Last year, scientists reported the coalescence of two astronomical objects known as neutron stars1. The event, called GW170817, produced gravitational waves, which had weakened to a faint ‘chirp’ by the time they reached us. In addition, some of the matter in the neutron stars was ejected into space. Moments later, this matter was hit by a powerful jet of material from the merged stars, resulting in a roaring outburst of radiation at all wavelengths2. However, despite a flood of data, the process by which this radiation was generated has not been certain. In a paper in Nature, Mooley et al.3 report that GW170817 still whispers to us in radio waves. These signals suggest that the observed radiation came from a relatively slow-moving ‘cocoon’ of matter that was energized by the jet.
The 1993 and 2017 physics Nobel prizes were awarded for the indirect4,5 and direct6 detection of gravitational waves, respectively. These studies concerned systems that can be well described using only Einstein’s theory of general relativity. But astrophysics is rarely so simple. For instance, when two neutron stars merge, they produce fireworks — they deform, splash, explode and radiate. Consequently, all the complexities of fluid dynamics, magnetic fields, nuclear reactions, particle acceleration and radiation come into play. Astronomers cannot create and tune experiments, but must make do with the messy ones performed by nature.
What astronomers can do, however, is take advantage of two of the biggest revolutions in the field since the invention of the telescope. First, in the twentieth century, astronomy became multi-wavelength: we can now detect radiation across the electromagnetic spectrum (from radio waves to γ-rays). Second, in this century, it became multi-messenger: we can now detect a broad range of emissions — from high-energy cosmic rays and neutrinos to gravitational waves. The discovery of GW170817 demonstrated the full potential of these advances for the first time.
After being alerted to the gravitational-wave signal, astronomers used just about every type of telescope available to try to view the event. As a result, a wide variety of data was obtained, potentially providing enough information to pin down a complete picture of what physically happened when the neutron stars merged. In particular, NASA’s Fermi Gamma-ray Space Telescope detected a flash of γ-rays that had formed within two seconds of the merger7. The properties of the flash were consistent with a γ-ray burst (a cosmic explosion long thought to be related to neutron-star mergers), which immediately increased interest in GW170817. However, the exact cause of the γ-ray emission became a matter of debate.
Standard γ-ray bursts can be produced only by a jet — an outflow of material moving at a speed at least 99.9% that of light. But the burst from GW170817 was about 10,000 times weaker than these bursts and seen only because it occurred relatively close to us7. Such a weak burst could have come from an off-axis jet (one that was aimed away from us), which would allow only the tiny fraction of light that it emitted sideways to be observed. But it could also have been produced by a comparatively slow-moving cocoon of matter, perhaps travelling at ‘only’ 95% of the speed of light (Fig. 1).
The initial papers2,8 concluded that both scenarios are possible, and that additional data should allow us to identify which one is correct. Mooley et al. now fulfil this promise. They show that although the outburst of radiation from GW170817 is dying down, the intensity of its radio emission is rising — a finding that is difficult to explain using the relatively simple jet models that they consider. The presence of an off-axis jet that breaks free of the surrounding material is not completely excluded, but the cocoon model is more consistent with the observational data.
Establishing the origin of the electromagnetic emission from GW170817 is key to gaining a detailed understanding of the relationship between gravitational-wave events, neutron-star mergers and γ-ray bursts. If a consideration of a greater number of jet and cocoon models than that of Mooley et al., and high-quality simulations of these models, support the authors’ conclusions, nature will once again have shown us that the range of phenomena possible is wider than our simplest thinking suggested. If the cocoon model is correct, this probably implies that many more gravitational-wave events have associated γ-ray bursts than was previously thought.
However, Mooley and colleagues’ explanation for the burst does not affect our basic understanding of what happens in a neutron-star merger — all of the models considered by the authors have great commonality. For instance, the merged stars are always surrounded by matter that is both inflowing (in the equatorial plane of the merged stars) and outflowing (in all other directions). And a faster and narrower outflow is always driven into this matter along the rotational axis of the merged stars.
The differences are in the precise outcome of the attendant fluid dynamics. How much mass is contained in slow parts of the surrounding matter, and how fast and narrow is the jet? Does the jet break out of the surrounding matter so that it can be seen by us, or does it share its energy with this material, producing a relatively slow and broad explosion?
The next few years will see fierce efforts to address these questions. Gravitational-wave events will be under surveillance by an army of telescopes, to find or exclude a fast jet. And elaborate computer simulations will be used to try to determine what we should expect to happen in neutron-star mergers from a theoretical standpoint. The next handful of well-observed events will bring us much closer to the answers.
Nature 554, 178-179 (2018)
Abbott, B. P. et al. (LIGO Scientific Collaboration and Virgo Collaboration) Phys. Rev. Lett. 119, 161101 (2017).
Abbott, B. P. et al. Astrophys. J. 848, L12 (2017).
Mooley, K. P. et al. Nature 554, 207–210 (2018).
Hulse, R. A. & Taylor, J. H. Astrophys. J. 195, L51–L53 (1975).
Taylor, J. H. & Weisberg, J. M. Astrophys. J. 253, 908–920 (1982).
Abbott, B. P. et al. (LIGO Scientific Collaboration and Virgo Collaboration) Phys. Rev. Lett. 116, 061102 (2016).
Goldstein, A. et al. Astrophys. J. 848, L14 (2017).
Kasliwal, M. M. et al. Science 358, 1559–1565 (2017).