Fermi-LAT Observations of LIGO/Virgo Event GW170817

We present the Fermi Large Area Telescope (LAT) observations of the binary neutron star merger event GW170817 and the associated short gamma-ray burst (SGRB) GRB 170817A detected by the Fermi Gamma-ray Burst Monitor. The LAT was entering the South Atlantic Anomaly at the time of the LIGO/Virgo trigger (tGW) and therefore cannot place constraints on the existence of high-energy (E > 100 MeV) emission associated with the moment of binary coalescence. We focus instead on constraining high-energy emission on longer timescales. No candidate electromagnetic counterpart was detected by the LAT on timescales of minutes, hours, or days after the LIGO/Virgo detection. The resulting flux upper bound (at 95% C.L.) from the LAT is 4.5 × 10−10 erg cm−2 s−1 in the 0.1–1 GeV range covering a period from tGW + 1153 s to tGW + 2027 s. At the distance of GRB 170817A, this flux upper bound corresponds to a luminosity upper bound of 9.7 × 1043 erg s−1, which is five orders of magnitude less luminous than the only other LAT SGRB with known redshift, GRB 090510. We also discuss the prospects for LAT detection of electromagnetic counterparts to future gravitational-wave events from Advanced LIGO/Virgo in the context of GW170817/GRB 170817A.


Introduction
Short gamma-ray bursts (SGRBs) have long been thought to be associated with the coalescence of binary compact objects, such as members of neutron star-black hole (NS-BH) and neutron star-neutron star (NS-NS) systems (Paczynski 1986;Eichler et al. 1989;Paczynski 1991;Narayan et al. 1992). This connection arises from considerations of their duration (Norris et al. 1984;Kouveliotou et al. 1993), redshift (Berger 2014a), and host galaxy distributions (Troja et al. 2008;D'Avanzo et al. 2009;). The observed absence of associated conventional supernovae (Fox et al. 2005;Hjorth et al. 2005aHjorth et al. , 2005bSoderberg et al. 2006;Kocevski et al. 2010;Berger et al. 2013;Troja et al. 2016) has further supported this paradigm. Deep HST observations of nearby SGRBs has also provided tantalizing evidence for their association to kilonovae (Metzger et al. 2010;Kouveliotou et al. 2012;Tanvir et al. 2013;Yang et al. 2015;Jin et al. 2016), short-lived infrared transients powered by the radioactive energy of the NS merger ejecta.
The merging of neutron star binaries is predicted to also result in the emission of gravitational waves (e.g., Kobayashi & Mészáros 2003 and references therein), making SGRBs promising candidates for joint gravitational wave (GW) and electromagnetic (EM) detections. For example, the observed orbital decay rate of the binary pulsar PSRB1913+16 is consistent with the predictions of general relativity to better than 10 −3 precision, setting the date of their merging as ∼300 million years from now (Taylor et al. 1979); see Weisberg & Huang (2016) for a recent update.
Strong evidence supporting the long-suspected association of SGRBs and compact binary coalescence was provided on 2017 August 17 when the Advanced Laser Interferometer Gravitational-wave Observatory (LIGO; Abbott et al. 2016b) and Advanced Virgo experiments (Virgo Collaboration 2009) triggered on a compact binary merger candidate (Abbott et al. 2017c) coincident in time with a Fermi Gamma-ray Burst Monitor (GBM) detected SGRB, GRB 170817A (Connaughton et al. 2017;Goldstein et al. 2017), which was also detected by INTEGRAL SPI-ACS (Savchenko et al. 2017a(Savchenko et al. , 2017b. The joint GW-EM detection has provided the first direct observational evidence connecting SGRBs to neutron star merger events, and has ushered in an exciting era of multi-messenger astronomy. Because gravitational waves probe the system, binding energy and gamma-ray bursts are driven by the configuration of matter 67 NASA Postdoctoral Program Fellow, USA. 68 Funded by contract FIRB-2012-RBFR12PM1F from the Italian Ministry of Education, University and Research (MIUR).
