Multimessenger astronomy with gravitational waves and high
Transcript Of Multimessenger astronomy with gravitational waves and high
Multimessenger astronomy with gravitational waves and high-energy neutrinos
S. Ando, B. Baret, B. Bouhou, E. Chassande-Mottin, A. Kouchner, L. Moscoso, V. Van Elewyck, I. Bartos, S. Ma´rka, Z. M´arka, et al.
To cite this version:
S. Ando, B. Baret, B. Bouhou, E. Chassande-Mottin, A. Kouchner, et al.. Multimessenger astronomy with gravitational waves and high-energy neutrinos. Reviews of Modern Physics, American Physical Society, 2013, 85, pp.1401-1420. <10.1103/RevModPhys.85.1401>.
HAL Id: in2p3-00681933 http://hal.in2p3.fr/in2p3-00681933v2
Submitted on 25 Jul 2013
HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destin´ee au d´epˆot et `a la diffusion de documents scientifiques de niveau recherche, publi´es ou non, ´emanant des ´etablissements d’enseignement et de recherche fran¸cais ou ´etrangers, des laboratoires publics ou priv´es.
Multimessenger astronomy with gravitational waves and high-energy
neutrinos
Shin’ichiro Ando,1 Bruny Baret,2 Imre Bartos,3 Boutayeb Bouhou,2 Eric Chassande-Mottin,2 Alessandra Corsi,4, 5 Irene Di Palma,6 Alexander Dietz,7 Corinne Donzaud,2, 8 David Eichler,9 Chad Finley,10 Dafne Guetta,11, 12 Francis Halzen,13 Gareth Jones,14 Shivaraj Kandhasamy,15 Kei Kotake,16, 17 Antoine Kouchner,2 Vuk Mandic,15 Szabolcs Ma´ rka,3 Zsuzsa Ma´ rka,3 Luciano Moscoso,2, ∗ Maria Alessandra Papa,6 Tsvi Piran,18 Thierry Pradier,19 Gustavo E. Romero,20, 21 Patrick Sutton,14 Eric Thrane,15 Ve´ ronique Van Elewyck,2, † and Eli Waxman22 1Gravitation AstroParticle Physics Amsterdam Institute (GRAPPA), The Netherlands 2AstroParticule et Cosmologie (APC), CNRS: UMR7164-IN2P3-Observatoire de Paris-Universite´ Denis Diderot-Paris VII-CEA: DSM/IRFU, France 3Department of Physics, Columbia University, New York, NY 10027, USA 4Physics Department, George Washington University, Washington, D.C. 20052, USA 5LIGO Laboratory, California Institute of Technology, Pasadena, CA 91125, USA 6Albert-Einstein-Institut, Max-Planck-Institut fu¨ r Gravitationsphysik, D-30167 Hannover, Germany 7Department of Physics and Astronomy of the University of Mississippi, Mississippi 38677-1848, USA 8Universite´ Paris-sud, Orsay, F-91405, France 9Department of Physics, Ben Gurion University, Beer-Sheva 84105, Israel 10Oskar Klein Centre & Dept. of Physics, Stockholm University, SE-10691 Stockholm, Sweden 11Department of Physics and Optical Engineering, ORT Braude, P.O. Box 78, Karmiel, Israel 12INAF-Observatory of Rome, Via Frascati 33, Monteporzio Catone, Italy 13Department of Physics, University of Wisconsin, Madison, WI 53706, USA 14School of Physics and Astronomy, Cardiff University,
2
Cardiff CF24 3AA, United Kingdom 15University of Minnesota, Minneapolis, MN 55455, USA 16Division of Theoretical Astronomy, National Astronomical Observatory of Japan, 2-21-1,Osawa, Mitaka, Tokyo, 181-8588, Japan 17Department of applied physics, Fukuoka University, Jonan, Fukuoka, 814-0189, Japan 18Racah Institute of Physics, Hebrew University of Jerusalem, Jerusalem 91904, Israel 19Universite´ de Strasbourg & Institut Pluridisciplinaire France 20Instituto Argentino de Radioastronomia (IAR, CCT La Plata, CONICET), C.C. No. 5, 1894, Villa Elisa, Buenos Aires, Argentina 21FCAyG, Observatorio de La Plata, Paseo del Bosque s/n, CP 1900 La Plata, Argentina. 22Department of Particle Physics & Astrophysics, The Weizmann Institute of Science, Rehovot 76100, Israel
Hubert
Curien,Strasbourg,
Many of the astrophysical sources and violent phenomena observed in our Universe are potential emitters of gravitational waves and high-energy cosmic radiation, including photons, hadrons, and presumably also neutrinos. Both gravitational waves (GW) and high-energy neutrinos (HEN) are cosmic messengers that may escape much denser media than photons. They travel unaffected over cosmological distances, carrying information from the inner regions of the astrophysical engines from which they are emitted (and from which photons and charged cosmic rays cannot reach us). For the same reasons, such messengers could also reveal new, hidden sources that have not been observed by conventional photon-based astronomy. Coincident observation of GWs and HENs may thus play a critical role in multimessenger astronomy. This is particularly true at the present time owing to the advent of a new generation of dedicated detectors: the neutrino telescopes IceCube at the South Pole and ANTARES in the Mediterranean Sea, as well as the GW interferometers Virgo in Italy and LIGO in the United States. Starting from 2007, several periods of concomitant data taking involving these detectors have been conducted. More joint datasets are expected with the next generation of advanced detectors that are to be operational by 2015, with other detectors, such as KAGRA in Japan, joining in the future. Combining information from these independent detectors can provide original ways of constraining the physical processes driving the sources, and also help confirm the astrophysical origin of a GW or HEN signal in case of coincident observation. Given the complexity of the instruments, a successful joint analysis of this combined GW+HEN observational dataset will be possible only if the expertise and knowledge of the data is shared between the two communities. This review aims at providing an overview of both theoretical and experimental state of the art and perspectives for GW+HEN multimessenger astronomy.
∗ deceased † Corresponding author’s electronic address: [email protected]
3
CONTENTS
I. Introduction
3
II. The science case for multimessenger GW+HEN searches
5
A. Potential emitters of GW and HEN
5
1. Galactic sources: soft gamma-ray repeaters
5
2. Gamma-ray bursts
5
B. Bounds on the GW+HEN time delay
8
III. GW and HEN detection: status and prospects
9
A. Interferometric Gravitational Wave detectors
9
1. Detection principle and state of the art
9
2. Multimessenger strategies
10
B. High-energy neutrino telescopes
11
1. Detection principle and state of the art
11
2. Multimessenger strategies
12
IV. Perspectives for the joint data analysis
13
A. GW data analysis
13
1. Multi-detector coherent analysis and background rejection
14
2. Source localization
14
B. HEN data analysis
14
C. HEN-triggered GW searches
15
D. Baseline search with combined HEN and GW events lists
17
V. Conclusions
18
VI. Acknowledgments
19
References
19
I. INTRODUCTION
High-energy multimessenger astronomy has entered an exciting era with the development and operation of new detectors offering unprecedented opportunities to observe cosmic radiation in the Universe in all its variety. Gamma-ray astronomy has been an illustrative example of the synergy between particle physics and astronomy. Soon after the discovery of the neutral pion, Hayakawa (1952) suggested that the interaction of cosmic rays with the neutral gas in the Galactic Plane should produce a diffuse, extended gamma-ray emission. Almost simultaneously, Hutchinson (1952) calculated the contribution to such an emission from relativistic bremsstrahlung. Prospects for gamma-ray astronomy were set up shortly after by Morrison (1958). The first gamma-ray sources were discovered during the late 1960s and 1970s, but even until the 1990s they were difficult to identify, calling for multi-wavelength observational efforts (see Cheng and Romero (2004) for an overview of the subject). During the past decade, this approach has revealed itself to be fruitful in the identification of several types of sources from MeV to TeV energy scales.
