Detection of the trace neutral fraction of hydrogen gas that stretches between the nearby Andromeda and Triangulum galaxies has allowed resolved spectral imaging of this elusive intergalactic medium. See Letter p.224
The historical concept that galaxies are 'island universes' that formed in the distant past and have since evolved in isolation from their surroundings has long been recognized as inadequate. That model was replaced by the hierarchical growth scenario1,2, which highlights the role of galaxy interactions, collisions and mergers in shaping the mature galaxies that we see today. But as Wolfe et al.3 report on page 224 of this issue, it could be argued that another model is now taking precedence, in which a diffuse 'cosmic web' between the discrete galaxies is the dominant reservoir for continued galaxy growth and evolution.
The impetus for looking further than the hierarchical growth scenario to better explain galaxy evolution has come from simple observational evidence. Over the past decade, it has become clear that the rate at which stars are forming in the Universe has varied markedly with time, first increasing to a peak some 10 billion years ago and then declining more than tenfold to the modest levels of today4,5. At the same time, the total amount of hydrogen in the atomic phase, which forms the basic fuel for star formation, has declined by only about half6,7. How can that be? To compound the mystery, measurements of the capacity of atomic-gas reservoirs and the rates of hydrogen-gas consumption by galaxies have shown that, surprisingly, the timescale for complete gas depletion is typically rather short, only a few billion years8. This is true for a wide range of galaxy types9, including spiral galaxies such as our own, which seem to have been consuming gas at a relatively constant rate over a large fraction of the Universe's 13.8-billion-year history10.
The answer to this mystery apparently lies on our Galactic doorstep, having been revealed by observations of our two nearest major galaxies — the Andromeda and Triangulum galaxies. The gas reservoir that fuels star formation through cosmic time may reside not in the galaxies themselves, but rather in their extended surroundings. What is more, in contrast to what has previously been assumed, this reservoir may not be in the form of atomic hydrogen at all, but instead in the form of ionized hydrogen. Numerical simulations of increasing precision11,12 suggest that, at the current cosmic epoch, about one-third of all baryonic (ordinary) matter in the Universe has been processed into the condensed form we know as galaxies, with two-thirds still residing in filaments of ionized gas in the intergalactic medium (Fig. 1). It is the gradual condensation of this ionized gas that seems to dominate the process of galactic-mass assembly, rather than the episodic interaction and merger of condensed (proto-) galaxies.
Although this solution to the galaxy-assembly enigma is appealing in its simplicity, it has proved surprisingly difficult to verify experimentally. The difficulty arises because diffuse ionized hydrogen is almost invisible. Regions of ionized hydrogen emit radiation called thermal Bremsstrahlung and recombination radiation, which has a brightness that is proportional to the square of the gas density. Although such emission from the high-density regions of massive star formation in our own and other galaxies is extremely bright, once the density has declined by a factor of 105 (the levels that apply to the intergalactic regime) the brightness has fallen by about 1010 and become almost undetectable. The first heroic efforts at detecting this medium directly are only beginning to achieve some, albeit limited, success13.
An alternative strategy, which Wolfe et al. employ, makes use of the fact that the amount of hydrogen ionization is never 100% under intergalactic conditions. Even in the case in which the hydrogen gas is smoothly distributed, so that it is illuminated by ionizing radiation from all directions, there will be a small but finite neutral fraction14 that can be as much as 1%. The neutral fraction would be further enhanced under conditions of gas clumping, which would provide directional shielding against the ionizing radiation. Wolfe and colleagues' strategy consists of mapping the 21-centimetre emission of the trace neutral fraction to allow resolved spectral imaging of this elusive medium between the Andromeda and Triangulum galaxies.
Although 21-cm emission has been used extensively to document the atomic-gas content and kinematics of galaxies in the nearby Universe15, this has normally been undertaken in the high-density, fully neutral gas regime. Extending such studies to the low-density, highly ionized realm, in which atomic densities are a factor of 104 or more lower, is extremely challenging. What makes this approach feasible at all is the fact that the brightness of 21-cm emission varies only linearly with gas density, rather than as the square of it that applies to the ionized emission tracers. Even so, the required precision for its detection is at the very limits of what is possible with current technology.
Wolfe and colleagues' confirmation of the first detection16 of 21-cm radiation in this regime represents a milestone in the application of this approach to cosmic-web imaging. Their demonstration opens the door for broader application of the method to document the location and kinematics of the densest filaments of ionized gas in the environments of other nearby galaxies. Such imaging should give a clear picture of the evolutionary state of individual galaxies and galaxy groups. It should also provide evidence for feedback processes, such as galactic outflows, that may have shaped the galaxies' broader environment.
However, the pinnacle of this research endeavour — confirmation of the total mass of gas in the intergalactic volume — will require further forms of observation to supplement the 21-cm imaging. Also required is determination of the amount of ionization, so that the total mass can be deduced. Although it is unlikely that high precision would be achieved in such a combined observation, it would still be a major advance in overcoming our current ignorance about two-thirds of the baryonic matter in the cosmos.
Davis, M. et al. Astrophys. J. 292, 371–394 (1985).
Stewart, K. R., Bullock, J. S., Wechsler, R. H., Maller, A. H. & Zentner, A. R. Astrophys. J. 683, 597–610 (2008).
Wolfe, S. A., Pisano, D. J., Lockman, F. J., McGaugh, S. S. & Shaya, E. J. Nature 497, 224–226 (2013).
Madau, P., Ferguson, H. C. & Dickinson, M. E. Mon. Not. R. Astron. Soc. 283, 1388–1404 (1996).
Hopkins, A. M. & Beacom, J. F. Astrophys. J. 651, 142–154 (2006).
Noterdaeme, P., Petitjean, P., Ledoux, C. & Srianand, R. Astron. Astrophys. 505, 1087–1098 (2009).
Braun, R. Astrophys. J. 749, 87 (2012).
Wong, T. & Blitz, L. Astrophys. J. 569, 157–183 (2002).
Bigiel, F. et al. Astrophys. J. 730, L13 (2011).
Ade, P. A. R. et al. Preprint at http://arxiv.org/abs/1303.5076 (2013).
Davé, R. et al. Astrophys. J. 552, 473–483 (2001).
Kereš, D., Katz, N., Fardal, M., Davé, R. & Weinberg, D. H. Mon. Not. R. Astron. Soc. 395, 160–179 (2009).
Simionescu, A. et al. Science 331, 1576–1579 (2011).
Dove, J. B. & Shull, J. M. Astrophys. J. 423, 196–206 (1994).
Walter, F. et al. Astron. J. 136, 2563–2647 (2008).
Braun, R. & Thilker, D. A. Astron. Astrophys. 417, 421–435 (2004).
Vogelsberger, M., Sijacki, D., Kereš, D., Springel, V. & Hernquist, L. Mon. Not. R. Astron. Soc. 425, 3024–3057 (2012).