Modeling the Universe

(Originally published by MIT)

January 31, 2018

Rendering of the gas velocity in a thin slice of 100 kiloparsec thickness (in the viewing direction), centered on the second most massive galaxy cluster in the TNG100 calculation. Where the image is black, the gas is hardly moving, while white regions have velocities that exceed 1,000 kilometers per second. The image contrasts the gas motions in cosmic filaments against the fast chaotic motions triggered by the deep gravitational potential well and the supermassive black hole sitting at its center.
Rendering of the gas velocity in a thin slice of 100 kiloparsec thickness (in the viewing direction), centered on the second most massive galaxy cluster in the TNG100 calculation. Where the image is black, the gas is hardly moving, while white regions have velocities that exceed 1,000 kilometers per second. The image contrasts the gas motions in cosmic filaments against the fast chaotic motions triggered by the deep gravitational potential well and the supermassive black hole sitting at its center. (Image courtesy of the IllustrisTNG collaboration)

A supercomputer simulation of the universe has produced new insights into how black holes influence the distribution of dark matter, how heavy elements are produced and distributed throughout the cosmos, and where magnetic fields originate.

Astrophysicists from MIT, Harvard University, the Heidelberg Institute for Theoretical Studies, the Max-Planck Institutes for Astrophysics and for Astronomy, and the Center for Computational Astrophysics gained new insights into the formation and evolution of galaxies by developing and programming a new simulation model for the universe — “Illustris - The Next Generation” or IllustrisTNG.

Mark Vogelsberger, an assistant professor of physics at MIT and the MIT Kavli Institute for Astrophysics and Space Research, has been working to develop, test, and analyze the new IllustrisTNG simulations. Along with postdocs Federico Marinacci and Paul Torrey, Vogelsberger has been using IllustrisTNG to study the observable signatures from large-scale magnetic fields that pervade the universe.

Vogelsberger used the IllustrisTNG model to show that the turbulent motions of hot, dilute gases drive small-scale magnetic dynamos that can exponentially amplify the magnetic fields in the cores of galaxies — and that the model accurately predicts the observed strength of these magnetic fields.

“The high resolution of IllustrisTNG combined with its sophisticated galaxy formation model allowed us to explore these questions of magnetic fields in more detail than with any previous cosmological simulation," says Vogelsberger, an author on the three papers reporting the new work, published today in the Monthly Notices of the Royal Astronomical Society.

Thin slice through the cosmic large-scale structure in the largest simulation of the IllustrisTNG project. The image brightness indicates the mass density, and color visualizes the mean gas temperature of ordinary (“baryonic”) matter. The displayed region extends by about 1.2 billion light years from left to right. The underlying simulation is presently the largest magneto-hydrodynamic simulation of galaxy formation, containing more than 30 billion volume elements and particles.

Thin slice through the cosmic large-scale structure in the largest simulation of the IllustrisTNG project. The image brightness indicates the mass density, and color visualizes the mean gas temperature of ordinary (“baryonic”) matter. The displayed region extends by about 1.2 billion light years from left to right. The underlying simulation is presently the largest magneto-hydrodynamic simulation of galaxy formation, containing more than 30 billion volume elements and particles. (Image courtesy of the IllustrisTNG collaboration)

Modeling a (more) realistic universe

The IllustrisTNG project is a successor model to the original Illustris simulation developed by this same research team but has been updated to include some of the physical processes that play crucial roles in the formation and evolution of galaxies.

Like Illustris, the project models a cube-shaped piece of the universe. This time, the project followed the formation of millions of galaxies in a representative region of the universe with nearly 1 billion light years on a side (up from 350 million light years on a side just four years ago). lllustrisTNG is the largest hydrodynamic simulation project to date for the emergence of cosmic structures, says Volker Springel, principal investigator of IllustrisTNG and a researcher at Heidelberg Institute for Theoretical Studies, Heidelberg University, and the Max-Planck Institute for Astrophysics.

The cosmic web of gas and stars predicted by IllustrisTNG produces galaxies quite similar to the shape and size of real galaxies. For the first time, hydrodynamical simulations could directly compute the detailed clustering pattern of galaxies in space. In comparison with observational data — including the newest large galaxy surveys such as the Sloan Digital Sky Survey — IllustrisTNG demonstrates a high degree of realism, says Springel.

In addition, the simulations predict how the cosmic web changes over time, in particular in relation to the underlying backbone of the dark matter cosmos. “It is particularly fascinating that we can accurately predict the influence of supermassive black holes on the distribution of matter out to large scales,” says Springel. “This is crucial for reliably interpreting forthcoming cosmological measurements.”

Visualization of the intensity of shock waves in the cosmic gas (blue) around collapsed dark matter structures (orange/white). Similar to a sonic boom, the gas in these shock waves is accelerated with a jolt when impacting on the cosmic filaments and galaxies.

Visualization of the intensity of shock waves in the cosmic gas (blue) around collapsed dark matter structures (orange/white). Similar to a sonic boom, the gas in these shock waves is accelerated with a jolt when impacting on the cosmic filaments and galaxies. (Image courtesy of the IllustrisTNG collaboration)

Astrophysics via code and supercomputers

For the project, the researchers developed a particularly powerful version of their highly parallel moving-mesh code AREPO and used it on the "Hazel-Hen" machine at the Supercomputing Center in Stuttgart, Germany's fastest mainframe computer.

To compute one of the two main simulation runs, more than 24,000 processors were used over the course of more than two months.

“The new simulations produced more than 500 terabytes of simulation data,” says Springel. “Analyzing this huge mountain of data will keep us busy for years to come, and it promises many exciting new insights into different astrophysical processes."

Supermassive black holes squelch star formation

In another study, Dylan Nelson, researcher at the Max-Planck Institute for Astrophysics, was able to demonstrate the important impact of black holes on galaxies.

Star-forming galaxies shine brightly in the blue light of their young stars until a sudden evolutionary shift quenches the star formation, such that the galaxy becomes dominated by old, red stars, and joins a graveyard full of old and dead galaxies.

“The only physical entity capable of extinguishing the star formation in our large elliptical galaxies are the supermassive black holes at their centers,” explains Nelson. “The ultrafast outflows of these gravity traps reach velocities up to 10 percent of the speed of light and affect giant stellar systems that are billions of times larger than the comparably small black hole itself.“

New findings for galaxy structure

IllustrisTNG also improves researchers' understanding of the hierarchical structure formation of galaxies. Theorists argue that small galaxies should form first, and then merge into ever-larger objects, driven by the relentless pull of gravity. The numerous galaxy collisions literally tear some galaxies apart and scatter their stars onto wide orbits around the newly created large galaxies, which should give them a faint background glow of stellar light.

These predicted pale stellar halos are very difficult to observe due to their low surface brightness, but IllustrisTNG was able to simulate exactly what astronomers should be looking for.

“Our predictions can now be systematically checked by observers,” says Annalisa Pillepich, a researcher at Max-Planck Institute for Astronomy, who led a further Illustris-TNG study. “This yields a critical test for the theoretical model of hierarchical galaxy formation.”

Astrophysics