The mechanisms that produce the high-energy radiation have, however, remained elusive, requiring the development of multimessenger techniques and programs that would explore all components of the cosmic radiation (see collective references in Paredes et al. (2007)). In the study of transient sources, which involve compact objects and ultra-violent phenomena (such as gammaray bursts and magnetars), multimessenger techniques are in fact the only approaches that might lead to a full understanding of the underlying processes. In this context, high-energy ( GeV) neutrinos (HENs) and gravitational waves (GWs) could play an important role. These messengers share interesting astronomical properties: HENs can escape from much denser, hence deeper, environments than photons, and GWs propagate virtually freely in any region of space. Moreover, and contrary to high-energy photons (which can be absorbed by intervening photon backgrounds) and charged cosmic rays (which are deflected by ambient magnetic fields), both GW and HEN propagate at the speed of light through magnetic fields and matter without being altered. Therefore, they are expected to provide important information about the processes taking place in astrophysical engines, and could even reveal the existence of sources opaque to hadrons and photons, sources that would thus far have remained undetected. While neither HENs nor GWs have been directly observed to date, it is widely believed that a first detection could plausibly occur in the near future; see e.g. Becker (2008) and Ma´rka et al. (2011) for reviews on these subjects. This colloquium is dedicated to the prospects for astronomy using these two cosmic messengers.
Many astrophysical sources, the majority of which originate from cataclysmic events, are expected to produce both GWs and HENs. While GWs are linked to the dynamics of the bulk motion of the source progenitor, HENs trace the interactions of accelerated protons (and possibly heavier nuclei) with matter and radiation in and around the source. An overview of the
4
most plausible sources of HENs and GWs is presented in Section II.A of this article, along with relevant references. It includes transient sources such as extra-galactic gamma-ray bursts (GRBs), for which popular progenitor models involve either the collapse of a highly-rotating massive star or the merger of a binary system of compact objects (neutron star/neutron star or black hole/neutron star); both of these scenarios are expected to be associated with the emission of GWs. The presence of accelerated hadrons in the jets emitted by the source would ensure the subsequent production of HENs. Magnetars, though less powerful sources, are closer (galactic) and more frequently occurring; they are also considered as possible GW+HEN emitters. Observation-based phenomenological arguments bounding the time delay between the GW and HEN emission in the sources are presented in Section II.B.
The current efforts carried out for the detection of GWs and HENs are described in Section III. Concerning the detection of neutrinos, huge (∼km3) volumes of target material need to be monitored to compensate for the feeble signal expected from plausible astrophysical sources. Current neutrino telescopes are in-water or in-ice Cherenkov detectors which rely on the construction of 3D arrays of photomultiplier tubes. IceCube1 is a km3-scale detector located at the geographic South Pole, while ANTARES2, with an instrumented volume ∼ 0.02 km3, is deployed undersea, 40 km off the French coast and serves as a prototype for a future km3-scale detector in the Mediterranean. The combination of the two detectors provides full coverage of the sky and partial redundancy. The direct detection of GWs is performed through the operation of large (∼km scale) laser interferometers. Several GW observatories have been operating recently: the two LIGO3 detectors in the USA (one in Livingston, Louisiana, another in Hanford, Washington), Virgo4 near Pisa (Italy) and GEO5 near Hanover (Germany), collectively form a network of detectors that allows for the localisation of astrophysical sources. Both HEN and GW detectors have been developing multimessenger strategies that involve other cosmic probes, in particular electromagnetic radiation in a wide range of wavelength bands. These are typically based either on the use of external triggers (such as GRB events) or on follow-up programs; more detail is given in Sections III.A.2 and III.B.2.
Neither GWs nor cosmic HENs have been individually detected so far. The detection of coincident GW and HEN events would hence be a landmark observation and help confirm the astrophysical origin of both signals. Coincident searches are also a way to enhance the sensitivity of the joint detection channel by exploiting the correlation between HEN and GW significances, taking advantage of the fact that the two types of detectors have uncorrelated backgrounds. Since a joint analysis requires a consistent signal to be observed in both instruments in space as well as time, there is a significant background suppression relative to each individual analysis, hence an increased discovery potential. Preliminary investigations of the feasibility of such searches have been performed by Aso et al. (2008) and Pradier (2009) and indicate that, even if the constituent observatories provide several triggers a day, the false alarm rate for the combined detector network can be maintained at a very low level, e.g. 1/(600 yr) for some realistic parameters.
A major challenge for the analysis lies in the combined optimisation of the selection criteria for the different detection techniques. Section IV starts with laying the basics of the data analysis procedures used in each experiment, including the performance of the detectors, and concentrating on the important aspects connected to GW+HEN searches such as the accuracy of the source sky position reconstruction. Different options for a combined GW+HEN analysis are then presented. Section IV.C describes a method for a HEN-triggered GW search: in this case, the search for GW signals is performed only in parts of the sky defined by neutrino candidate events, and within a time window defined by the observational and phenomenological considerations discussed in Section II.B. The outcome of such an analysis, performed recently with data from the construction phase of ANTARES and from the initial LIGO/VIRGO detectors, is also presented.Alternatively, comprehensive searches for space-time coincidences between independent lists of neutrino and GW events can also be performed, as illustrated in Section IV.D through an example baseline search that could be performed with IceCube and LIGO/Virgo. In this case, time-coincident signals are tested for correlation using a combined GW+HEN likelihood skymap, as well as additional information on the individual significance of the HEN and GW candidates (such as their spatial correlation with large-scale matter distribution). This second, more symmetric and comprehensive option requires the existence of two independent analysis chains scanning the whole phase space in search of interesting events. Both Sections IV.C and IV.D illustrate how these searches can be used to infer limits on the population of astrophysical GW+HEN sources, while the Conclusion presents general perspectives for the astrophysical reach of GW+HEN searches.
The joint search activities described in this paper are performed in the framework of a dedicated GW+HEN working group involving collaborators from all the previously mentioned experiments. The data-exchange policies are regulated by specific bilateral Memoranda of Understanding.
1 Halzen and Klein (2010); see also http://icecube.wisc.edu 2 Ageron et al. (2011); see also http://antares.in2p3.fr 3 Abbott et al. (2009); see also http://www.ligo.org 4 Accadia et al. (2012); see also http://www.virgo.infn.it 5 Grote (2010); see also http://geo600.aei.mpg.de
5
II. THE SCIENCE CASE FOR MULTIMESSENGER GW+HEN SEARCHES
A. Potential emitters of GW and HEN
1. Galactic sources: soft gamma-ray repeaters
Soft gamma-ray repeaters (SGRs) are X-ray pulsars which have quiescent soft (2-10 keV) periodic X-ray emissions with periods ranging from 5 to 10 s. They exhibit repetitive erratic bursting episodes lasting a few hours each and composed of numerous very short (∼ ms) pulses. Every once in a while they emit a giant flare in which a short (< 0.5 sec) spike of harder radiation is observed; such flares can reach peak luminosities of ∼ 1047 erg/s, in X-rays and γ-rays. A handful of SGR sources are known, most of them in the Milky Way and one in the Large Magellanic Cloud. Their detected population has been increasing in the last years, thanks to more sensitive instruments and better monitoring6. Three of the known SGRs have had hard spectrum (∼ MeV energy) giant flares: one with a luminosity of 1047 erg/s, the two others being two orders of magnitude weaker.
