High-energy astrophysics and the search for sources of gravitational waves

Abstract

The dawn of the gravitational-wave (GW) era has sparked a greatly renewed interest into possible links between sources of high-energy radiation and GWs. The most luminous high-energy sources—gamma-ray bursts (GRBs)—have long been considered as very likely sources of GWs, particularly from short-duration GRBs, which are thought to originate from the merger of two compact objects such as binary neutron stars and a neutron star–black hole binary. In this paper, we discuss: (i) the high-energy emission from short-duration GRBs; (ii) what other sources of high-energy radiation may be observed from binary mergers; and (iii) how searches for high-energy electromagnetic counterparts to GW events are performed with current space facilities. While current high-energy facilities, such as Swift and Fermi, play a crucial role in the search for electromagnetic counterparts, new space missions will greatly enhance our capabilities for joint observations. We discuss why such facilities, which incorporate new technology that enables very wide-field X-ray imaging, are required if we are to truly exploit the multi-messenger era.

This article is part of a discussion meeting issue ‘The promises of gravitational-wave astronomy’.

1. Introduction

The direct observation of gravitational-waves (GWs) has opened a new era in observational astrophysics [1,2]. The first GW sources detected were binary black hole systems. Although thought less likely than binary neutron-star mergers to produce a strong electromagnetic signal, the first GW events resulted in a large observational effort to search for an electromagnetic counterpart. A significant part of this effort involved the search for a high-energy photon signal via either a joint trigger with a wide-field observatory or a more targeted search for a longer-lived electromagnetic source. For the first binary black hole system detected, a possible near-simultaneous high-energy signal was claimed [3], although it remains unclear whether this was just the chance detection of an unrelated source.

Many lessons were learned from the initial electromagnetic searches during the first run (O1) of the Advanced LIGO era, and these led to a significantly improved capability which was deployed during the second (O2) run which ended in August 2017. In this brief review, we discuss the possible sources of high-energy radiation that may be associated with GW events, the observational techniques currently deployed to search for an electromagnetic counterpart, and look forward to the new generation of high-energy observing facilities which will make multi-messenger detections much more likely.

In §2, we outline the high-energy observations of gamma-ray bursts (GRBs), followed by a discussion of other possible high-energy emission components in §3. In §4, we summarize how searches for GW counterparts have proceeded and then in §5 several future missions designed to enhance the electromagnetic search and characterization capability are described. The first jointly observed GW and electromagnetic binary neutron-star merger, GW170817, was announced close to the completion of this paper. The implications of this event on future requirements are discussed in §6, followed by the conclusions in §7.

2. High-energy observations of gamma-ray bursts

One of the most likely electromagnetic counterparts expected for a GW event is a GRB. Most GRBs detected are the so-called long GRBs. It is standard to refer to a GRB as long if it has a duration of more than 2 s for the period over which gamma rays are detected—the so-called T90 period over which 90% of the gamma rays were detected. Short GRBs are likewise those with a T90 of less than 2 s [4]. This is not an ideal definition, being both waveband and detector dependent, and examples exist of GRBs which challenge the traditional long/short divide [5]. The luminosity of GRBs strongly implies that a highly relativistic outflow is present in the form of a narrow jet with an opening angles of a few to a few tens of degrees powered by a compact central engine. The ‘prompt emission’ comes from the jet, dominated by either internal shock or magnetic recombination processes, followed by the so-called afterglow emission as the jet interacts with the surrounding material. The formation of a jet is required for both long and short GRBs, but the progenitor may differ [5].

Thanks primarily to the Swift satellite [6], we now have a much better appreciation of the properties of short-duration GRBs, and there are differences in the general properties of long and short bursts which argue for at least two types of progenitor.

Long GRBs are associated with the death of massive stars—the collapsar model—whereas short GRBs differ in lacking clear supernova signatures in the optical/IR and in having a wider diversity in host galaxy properties [5]. The long-standing proposition that short GRBs are due to binary mergers has (so far) survived, but it is certainly not as well evidenced as the collapsar model for long GRBs. Indeed, the joint detection of GWs may provide the crucial evidence.

Swift has revealed intriguing features in the light curves of both long and short GRBs which point to central engine activity in both cases far beyond the T90 duration. Periods of flaring activity, changes in the rate of flux decay and indeed periods where the flux does not decay are observed [5]. A longer duration for central activity is important for the detection of both on- and off-axis emission from GW electromagnetic counterparts in all wavebands, as a longer-lived engine may power longer-lasting emission, raising the chances of joint detections at later times in addition to a joint prompt detection.

