CHAMPAIGN, Ill. - By comparing high-resolution, multi-wavelength observations with advanced computer simulations, astronomers at the University of Illinois are obtaining a clearer understanding of how stars form when galaxies collide.
"Galaxy collisions are considered rare events today, but were much more common in the early universe," said Susan Lamb, a U. of I. professor of physics and of astronomy who presented her team's latest findings at the American Astronomical Society meeting in January. "Our current study is providing additional insight into the kinds of events that produced vast numbers of stars in the past."
The particular system being studied is composed of two interacting galaxies, an elliptical galaxy (NGC 1143) and a disturbed disk galaxy (NGC 1144) that contains a very extended starburst ring. Bursts of star formation occur as a result of density increases and shocks in the gas that are created as one galaxy plows through the other.
"The collision generated a combined density/material wave in the disk, which moves outward and produces clumping in the gas on a large scale," Lamb said. "New stars tend to be formed in these large clumps. By combining multi-wavelength observations with our detailed models, we not only can examine the individual clumps of stars that are now seen, but also get a much more complete sense of the time history and evolution of the different star-formation events that took place after the collision."
To investigate the sequence of star formation triggered by a galaxy collision, Lamb and her colleagues -- visiting postdoctoral research associate Yu Gao and graduate student Nathan Hearn -- used a series of three-dimensional numerical simulations of collisions between a gas-rich disk galaxy and a gas-free spherical galaxy. The models were computed at the National Center for Supercomputing Applications (now part of the National Computational Science Alliance) at the U. of I.
"On the computer, we can watch the movement of the gas and stars as the collision takes place," Lamb said. "Then, by matching the models at different stages of the simulation with observations taken at different wavelengths, we can determine the sequencing of the various star-formation episodes."
Regions of very strong radio emission, for example, indicate where the earliest bursts of star formation occurred, Lamb said. Those stars have now aged, and there has been sufficient time for them to reach the supernova stage. Regions of intense hydrogen-alpha emission show where stars have formed more recently, while observations of carbon monoxide provide information about those regions of the gas that are currently dense and likely to be about to form stars.
"The intensity and location of the starburst at any particular time will depend upon the speed with which the density/material waves are propagating through the expanding disk," Lamb said. "Because these quantities can now be predicted quite accurately from the current models of colliding galaxies, global star formation can be investigated much more thoroughly than previously."