Scientists develop virtual universe machine to study galaxy formation

Scientists and astronomers have been studying how galaxies form for a long time and have a new tool to try and figure out how formation happens. A team from the University of Arizona is closer to answering those questions thanks to a new computer simulation they have developed. One challenge for researchers is that studying real galaxies only gives a snapshot in time.

Scientists wanting to study the evolution of galaxies over billions of years have to use computer simulations. Scientists Peter Behroozi and his team generated millions of different universes on a supercomputer with each obeying different physical theories for how galaxies form.

The findings of the team from the simulations challenge the fundamental ideas about the role of dark matter in galaxy formation and how the galaxy evolves over time and how birth is given to stars. The simulations allow the team to create many universes and then compare them to the actual universe to see which rules lead to what we see.

This study is the first to create self-contained universes that are such exact replicas of a real galaxy. The simulations each represent a big chunk of the actual cosmos with 12 million galaxies spanning the time from 400 million years after the Big Bang to today. Each of the "Ex-Machina" universes was put through a series of tests to evaluate how similar galaxies appeared in the generated universe compared to the real universe.

Each of the simulated universes most similar to our own had similar underlying physical rules. The research aimed to help explain why galaxies cease to form new stars when they retain plenty of hydrogen gas and raw material to continue creating stars. The team found in some simulations that galaxies were much redder than those we see in the real world and found that had to do with two things. One is the age of the galaxy; those further away had shifted into the red spectrum, known as redshift.

The other is that if the galaxy has stopped star formation, it will contain fewer blue stars, which die out sooner, leaving the older red stars. However, the team didn't see that indicating that supermassive black holes and exploding stars are less efficient at stopping star formation than previously believed. The team plans to expand the simulation further for additional study.