An example of a short GRB light curve, GRB 090515, is shown in figure 1 using data from the Swift BAT and XRT instruments [7]. This burst was faint and very short in gamma rays, yet very bright in X-rays, detected once Swift had slewed to point its narrow-field X-ray telescope. However, this bright X-ray emission lasted only a short while before rapidly fading. It has been suggested that such a GRB could have resulted from a binary merger of two neutron stars which initially formed a magnetar which later collapsed to a black hole [8]. Other short GRBs show a more gradual decline in X-rays [9], suggesting possible multiple routes to the final compact central engine configuration. Whatever the engine or the route, the complex nature of such light curves illustrates how joint electromagnetic and GW observations could provide powerful constraints, if we can jointly observe a population of objects.

Figure 1.

The Swift 10 keV light curve for GRB 090515. The early (black) points are derived from the BAT and the later (blue and red points, which are detections and upper limits, respectively) are from the XRT. The lower panel shows the photon index. (Online version in colour.)

3. Other high-energy emission components

Detection of a simultaneous on-axis event which triggers a high-energy facility would provide a wealth of information. However, such events will probably be rare given the implied GRB jet opening angles and the volume of space currently probed by the GW facilities. Thus, other more isotropic emission components must also be considered, which would be detectable following GW events seen off-axis. This is a complex problem as relatively little is known about the process by which a jet emerges from the merger, how the jet interacts with its surroundings and what other emission may come from material left over after the merger. Here, we briefly discuss some examples to illustrate the range of high-energy emission that may be detectable.

(a). Fallback accretion

It was noted by Kisaka et al. [10] that the kilonova [11] candidate GRB 130603B had excess X-ray emission in addition to the IR emission possibly associated with a kilonova [12]. This excess X-ray emission could be due to fallback accretion which continues to power the central engine. The resultant high-energy emission could then be further reprocessed by ejecta around the centre (left over after the merger) into longer-wavelength thermal emission and hence contribute to the observed IR excess attributed to a kilonova. The required ejecta masses are reasonable (10⁻² to 10⁻³ solar masses) for a binary neutron-star merger. The model from Kisaka et al. [10] implies an X-ray luminosity of 10⁴¹ erg s⁻¹ which would be readily detectable from within the current GW search volume by a facility such as Swift. If fallback accretion does contribute, we may see sources with a contribution from both kilonova and fallback accretion. Distinguishing between such emission processes will require detailed modelling of the light curves and/or undertaking IR spectroscopy to search for discrete emission features in the spectrum expected from a kilonova.

(b). Magnetar and ejecta

The possibility of additional energy from a longer-lived central engine has led to a number of predictions, which are complicated by a lack of knowledge of both the amount and geometry of the ejecta around a binary merger. In [13], it is argued that the observed emission will depend on the angle at which the observer views the merger (something that may be derivable from a GW signal if the signal-to-noise ratio is sufficient). In [13], the central engine is a magnetar which emits an isotropic wind. When viewed down the jet, the observer sees high-energy emission from a GRB. When viewed slightly off-axis, the observer may see into what Sun et al. [13] call the ‘free zone’, and see high-energy emission from magnetar wind internal dissipation. Based on existing observations, Sun et al. [13] suggest this zone includes off-axis angles of at most a few times the opening angle of the jet. Further off-axis, towards the equatorial zone of the merger, the observer’s direct view will be blocked by the ejecta (the ‘trapped zone’), where the observed soft X-ray emission is likely to be reduced in luminosity. In practice, the ejecta may be clumpy, so the viewing angle dependence may be more complex, but this model well illustrates how additional information from the GW signal may help disentangle the geometry and emission processes.

(c). Cocoon emission

As a jet emerges from a merger, it will interact with its surroundings, so even if the jet is not viewed directly it can have an observable effect on the emission seen off-axis. One scenario is of a high-pressure cocoon around a jet [14], formed as the jet emerges through a baryon-contaminated region around the merger, with additional non-relativistic shocked material. This is a very complex situation to model, but Lazzati et al. [14] argue that the presence of cocoon material may result in a prompt, short-duration (few seconds) X-ray flash which could be detectable by high-energy facilities. The predicted spectrum, such as the peak photon energy, depends on the viewing angle. As the observer views more off-axis from the jet, the predicted peak photon energy falls. Thus, the observed luminosity in the bandpass of a gamma-ray instrument, such as the Fermi/GBM, would fall, although still detectable at off-axis angles of a few tens of degrees. This type of model favours—for off-axis events—observing in the X-ray band (few keV) if sufficient sky coverage can be achieved, as discussed below.

