Survey of Andromeda’s environs reveals relics of galactic cannibalism

A recent paper reporting the first major results of the Pan-Andromeda Archaeological Survey provides a striking look at what can be learned from the relatively young field of galactic archaeology. ¹ The PANDAS collaboration has undertaken what it calls a “panoramic survey” of the Andromeda galaxy and its extended environs, using a wide-field camera on the 3.6-m-diameter Canada-France-Hawaii Telescope (CFHT) atop Mauna Kea. The survey is archaeology in the sense that a principal goal is to find relics of past cannibalistic encounters between Andromeda and its satellite dwarf galaxies. Favored hierarchical theories of galaxy formation predict the existence of such relics.

At a distance of only 2.5 million light-years, Andromeda, a 10¹²-star spiral galaxy very much like the Milky Way, is our nearest full-grown galactic neighbor. As such, it presents a unique opportunity for a systematic search for remnants of galaxy formation unbiased by the inevitable limitations of an insider’s view. The hierarchical theories posit that big galaxies grow by the continual accretion of smaller ones. Among the expected remnants would be filamentary streams of stars from the tidal rupture of gravitationally captured dwarf galaxies that ventured too close to Andromeda’s center. But captured dwarfs whose orbits avoided such close encounters would remain largely intact. One goal of PANDAS is an unbiased census of the number of such survivors. A nagging problem for galaxy-formation theorists has been the scarcity—relative to what’s expected from dark-matter simulations—of observed dwarf galaxies orbiting big ones.

PANDAS

Within the diffuse spherical halo of very old stars that envelops the bright disk of the Milky Way, the Sloan Digital Sky Survey has, in recent years, revealed a number of such tidal streams and intact survivors. ² Organized last year by Alan McConnachie (Herzberg Institute of Astrophysics, Victoria, British Columbia) and collaborators from six countries, PANDAS is a systematic survey of Andromeda’s halo intended to reveal such structures while mapping the halo’s stellar population density.

Assuming that the radius of Andromeda’s halo, like the Milky Way’s, is more than 10 times bigger than the 50 000-light-year radius of the galaxy’s familiar luminous disk, McConnachie and company undertook to survey the area in two large circular patches of sky (see figure 1). One patch covers a radius of about 500 kly around Andromeda’s center, and the other extends one third that far from the center of the nearby Triangulum galaxy—the third largest member of our local group. Triangulum’s mass is about 10% that of Andromeda. The dwarf galaxies in question are two or three orders of magnitude smaller still.

So ambitious an undertaking dictates two requirements for the observing instrument. To cover the requisite 300 square degrees of sky in reasonable observing time, the camera must have an unusually wide field of view. And the telescope’s aperture must be big enough to resolve individual red giant stars at a distance of 2.5 Mly. Red giants are very old and hundreds of times more luminous than the Sun. They are good representatives of the halo’s geriatric population. At Andromeda’s distance they appear very faint; but stars any less luminous would be too faint. “Given those requirements,” says McConnachie, “the CFHT with its Megacam camera is uniquely suited to our task.” Several of the CFHT’s neighbors on Mauna Kea have apertures of 8 to 10 meters. But none of the bigger telescopes anywhere can, as yet, match the Megacam’s 1-degree-wide field of view.

Figure 1 is the PANDAS population-density map of red giants, covering a bit more than half the halo area ultimately to be surveyed. The two survey circles dwarf the superposed conventional images of the central disks of Andromeda and Triangulum, both seen obliquely. Those well-studied luminous disks and their immediate vicinities (shown black, as are the far outer areas still to be surveyed) are not included in PANDAS, whose principal concern is studying the halos. “And everywhere we’ve looked out to half a million light-years from Andromeda’s center,” says McConnachie, “we find a red-giant population indicating that the halo extends even further out.”

Halos and dark matter

It’s known from the orbital speeds of stars circling the centers of large galaxies that the galactic halos contain an order of magnitude more mass than can be accounted for by stars, gas, and dust. Modern cosmology presumes that the dark matter dominating the halos is nonbaryonic and that spherical accumulations of dark matter initiate galaxy formation by gravitationally trapping hydrogen gas, which is then further concentrated by dissipative interactions and angular-momentum conservation into flattened, spinning disks where star formation begins in earnest.

The more three-dimensional distribution of halo stars is attributed to cannibalism, with varying levels of violence: In the outer precincts of the halo, the tidal forces are fairly gentle with the captive dwarf galaxies, pulling out only their loosely bound outermost stars. But some of the dwarf galaxies—those whose eccentric orbits take them too close to the host galaxy’s central disk—have it worse. Eviscerated by much stronger tidal forces, they leave behind streams of stars that remain discernibly coherent for billions of years. As James Binney (University of Oxford) has put it, “Galaxies, like elephants, have long memories.” In either case, extracted stars, like the dark matter itself, are not subject to the dissipative forces that flatten the gas. So they remain distributed throughout the spherical halo.