Original content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. outside the ultimate event horizon, the combination of such information opens up discovery space concerning tidal disruption of infalling neutron stars, and associated accretion disk and jet formation. It perhaps also offers the prospect for measuring the mass-to-radius ratio and equation of state of a participating neutron star prior to destruction. Shortly after the trigger, the detection of an optical counterpart in the elliptical galaxy NGC 4993 was announced  (Abbott et al. 2017a;Coulter et al. 2017aCoulter et al. , 2017b at the position (R.A., decl.=197°.450354, −23°.381484, J2000). We will assume these coordinates as the location of GW170817 for the remainder of the paper.
The Fermi Large Area Telescope (LAT) unfortunately was not collecting data at the time of the LIGO/Virgo and GBM triggers, due to a passage through the South Atlantic Anomaly (SAA), and was thus unable to observe the prompt emission phase of the GRB. The LAT resumed collecting science data ∼10 3 s later, so we focus instead on constraining high-energy emission on longer timescales. We use these limits and the estimated distance of 42.9±3.2 Mpc to NGC 4993 (Abbott et al. 2017b) to set upper limits on the luminosity of the latetime emission of GRB 170817A above 100 MeV.
We describe the details of the data analysis in Section 2, compare GRB 170817A to other LAT-detected SGRBs in Section 3, discuss the prospects of detecting SGRBs associated with future GW triggers in Section 4, and conclude in Section 5.

LAT Observations of GW170817
The Fermi Gamma-ray Space Telescope consists of two primary science instruments: the GBM (Meegan et al. 2009) and the LAT (Atwood et al. 2009). The GBM comprises 14 scintillation detectors designed to study the gamma-ray sky in the energy band of ∼8 keV-40 MeV. The LAT is a pair conversion telescope consisting of a 4×4 array of silicon strip trackers and tungsten converters together with Cesium Iodide (CsI) calorimeters covered by a segmented anticoincidence detector to reject charged-particle background events. The LAT is sensitive in an energy range covering 20MeV to more than 300GeV, with a field of view (FOV) of 2.4sr, observing the entire sky every two orbits (∼3 hr) by rocking north and south about the orbital plane on alternate orbits (Atwood et al. 2009). The LAT detects roughly 15 GRBs per year above 100 MeV, of which ∼1-2 are SGRBs, with localization precisions of ∼10 arcmin (Vianello et al. 2017). The high-energy emission associated with SGRBs is substantially longer in duration with respect to their keV-MeV emission as observed by the GBM, having been detected on timescales of >100 s in two burst sources, and has been proposed to be related to the afterglow phase of the burst (Kumar & Barniol Duran 2009;Ackermann et al. 2013bAckermann et al. , 2014Kouveliotou et al. 2013). The LAT is currently the only instrument that has detected and localized long-lived, high-energy emission from SGRBs, and can substantially reduce the localization uncertainties with respect to GBM, aiding follow-up at other wavelengths.
Fermi-LAT was entering the SAA at the time of the LIGO/ Virgo trigger (t GW =2017 August 17 12:41:04.444 UTC). During SAA passages, the LAT and the GBM do not collect data due to the high-charged-particle background in this region. Because of the higher susceptibility of the LAT to the charged particles in this region, the SAA boundary employed by the LAT encompasses a ∼14% larger area than the boundary used by the GBM, resulting in slightly different times at which the two instruments do not collect data. The GBM and LAT SAA boundaries are illustrated in Figure 1. At the time of the GBM Figure 1. Position of Fermi at the trigger time of GRB 170817A (green dot) and its orbital path from west to east. The dark and light red regions define the boundaries of the SAA for the GBM and LAT instruments, respectively. Both instruments do not collect data inside their respective SAA boundaries, due to an elevated chargedparticle background.
trigger (t EM =2017 August 17 12:41:06.474598 UTC), the centroid of the final LIGO/Virgo LALInference map (Veitch et al. 2015;The LIGO Scientific and Virgo Collaboration et al. 2017) using data from all three gravitational-wave observatories (H1, L1, and V1) was located at R.A.=197°.25, decl.=−25°.62 (J2000), or Galactic l=307°.9, b=37°.1. This position was θ∼90°from the LAT boresight and outside the nominal θ<65°LAT FOV. The LAT resumed data taking upon exiting the SAA at t GW +1153 s. At that time, the entire 90% credible region of the LALInference map was within the LAT FOV, and the region subsequently exited at t GW +2027 s. Figure 2 shows the sky coverage of the LAT at t GW + 1153 s, when the entire localization region was observed.

