even more important -- effect from massive stars takes place after their short lives end, and they explode as supernovae. Each supernova sends a blast wave into the surrounding ISM that travels over hundreds of light years, and the combination of shocks from many supernova blasts stirs up turbulence in the ISM. These shocks and turbulence can turn around cold clouds as they fall toward the galactic midplane. While the properties of the ISM and the formation and properties of young stars have been observed for many years, it has been a puzzle to understand what sets the rate at which new stars form, and exactly how the UV radiation and supernova shock waves produced by massive stars control the temperature, density, and turbulence levels in the ISM. In this project, we developed theoretical and computational models that explain quantitatively, for a wide range of environments, how star formation rates are controlled, and how at the same time star formation ends up controlling the properties of the ISM. In our models, we show how ongoing cycles of local gravitational collapse followed by dispersal by supernova shock waves mediates the fractions of the ISM that are in diffuse gas versus gas that is actively forming stars. Too much star formation produces excessive heating and shock waves that cause gas to be dispersed away from the midplane, but this reduces the amount of gas in GMCs and thus the star formation rate drops. When there is too little star formation, more and more clouds will fall to the midplane, until there is enough dense gas to cause a new round of gravitational collapse. Our models show that the heating and stirring of the ISM produced by star formation "feedback" is just what it needs to be in order to explain the detailed properties of the ISM observed in the local neighborhood of our own Sun. In addition, we show that the levels of star formation predicted by our models in other environments -- either with much less ISM gas, at galactic "fringes," or much more ISM gas, at galactic centers -- are also in good agreement with observations. Remarkably, our models also appear able to explain star formation at much earlier epochs of the Universe, when galaxies were young and our own Sun was born. A number of different researchers were involved in this project. In addition to the Principal Investigator, Professor Eve Ostriker, several graduate students, undergraduates, and postdoctoral researchers were trained to use advanced computational techniques, and worked on developing scientific models. The team also worked closely in developing and testing models with researchers at other universities and laboratories in the United States, Korea, and Canada, as well as partnering in international collaborations that used the NSF ALMA facility to observe and interpret star-forming systems at unprecedented resolution. New computational tools developed by the team for use in this project have been released for free use in the astrophysics community, and have already been adopted by other researchers for novel applications.
