CDS&E: The development of open-source software with an application to modeling the formation of globular star clusters

Columbia University, New York NY

Investigators

Abstract

Globular star clusters are among the oldest objects in our Universe, and understanding their origin has remained a great challenge to astrophysicists for many decades. This project will develop computer software employing a new numerical method to perform the calculations modeling the formation of these clusters of stars. This new computer software will enable better computer modeling of the astrophysics governing a variety of physical applications. It will also be the first such software to be made publicly available. The investigators will then use this software to create computer simulations that model the formation of globular star clusters, and they specifically seek to understand the apparent creation of multiple generations of stars in these globular clusters. This project's goals will advance the national interests of developing US leadership in physics and enhancing the US cyber infrastructure. In addition, the investigators will address the national underrepresentation of minorities in the sciences by funding a graduate student to participate in Columbia University's Bridge to the PhD in the Natural Sciences program. The investigators will train the student in the research program and will mentor the student through the early stages of his or her science career. More technically, the first part of this project lays out a systematic plan to develop a new open source moving-mesh code. Recent development of the moving-mesh method as a viable alternative to smoothed particle hydrodynamics and adaptive mesh refinement is a promising development. Moving-mesh hydrodynamics brings many desirable features including low-diffusion in bulk flows, Galilean invariance, and good shock handling. The software development uses modern design and coding techniques to produce an efficient, parallel code that will be capable of using a wide range of physics and that is easy to use and modify. It will use the widely used Python language combined with Cython (a static Python compiler) in computationally intensive parts. Coding will be modular and designed to allow easy and flexible extension to all parts of the code. The second part of this project is to simulate the formation of globular clusters. The code will be applied to a three-part plan for understanding globular cluster formation: (1) modeling the fragmentation and collapse of idealized low-metallicity gas clouds to understand the relative importance of global collapse and fragmentation; (2) using cosmological simulations to identify the physical properties of gas undergoing rapid colliding flows, a prerequisite to bring together a large amount of gas in a small region for GC formation; and (3) exploring the ability of gas outflows from the proposed sources of enriched material to be re-incorporated into a second generation of stars.