Constraints on the High-energy, Gamma-ray Flux of GRB 170817A
We searched the LAT data for a high-energy, gamma-ray counterpart on different timescales before and after the trigger time, and we computed upper bounds on its flux using an unbinned likelihood analysis described in further detail in Ackermann et al. (2016), Racusin et al. (2017), and Vianello et al. (2017). For all the analyses presented here, we used the P8_TRANSIENT010E_V6 events class and the corresponding instrument response functions, and the Fermi Science Tools version v10r0p5. 69 We furthermore assume a flat ΛCDM cosmology with H 0 =67.8 kms −1 Mpc −1 , Ω Λ =0.692 and Ω m =0.308 (Planck Collaboration et al. 2016), and a distance to the host of GW170817, NGC 4993, of 42.9±3.2 Mpc (Abbott et al. 2017b), i.e., z≈0.0098. For the computation of the upper bounds on gamma-ray flux and luminosity, we assume a power-law spectrum with a photon index of −2.
We first performed a search for a transient counterpart within the 90% contour of the final LIGO/Virgo LALInference map in the time window from t GW +1153 s to t GW +2027 s, the earliest interval in which the region was observed by the LAT, and no significant new sources were found. At the position of the optical counterpart (Abbott et al. 2017a;Coulter et al. 2017a) the value for the flux upper bound over this interval and in the 0.1-1 GeV energy range is 4.5×10 −10 erg cm −2 s −1 (95% confidence level), corresponding to an equivalent isotropic luminosity of approximately 9.7×10 43 erg s −1 .
For typical GRBs where the viewing angle is smaller than the jet opening angle, the afterglow is coincident with the end of the prompt emission (Berger 2014a; D'Avanzo 2015), while for jets viewed off-axis, the onset of the afterglows is predicted to be of the order of few days up to 100 days van Eerten & MacFadyen 2012). As reported in Troja et al. (2017a;, an X-ray source positionally coincident with the optical transient was detected at 8.9 days after the GW event, followed by a radio source detection Mooley et al. (2017), suggesting the detection of an afterglow from a possible off-axis jet (Abbott et al. 2017a).
Regardless of the origin of this emission, we monitored the source by performing a likelihood analysis in every interval of time after the trigger when the source was in the LAT FOV. The values of the flux upper bounds vary due to differences in exposures, as shown in Figure 3. In a time period spanning from t EM − 1 day to t EM +45 days, these limits range between 9.7×10 −11 to 3.7×10 −8 erg cm −2 s −1 , corresponding to a luminosity range of 2.1×10 43 to 8.1×10 45 erg s −1 (0.1-1 GeV).
We also examined intervals spanning years before the trigger time. During normal operations, the LAT surveys the entire sky continuously, and has observed the position of GRB 170817A every ∼3 hr since 2008, for a total of ∼55 Ms of exposure. A search for a counterpart using data collected over the entire lifetime of the mission yielded no detection, returning an upper bound for the average 9 years flux of 1.32×10 −12 erg cm −2 s −1 (0.1-1 GeV), corresponding to a luminosity limit of 2.9× 10 41 erg s −1 .
We have also looked into the results of the diverse and complementary automatic techniques developed by the Fermi-LAT team to continually search for transient events over a variety of timescales. For example, the Fermi Autonomous Science Processing (ASP) pipeline performs a search on six hour, one day, and one week timescales to identify candidate gamma-ray sources (Chiang et al. 2007). The candidate sources are subsequently reviewed through a more rigorous likelihood analysis by the Flare Advocate/Gamma-ray Sky Watcher (FA-GSW) pipeline, and positionally cross-checked with known cataloged gamma-ray objects. ASP allows for the detection of flux variations of known cataloged sources, as well as the detection of new unassociated gamma-ray transients. Similar to ASP, the Fermi All-sky Variability Analysis (FAVA; Ackermann et al. 2013a) searches for transients over 24 hr and one week timescales. 70 FAVA is a blind photometric technique that looks for deviations from the expected flux in a grid of regions covering the entire sky. The observed long-term, mission-averaged emission serves as reference for the expected flux, allowing the FAVA technique to be independent of any model of the gammaray sky. We searched through the full-mission data set for transients detected by ASP and FAVA, positionally consistent with the optical counterpart. No sources with significance greater than 5σ were found. The closest ASP transient is a known gamma-ray source, 3FGL J1312.7−2349, that is associated with the blazar NVSS J131248−235046, roughly 2 • from the optical counterpart and therefore unrelated.