4. Searching for high-energy counterparts to gravitational-wave events

There are two basic methods in use to find a high-energy counterpart to a GW event: (1) a joint prompt detection from a wide-field instrument, a method currently used to find GRBs as a real-time trigger or in ground processing; and (2) using a narrow-field instrument to search for a transient high-energy source within the GW skymap region.

Method (1) is used for several instruments, including Fermi/GBM, Swift/BAT, Integral/SPI-ACS, ASTROSAT/CZTI and HAWC. These have large sky area (e.g. Fermi/BAT sees approx. 70% of the sky instantaneously), and are sensitive to an on-axis GRB within the GW search volume (and indeed can search for lower-luminosity events within that volume than that of the known GRB population). The detection can occur either on board or during ground processing. The availability of the GW trigger time allows for the use of a lower threshold around that time than that used for normal triggers, enabling searches for fainter signals with acceptable (low false-alarm rate) significance. This method was used during the LVC O1 run to identify a low-fluence high-energy event by the Fermi/BAT team [3] that was coincident in time with GW150914. It is unclear whether the events are physically related as the GBM detection false-alarm significance is less than 3σ [3]. Joint triggers have also been searched for during the LVC O2 run.

Given the large GW skymaps (can cover several hundred square degrees), method (2) is only practicable for an initial search of the GW skymap when using the Swift satellite, which is designed to enable a large number (hundreds) of slews per day. The Swift team have evolved the search method over the LVC O1 and O2 runs [15–17], but the principle is similar. The satellite is commanded to point at individual locations on the sky for about 60 s each within the LVC skymap region. During the O1 run, each GW skymap was convolved with the GWGC galaxy catalogue [18] and the resultant galaxy list was then ranked by brightness and probability. During the O2 run, the 2MPZ galaxy catalogue [19] was used together with additional GW distance information provided by the LVC [20].

Figure 2 shows the actual sky coverage achieved by the Swift/XRT during the follow-up of GW150914 [16]. The coverage is relatively low as the GW trigger was issued late and the trigger came before the official start of the O1 science run and hence the Swift system was not fully operational. Only a relatively low number of spacecraft pointings could be carried out, and, as the skymap overlapped the Large Magellanic Cloud, that was chosen to concentrate on. Subsequent Swift system testing used the GW150914 skymap for reference, and figure 3 shows the dramatic improvement in coverage that was achieved. In the test, Swift observed 426 fields in 24 h, covering some 50% of the galaxy-weighted skymap. This type of coverage was achieved during the LVC O2 run in the first 2 days following the GW triggers.

Figure 2.

Swift observations of LVC150914 (red points) superimposed on the LIGO skymap. Black areas show desired pointings based on convolution with the galaxy catalogue as described in the text.

Figure 3.

Test Swift observations of LVC150914 (green points) superimposed on the LIGO skymap. Black areas show the LIGO error region and green points show the observations actually performed during the test run in 24 h.

Following the search for a fading transient, the Swift strategy also allows for additional, longer observations (500 s) after a few days to search for off-axis emission, which may not emerge until later [15,17]. Either the initial or subsequent search pattern can be interrupted if a promising electromagnetic counterpart candidate is identified by Swift or another facility.

5. Future high-energy missions

The current high-energy missions have provided important data in the hunt for an electromagnetic signal from GW events. But, to maintain capability, to be more efficient and to improve the rate of detecting a counterpart requires the construction of a new generation of facilities. In this section, we describe some of the facilities under construction and some of those proposed for the longer term.

New facilities could include technology similar to that in current wide-angle detectors, such as the Fermi/GBM instrument, which uses scintillator technology, or the Swift/BAT coded-mask instrument. A GBM-like device is rugged and has the advantage of providing a very wide field of view (roughly half-sky in low Earth orbit) at the cost of moderate localization accuracy (a few degrees at best). Coded-mask instruments can provide a few arcminute location accuracy with good sky coverage, but, as with scintillators, have limited sensitivity compared to using focusing optics as in an X-ray telescope. To improve the localization accuracy and sensitivity, while providing a large field of view, requires a different technology.

The facilities discussed below have a common characteristic: they use microchannel plate optics (MPOs), incorporating glass plates with micropores etched in them (usually a few tens of micrometres in size) which are coated to reflect X-rays. These are often referred to as ‘lobster’ optics as they resemble the structure of the eyes of a lobster which have evolved square pores to provide a wide-field optical view.

