3-D Simulations of i-Process Nucleosynthesis in the Early Universe

Astronomical observations are now probing the early stages of the universe, in which galaxies developed and merged, and the formation of the structures in the universe at the present era was set in motion. An important part of this story is the chemical evolution of galaxies. The formation of the elements in the first stars and their subsequent dispersal are powerful tracers of structure formation and evolution in the early universe. The elements are synthesized deep in the interiors of stars well beyond the reach of direct observation. For this reason, understanding their synthesis and the environment which enables it is clearly in the realm of theory. Although stellar evolution theory is one of the oldest branches of computational science, it has only recently become possible to accurately simulate

some of the key events and processes inside stars that give rise to the heavy elements that are injected back into the interstellar environment rather than being swallowed up in collapsed objects. These events can be quite brief in some cases, lasting only years, days, or even seconds in the case of explosive phenomena. The brevity of these events makes it possible for us now to simulate them in full 3-D detail. This project aims to build a new and powerful capability to simulate these events by coupling together 3-D simulations that cover brief time intervals with 1-D simulations spanning much longer times.

This project will simulate processes deep inside stars of the early universe that involve hydrogen ingestion events caused by mixing of material across the boundaries of convection zones. In 1-D stellar evolution simulations, these give rise to rapid hydrogen burning, with enormous energy release and causing nucleosynthesis by neutron capture from neutron fluxes intermediate between the slow (s) and rapid (r) processes. This is the intermediate, or i-process nucleosynthesis that is targeted in this study. Accurate treatment of the convective boundary mixing that drives this i-process requires full 3-D simulation. In some cases that will be addressed, it is possible to simulate the entire event in 3-D. In others, the 3-D simulations performed at intervals as part of a coupled 1-D and 3-D stellar evolution calculation will serve to recalibrate the coefficients in approximate 1-D models for dynamic mixing processes. Such coupled calculations will use these periodic model recalibrations to validate the use of the models as the calculation progresses. When unacceptably large revisions of model 'constants' are indicated, a recomputation with 3-D recalibrations taken at shorter time intervals will result. This new process called autocorrecting modeling involves periodic generation of large amounts of 3-D data from simulation codes. This data generation is followed by a detailed analysis of the data in terms of 1-D models. The results of this analysis are then fed back into the 1-D stellar evolution computation. In the process of carrying out the simulations a large database of results from 3-D simulations of convective boundary mixing in stellar interiors will be generated. This database will be organized so that the simulation codes can mine it automatically, comparing it to a variety of potentially useful 1-D models of the mixing process. This database, together with the tools that mine it, analyze it, and display results, will be unique. It will represent a significant investment, not only in our time developing the simulation codes and data analysis and visualization tools, but also in computer time on one of the most powerful computing systems in the world. A broad impact of this project will therefore be this database and tools that make it useful in stellar evolution research. The project will result in the development of new, automated techniques for data-enabled science but in addition to these techniques an important outcome will be the data itself which will be made available to the community on-line in a useful format. The hydrodynamics simulations will enable the clarification of the nucleosynthetic yields of the i-process in early generations of stars. The PI will work with the NuGrid collaboration to see that the results become incorporated into data sets that are available to the community for chemical evolution simulations of galaxies and structures in the early universe. The PI is also collaborating with the scientists at NSF?s Joint Institute for Nuclear Astrophysics (JINA). A further impact of this work will be to perform the detailed simulations that allow the assessment of the impact of these new rate measurements on nucleosynthesis, so that it can be compared with observables such as spectra of metal-poor stars in globular clusters and isotopic compositions of pre-solar grains.