On the wrong side of 3 sigma

The claim was made just over ten years ago, that we could see matter in the grip of the supermassive black hole at the center of our galaxy. This black hole is called Sagittarius A*, or Sgr A* for short. A 2003 paper in The Astrophysical Journal (preprint on arXiv) by Fred Baganoff and colleagues analyzed images from NASA's Chandra X-ray Observatory and showed that there was a slightly extended X-ray source coinciding with the position of the black hole. They argued that the best explanation for this extended source was matter caught in the gravitational hold of the black hole and falling inwards. Some of this matter would likely make the ultimate one-way trip and fall over the black hole’s event horizon, never to be seen again.

The central region of the Milky Way galaxy. The large image contains X-ray from the Chandra X-ray Observatory (blue) and infrared data from the Hubble Space Telecope (red and yellow). The inset shows a close-up in X-rays only, covering a region only half a light year wide centered on Sgr A*, the supermassive black hole at the center of our galaxy. Credit: X-ray: NASA/UMass/D.Wang et al., IR: NASA/STScI.

Jump forward 8 years and - based partly on their 2003 paper - Baganoff and team are awarded a 3 million second long observation with Chandra, to make a much more detailed study of Sgr A*. In their proposal they argued that this large dataset would tell them how matter falls towards the black hole and how much and where some of it flows back. This is one of the largest Chandra programs ever, so the Time Assignment Committee obviously thought the science was compelling.

However, before even a single photon was obtained in this program, a paper appeared with a different spin on the X-ray source at the center of the galaxy. Sergey Sazonov and co-authors argued that a large population of volatile stars with masses less than the Sun may be orbiting around the black hole. If they were spun up by interactions with other stars in their crowded neighborhood, these young stars could be very active, producing flares and copious amounts of X-rays. Sazonov et al. suggested these X-rays could produce much of the extended emission seen near Sgr A*. Baganoff et al. had briefly mentioned a similar idea in their paper but pointed out there is no evidence for such a population at other wavelengths, including infrared and radio data.

Sazonov et al. explained that X-ray data by itself could help decide between the two competing arguments. The emission of X-rays at a specific energy would support the volatile star idea, because this signal is seen in young stars located nearby, but shouldn't be seen when matter falls onto black holes. Also, giant flares should regularly be seen in the extended X-ray source, because similar events are also observed in nearby stars.

Their paper even provided support for the first piece of evidence based on Chandra data. A hint of X-ray emission was seen at exactly the predicted energy - 6.4 keV - with a significance of “~3 sigma” (actually 2.75 sigma). They mentioned that the long observation, coming in 2012, could help confirm this possible detection.

Figure 7 from Sazonov et al. (2012) showing a Chandra spectrum of the extended source (within 1.5 arcseconds) of Sgr A*. Note the small amount of X-ray emission possibly detected at 6.4 keV, a hint that the X-ray source is dominated by emission from low-mass stars rather than hot gas captured by the black hole. The figure is taken from the arXiv version of the paper.

This was an intriguing result, and it was tempting to do a press release on the paper. It provided a different take on a familiar object, with possible evidence for a cocoon of interacting, overactive stars around our black hole that hadn't been detected by any other observatory. It would also imply that the X-ray emission generated by material falling towards the black hole was even fainter than previously thought, deepening a well-known mystery. I didn't know much about the first author, but the second author, Rashid Sunyaev, is one of the most outstanding astrophysicists in the world (here’s an interview I did with him in 2012) and I was familiar with some of the fine work done by Mikhail Revnivtsev, the third author.

However, the strength of the possible signal was on the wrong side of 3 sigma, meaning that it was slightly less significant than a commonly used threshold for evidence (3 sigma) and much less significant than the threshold for a discovery (5 sigma). Even significance levels this high are not an absolute guarantee. For example, the result claiming faster than light speed for neutrinos involved an apparent detection at 6 sigma. But that claim famously proved to be wrong.

There was more to the Sazonov paper than this hint of a signal at ~3 sigma, so a release would have been reasonable, but I thought it would have to be carefully worded with some speculation. Also, we knew that a much deeper observation was already scheduled and that doesn't happen too often. So, it seemed best to wait for more data. It's good that we did, because a definitive answer did come from the 3 million second observation. The apparent signal seen before at ~3 sigma did not survive and flaring from the extended source was not seen, as described in this Science paper (arXiv) by Q. Daniel Wang and collaborators, including Fred Baganoff. We did a press release on the paper and Wang wrote a blog post giving more details.

Figure S.3 from Wang et al. (2013) showing the X-ray spectrum of the point source corresponding to Sgr A* (black) and the extended X-ray source around it (red; 2-5 arcseconds annulus). Neither of these shows significant evidence for X-ray emission at 6.4 keV, ruling out the hints reported in Sazonov et al. (2012). The figure is taken from the arXiv version of the paper. 

The basic claim made in 2003 by Baganoff et al. was confirmed. The picture of a disk of hot gas surrounding the black hole is described by the artist's impression shown below. Despite the reputation black holes have for engulfing everything that's nearby, less than 1% of the material that is captured by the black hole ends up being pulled across the event horizon, and the rest is expelled in an outflow.

By holding back we avoided having to do a correction. Sometimes it's good to be patient.

This artist's illustration shows the environment around Sgr A*. The red disk shows hot gas that has been captured by the black hole and is being pulled inwards. The source of the hot gas is young, massive stars, shown in blue, orbiting around Sgr A*. The illustration also shows a large amount of material being thrown outwards, a key factor in explaining why there is so little radiation from material near Sgr A*. Credit: NASA/CXC/M.Weiss