The magnetar model describes these objects as a neutron star with an enormous magnetic field B >∼ 1015 G which can be subject to star-quakes that are thought to fracture the rigid crust, causing outbursts (Duncan and Thompson, 1992; Thompson and Duncan, 1995, 1996) . The giant flares result from the formation and dissipation of strong localized currents due to magnetic field rearrangements that are associated with the quakes, and liberate a high flux of X- and γ-rays. Sudden changes in the large magnetic fields would accelerate protons or nuclei that produce neutral and charged pions in interactions with thermal radiation. These hadrons would subsequently decay into TeV or even PeV energy γ-rays and neutrinos (Halzen et al., 2005; Ioka et al., 2005), making flares from SGRs potential sources of HENs. An alternative model involving a large scale rearrangement of the magnetic field has also been proposed by Eichler (2002), which allows for huge energy releases, and detectable HEN fluxes from Galactic magnetars even for relatively small HEN efficiencies.
During the crustal disruption, a fraction of the initial magnetic energy is annihilated and released as photons, and the stored elastic energy is also converted into shear vibrations. SGR flares may excite to some extent the fundamental or f-modes of the star, which radiate GW with damping times of ∼ 200 ms, as described e.g. by Lindblom and Detweiler (1983), de Freitas Pacheco (1998) and Gualtieri et al. (2004). These timescales are shorter than other relevant ones, except for the Alfve´n-wave crossing time of the star, to which they are comparable. If much of the flare energy goes into exciting the f-modes, they might emit GW energy exceeding the emitted EM energy.
Detailed predictions about the GW amplitude are difficult to obtain. An upper limit of ∼ 1049 erg on the maximum total energy release in an SGR giant flare can be derived from one of the most optimistic models (Ioka (2001), Fig. 3) of a giant flare associated with a global reconfiguration of the internal magnetic field (Eichler, 2002). Similarly, Corsi and Owen (2011) estimated a maximum total energy release of ∼ 1048-1049 ergs in a fraction of the parameter space, within the model originally proposed by Ioka (2001). To date, the best LIGO f-mode limit is 1.4 × 1047 erg (at 1090 Hz and a nominal distance of 1 kpc) for SGR 0501+451 (Abadie et al., 2011a). This upper limit probes below the most optimistic total energy estimates, but most likely only a small fraction of the total available energy actually goes into GWs. Recent works by Kashiyama and Ioka (2011), Levin and van Hoven (2011), Ciolfi and Rezzolla (2012), Lasky et al. (2012) and Zink et al. (2012) suggest that the fraction of flare energy that goes into exciting the f-mode is very small (EGW 1045 erg for magnetic fields smaller than 1016 G), making the prospect for a detection of GWs from SGR f-modes with the advanced LIGO and Virgo unlikely. However, this question may be open to further investigations; see e.g. Kashiyama and Ioka (2011), Levin and van Hoven (2011), Lasky et al. (2012) and Zink et al. (2012).
2. Gamma-ray bursts
Gamma-ray Bursts (GRBs) are detected as an intense and short-lived flash of gamma-rays with energies ranging from tens of keVs to tens of GeVs. The morphology of their light curves is highly variable and typically exhibits millisecond variability, suggesting very compact sources and relativistic expansion. GRBs are divided into two classes depending on the duration of their prompt gamma-ray emission, which appears to be correlated with the hardness of their spectra and are believed to arise from different progenitors: short-hard bursts last less than 1 – 2 seconds (depending on the observing detector) while long-soft bursts can last up to dozens of minutes.
The BATSE detector, launched in 1991 on board the Compton Gamma-Ray Observatory, was the first mission to accumulate observations on more than a thousand GRBs, establishing the isotropy of their sky distribution and characterizing their light curve and broken power-law spectra (Paciesas et al., 1999). The detection of X-ray and optical counterparts pertaining to the afterglow
6 See e.g. Hurley et al. (1999a); Hurley et al. (1999b); Cline et al. (2000); Kulkarni et al. (2003); Palmer et al. (2005); Mereghetti (2008); Aptekar et al. (2009); Hurley (2010); Go¨g˘u¨s¸ et al. (2010); Kaneko et al. (2010); van der Horst et al. (2010).
6
NeuCosmA 2012
NFC prediction
100
IC40
GRB, all
GRB, z known 50 stat. error
astrophysical
10 9
IC40 59
uncertainties
20
1 10
fe
5
EΝ2 ΦΝ E GeV cm 2 s 1 sr 1
10 10
IC86, 10y extrapolated
105
106
107
108
109
EΝ GeV
FIG. 1 Left: Schematic depiction of the fireball mechanism, with the characteristic time and distance scales associated with the different phases. The prompt (burst) phase is due to internal shocks in the relativistically expanding fireball, producing strong gamma-ray and X-ray emission. The afterglow arises from the cooling fireball and its interaction with the surrounding medium; it is associated with X-ray, optical and radio emission. From www.swift.ac.uk . Right: Current limits set by IceCube in its 40-string and 59-string configurations (IC40 and IC40+59) on the quasi-diffuse GRB neutrino flux, together with predictions based on recent numerical calculations taking into account uncertainties on the astrophysical parameters and those due to the limited statistics of bursts. Also shown is the extrapolated limit that one can expect from 10 years of operation with the full IceCube detector (IC86). From Hummer et al. (2012).
phase of several GRBs, triggered by the first observation of an X-ray transient emission from GRB970228 by the BeppoSAX satellite (Costa et al., 1997), subsequently confirmed their extragalactic origin by allowing more accurate localization of the source and redshift determination. Currently operating GRB missions include Swift (Gehrels et al., 2004), hosting a wide-field hard X-ray (15 keV - 350 keV) burst alert telescope (BAT) coupled to softer X-ray, ultraviolet and optical telescopes and the GBM on the Fermi Gamma-Ray Space Telescope (Atwood et al., 2009) which focuses on the high-energy (15 keV - 300 GeV) emission from GRBs.
In the standard picture, the mechanism responsible for the enormous, super-Eddington energy release (∼ 1050 − 1052 ergs) in the prompt emission and in the afterglow is the dissipation (via internal shocks, magnetic reconnection and external shocks) of bulk kinetic or Poynting flux into highly relativistic particles; see e.g. Me´sza´ros and Rees (1993) and the review by Piran (2004). The particles are accelerated to a non-thermal energy distribution via the Fermi mechanism in a relativistically expanding fireball ejected by the GRB central engine, as sketched in Fig. 1 (left panel). The accelerated electrons (and positrons) in the intense magnetic field emit non-thermal photons via synchrotron radiation and inverse Compton scattering. The plasma parameters inferred from observations to characterize GRB baryonic fireballs are such that proton acceleration to energies exceeding 1020eV is likely to be possible in these sources. Moreover, the time averaged energy output of GRBs in photons is comparable to the proton energy production rate required to produce the UHECR flux7. Therefore, the canonical baryonic fireball also suggests that GRBs are a prime candidate source for the UHECR, observed at energies E ∼ 1018 − 1020 eV (Levinson and Eichler, 1993; Vietri, 1995; Waxman, 1995).
In a baryonic outflow, the internal or external shocks accelerate protons that interact with the gamma-rays and/or other protons inside the fireball, producing charged pions and kaons that subsequently decay into HENs ( π±, K± → µ± + νµ/νµ → e± + νe/νe + νµ/νµ)8. Such neutrinos are emitted in spatial and temporal coincidence with the GRB prompt electromagnetic signal; their energy is typically in the range ∼ TeV to PeV. Neutrinos with higher (up to ∼ 1010 GeV) energy can also be emitted at the beginning of the afterglow phase, when the outflow is decelerated by external shocks with ambient material and the accelerated protons undergo interactions with the matter outside of the jet (Waxman and Bahcall, 2000). An alternative mechanism for neutrino production in fireballs suggested by Levinson and Eichler (2003) involves neutral particles that are picked up by the stream when they acquire a charge, such as a decaying neutron or, further downstream, a neutral atom that is ionized. Such a particle will be extremely energetic in the jet frame, and immediately attains an energy of a PeV. The associated neutrinos would come within an order of magnitude of that energy (∼ 100 TeV), providing a harder spectrum than the one expected from shock acceleration.