GRB 170817A in the Context of Other LAT Detected SGRBs
We can compare the properties of GRB 170817A in the context of other GBM and LAT detected bursts. The GBM observations reported in Goldstein et al. (2017) show that the gamma-ray emission from GRB 170817A was softer than that of typical SGRBs, with a peak in its νF ν spectrum of E peak =215±54 keV, a value falling in the lowest ∼15th percentile of the SGRB distribution. The burst fluence as measured in the 10-1000 keV energy range is dimmer than typical long bursts, but is consistent with those obtained for other GBM-detected SGRBs.
The GBM has detected over 2000 GRBs in over 9 years of science operations, of which ∼8% have been detected at energies greater than 30 MeV by the LAT. Roughly 17% of the GBM population is made up of SGRBs, with durations  (Bhat et al. 2016) and duration of the GBM and LAT detected populations can be seen in Figure 4(a). Fluences of SGRBs are generally lower than for long bursts, owing to their shorter durations. The LAT has detected 14 SGRBs with durations shorter than GRB 170817A, eight with comparably low fluence. Figure 4(a) shows the GBM fluence of the GBM and LAT detected populations plotted versus the angle at which the burst occurred with respect to the LAT boresight at the time of the GBM trigger (T0). Because the LAT sensitivity decreases significantly as a function of increasing boresight angle, the lowest fluence bursts are only detected by the LAT when they occur close to the instrument boresight. The location of the optical counterpart associated with GRB170817A occurred ∼90°away from the LAT boresight at the time of the GBM trigger. Eight other LAT detected bursts have fluence values as low as GRB170817A, but a majority of them occurred within θ<25°of the LAT boresight, where the instrument is most sensitive. LAT bursts with boresight angles greater than θ>65°w ere either detected using the LAT Low Energy, which provides an additional sensitivity to off-axis photons, or were brought into the FOV through an automatic re-point request of the spacecraft initiated by the GBM for high peak flux.
The LAT detects 1-2 SGRBs per year in the 100MeV-100GeV energy range (Ackermann et al. 2013c). Their highenergy emission lasts longer than the prompt emission observed by the GBM, up to ∼200 s after T 0 for the brightest cases. In order to compare GRB 170817A with other LAT detected SGRBs, we compute the flux upper bound in time interval t GW +1153 s to t GW +2027 s in the 100MeV-100GeV energy range, obtaining 2×10 −9 ergcm −2 s −1 (corresponding to an isotropic luminosity of 4.3 × 10 44 erg s −1 ). The flux upper bound value is above the expected flux at this time from an extrapolation of the power-law temporal decay of the brightest observed SGRB detected by the LAT to date, GRB 090510 (Ackermann et al. 2010;Razzaque 2010). Therefore, the lack of a detection of GRB 170817A at >1000 s is consistent with previous non-detections at this time. We can compare the inferred luminosity upper bound of 4.3× 10 44 erg s −1 for GRB170817A with the luminosity of the extended emission for GRB090510, also the only LAT-detected SGRB with known redshift (z=0.903). SGRBs in the LAT energy range are typically characterized by a power-law spectrum ∝E −2 and a flux decaying as ∝t −1 (Ackermann et al. 2013c). By extrapolating the late-time light curve of GRB090510(Ackermann et al. 2013c), we estimate the high-energy emission at T 0 +1153s in its source frame to be roughly 2×10 48 erg s −1 . This effectively rules out late-time emission from GRB170817A as luminous as GRB090510.