Lobster optics provide an affordable and low-mass solution to the provision of wide-field, focusing X-ray imaging systems. The assembly of a large number of plates side by side creates an optic without a single optical axis. Hence, light can be focused from a wide range of directions simultaneously. In principle, the entire sky could be viewed by the appropriate deployment of multiple lobster modules. In practice, the field of view is limited by budget and instrument accommodation issues. We note that MPOs can also be used to provide a lighter narrow-field X-ray telescope.

(a). Space-based variable object monitor

The Swift satellite has been a remarkable success since its launch in 2004. The availability of an orbiting observatory with on-board multi-wavelength capability, coupled with rapid response to new transients discovered on board or loaded as a new observing sequence, has transformed many areas of observational astronomy. To continue and expand this legacy requires a new facility both to complement Swift and to continue its work. China and France are currently constructing such a facility, in collaboration with partners in the UK (Leicester) and Germany (MPE), called SVOM (space-based variable object monitor).

Due for launch in late 2021, SVOM (PIs: Jianyan Wei/NAOC and Bertrand Cordier/CEA) consists of a rapid response (few minutes) spacecraft carrying four instruments: (i) the ECLAIRS coded-mask telescope, which locates transients in the 4–150 keV band over 2 sr; (ii) the GRM scintillator instrument, providing higher-energy spectra (15–5000 keV) over 2 sr field of view; (iii) the MXT X-ray (0.2–10 keV) telescope that uses MPOs and a CCD detector to provide a 1 degree field of view for X-ray follow-up of transients and to search for GW counterparts; and (iv) the VT visual telescope with a 25×25 arcminute field of view and two channels (blue and red) covering the 0.4–1 μm band. Unlike Swift, the SVOM project also has several dedicated ground-based optical/IR telescopes. Communicating via a VHF network, SVOM will rapidly provide targets for other facilities as well as searching for counterparts to multi-messenger signals.

(b). Einstein Probe

The Einstein Probe (EP) is a Chinese mission (PI: Weimin Yuan/NAOC) due for launch in 2022 (figure 4). It is a revolutionary concept consisting of 12 wide-field lobster modules providing a total field of view of 3600 square degrees in the soft X-ray band (0.5–5 keV). This wide-field X-ray telescope exceeds current fields of view by more than three orders of magnitude. Carried on a rapid-response spacecraft, EP will be able to monitor the sky on both short and long time scales, and observe new transients using an on-board narrow-field (30 arcminute) follow-up X-ray telescope (0.3–10 keV). Placed into low Earth orbit, EP will be able to observe the entire available sky every few orbits (several times per day), providing a unique capability in the era of LSST, CTA, SKA, ELT and other multi-messenger facilities.

Figure 4.

Schematic of the Einstein Probe mission (Weimin Yuan/NAOC) showing the 12 WFT modules and the FXT on a rapid-response spacecraft. (Online version in colour.)

Based on the projected GW error regions to be provided in the early 2020s, EP will be able to cover an entire GW skymap in very few pointings. The X-ray focusing optics also provide a sensitive real-time trigger capability for those events which go off in the field of view. Although EP has a smaller field of view than scintillator instruments, such as the Fermi/GBM or SVOM/GRM, it has the advantage of being both sensitive and able to follow-up in the same waveband as the discovery.

(c). International Space Station-Transient Astrophysical Observatory

While SVOM and Einstein Probe are under construction, other facilities are also being studied to provide additional and longer-term capability. The International Space Station-Transient Astrophysical Observatory (ISS-TAO) is planned to have an X-ray wide-field imager (0.3–6 keV), consisting of a lobster module mounted on a pivot arm attached to the ISS, along with a gamma-ray transient monitor (50 keV–1 MeV). With such a flexible mount and a WFI field of view of approximately 400 square degrees, ISS-TAO would be able to quickly ‘tile the sky’ in a search for a GW counterpart. ISS-TAO is currently under study by NASA (PI: Jordan Camp/GSFC), and if selected could fly in 2022. While smaller than the Einstein Probe, ISS-TAO would provide invaluable complementary capability in a different orbit and hence increase the sky fraction accessible for prompt follow-up.

(d). Transient High-Energy Sky and Early Universe Surveyor

The Transient High-Energy Sky and Early Universe Surveyor (THESEUS) (PI: Lorenzo Amati/INAF) is a candidate mission in the ESA M5 call (figure 5), scheduled for launch in the late 2020s. THESEUS consists of a rapid-response spacecraft carrying three instruments: (i) the wide-field Soft X-ray Imager (0.3–5 keV), consisting of four MPO modules providing a total field of view of 3200 square degrees; (ii) the X-Gamma rays Imaging Spectrometer (2 keV–20 MeV) using a set of three coded-mask cameras providing a field of view of 7800 square degrees, including complete overlap with the SXI field of view; and (iii) a 0.7 m IR telescope (0.7–1.8 μm) with a 10×10 arcminute field of view with both imaging and spectroscopic capability (resolutions of 20 or 500).