It is expected that in the next few years neutrino telescopes will be sufficiently sensitive to test and distinguish between GRB models with different physics, and to constrain the parameters of such models. The current non-detection of neutrinos by
7 The latter statement has been criticized by several authors recently, arguing that the GRB energy production rate is too small to account for the flux of ultrahigh-energy cosmic rays (UHECRs); see e.g. (Eichler et al., 2010). However, the validity of this criticism has been challenged by Waxman (2010).
8 Relevant references on these mechanisms include Eichler (1994), Paczynski and Xu (1994), Waxman and Bahcall (1997), Rachen and Me´sza´ros (1998), Alvarez-Mun˜iz et al. (2000), Me´sza´ros and Waxman (2001), Me´sza´ros and Waxman (2001), Guetta and Granot (2003), Razzaque et al. (2003a), Razzaque et al. (2003b), Dermer and Atoyan (2003), Guetta et al. (2004), Ando and Beacom (2005), Murase and Nagataki (2006), Murase et al. (2006).
7
IceCube (Abbasi et al. (2012a); see also Sec. III.B) already questions the viability of models in which ultra high energy cosmic rays are the decay products of neutrons that have escaped the fireball with high energy. The current upper limit is still consistent with the “standard” (i.e., following Waxman and Bahcall (1997)) predictions of neutrino emission from GRB fireballs (Hummer et al., 2012; Li, 2012). Given current uncertainties, significant constraints on this model will be obtained within 5-10 yrs of full IceCube operation, as can be seen from Fig. 1 (right panel).
While gamma-ray and HEN emissions from GRBs are related to the mechanisms driving the relativistic outflow, GW emission is closely connected to the central engine and hence to the progenitor of the GRB. Short-hard GRBs are thought to be driven by neutron star–neutron star or neutron star–black hole mergers9. GW detectors can ideally observe those binary systems up to a distance of ∼ 30 Mpc and ∼ 440 Mpc for initial and advanced detectors respectively (Abadie et al., 2010b). These distances coincide with the range where the HEN flux is thought to be large enough for detection with current HEN detectors. Note that these short GRBs are beamed and so is the expected HEN emission. Hence one can expect cases in which GWs will be observed from such sources without an observed GRB or HEN signal. However, an orphan afterglow (Levinson et al., 2002; Nakar and Piran, 2003), macronovae (Li and Paczynski, 1998) or radio flares (Nakar and Piran, 2011) might be observed in these cases.
Weaker GW signals are expected in any source that accelerates relativistic jets (Piran, 2013). A few mechanisms have been suggested for GW generation in long-soft GRBs, which arise during the collapse of a massive star; see e.g. Woosley and Bloom (2006) and references therein. According to the collapsar model proposed by Woosley and MacFadyen (1999) a long GRBs arises when a relativistic jet, produced by an central inner engine penetrates the stellar envelope of the collapsing star. The inner engine driving the jet can be either an accreting newborn black hole (Woosley and MacFadyen, 1999) or a newborn rapidly rotating magnetar (Usov, 1992). For accretion models, the high rotation rate required to form the accretion disk that powers the GRB may also lead to the production of GWs via bar or fragmentation instabilities in the accretion disks and also via the precession of the disks due to general relativistic effects10. Asymmetrically infalling matter produces the burst GW signals not only at the moment of the core bounce when the central density exceeds nuclear density (Kotake et al., 2006; Ott, 2009), but also at the moment of the black hole formation, followed by the subsequent ring-down phases (Ott et al., 2011). Optimistic estimates based on semi-analytical calculations suggest that the GW signals from some of these mechanisms are high enough to be visible in Advanced LIGO class detectors up to a 100 Mpc distance scale; see collective references in Kotake et al. (2012b). However, to obtain more quantitative predictions, full 3D simulations using general-relativistic magnetohydrodynamics and sophisticated neutrino transport schemes are needed. Much effort has been recently dedicated to this issue with encouraging results; see. e.g., Kuroda et al. (2012); Muller et al. (2012); Ott et al. (2012), and Kotake et al. (2012a) for a review. Unfortunately, such studies suggest a very weak GW emission and, given the fact that the long GRB population is distributed over cosmological distances, seriously challenge the prospects for long GRB detection even in the next generation of GW detectors.
“Low-luminosity GRBs” (llGRBs) are a subclass of long-soft GRBs. llGRBs are characterized by luminosities lower by a few orders of magnitude than typical gamma-ray luminosities, a smooth, single-peaked light curve, and a soft spectrum. These bursts are associated with particularly energetic type Ibc core-collapse supernovae as observed in GRB 980425/SN 1998bw (Galama et al., 1998; Kulkarni et al., 1998), GRB 031203/SN 2003lw (Malesani et al., 2004; Soderberg et al., 2004b), and GRB 060218/SN 2006aj (Campana et al., 2006; Cobb et al., 2006; Pian et al., 2006; Soderberg et al., 2006). It is interesting to note that most of the GRB-SNe associations are with llGRBs and not with regular long GRBs. Less luminous than typical long GRBs, these events are (not surprisingly) discovered at much smaller distances (SN 1998bw at redshift z = 0.0085, about 40 Mpc away from Earth; SN 2003lw at z = 0.105, and SN 2006aj at z = 0.033). Remarkably, the event rate of llGRBs per unit local volume is more than one order of magnitude larger than that of conventional long GRBs; see e.g. Coward (2005), Guetta and Della Valle (2007), Soderberg et al. (2006), Daigne and Mochkovitch (2007) and Liang et al. (2007). This makes this source population an interesting target of study from the GW+HEN point of view as well, as discussed by Razzaque et al. (2004), Murase et al. (2006), Gupta and Zhang (2007), and Wang et al. (2007).
Bromberg et al. (2011) have recently argued that, given their apparently low power, these llGRBs cannot arise from the regular collapsar model because the time needed for the jet to bore an escape channel through the host envelope would, in most reported cases, exceed the GRB duration. Rather, they may be gamma rays from shock break-out imparted to the host envelope by jets that fail to emerge (“choked jets”)11. The smooth light curve and soft spectra of these events are indeed expected from shock breakout; see Katz et al. (2010); Nakar and Sari (2012); Waxman et al. (2007). Other suggested models, which produce smooth, soft emission, include scattering of the gamma-rays off an accelerating envelope or wind material (Eichler and Levinson, 1999), or gamma rays that are released from baryon-rich jet material (dirty fireballs) only after some adiabatic loss (Mandal and Eichler, 2010). It has also been suggested that llGRBs are associated with the formation of magnetars rather than black holes, as argued
9 This mechanism has been described e.g. by Eichler et al. (1989); Kochanek and Piran (1993); Nakar (2007); Bloom et al. (2007); Lee and Ramirez-Ruiz (2007); Etienne et al. (2009).
10 See e.g. Fryer and Woosley (1998); Davies et al. (2002); Fryer et al. (2002); Kobayashi and Me´sza´ros (2003); Piro and Pfahl (2007), Romero et al. (2010), Sun et al. (2012).
11 Relevant references include MacFadyen et al. (2001); Tan et al. (2001); Campana et al. (2006); Wang et al. (2007); Waxman et al. (2007); Katz et al. (2010); Nakar and Sari (2012).
8
central engine active
central engine active
(a) 100s
(b)
(c) 250s gamma HEN GW
(a)
(e) 500s
100s
150s
(b) (d) gamma > 100 MeV
FIG. 2 Summary of the upper bounds on the duration of GRB emission processes taken into account in the total GW+HEN coincidence time window. (a) active central engine before the relativistic jet has broken out of the stellar envelope (note that recent estimates of jet propagation within a stellar envelope suggest that this phase lasts typically 10-20 sec; see Bromberg et al. (2011, 2012)); (b) active central engine with the relativistic jet broken out of the envelope; (c) delay between the onset of the precursor and the main burst; (d) duration corresponding to 90% of GeV photon emission; (e) time span of central engine activity. The top of the figure shows a schematic drawing of a plausible emission scenario. Figure taken from Baret et al. (2011).
for GRB060218 by Mazzali et al. (2006), a scenario that might give rise to somewhat longer GW signals (Corsi and Me´sza´ros, 2009; Piro and Ott, 2011).
