Massive star formation not so different after all?

Reconstructed image from near-IR interferometric observations of IRAS 13481-6124 using VLTI/AMBER

ResearchBlogging.orgIn my previous post on the Zooniverse Project IX I’m involved in, I talked about the importance of star formation in the Universe and some of the difficulties we face in studying it. Some big unanswered question particularly remain in our understanding of how massive stars form. Fittingly, the latest edition of Nature has a paper on a nice result in the study of massive star formation: a detection by direct imaging of an accretion disk around a massive young star.

Massive stars form in far smaller numbers than the regular run-of-the-mill stars, like our Sun. They also have far shorter lifetimes, exploding in cataclysmic supernovae after just millions, rather than billions, of years. Another difference between high and low mass stars lies in the sequence of events in their earliest years: low mass stars stop accreting new material once the stars have “turned on”, and the radiation from the new star blows away the surrounding envelope. High mass stars, however, manage somehow to continue accreting new material even after they’ve started emitting the ionising radiation that will slowly dissipate their natal clouds.

How large amounts of gas can find their way onto the star despite the outward force of the emitted radiation is tough to understand. A paper by Krumholz et al (2009) showed via modelling how non-spherical accretion scenarios, like filaments or disks, could play a role in channeling the accretion flow into the star. Accretion disks are ubiquitous around young low-mass stars, but around massive forming stars they had, despite ample circumstantial evidence, never been directly imaged.

In their Nature paper, Stefan Kraus of UMichigan at Ann Arbor and collaborators from Germany, France and Italy, describe how they combined observations at different wavelengths covering different spatial scales to build a model of the massive young stellar object (YSO) IRAS 13481-6124, which includes a massive central star, a dust disk heated by the central star, and a bipolar collimated outflow of molecular material, which is strongly indicative of ongoing accretion via a disk.

Using the near-infrared imager AMBER on ESO’s VLTI telescope – the interferometric mode offered by the array of 4 VLT unit telescopes and a number of smaller 1.8-m auxiliary telescopes – and the 3.5-m New Technology Telescope, they produced resolved images of 2 distinct components to the source. Combing their data processing and modelling, they identify these as a compact hot dusty disk that is being heated and slowly evaporated by the stellar radiation, surrounded by the extended remainder of the dense envelope from which the new star formed.

This is the second time in a few months that obervations with near-IR interferometry result in a Nature paper (see also the eps Aurigae paper in April) – a welcome sign perhaps that this technique is slowly coming into maturity and starting to produce good science. Interferometry, while long a staple in the radio astronomer’s toolbox, is really tough at shorter optical or infrared wavelengths, often requiring control of aberrations and deformations in the instruments to sub-micrometer scales. Converting the output from interferometric instruments to images or spectra also requires additional processing steps, complicating the data reduction procedures. If you’re interested in how that works, the gory details are described in the Supplementary information at Nature.

Is this the first ever detection of an accretion disk around a massive YSO? I don’t know everything that’s been published on the subject but I don’t think so. Several papers have been published in recent years with other kinds of disk detections that seem pretty conclusive to me, e.g. Davies et al’s paper of last year using integral field spectroscopy, showing clear signs of a rotating disk-like structure around the massive YSO W33A – so we had a good idea that they were there. But as for an actual image of such a disk, well, I guess Kraus et al may now take that prize.

Interestingly, Kraus et al fitted the observed characteristics of the disk with the analytical relations that low-mass YSO disks are observed to follow. They find that the disk of IRAS 13481-6124 is consistent with these same relations. This result suggests that the process of massive star formation is perhaps not as fundamentally different from “regular” low-mass star birth as astronomers thought, and give a nice vote of confidence to our currently accepted models of star formation.

A model that works: isn’t that nice for a change?


Kraus S, Hofmann KH, Menten KM, Schertl D, Weigelt G, Wyrowski F, Meilland A, Perraut K, Petrov R, Robbe-Dubois S, Schilke P, & Testi L (2010). A hot compact dust disk around a massive young stellar object. Nature, 466 (7304), 339-42 PMID: 20631793

Krumholz MR, Klein RI, McKee CF, Offner SS, & Cunningham AJ (2009). The formation of massive star systems by accretion. Science (New York, N.Y.), 323 (5915), 754-7 PMID: 19150809

Davies B et al (2009). The circumstellar disk, envelope, and bi-polar outflow of the Massive Young Stellar Object W33A. MNRAS, Vol. 402, Issue 3, pp. 1504-1515. (arxiv)

Image: Kraus et al (2010)


  1. There have been a number of very strong indications that massive star formation is very much like low-mass star formation over the past few years to a decade, and so it always makes me wince when I hear massive-star researchers say “No-one really knows how massive stars form… but this is how low-mass stars form”. I think we do know quite well how massive stars form, but that would make for a boring introduction to a talk/paper, wouldn’t it?! I think the field now needs to move on from this gimmick.

  2. Paul – yes I do agree with you, the story around massive star formation tends to be very formulaic, like you say – I guess I’m a little guilty of it myself on here. Maybe we should start a facebook group? ;-)

    I don’t think this paper gives us any *fundamentally* new information about (massive) star formation, but it’s neat that they got the image with AMBER. As an instrumentalist I enjoy reading about methods as much as results….

  3. Paul Crowther says:

    Historically, issues relating to accretion have been thought to arise once forming massive protostars develop strong radiative winds, i.e cases well above 20 solar masses. Witnessing the formation of these stars tends to be challenging since they originate within ultra-compact HII regions (far away, deeply embedded) that are forming stellar clusters, similar to the Orion Nebula Cluster. All cases studied so far, including Kraus et al. relate to (atypical) isolated protostars below circa 20 solar masses, so the question should perhaps be rephrased as “do very massive stars form in the same way as low and intermediate mass stars?”

  4. Cheers Paul, that’s good to note. I had wondered why there was no mention of cluster companions – this is more of an intermediate mass protostar, and it’s isolated. So how *did* your 300 solar mass star form?!

  5. Paul Crowther says:

    Dunno, but recent simulations of 30 Dor-type cluster formation suggest central densities of 100,000,000 solar masses per pc^3 which is high enough for collisions/mergers to take place. Alternatively, for a star to accrete 300 solar masses in a few hundred thousand years requires an accretion rate of order 0.001 solar masses per yr that 3D simulations by Krumholz et al. (2008, Science) seem to be able to produce.


  1. […] This post was mentioned on Twitter by Sarah Kendrew, Rafael Martinez. Rafael Martinez said: RT @sarahkendrew: blog post: Massive star formation not so different after all? […]