Figure 1 shows seven relic star streams in the halos of Andromeda and Triangulum, four of them newly discovered by PANDAS. “Their enormous sizes and distances from the galactic centers are remarkable,” comments McConnachie. None of them appear near known halo concentrations of hydrogen gas, where they would be expected if the stars had been formed in place rather than acquired by cannibalism.

One of the new finds (labeled 1), the stream extending from Triangulum’s disk in the direction of Andromeda, is particularly intriguing. It could—as one supposes of the streams found in the Andromeda halo—simply be the relic of a captured and shredded dwarf galaxy. But the PANDAS collaboration speculates that this structure, which extends both north and south from Triangulum, may well be the scar of a more gargantuan encounter—an extensive tidal disruption of the galaxy by its putative passage, some 3 billion years ago, through an outer precinct of Andromeda’s disk. That scenario could also account for a well-known but puzzling warp of Triangulum’s hydrogen disk.

To test the hypothesis, the group simulated such an encounter with possible orbits of Triangulum about its more massive neighbor. Figure 2 (and the animation available online ³ ) shows the orbit that best reproduces Triangulum’s present position and velocity, and the disheveled appearance revealed by PANDAS. That orbit would have taken the galaxy through the plane of Andromeda’s disk 2.6 billion years ago at a distance of 170 kly from its center, well beyond the luminous inner disk, but close enough—so the simulation finds—to rip the newly discovered tails out of the smaller galaxy. The encounter may also have left traces on Andromeda itself, possibly the small warp previously observed in the outer reaches of its disk.

Counting survivors

Though it’s not yet known what non-baryonic dark matter is made of, the cosmological evidence favors weakly interacting particles massive enough to be “cold,” that is, nonrelativistic. The physics required to simulate the accumulation of such cold dark matter into halos small and large is relatively simple; gravity is the only appreciable force at play. So astrophysicists are inclined to believe the simulations that predict about a thousand captured dark-matter “minihalos” orbiting Andromeda—each big enough to house a dwarf galaxy.

But forming a visible dwarf galaxy inside a minihalo is much harder to simulate. “Star and galaxy formation involves all sorts of physics on all sorts of scales,” says McConnachie. So one wants, first of all, a reliable count of how many dwarf galaxies there are in the halo of a large galaxy like Andromeda and, second, some common property of the observed dwarf satellites that might reveal what it takes to keep a minihalo lit up with stars. It might, for example, require some minimum accumulation of dark matter—say 10⁸ solar masses—to form a gravitational well deep enough to keep the trapped hydrogen gas from boiling out when heated by UV radiation from young stars.

“Finding such a common property is a long-term goal still to be achieved,” says McConnachie. “But we can already quote a useful upper limit on the number of surviving dwarf galaxies.” Figure 1 shows some of the 22 dwarf galaxies discovered in Andromeda’s halo by PANDAS and by earlier observations. All are at least 100 kly from the galaxy’s center. Dwarf galaxies much closer in would be unlikely to survive the host’s tidal forces. In any case, such faint objects become increasingly hard to discern against background glare as one approaches Andromeda’s dense central population.

Beyond 100 kly, the abundance of dwarf galaxies per square degree of sky appears to remain roughly constant as a function of radial distance, out to the 500-kly survey limit. By assuming that the density remains constant all the way out to the theoretical maximum halo radius of about 1 Mly for a galaxy of Andromeda’s estimated mass, one gets an upper limit of 88 ± 20 for the total number of dwarf galaxies—an order of magnitude less than one would expect from the dark-matter simulations if one ignored the complex issue of actual star formation.

To map Andromeda’s actual dark-matter profile, one wants to study orbits of objects trapped in the halo. To that end, PANDAS has been using the spectrographic capability of the 10-m Keck II telescope on Mauna Kea to determine Doppler shifts of dwarf galaxies and of individual stars in the tidal structures and the general halo population. That exercise also gives information about what fraction of the dwarf galaxies have the high-eccentricity “plunging orbits” that lead to early tidal disaster. The spread of stellar Doppler shifts within a given dwarf galaxy is also a clue to its total mass—baryonic plus dark.

Plans are afoot for fitting the 8-m Subaru telescope on Mauna Kea with a camera whose field of view is twice that of the CFHT’s Megacam. “With such a survey camera on the bigger telescope,” says McConnachie, “we could map Andromeda’s halo with much greater sensitivity.”