Regardless of the question of whether or not they produce llGRBs, such “choked jet” events are interesting objects on their own, as pointed out e.g. by Eichler and Levinson (1999), Me´sza´ros and Waxman (2001) and Ando and Beacom (2005). In fact, late-time radio emission of some type Ic supernovae indeed suggests the presence of mildly relativistic outflow (Granot and Ramirez-Ruiz, 2004; Mazzali et al., 2005; Soderberg et al., 2004a, 2010) that may indicate the activity of a jet in these cases. The expected overall energy budget of choked jets is comparable to the energy budget observed in regular GRBs12. Therefore, these choked jets could produce GWs and HENs13 in a way comparable with regular GRBs; see Eichler and Levinson (1999), Me´sza´ros and Waxman (2001), Ando and Beacom (2005), Koers and Wijers (2007), Horiuchi and Ando (2008). As they are likely to be more numerous than regular GRBs (as is for example the case of llGRBs), they could be observed from nearer distances. In such a case, optimistic estimates predict potentially detectable levels of both GW and HEN signals, and an observable occurrence rate in the volume probed by current GW and HEN detectors. For example, according to Ando and Beacom (2005), an ejected mass with a kinetic energy of 3 × 1051 erg and a Lorentz factor of 3 at 10 Mpc would generate ∼30 neutrino events detected in a km3 detector. These events should be seen accompanying some specific local core collapse SNe. In this context, HEN and GW could play an interesting role in revealing the inner “choked jets” nature of these sources.
B. Bounds on the GW+HEN time delay
The possible time delay between the arrival of GWs and HENs from a given source defines the coincidence time window to apply in a multimessenger search algorithm. This window should not be too small, which could lead to the exclusion of potential emission mechanisms, nor too large, which would decrease the detection sensitivity by including non-physical coincidences. Upon detection, the difference between the times of arrival of GW and HEN signals can give us important clues about the emission mechanism. For instance detecting a HEN prior to a GW signal may indicate that the strongest GW emission from the source is not connected to the onset of the activity of the central engine that one might expect from core-collapse models.
12 Note that if indeed a choked jet produces a llGRB via shock breakout, the prompt γ-rays involve only a small fraction of the total energy (Bromberg et al., 2011; Nakar and Sari, 2012).
13 The production of HEN is closely related to the efficiency of proton acceleration inside the jet, an issue which is still debated considering that the relevant shocks in these choked GRBs are expected to be radiation-dominated (Levinson and Bromberg, 2008).
9
FIG. 3 Optical scheme of a gravitational-wave interferometric detector, consisting of two twin laser beams propagating in km-long arms oriented at 90◦ to each other. The Fabry-Pe´rot cavities enable the storage of the beams, thereby increasing by a significant factor their effective path length; the suspended, highly reflective mirrors play the role of test masses. A gravitational wave would cause a difference in the optical pathlengths in each arm, which can be inferred by measuring the interference pattern at the photodiode. Figure adapted from Y. Aso (GECo, Columbia University).
Baret et al. (2011) used model-motivated comparisons with GRB observations to derive a conservative coincidence time window for joint GW+HEN searches. Various GRB emission processes were considered, assuming that GW and HEN emission are connected to the activity of the central engine. Considered processes include prompt gamma-ray emission of GRBs, with a duration upper limit (∼ 150 s) based on BATSE observations (Paciesas et al., 1999), as well as GRB precursor activity, with an upper limit on the time difference (as compared to the onset of the main burst) of ∼ 250 s, following the analysis of Burlon et al. (2009). Further processes considered include precursor neutrino emission, as well as >∼ 100 MeV photon emission from some GRBs, as detected by Fermi LAT (Atwood et al., 2009). The authors conclude that GW and HEN signals are likely to arrive within a time window of ± 500 s, as illustrated in Fig. 2.
The time-delay between HENs and GWs could be much smaller for binary mergers which are often mentioned as the possible progenitor of short-hard GRBs. The amount of accreted/ejected matter involved in such case is very small, and the outflowing matter can expand unhindered, adding almost nothing to the time delay. A semi-analytical description of the final stage of such mergers indicates that most of the matter is accreted within 1 second (Davies et al., 2005), and numerical simulations of the mass transfer suggest time scales of milliseconds (Shibata and Taniguchi, 2008) to a few seconds maximum (Faber et al., 2006). Therefore, the GW signal is expected to arrive very close to HENs. A window of [−5, +1] seconds around the trigger time, as used for (short) GRB-GW searches in Abbott et al. (2008) and Abadie et al. (2010c), seems reasonable.
III. GW AND HEN DETECTION: STATUS AND PROSPECTS
A. Interferometric Gravitational Wave detectors
1. Detection principle and state of the art
The first generation of interferometric GW detectors included a total of six large-scale instruments. The US-based Laser Interferometer Gravitational-Wave Observatory (LIGO, see Abbott et al. (2009)) was comprised of three kilometer-scale instruments located in Livingston, Louisiana and Hanford, Washington (the latter hosted two interferometers in the same vacuum enclosure). The French-Italian project Virgo (Accadia et al., 2012) had one instrument of the same class located in Cascina near Pisa, Italy. This set of kilometer-scale instruments was complemented by a couple of detectors with more modest dimensions (several hundreds of meters): GEO (Grote, 2010), a German-British detector in operation near Hanover, Germany and the Japanese prototype CLIO (Agatsuma et al., 2010) located in the Kamioka mine.
Despite major differences in the technologies, all past and upcoming ground-based km-scale detectors measure gravitational waves through the same principle (see Fig. 3 for an illustration of the general scheme). They all sense the strain that a passing GW exerts on space-time by monitoring the differential length δ of the optical path followed by two laser beams propagating along orthogonal directions. Measurement noises (mainly the thermal noise due to the Brownian agitation of the atoms constitutive of the main optics and the shot noise due to the quantum nature of light) can be reduced to reach the level of h ≡ δ /L ∼ 10−21,
S. Ando, B. Baret, B. Bouhou, E. Chassande-Mottin, A. Kouchner, L. Moscoso, V. Van Elewyck, I. Bartos, S. Ma´rka, Z. M´arka, et al.
To cite this version:
S. Ando, B. Baret, B. Bouhou, E. Chassande-Mottin, A. Kouchner, et al.. Multimessenger astronomy with gravitational waves and high-energy neutrinos. Reviews of Modern Physics, American Physical Society, 2013, 85, pp.1401-1420. <10.1103/RevModPhys.85.1401>.
HAL Id: in2p3-00681933 http://hal.in2p3.fr/in2p3-00681933v2
Submitted on 25 Jul 2013
HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destin´ee au d´epˆot et `a la diffusion de documents scientifiques de niveau recherche, publi´es ou non, ´emanant des ´etablissements d’enseignement et de recherche fran¸cais ou ´etrangers, des laboratoires publics ou priv´es.