The lack of such emission is not surprising, though, given that the prompt isotropic equivalent energy observed in the GBM for GRB 170817A of E 3.0 0.6 10 iso 46 = ´erg ) is six orders of magnitude lower than the value estimated for GRB090510. GRB090510 radiated an isotropic equivalent energy of approximately 1.1×10 53 erg during its prompt emission observed by GBM, and an additional E iso =5.5×10 52 erg in the high-energy extended emission detected by the LAT (Ackermann et al. 2013c). This indicates that the reason for the underluminous nature of GRB170817A must also affect the energetics of the component responsible for the high-energy emission observed by the LAT. Purposed explanations include a relativistic jet viewed off-axis of up to 56°or with an intrinsically low Lorentz factor (Abbott et al. 2017b), a mildly relativistic shock breakout of a cocoon associated with the merger ejecta (Gottlieb et al. 2017;Kasliwal et al. 2017), or the high velocity tail of the neutron rich material dynamically ejected during the merger (Mooley et al. 2018). If we assume a similar ratio between the prompt and extended E iso for GRB170817A as was determined for GRB090510 (∼50%), then we would expect a luminosity for the extended emission for GRB170817A at t GW +1153 s to be on the order of 5×10 41 erg s −1 , well within our estimated luminosity upper bound.

Prospects for Future LAT Detections and Sensitivity Study
We now proceed to characterize the sensitivity of the LAT to SGRBs in general. This sensitivity is a function of the characteristics of the LAT, of its background, and of our search methodology. The background level and, consequently, the sensitivity varies across the sky, in particular as a function of Galactic latitude because of the bright diffuse emission associated with the plane of the Milky Way. We first present the computation of the sensitivity for the optical position of GRB170817A, which corresponds to a middle Galactic latitude b=39°.296. We will then show how the sensitivity changes for typical positions in the sky at lower and higher Galactic latitude. As with all frequentist statistical tests, our method is characterized by the probability of false positives (Type I error, α) and the probability of false negatives (Type-II error, β). We consider an SGRB detected when it has a significance of at least 5σ (TS>25), i.e., we fix α=2.86×10 −7 . Following Kashyap et al. (2010), we can then compute the minimum flux that a point source must have in order to be detected above the 5σ level with a given Type-II error probability. We fix for this computation β=0.5. We assume an exposure of 100s, which is roughly the longest duration over which an SGRB has been detected by the LAT in the observer frame. We also assume, for simplicity, that the GRB is close to the axis of the LAT, where the effective area is maximized, and that the pointing is not changing with time. We use as background the Galactic and the Isotropic template provided by the LAT collaboration, 71 representing, respectively, the diffuse emission coming from the Milky Way and the isotropic component generated by particles misclassified as photons and by unresolved sources. We also include all known gamma-ray sources from the 3FGL catalog (Acero et al. 2015) as part of the background. We then simulate repeatedly a point source with a spectrum dN/dE∝E −2 , and vary the flux (averaged over the observation) until 50% of the realizations of the simulated source are detected above 5σ. We ascertain that an average flux of F s =9.5×10 −9 ergcm −2 s in the 0.1-100 GeV energy range is required to detect a source at mid-latitude with a type II error probability of β=50% in a 100 s observation. The required flux increases (i.e., the sensitivity decreases) for a source on the Galactic plane by a factor of ∼2-3 (depending on the Galactic longitude), and decreases (i.e., the sensitivity increases) by a factor of ∼2 for a source at the Galactic poles, where the background is lower.
The starting time of a GRB observation is critical because the GRB flux fades rapidly, typically ∝t −1 . Therefore, we can detect fainter SGRBs the earlier we start observing. In Figure 5, we show how the sensitivity of the LAT for sources at mid-Galactic latitudes changes for five different starting times. The first four shaded regions are for observations with a duration of 100s, starting, respectively, at T 0 , T 0 +2 s, T 0 +10 s, T 0 +100 s, while the last is between T 0 +1153 s and T 0 +2027 s after the trigger time as for GRB170817A. For reference, we also report the measurements for other SGRBs detected by the LAT, as well as the upper bound for GRB170817A in the 100MeV-100GeV energy range. Among the sample of SGRBs detected by the LAT, we note that the fluences of GRB081024B and GRB140402A measured by the GBM in the 10 keV-1 MeV energy band are similar to the one of GRB170817A. GRB090510 is the brightest SGRB detected by the LAT so far and resulted in the detection of both its prompt and extended emission. The much dimmer GRB130804A, on the other hand, was in the field of view at the time of the trigger, but was only detected by the LAT at T 0 ∼200 s, constituting an example of delayed highenergy emission. The statistics are limited, but we can conclude that the LAT needs to start observing a source within 100-200 s to have a chance at detecting even the brightest of the LATdetected SGRBs.
