(ORDO NEWS) — Active galactic nuclei (AGNs) play a major role in the evolution of galaxies. Astronomers from SRON and RuG used a record sample of galaxies to confirm that galactic mergers have a positive effect on the work of AGNs. They were able to collect about 10 times more images of merging galaxies than previous studies using the machine learning algorithm.
One of the biggest questions in astronomy is how galaxies evolve from clouds of gas and dust into beautiful spiral structures observed in our galactic neighborhood. The so-called active galactic nuclei (AGNs) are interesting research objects to answer part of the question, because, apparently, there is co-evolution between AGNs and galaxies.
In the AGN there are supermassive black holes that release a huge amount of energy after gas accretion from the environment. Some have large enough magnetic or gravitational fields to spit out jets from poles thousands of light-years long.
Co-evolution is a two-way street. On the one hand, the stage of evolution of the galaxy affects the activity of AGN. It seems that AGNs flourish at a certain stage in the evolution of the galaxy, because we observe the peak activity of AGNs in galaxies at a certain distance and, therefore, at a certain time in the past. On the other hand, the activity of AGN affects the star formation of the galaxy.
This process can go either way. The AGN jet repels gas, spreading through the galaxy, causing the gas to collide with another gas and, thus, creating clots – seeds for small stars. But AGN also release energy, heating the gas and thereby preventing its cooling and condensation into lumps.
Astronomers from the SRON of the Netherlands Institute for Space Research and the University of Groningen (RuG) used a sample with a record number of galaxies to study one of the factors that are believed to positively influence the ignition of AGN: fusion between galaxies. Indeed, the researchers found a correlation. In mergers, they accounted for 1.4 times more AGN than in non-mergers. Conversely, researchers found 1.3 times more mergers in galaxy samples with AGN compared with galaxy samples without AGN.
The research team used machine learning to recognize mergers. He provided a sample that is about an order of magnitude larger than in previous studies, which makes correlation much more reliable. “We created a network to train the merger recognition system on a large number of images,” says Fangyu Gao, the first author.
“This allows us to use a large sample of two telescopic images with tens of thousands of galaxies. AGNs are relatively easy to recognize by their spectrum. But mergers must be classified by image, which is usually the work of man. With machine learning, we can now have computers do it for us.”
Contact us: [email protected]