Multimessenger astronomy with gravitational waves and high-energy
neutrinos
Shin’ichiro Ando,1 Bruny Baret,2 Imre Bartos,3 Boutayeb Bouhou,2 Eric Chassande-Mottin,2 Alessandra Corsi,4, 5 Irene Di Palma,6 Alexander Dietz,7 Corinne Donzaud,2, 8 David Eichler,9 Chad Finley,10 Dafne Guetta,11, 12 Francis Halzen,13 Gareth Jones,14 Shivaraj Kandhasamy,15 Kei Kotake,16, 17 Antoine Kouchner,2 Vuk Mandic,15 Szabolcs Ma´ rka,3 Zsuzsa Ma´ rka,3 Luciano Moscoso,2, ∗ Maria Alessandra Papa,6 Tsvi Piran,18 Thierry Pradier,19 Gustavo E. Romero,20, 21 Patrick Sutton,14 Eric Thrane,15 Ve´ ronique Van Elewyck,2, † and Eli Waxman22 1Gravitation AstroParticle Physics Amsterdam Institute (GRAPPA), The Netherlands 2AstroParticule et Cosmologie (APC), CNRS: UMR7164-IN2P3-Observatoire de Paris-Universite´ Denis Diderot-Paris VII-CEA: DSM/IRFU, France 3Department of Physics, Columbia University, New York, NY 10027, USA 4Physics Department, George Washington University, Washington, D.C. 20052, USA 5LIGO Laboratory, California Institute of Technology, Pasadena, CA 91125, USA 6Albert-Einstein-Institut, Max-Planck-Institut fu¨ r Gravitationsphysik, D-30167 Hannover, Germany 7Department of Physics and Astronomy of the University of Mississippi, Mississippi 38677-1848, USA 8Universite´ Paris-sud, Orsay, F-91405, France 9Department of Physics, Ben Gurion University, Beer-Sheva 84105, Israel 10Oskar Klein Centre & Dept. of Physics, Stockholm University, SE-10691 Stockholm, Sweden 11Department of Physics and Optical Engineering, ORT Braude, P.O. Box 78, Karmiel, Israel 12INAF-Observatory of Rome, Via Frascati 33, Monteporzio Catone, Italy 13Department of Physics, University of Wisconsin, Madison, WI 53706, USA 14School of Physics and Astronomy, Cardiff University,
2
Cardiff CF24 3AA, United Kingdom 15University of Minnesota, Minneapolis, MN 55455, USA 16Division of Theoretical Astronomy, National Astronomical Observatory of Japan, 2-21-1,Osawa, Mitaka, Tokyo, 181-8588, Japan 17Department of applied physics, Fukuoka University, Jonan, Fukuoka, 814-0189, Japan 18Racah Institute of Physics, Hebrew University of Jerusalem, Jerusalem 91904, Israel 19Universite´ de Strasbourg & Institut Pluridisciplinaire France 20Instituto Argentino de Radioastronomia (IAR, CCT La Plata, CONICET), C.C. No. 5, 1894, Villa Elisa, Buenos Aires, Argentina 21FCAyG, Observatorio de La Plata, Paseo del Bosque s/n, CP 1900 La Plata, Argentina. 22Department of Particle Physics & Astrophysics, The Weizmann Institute of Science, Rehovot 76100, Israel
Hubert
Curien,Strasbourg,
Many of the astrophysical sources and violent phenomena observed in our Universe are potential emitters of gravitational waves and high-energy cosmic radiation, including photons, hadrons, and presumably also neutrinos. Both gravitational waves (GW) and high-energy neutrinos (HEN) are cosmic messengers that may escape much denser media than photons. They travel unaffected over cosmological distances, carrying information from the inner regions of the astrophysical engines from which they are emitted (and from which photons and charged cosmic rays cannot reach us). For the same reasons, such messengers could also reveal new, hidden sources that have not been observed by conventional photon-based astronomy. Coincident observation of GWs and HENs may thus play a critical role in multimessenger astronomy. This is particularly true at the present time owing to the advent of a new generation of dedicated detectors: the neutrino telescopes IceCube at the South Pole and ANTARES in the Mediterranean Sea, as well as the GW interferometers Virgo in Italy and LIGO in the United States. Starting from 2007, several periods of concomitant data taking involving these detectors have been conducted. More joint datasets are expected with the next generation of advanced detectors that are to be operational by 2015, with other detectors, such as KAGRA in Japan, joining in the future. Combining information from these independent detectors can provide original ways of constraining the physical processes driving the sources, and also help confirm the astrophysical origin of a GW or HEN signal in case of coincident observation. Given the complexity of the instruments, a successful joint analysis of this combined GW+HEN observational dataset will be possible only if the expertise and knowledge of the data is shared between the two communities. This review aims at providing an overview of both theoretical and experimental state of the art and perspectives for GW+HEN multimessenger astronomy.
∗ deceased † Corresponding author’s electronic address: [email protected]
3
CONTENTS
I. Introduction
3
II. The science case for multimessenger GW+HEN searches
5
A. Potential emitters of GW and HEN
5
1. Galactic sources: soft gamma-ray repeaters
5
2. Gamma-ray bursts
5
B. Bounds on the GW+HEN time delay
8
III. GW and HEN detection: status and prospects
9
A. Interferometric Gravitational Wave detectors
9
1. Detection principle and state of the art
9
2. Multimessenger strategies
10
B. High-energy neutrino telescopes
11
1. Detection principle and state of the art
11
2. Multimessenger strategies
12
IV. Perspectives for the joint data analysis
13
A. GW data analysis
13
1. Multi-detector coherent analysis and background rejection
14
2. Source localization
14
B. HEN data analysis
14
C. HEN-triggered GW searches
15
D. Baseline search with combined HEN and GW events lists
17
V. Conclusions
18
VI. Acknowledgments
19
References
19
I. INTRODUCTION
High-energy multimessenger astronomy has entered an exciting era with the development and operation of new detectors offering unprecedented opportunities to observe cosmic radiation in the Universe in all its variety. Gamma-ray astronomy has been an illustrative example of the synergy between particle physics and astronomy. Soon after the discovery of the neutral pion, Hayakawa (1952) suggested that the interaction of cosmic rays with the neutral gas in the Galactic Plane should produce a diffuse, extended gamma-ray emission. Almost simultaneously, Hutchinson (1952) calculated the contribution to such an emission from relativistic bremsstrahlung. Prospects for gamma-ray astronomy were set up shortly after by Morrison (1958). The first gamma-ray sources were discovered during the late 1960s and 1970s, but even until the 1990s they were difficult to identify, calling for multi-wavelength observational efforts (see Cheng and Romero (2004) for an overview of the subject). During the past decade, this approach has revealed itself to be fruitful in the identification of several types of sources from MeV to TeV energy scales.
The mechanisms that produce the high-energy radiation have, however, remained elusive, requiring the development of multimessenger techniques and programs that would explore all components of the cosmic radiation (see collective references in Paredes et al. (2007)). In the study of transient sources, which involve compact objects and ultra-violent phenomena (such as gammaray bursts and magnetars), multimessenger techniques are in fact the only approaches that might lead to a full understanding of the underlying processes. In this context, high-energy ( GeV) neutrinos (HENs) and gravitational waves (GWs) could play an important role. These messengers share interesting astronomical properties: HENs can escape from much denser, hence deeper, environments than photons, and GWs propagate virtually freely in any region of space. Moreover, and contrary to high-energy photons (which can be absorbed by intervening photon backgrounds) and charged cosmic rays (which are deflected by ambient magnetic fields), both GW and HEN propagate at the speed of light through magnetic fields and matter without being altered. Therefore, they are expected to provide important information about the processes taking place in astrophysical engines, and could even reveal the existence of sources opaque to hadrons and photons, sources that would thus far have remained undetected. While neither HENs nor GWs have been directly observed to date, it is widely believed that a first detection could plausibly occur in the near future; see e.g. Becker (2008) and Ma´rka et al. (2011) for reviews on these subjects. This colloquium is dedicated to the prospects for astronomy using these two cosmic messengers.