Because the LAT detection efficiency for SGRBs appears to decrease after 100-200 s, we next estimate the probability that the LAT will observe an SGRB within 100 s of the trigger time during normal survey mode. The Fermi-GBM observes ∼65% of the sky, with the rest being occulted by the Earth. The LAT observed ∼35% of SGRBs detected by the GBM within ∼100 s of the trigger. Hence, a continuation of this survey strategy indicates that on average, the LAT can observe (either a detection or upper bound) ∼23% of the full-sky SGRB population within 100 s from their GBM trigger. While the population of SGRBs is isotropically distributed, the population of SGRBs within the LIGO/Virgo horizon is likely not. Indeed, the expected horizon for LIGO/Virgo at design sensitivity to BH-NS and NS-NS mergers is estimated to be ∼190Mpc (Abbott et al. 2017b), and the distribution of galaxies that can potentially host an SGRB is not isotropic within such volume (Bilicki et al. 2014). However, we have performed simulations that showed that such anisotropy does not appreciably affect our observation rate.
The LAT detects ∼5% of all GBM-detected SGRBs. If we assume the LAT will have the same efficiency for GRB/GW triggers and a rate of joint GBM/GW events of 1 (2) per year, we obtain at most a ∼5% (∼10%) probability of detecting one or more GRB/GW with the LAT in one year, respectively. This assumes that GRB/GW events will be representative of the entire GBM-detected SGRB population when observed in gamma-rays. We also computed the sensitivity curve for an observation between 1153 and 2027 s, as for GRB170817A. This is also extrapolated back in time according to a t −1 afterglow decay law (dashed gray line).
Two important factors pertaining to the detectability of such event by LAT are as follows. (i) Off-axis viewing. Since GRB jets propagate at ultra-relativistic speeds with bulk Lorentz factor Γ100, the prompt and early afterglow high-energy emission is narrowly beamed with a mean beaming fraction f 1 cos 0.015 b 0  q á ñ = á -ñ corresponding to a mean jet opening angle 10 0  q á ñ  (e.g., Berger 2014b). The collimation of SGRBs into narrow jets implies that most GW detections of NS-NS or NS-BH mergers would correspond to jets that do not point toward us, and are instead viewed off-axis, at rather large angles θ obs from the jet symmetry axis. For such off-axis viewing angles the peak afterglow flux at any given time drops rapidly with increasing θ obs (Nakar et al. 2002), which makes their detection challenging. The brightness of the off-axis afterglow emission strongly depends on the jet's angular structure, which is not well known for SGRBs (or even for LGRBs for that matter; e.g., Granot 2007). Moreover this brightness should differ between the GBM and LAT wavebands. (ii) Proximity. However, given the current sensitivity of the LIGO/Virgo GW detectors, any SGRBs that are observed in coincidence with GW triggers in the future must be relatively nearby with z<0.1 (Abbott et al. 2016a), which improves their detectability (since for a given isotropic equivalent luminosity along our line of sight, F ν ∝D −2 where D is the source distance). These two effects work in the opposite directions and a careful study is needed to determine the likelihood of such future detections, accounting for D, f b , and the jet's angular structure.

