Archive
Galaxy collisions take a lesson from dance
By Julie Comerford, University of California Berkeley
As expansive as the Universe is, galaxies still manage to crowd together and interact. Mergers between galaxies are relatively common events, with even the Milky Way expected to sustain a head-on collision with the Andromeda galaxy in 3 billion years. However, it can be difficult to accurately identify merging galaxies using current methods of spotting galaxy mergers in optical images.
Image: The Antennae galaxies, a pair of merging galaxies named for the long tidal tails of stars, gas, and dust ejected during the merger. In the next several hundred million years, the two galaxy nuclei seen here will combine to form a single merger-remnant galaxy. Credits: NOAO/AURA/NSF, B. Twardy, B. Twardy, and A. Block (NOAO)
To address this challenge, we developed a new observational technique for identifying galaxy collisions using their inspiralling black holes. Our approach uses galaxy spectra rather than imaging, which is an advantage for building a large sample of galaxy mergers since many more telescopes are capable of spectroscopy than of high-resolution imaging.
All or nearly all galaxies — including the Milky Way — have a central, supermassive black hole that is a million to a billion times the mass of the Sun, and two colliding galaxies bring their black holes to the resultant merger-remnant galaxy. The black holes gradually inspiral toward the center of the merger-remnant galaxy, engaging in a gravitational tug-of-war with the surrounding stars. The result is a black hole dance, choreographed by Newton himself.
Imagine two dancers waltzing in a darkened ballroom. An observer standing in the ballroom would not be able to see the dancers at all. Similarly, a black hole is not visible because it absorbs all light that comes within its Schwarzschild radius —the gravitational radius where no light or matter can escape from a black hole. For a supermassive black hole, this radius is similar in size to the radius of Earth.
However, if the black hole is surrounded by gas from the host galaxy, this gas can fall onto the black hole and power an active galactic nucleus, or AGN, which may produce more radiation than the host galaxy itself. The effect is similar to the two ballroom dancers picking up sparklers. Suddenly their dance is very visible.
We searched for such inspiralling AGNs in the DEEP2 Galaxy Redshift Survey, a collaborative research program between the University of California Berkeley and the University California Santa Cruz led by Marc Davis and Sandy Faber. The goal of DEEP2 was to obtain spectra for 50,000 galaxies in the middle-aged Universe, when the cosmos was roughly seven billion years old (equivalent to redshift z=1).
The DEep Imaging Multi-Object Spectrograph, or DEIMOS instrument, on Keck II was built to enable DEEP2 to accomplish this task. To observe such a large number of galaxies, quarter-inch long slits were cut into sheets of aluminum 3 feet long, called slitmasks. Each slitmask has roughly 150 such slits and each slit has been carefully aligned so that when a slitmask is inserted into DEIMOS, each slit allows in light from one particular galaxy.
Because it could observe about 150 galaxies at the same time with slitmasks on DEIMOS, the DEEP2 survey obtained spectra for 50,000 galaxies in only 90 Keck nights—Julie Comerford, UC Berkeley astronomer.
Image: The DEep Imaging Multi-Object Spectrograph (DEIMOS), shown here at Lick Observatory where it was built, was mounted on the Keck II telescope in 2002. We used data taken with DEIMOS to identify galaxy mergers by their inspiralling black holes. Photo courtesy of Lick Observatory.
From the tens of thousands of galaxies surveyed in this program, we selected red galaxies at redshifts between z= 0.3 and z=0.8, when the Universe was four billion to seven billion years younger than it is today. The galaxy spectra show emission lines, chemical fingerprints that provide clues about the nature of the objects we are observing. The galaxy spectra of interest to us show emission lines such as the forbidden line transition of doubly ionized oxygen [O III], which indicates AGN activity. We restricted our sample to red galaxies to ensure that the emission lines were produced by AGNs and not star formation, which makes galaxies blue.