Many astrophysical sources, the majority of which originate from cataclysmic events, are expected to produce both GWs and HENs. While GWs are linked to the dynamics of the bulk motion of the source progenitor, HENs trace the interactions of accelerated protons (and possibly heavier nuclei) with matter and radiation in and around the source. An overview of the
4
most plausible sources of HENs and GWs is presented in Section II.A of this article, along with relevant references. It includes transient sources such as extra-galactic gamma-ray bursts (GRBs), for which popular progenitor models involve either the collapse of a highly-rotating massive star or the merger of a binary system of compact objects (neutron star/neutron star or black hole/neutron star); both of these scenarios are expected to be associated with the emission of GWs. The presence of accelerated hadrons in the jets emitted by the source would ensure the subsequent production of HENs. Magnetars, though less powerful sources, are closer (galactic) and more frequently occurring; they are also considered as possible GW+HEN emitters. Observation-based phenomenological arguments bounding the time delay between the GW and HEN emission in the sources are presented in Section II.B.
The current efforts carried out for the detection of GWs and HENs are described in Section III. Concerning the detection of neutrinos, huge (∼km3) volumes of target material need to be monitored to compensate for the feeble signal expected from plausible astrophysical sources. Current neutrino telescopes are in-water or in-ice Cherenkov detectors which rely on the construction of 3D arrays of photomultiplier tubes. IceCube1 is a km3-scale detector located at the geographic South Pole, while ANTARES2, with an instrumented volume ∼ 0.02 km3, is deployed undersea, 40 km off the French coast and serves as a prototype for a future km3-scale detector in the Mediterranean. The combination of the two detectors provides full coverage of the sky and partial redundancy. The direct detection of GWs is performed through the operation of large (∼km scale) laser interferometers. Several GW observatories have been operating recently: the two LIGO3 detectors in the USA (one in Livingston, Louisiana, another in Hanford, Washington), Virgo4 near Pisa (Italy) and GEO5 near Hanover (Germany), collectively form a network of detectors that allows for the localisation of astrophysical sources. Both HEN and GW detectors have been developing multimessenger strategies that involve other cosmic probes, in particular electromagnetic radiation in a wide range of wavelength bands. These are typically based either on the use of external triggers (such as GRB events) or on follow-up programs; more detail is given in Sections III.A.2 and III.B.2.
Neither GWs nor cosmic HENs have been individually detected so far. The detection of coincident GW and HEN events would hence be a landmark observation and help confirm the astrophysical origin of both signals. Coincident searches are also a way to enhance the sensitivity of the joint detection channel by exploiting the correlation between HEN and GW significances, taking advantage of the fact that the two types of detectors have uncorrelated backgrounds. Since a joint analysis requires a consistent signal to be observed in both instruments in space as well as time, there is a significant background suppression relative to each individual analysis, hence an increased discovery potential. Preliminary investigations of the feasibility of such searches have been performed by Aso et al. (2008) and Pradier (2009) and indicate that, even if the constituent observatories provide several triggers a day, the false alarm rate for the combined detector network can be maintained at a very low level, e.g. 1/(600 yr) for some realistic parameters.
A major challenge for the analysis lies in the combined optimisation of the selection criteria for the different detection techniques. Section IV starts with laying the basics of the data analysis procedures used in each experiment, including the performance of the detectors, and concentrating on the important aspects connected to GW+HEN searches such as the accuracy of the source sky position reconstruction. Different options for a combined GW+HEN analysis are then presented. Section IV.C describes a method for a HEN-triggered GW search: in this case, the search for GW signals is performed only in parts of the sky defined by neutrino candidate events, and within a time window defined by the observational and phenomenological considerations discussed in Section II.B. The outcome of such an analysis, performed recently with data from the construction phase of ANTARES and from the initial LIGO/VIRGO detectors, is also presented.Alternatively, comprehensive searches for space-time coincidences between independent lists of neutrino and GW events can also be performed, as illustrated in Section IV.D through an example baseline search that could be performed with IceCube and LIGO/Virgo. In this case, time-coincident signals are tested for correlation using a combined GW+HEN likelihood skymap, as well as additional information on the individual significance of the HEN and GW candidates (such as their spatial correlation with large-scale matter distribution). This second, more symmetric and comprehensive option requires the existence of two independent analysis chains scanning the whole phase space in search of interesting events. Both Sections IV.C and IV.D illustrate how these searches can be used to infer limits on the population of astrophysical GW+HEN sources, while the Conclusion presents general perspectives for the astrophysical reach of GW+HEN searches.
The joint search activities described in this paper are performed in the framework of a dedicated GW+HEN working group involving collaborators from all the previously mentioned experiments. The data-exchange policies are regulated by specific bilateral Memoranda of Understanding.
1 Halzen and Klein (2010); see also http://icecube.wisc.edu 2 Ageron et al. (2011); see also http://antares.in2p3.fr 3 Abbott et al. (2009); see also http://www.ligo.org 4 Accadia et al. (2012); see also http://www.virgo.infn.it 5 Grote (2010); see also http://geo600.aei.mpg.de
5
II. THE SCIENCE CASE FOR MULTIMESSENGER GW+HEN SEARCHES
A. Potential emitters of GW and HEN
1. Galactic sources: soft gamma-ray repeaters
Soft gamma-ray repeaters (SGRs) are X-ray pulsars which have quiescent soft (2-10 keV) periodic X-ray emissions with periods ranging from 5 to 10 s. They exhibit repetitive erratic bursting episodes lasting a few hours each and composed of numerous very short (∼ ms) pulses. Every once in a while they emit a giant flare in which a short (< 0.5 sec) spike of harder radiation is observed; such flares can reach peak luminosities of ∼ 1047 erg/s, in X-rays and γ-rays. A handful of SGR sources are known, most of them in the Milky Way and one in the Large Magellanic Cloud. Their detected population has been increasing in the last years, thanks to more sensitive instruments and better monitoring6. Three of the known SGRs have had hard spectrum (∼ MeV energy) giant flares: one with a luminosity of 1047 erg/s, the two others being two orders of magnitude weaker.
The magnetar model describes these objects as a neutron star with an enormous magnetic field B >∼ 1015 G which can be subject to star-quakes that are thought to fracture the rigid crust, causing outbursts (Duncan and Thompson, 1992; Thompson and Duncan, 1995, 1996) . The giant flares result from the formation and dissipation of strong localized currents due to magnetic field rearrangements that are associated with the quakes, and liberate a high flux of X- and γ-rays. Sudden changes in the large magnetic fields would accelerate protons or nuclei that produce neutral and charged pions in interactions with thermal radiation. These hadrons would subsequently decay into TeV or even PeV energy γ-rays and neutrinos (Halzen et al., 2005; Ioka et al., 2005), making flares from SGRs potential sources of HENs. An alternative model involving a large scale rearrangement of the magnetic field has also been proposed by Eichler (2002), which allows for huge energy releases, and detectable HEN fluxes from Galactic magnetars even for relatively small HEN efficiencies.
During the crustal disruption, a fraction of the initial magnetic energy is annihilated and released as photons, and the stored elastic energy is also converted into shear vibrations. SGR flares may excite to some extent the fundamental or f-modes of the star, which radiate GW with damping times of ∼ 200 ms, as described e.g. by Lindblom and Detweiler (1983), de Freitas Pacheco (1998) and Gualtieri et al. (2004). These timescales are shorter than other relevant ones, except for the Alfve´n-wave crossing time of the star, to which they are comparable. If much of the flare energy goes into exciting the f-modes, they might emit GW energy exceeding the emitted EM energy.