Summary
We present the Fermi-LAT observations of the first confirmed LIGO/Virgo binary neutron star merger event GW170817 and the associated SGRB GRB170817A. Because the LAT was entering the SAA at the time of the LIGO/Virgo trigger, we cannot place constraints on the high-energy (E>100 MeV) emission associated with the moment of binary coalescence. Instead we focus on constraining highenergy emission on longer timescales and report no candidate electromagnetic counterpart above >100 MeV on timescales of minutes, hours, and days after t GW . The resulting flux upper bound from the LAT is 4.5×10 −10 erg cm −2 s −1 in the 0.1-1 GeV range (and 2×10 −9 erg cm −2 s −1 in the 0.1-100 GeV energy range) covering the time interval t GW +1153 s to t GW +2027 s. This limit is above the expected flux at this time from an extrapolation of the power-law temporal decay of the brightest observed SGRB detected by the LAT to date, GRB090510, which is also the only LAT SGRB with a measured redshift and much more distant (z=0.903). The upper bound on the flux of GRB170817A corresponds to a luminosity upper bound in the 0.1-1 GeV energy range of 9.7×10 43 erg s −1 (or 4.3 × 10 44 erg s −1 in the 0.1-100 GeV energy range), significantly lower than the luminosity for GRB090510 obtained by extrapolating its light curve to 1153 s after the trigger time (4 × 10 49 erg s −1 ). This effectively rules out emission for GRB170817A as luminous as GRB090510. However, this is not surprising as the total energy output (E iso ) during the prompt emission, as measured by the GBM, was 6 orders of magnitude lower for GRB170817A than for GRB090510. This might indicate an intrinsically less energetic event or a jet observed off-axis instead of on-axis, or both.
We also looked for a possible high-energy signal in the ∼55 Ms of exposure that the LAT has accumulated at the position of GRB170817A over the life of the Fermi mission. We do not detect any signal from the host galaxy over the entire mission, placing a luminosity upper bound of 2.9×10 41 erg s −1 (0.1-100 GeV). We also did not detect any signal on timescales of minutes, hours or days in the ∼9 years before the trigger time.
The LAT has detected SGRBs with GBM fluences that are comparable to GRB170817A. Therefore, it is possible for the LAT to detect GRBs characterized by a low-energy prompt emission as faint as GRB170817A, provided that they have a ratio between low-energy prompt emission and high-energy extended emission similar to previously detected bursts and are observed relatively close to the boresight of the instrument. We have estimated the sensitivity of the LAT to SGRBs and compared it to these previous observations, concluding that the LAT would need to begin observations within ∼100 s from the trigger time.
Finally, under the current observation strategy the LAT should observe ∼23% of the population of SGRBs within 100 s from the trigger, providing meaningful constraints. At the current LAT detection efficiency of ∼5% of all GBM-detected SGRBs (Ackermann et al. 2013c), we estimate that a joint GRB/GW rate of 1 (2) per year would result in at most a ∼5% (∼10%) probability of LAT detecting one or more GRB/GW in one year. This assumes that GRB/GW events are similar to the previously detected population of SGRBs when observed in gamma-rays.
The detection of gravitational waves and an associated EM signal from the coalescence of compact binary system has initiated a new phase in high-energy astrophysics. The discovery of GW170817/GRB170817A strongly supports the long-suspected association of SGRBs with merging neutron stars. Multi-messenger study of future events can verify this and should impact our understanding of many astrophysical processes-general relativistic dynamics, the equation of state of cold nuclear matter, relativistic jet formation by compact objects, particle acceleration, radioactive heating of the expanding debris (as in kilonovae) and cosmography. We already know that the observations depend strongly upon the orientation of the axes of the binary merger and/or the jet, which can now be determined, at least in principle. Other factors, in particular the initial neutron star masses, spins and magnetization and the environment may also play a role but ought not to be very important. Presuming that there will be an improvement in LIGO-Virgo sensitivity which results in a much larger detection rate and given what we already know from GBM and LAT, there is every reason to believe that gamma-ray observations will drive future developments in this exciting field.
We dedicate this paper to our late friend and colleague, Neil Gehrels, whose curiosity and enthusiasm inspired so much of our work.
The Fermi LAT Collaboration acknowledges generous ongoing support from a number of agencies and institutes that have supported both the development and the operation of the LAT as well as scientific data analysis. These include the National Aeronautics and Space Administration and the Department of Energy in the United States, the Commissariat à l'Energie Atomique and the Centre National de la Recherche Scientifique/Institut National de Physique Nucléaire et de