A large simulation reveals that most of the detectable signal from dark matter in our Milky Way probably comes from the main, smooth Galactic halo, rather than from small clumps.
Most of the mass of the Universe is believed to be in the form of dark matter — an invisible component that has so far been only indirectly detected through the effects of its gravity on visible matter. In the theory of supersymmetry in particle physics, there is a corresponding dark-matter-particle candidate that interacts only very weakly with the rest of the Universe, and is thus very difficult to detect directly. There is, however, a general feeling in the astronomical community that the search for dark matter is now at a turning point. This feeling stems from the recent start of the largest particle accelerator in the world (the Large Hadron Collider), which could provide clues about the nature of dark matter, and from the advent of high-energy astrophysics observations, such as γ-ray observations carried out by NASA's Fermi Gamma-ray Space Telescope. Such observations are potentially able to detect dark-matter particles indirectly through their annihilation radiation. On page 73 of this issue, Springel et al.1 show that the primary and probably most easily observable annihilation signal is produced by the diffuse dark-matter component rather than the very small clumps in the main halo of our Galaxy (Fig. 1).
The challenges in determining the nature of dark matter are not only experimental. At a time when observations are about to start providing data, it is necessary to understand in detail how dark matter is distributed in our neighbourhood, in particular in the halo surrounding our Galaxy (an extended, ellipsoid-shaped dark-matter structure), in order to make predictions about the expected annihilation signal. During the past few years, there has been controversy about the nature of the clustering of dark matter inside galactic haloes, and particularly the mass and distance from our Solar System of the dark-matter structures that are likely to contribute to the annihilation signal that could be measured by the high-energy astrophysics experiments.
There are two ways to predict the properties of dark matter inside galactic haloes. The first involves simplifying the geometry of the problem, and making predictions using relatively simple but robust analytical calculations2. Although these calculations are rigorous and free from any numerical artefacts, the oversimplification of the geometry can lead to questionable results. The second approach, used by Springel et al.1, involves performing sophisticated numerical experiments on supercomputers.
Springel et al. study the dynamics of dark matter in a cosmological background — the expanding Universe — by modelling the dark-matter distribution with a set of macroparticles that interact which each other only through gravitational forces. Each macroparticle represents a huge number of actual dark-matter particles. Because the gravitational force is very long-range in nature, the authors simulate a large volume of the Universe and zoom in on a region where a halo similar to that of our Galaxy is formed. In that smaller region, the resolution of the simulation is increased, enabling many macroparticles of smaller mass to trace all the fine details of dark-matter dynamics.
Springel and colleagues' simulations are developed in the framework of the cold dark matter (CDM) hypothesis, which is now the commonly accepted model for the formation of large-scale structures in the Universe. One of the hurdles to performing simulations in CDM models is achieving numerical convergence at small scales, or equivalently at small masses. Within the CDM hypothesis, the consensus is that the smallest dark-matter structures formed have sizes comparable to that of the Solar System and masses equivalent to that of Earth3. These small structures (substructures) would then have merged together to form larger ones, and so on, forming a full hierarchy of structures within structures. The largest structures correspond to haloes of rich clusters of galaxies.
So the question is whether or not dark-matter simulations have enough resolution to resolve the smallest structures. The bigger the number of dark-matter particles used in the simulations, the larger the number of substructures detected. But at what stage can we be sure that numerical convergence is achieved?
Springel et al. answer this question to a large extent4 by identifying and tracing dark-matter structures and substructures in a very robust way. To achieve that end, they perform several simulations with various resolutions — that is, with different numbers of particles, but with the same initial configuration. They are then able to cross-identify the substructures found in the different simulations and perform a quantitative, unprecedented convergence study of the fine details in the distribution of dark matter in our Galactic halo.
They conclude that, in fact, the main contribution for indirect dark-matter detection should come from the smooth component of the halo of our Galaxy instead of its substructures, at variance with some earlier analyses5,6. If these results are confirmed, astronomers should take them into account in future analyses of γ-ray observations, particularly when trying to disentangle the contribution of dark matter from that of other γ-ray sources, such as those found in the plane of our Milky Way. The debate about the nature and small-scale distribution of dark matter remains open. At the very least, however, Springel and colleagues have made a great advance in the field of computational cosmology.
Springel, V. et al. Nature 456, 73–76 (2008).
Mohayaee, R. & Salati, P. Preprint at http://lanl.arxiv.org/abs/0801.3271 (2008).
Diemand, J., Moore, B. & Stadel, J. Nature 433, 389–391 (2005).
Springel, V. et al. Preprint at http://lanl.arxiv.org/abs/0809.0898 (2008).
Calcáneo-Roldán, C. & Moore, B. Phys. Rev. D 62, 123005 (2000).
Berezinsky, V., Dokuchaev, V. & Eroshenko, Y. Phys. Rev. D 68, 103003 (2003).
About this article
Astronomy & Astrophysics (2011)
The Astrophysical Journal (2009)