The First Stars, As Seen By Supercomputers

This is the title of an article by Tom Abel of the Kavli Institute, Stanford and appearing in the April 2011 issue of Physics Today. The article may be downloaded for personal use free of charge. The on-line article contains some animations that are well worth a look.

Much of the first billion years of the history of the Universe, when the first stars and galaxies were formed, has been inaccessible to direct observation. That will begin to change in the coming years, as powerful new telescopes will begin to lift the shroud on the Dark Ages, as this era is dubbed by astronomers. These telescopes include the Atacama Large Telescope Array (ALMA), the James Webb Space Telescope  (JWST) (successor to the Hubble) and the European Extremely Large Telescope (EELT).

Evolving Computational Grids. Tom Abel writes: With adaptive mesh refinement, a numerical simulation generates ever smaller grids to accurately capture the physics of contracting cosmic gas as it gets denser. The image here corresponds to a time 200 million years after the Big Bang and represents a region 2000 light-years across. The bright areas show where the cosmic gas density is greatest. Superimposed on the simulation frame is a hierarchy of grids. The smallest grids tend to track the bright regions of high density. (Simulation by John Wise and Tom Abel; image by Ralf Kaehler and Tom Abel. Courtesy Physics Today.)

Abel describes how an “adaptive mesh refinement technique,” encoded in a community supported numerical code called Enzo, has been used to model the formation of the first stars. The simulations will guide the design of experiments to measure structures in the early Universe, and provides an example of the confrontation between massive computer simulations and PB-scale data sets that will become commonplace in astronomy. The Enzo code has been written to accommodate the extraordinary dynamic range of spatial and temporal steps needed to model the volume of the Universe.

The adaptive mesh technique starts out with as large a computational box as possible, and the computer stores all the relevant physical information. Equations are cast in a co-moving form so that cosmic expansion is absorbed in the coordinates. Once the motion of material that is collapsing to form galaxies and stars begins to deviate from the uniform cosmic expansion, the simulations create an hierarchy of grids that capture the length scale of the structure, optimally constructed to capture all the physical processes working in it. The figure above, Evolving Computational Grid, shows this process.

Blown away. Tom Abel writes: Massive stars are hot and intensely bright. The radiation they emit ionizes hydrogen out to thousands of light-years, and the associated pressure clears material away from the star. In the simulation shown here, a bright star sits roughly in the cen- ter of the frame. Except for a small angular region to its lower right, the star has largely blown away its surroundings, which appear darker in the image. If the star is massive enough, it may end its life as a black hole in an environment where there is little to accrete. (Simulation by John Wise and Tom Abel; image by Ralf Kaehler and Tom Abel. Courtesy Physics Today)

What do the simulations show?  One of the most important results is that massive stars, of 100 times the mass of the Sun, predominate among the first luminous objects. These stars are as hot as 200,000 K, and photoionize surrounding material to distances of thousands of light years; see Figure “Blown Away” above.  Moreover, radiation pressure effectively evacuate the surroundings. If the star dies as a supernova,  ejecta may travel 1,000 light years in such low density surroundings.  Gravity collects these ejecta into star forming regions, which grow into galaxies over periods of a billion years.

The most massive early stars may end their lives as black holes rather than supernovae, but cannot accrete material in their rarefied surroundings,  and will likely remain on this state for millions of years.

This entry was posted in Astronomy, cosmology, cyberinfrastructure, High performance computing, Parallelization, programming, software engineering and tagged , , , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s