Figure 5.

Schematic of the THESEUS mission showing the four SXI modules, the central IRT and the three XGIS modules on a rapid-response spacecraft. (Online version in colour.)

THESEUS would monitor the sky over a broad area and over a broad energy range, with on-board spectral capability to classify transients and provide redshift estimates. The sensitivity of the wide-field instruments on THESEUS would greatly increase the discovery space of high-energy transient phenomena over the entirety of cosmic history.

An example illustrating the large field of view of the THESEUS SXI superimposed on an actual LIGO probability skymap is shown in figure 6. The availability of such a large field of view, sensitive focusing X-ray instrument would enable very rapid follow-up of GW alerts, and other multi-messenger triggers, compared with current facilities.

Figure 6.

Schematic of the field of view of the THESEUS SXI superimposed on the LVC151226 skymap. (Online version in colour.)

(e). Transient Astrophysics Probe

The Transient Astrophysics Probe (TAP) is a NASA concept study (PI: Jordon Camp/GSFC) for a probe-class mission to fly in the late 2020s. The concept study proposal baselines three instruments mounted on a rapid-response spacecraft: (i) a wide-field X-ray imager (0.3–10 keV), consisting of 4–6 MPO modules imaging 500 square degrees each; (ii) an optical/IR telescope (0.7 m diameter, 0.6–2.5 μm, 1 degree field of view), with imaging and low-resolution spectral capability; and (iii) a narrow-field X-ray telescope (0.3–10 keV, 1 square degree field of view) with a single-crystal silicon mirror for follow-up of transients. Like THESEUS, TAP would be able to both locate transients, classify them on board and provide rapid response and downlink.

6. The binary neutron-star merger event, GW170817

The first detection of GWs from a binary neutron-star merger occurred on 17 August 2017 [21]. This detection was accompanied by a short-duration high-energy electromagnetic signal detected by Fermi [22] and Integral [23]. The joint error region remained quite large, but significantly smaller than that from the GW signal alone, permitting a more focused search by other electromagnetic facilities. This led to the detection of an optical–IR counterpart, including emission consistent with a kilonova component, located in the nearby galaxy, NGC 4993 ([24] and references therein). Overall, the electromagnetic data reveal a complex and evolving spectrum extending from the radio to gamma ray.

The detection and characterization of the electromagnetic counterpart to GW170817 would have been faster had there been an early, more precise high-energy localization. This could be provided by having a wide-field focusing optics X-ray instrument, such as those described in §5. Rapid X-ray imaging may also have revealed more detailed characteristics during the first few hours when the source flux decayed below the sensitivity level of the wide-field gamma-ray instruments. Detection and characterization of the optical–IR emission would also have been enhanced by having an on-board IR telescope, such as that which would be provided by the proposed THESEUS or TAP missions.

7. Conclusion

The GW era has brought renewed interest in the search for electromagnateic transients and provided significant challenges given the large sky areas that must be searched for a GW counterpart. The predicted emission from several classes of GW sources, such as binary neutron-star mergers, includes detectable high-energy emission as proven in the case of GW170817. The high-energy domain also provides some significant advantages over other wavebands in terms of immediate coverage, as several existing facilities provide wide-area capability which may result in a joint, prompt detection. Narrow-field X-ray telescopes are less well matched in instantaneous sky area to GW skymaps, and hence require detailed observation planning to maximize the chances of follow-up observations being able to detect an electromagnetic counterpart. The current high-energy search strategies combine the advantages of both wide- and narrow-field facilities. Future facilities will increase capability, particularly in the soft X-ray band, taking advantage of focusing microchannel plate optics which enables wide-field and more sensitivity. Over the next few years, the development of these new facilities will enhance the regular observation of electromagnetic counterparts to GW events.

Data accessibility

The data shown in figure 1 are available at http://www.swift.ac.uk/burst_analyser.

Authors' contributions

P.T.O’B. drafted the text. P.E. edited the text and provided several of the figures. Both authors read and approved the manuscript.

Competing interests

The authors declare they have no competing interests.

Funding

We acknowledge support from the UK Science and Technology Facilities Council and the UK Space Agency which support transient science and high-energy facilities at the University of Leicester.