We then measured the redshift of each galaxy based on the wavelengths of its stellar absorption lines, and we measured the redshift of the AGN in each galaxy based on the wavelength of the AGN-fueled [O III] emission line. Any difference between the redshift of the host galaxy’s stars and the redshift of its AGN corresponds to a difference in the line-of-sight velocities of the host galaxy and the AGN.
Such velocity differences can be produced by outflows of gas from the AGN or the rotation or flows of gas within the host galaxy, but the galaxies in our sample are not consistent with the theoretical expectations and observations of galaxies with AGN outflows or gas kinematics. Rather, the most likely explanation for velocity differences in our sample is AGNs inspiralling within galaxies, which produce a velocity of the AGN that is different from the bulk velocity of the galaxy through space.
For the DEEP2 red galaxies hosting AGNs, we measure velocity differences between the host galaxy and the AGN that range from zero to 300 kilometers per second. For reference, the Solar System orbits the center of the Milky Way at a velocity of 200 kilometers per second.
Interestingly, we found two galaxies that have double-peaked [O III] emission lines. This suggests that each of these galaxies hosts two inspiralling AGNs!
Figure: DEIMOS spectra of merger-remnant galaxies show the signatures of inspiralling AGNs. An AGN produces [O III] emission lines at the wavelengths shown by the dashed lines if the AGN is at rest within the host galaxy. However, these emission lines peak at slightly different wavelengths, which correspond to velocity offsets of the AGNs moving within the host galaxies. The top two spectra (in black) are of galaxies hosting two AGNs that are inspiralling relative to each other at 630 km/s and 440 km/s, respectively. The bottom six spectra (in red) are examples of galaxies hosting one inspiralling AGN, where the AGN velocities range from -140 km/s to 170 km/s within their host galaxies. Here, negative velocities indicate motion towards our telescope and positive velocities indicate motion away from our telescope. Credit: Julia Comerford, UCB
Each AGN produces an [O III] emission line, and combined these two [O III] emission lines produce the double-peaked line structure we see in the spectrum of the galaxy. Each of these two galaxies has two AGNs swirling around in it at a few hundred kilometers per second. This is akin to both of the ballroom dancers holding sparklers, so that we can see both parts of the dance.
We also found 30 galaxies that have single-peaked [O III] emission lines that have velocities significantly different than the velocities of the host galaxies. These galaxies have two inspiralling black holes, where there is sufficient gas around one to power an AGN and not around the other. This is analogous to only one of the ballroom dancers holding a sparkler, so that we can see only half the dance. Although we cannot see the second dancer, we infer that dancer’s presence based on the motion of the lit dancer.
These inspiralling AGNs are a clear signature that the host galaxy has recently undergone a merger with another galaxy. From the inspiralling AGNs we found in the DEEP2 Galaxy Redshift Survey, we estimate that in half of the 0.3 < z < 0.8 red galaxies with AGNs, the AGNs are inspiralling due to a recent galaxy merger. This striking result, that half of red galaxies hosting AGNs in this redshift range are also merger remnants, suggests a strong link between AGN activity and galaxy mergers. Some numerical simulations of galaxy mergers show that mergers trigger gas flows that fuel inspiralling black holes as inspiralling AGNs, and our results lend observational support to this idea. Finally, our inspiralling AGNs offer an independent answer to the question of how often galaxies merge: at a cosmic time of four to seven billion years ago (0.3 < z < 0.8), red galaxies undergo on average three mergers every billion years.
Inspiralling AGNs are a powerful direct observational probe of not only the galaxy merger rate, but also the kinematics of black hole mergers. When black holes complete their inspiral and coalesce, they are expected to produce ripples in space called gravity waves. In 1993, a Nobel Prize for physics was awarded for indirect detections of gravity waves, and next-generation detectors are underway to make direct detections of the ripples from events such as black hole mergers.
Inspiralling AGNs are unique direct observational tracers of galaxy and black hole mergers that will continue to advance our understanding of galaxy and black hole evolution.