Ten thousand light years away, a monster is stirring and a storm is brewing. The 'monster' is a swirling cloud of gas and dust with a mass thousands of times greater than the Sun's. The 'storm' is the birth of massive stars from this cloud — within a million years a great outpouring of light and winds from these stars will be sweeping through space, followed eventually by the cataclysm of supernovae. Understanding this phenomenon is important not only for its effects on our Galaxy today, but also because a cluster of Sun-like stars is expected to form alongside these massive stars and we think our own Solar System was born in such an environment1. Writing in Astronomy & Astrophysics, Peretto et al.2 present new observations of this dark cloud, SDC335 (Fig. 1), including early data from the Atacama Large Millimeter Array (ALMA) located high in the Chilean Andes, and attempt to answer a basic question: how do massive stars form?

Figure 1: Collapsing cloud.
figure 1

NASA/JPL-CALTECH /UNIV. MANCHESTER

This infrared image of the SDC335 dark cloud was taken with the Spitzer telescope. Peretto et al.2 find two massive gas cores (dotted box) near the cloud centre, coinciding with infrared sources, which are likely to be forming massive stars. A web of surrounding filaments (dashed lines) is contracting towards the centre, providing clues to how these cores and stars are forming.

Full size image

Stars exist with a wide variety of masses, from about one-tenth that of the Sun to at least 100 times more massive. Although we have a well-developed understanding of the formation of low-mass, Sun-like stars3, how massive ones are assembled has been shrouded in mystery. Only one out of every few hundred stars is massive. This rarity means that, in terms of stars undergoing formation, the closest examples of massive stars are much more distant than those of low-mass stars — typically thousands of light years away rather than hundreds. Furthermore, massive-star birth seems to require high densities of interstellar gas, and dust grains that are mixed throughout this gas — galactic smog — obscure our view in the optical and even infrared wavebands. To peer into the birth clouds of massive stars, astronomers have had to develop telescopes that see fine detail at longer (far-infrared and radio) wavelengths. Owing to the effect of Earth's atmosphere, sometimes these telescopes need to be sent into space, for example the Spitzer and Herschel telescopes, or located at high altitude, like ALMA.

There are several theoretical ideas about how massive stars form. One, known as core accretion4,5, involves an interstellar gas cloud, compressed under its own self-gravity, fragmenting into many 'cores' that have a range of masses. Low-mass stars grow from low-mass cores, massive stars from massive cores. Once a core starts collapsing to form a star, there is relatively limited further accumulation of gas from the surrounding cloud to the core. We think low-mass stars form this way, on the basis of detailed observations of low-mass cores6. A second idea, competitive accretion7, involves all stars starting to form from low-mass cores but then competing for feeding of additional gas from the surrounding cloud. A star's ultimate mass is determined by its environment: stars grow massive if they are at the centres of clouds undergoing rapid global collapse, which maintains a rich supply of gas to their location. Moreover, massive stars in the process of forming should be seen surrounded by a cluster of lower-mass stars, which eventually comes to dominate the mass of the system and directs the flow of global collapse.

Peretto and colleagues' study bears upon this debate. First, from millimetre-wavelength observations with ALMA of thermal emissions from dust, they derive properties of two massive cores near the cloud centre. These cores are among the most extreme in terms of the concentration of a large mass, of several hundred solar masses, within a small, roughly spherical volume, of radius less than 6,000 astronomical units (AU), where 1 AU is the Earth–Sun distance. If confirmed, these would be very unusual conditions for an interstellar cloud, with 10% of its total mass concentrated in just 0.001% of its volume. However, as the authors describe, mass estimates from millimetre-wavelength dust emission are uncertain, at least by factors of a few. In addition, the small core sizes are not well resolved, but are instead inferred from a slight broadening of their angular profiles compared with the narrowest profile the telescope achieves of a point source. Nevertheless, it seems likely that massive, dense cores have been found that are capable of forming massive stars.

Second, the authors use infrared and far-infrared data from the Spitzer and Herschel telescopes to study the larger-scale cloud. It seems to be threaded by several filaments of gas and dust that extend radially from the centre. Peretto and colleagues argue that this is suggestive of a pattern of global feeding of the cores from the surrounding cloud. To search for motions associated with such feeding, Peretto et al. examine spectral-line emission from several molecular tracers, using both ALMA and the Australian Mopra telescope. They detect small Doppler shifts in frequency caused by the motion of these molecules either towards or away from us. From this information, they infer that the cloud is turbulent and also contracting, presumably owing to gravity. The maximum 'free-fall' rate of such collapse occurs if there is no resistance from internal pressure. However, the observed flow along the filaments implies a collapse rate of only about one-tenth of free fall.

Peretto et al. conclude that the observation of global collapse supports the theory of competitive accretion of massive-star formation. Although the evidence for collapse is compelling, this conclusion may be premature. Global collapse alone cannot be used to discriminate between core and competitive accretion, because both involve self-gravitating clouds that must have assembled in this way. Rather, it is the rate of collapse that is likely to be important in determining whether gas that forms massive stars has a chance to first become organized in massive cores or instead is delivered quickly to a central region and there apportioned among a forming cluster, as yet unseen in SDC335. An alternative possible interpretation of the observations is that global collapse is relatively slow and that massive cores have formed and achieved densities much larger than their surroundings, from which they are then effectively decoupled. Observations with a fully operational ALMA will see at least ten times finer detail and will be needed to resolve this debate decisively.