Detailed predictions about the GW amplitude are difficult to obtain. An upper limit of ∼ 1049 erg on the maximum total energy release in an SGR giant flare can be derived from one of the most optimistic models (Ioka (2001), Fig. 3) of a giant flare associated with a global reconfiguration of the internal magnetic field (Eichler, 2002). Similarly, Corsi and Owen (2011) estimated a maximum total energy release of ∼ 1048-1049 ergs in a fraction of the parameter space, within the model originally proposed by Ioka (2001). To date, the best LIGO f-mode limit is 1.4 × 1047 erg (at 1090 Hz and a nominal distance of 1 kpc) for SGR 0501+451 (Abadie et al., 2011a). This upper limit probes below the most optimistic total energy estimates, but most likely only a small fraction of the total available energy actually goes into GWs. Recent works by Kashiyama and Ioka (2011), Levin and van Hoven (2011), Ciolfi and Rezzolla (2012), Lasky et al. (2012) and Zink et al. (2012) suggest that the fraction of flare energy that goes into exciting the f-mode is very small (EGW 1045 erg for magnetic fields smaller than 1016 G), making the prospect for a detection of GWs from SGR f-modes with the advanced LIGO and Virgo unlikely. However, this question may be open to further investigations; see e.g. Kashiyama and Ioka (2011), Levin and van Hoven (2011), Lasky et al. (2012) and Zink et al. (2012).
2. Gamma-ray bursts
Gamma-ray Bursts (GRBs) are detected as an intense and short-lived flash of gamma-rays with energies ranging from tens of keVs to tens of GeVs. The morphology of their light curves is highly variable and typically exhibits millisecond variability, suggesting very compact sources and relativistic expansion. GRBs are divided into two classes depending on the duration of their prompt gamma-ray emission, which appears to be correlated with the hardness of their spectra and are believed to arise from different progenitors: short-hard bursts last less than 1 – 2 seconds (depending on the observing detector) while long-soft bursts can last up to dozens of minutes.
The BATSE detector, launched in 1991 on board the Compton Gamma-Ray Observatory, was the first mission to accumulate observations on more than a thousand GRBs, establishing the isotropy of their sky distribution and characterizing their light curve and broken power-law spectra (Paciesas et al., 1999). The detection of X-ray and optical counterparts pertaining to the afterglow
6 See e.g. Hurley et al. (1999a); Hurley et al. (1999b); Cline et al. (2000); Kulkarni et al. (2003); Palmer et al. (2005); Mereghetti (2008); Aptekar et al. (2009); Hurley (2010); Go¨g˘u¨s¸ et al. (2010); Kaneko et al. (2010); van der Horst et al. (2010).
6
NeuCosmA 2012
NFC prediction
100
IC40
GRB, all
GRB, z known 50 stat. error
astrophysical
10 9
IC40 59
uncertainties
20
1 10
fe
5
EΝ2 ΦΝ E GeV cm 2 s 1 sr 1
10 10
IC86, 10y extrapolated
105
106
107
108
109
EΝ GeV
FIG. 1 Left: Schematic depiction of the fireball mechanism, with the characteristic time and distance scales associated with the different phases. The prompt (burst) phase is due to internal shocks in the relativistically expanding fireball, producing strong gamma-ray and X-ray emission. The afterglow arises from the cooling fireball and its interaction with the surrounding medium; it is associated with X-ray, optical and radio emission. From www.swift.ac.uk . Right: Current limits set by IceCube in its 40-string and 59-string configurations (IC40 and IC40+59) on the quasi-diffuse GRB neutrino flux, together with predictions based on recent numerical calculations taking into account uncertainties on the astrophysical parameters and those due to the limited statistics of bursts. Also shown is the extrapolated limit that one can expect from 10 years of operation with the full IceCube detector (IC86). From Hummer et al. (2012).
phase of several GRBs, triggered by the first observation of an X-ray transient emission from GRB970228 by the BeppoSAX satellite (Costa et al., 1997), subsequently confirmed their extragalactic origin by allowing more accurate localization of the source and redshift determination. Currently operating GRB missions include Swift (Gehrels et al., 2004), hosting a wide-field hard X-ray (15 keV - 350 keV) burst alert telescope (BAT) coupled to softer X-ray, ultraviolet and optical telescopes and the GBM on the Fermi Gamma-Ray Space Telescope (Atwood et al., 2009) which focuses on the high-energy (15 keV - 300 GeV) emission from GRBs.
In the standard picture, the mechanism responsible for the enormous, super-Eddington energy release (∼ 1050 − 1052 ergs) in the prompt emission and in the afterglow is the dissipation (via internal shocks, magnetic reconnection and external shocks) of bulk kinetic or Poynting flux into highly relativistic particles; see e.g. Me´sza´ros and Rees (1993) and the review by Piran (2004). The particles are accelerated to a non-thermal energy distribution via the Fermi mechanism in a relativistically expanding fireball ejected by the GRB central engine, as sketched in Fig. 1 (left panel). The accelerated electrons (and positrons) in the intense magnetic field emit non-thermal photons via synchrotron radiation and inverse Compton scattering. The plasma parameters inferred from observations to characterize GRB baryonic fireballs are such that proton acceleration to energies exceeding 1020eV is likely to be possible in these sources. Moreover, the time averaged energy output of GRBs in photons is comparable to the proton energy production rate required to produce the UHECR flux7. Therefore, the canonical baryonic fireball also suggests that GRBs are a prime candidate source for the UHECR, observed at energies E ∼ 1018 − 1020 eV (Levinson and Eichler, 1993; Vietri, 1995; Waxman, 1995).
In a baryonic outflow, the internal or external shocks accelerate protons that interact with the gamma-rays and/or other protons inside the fireball, producing charged pions and kaons that subsequently decay into HENs ( π±, K± → µ± + νµ/νµ → e± + νe/νe + νµ/νµ)8. Such neutrinos are emitted in spatial and temporal coincidence with the GRB prompt electromagnetic signal; their energy is typically in the range ∼ TeV to PeV. Neutrinos with higher (up to ∼ 1010 GeV) energy can also be emitted at the beginning of the afterglow phase, when the outflow is decelerated by external shocks with ambient material and the accelerated protons undergo interactions with the matter outside of the jet (Waxman and Bahcall, 2000). An alternative mechanism for neutrino production in fireballs suggested by Levinson and Eichler (2003) involves neutral particles that are picked up by the stream when they acquire a charge, such as a decaying neutron or, further downstream, a neutral atom that is ionized. Such a particle will be extremely energetic in the jet frame, and immediately attains an energy of a PeV. The associated neutrinos would come within an order of magnitude of that energy (∼ 100 TeV), providing a harder spectrum than the one expected from shock acceleration.
It is expected that in the next few years neutrino telescopes will be sufficiently sensitive to test and distinguish between GRB models with different physics, and to constrain the parameters of such models. The current non-detection of neutrinos by
7 The latter statement has been criticized by several authors recently, arguing that the GRB energy production rate is too small to account for the flux of ultrahigh-energy cosmic rays (UHECRs); see e.g. (Eichler et al., 2010). However, the validity of this criticism has been challenged by Waxman (2010).
8 Relevant references on these mechanisms include Eichler (1994), Paczynski and Xu (1994), Waxman and Bahcall (1997), Rachen and Me´sza´ros (1998), Alvarez-Mun˜iz et al. (2000), Me´sza´ros and Waxman (2001), Me´sza´ros and Waxman (2001), Guetta and Granot (2003), Razzaque et al. (2003a), Razzaque et al. (2003b), Dermer and Atoyan (2003), Guetta et al. (2004), Ando and Beacom (2005), Murase and Nagataki (2006), Murase et al. (2006).
7
IceCube (Abbasi et al. (2012a); see also Sec. III.B) already questions the viability of models in which ultra high energy cosmic rays are the decay products of neutrons that have escaped the fireball with high energy. The current upper limit is still consistent with the “standard” (i.e., following Waxman and Bahcall (1997)) predictions of neutrino emission from GRB fireballs (Hummer et al., 2012; Li, 2012). Given current uncertainties, significant constraints on this model will be obtained within 5-10 yrs of full IceCube operation, as can be seen from Fig. 1 (right panel).
While gamma-ray and HEN emissions from GRBs are related to the mechanisms driving the relativistic outflow, GW emission is closely connected to the central engine and hence to the progenitor of the GRB. Short-hard GRBs are thought to be driven by neutron star–neutron star or neutron star–black hole mergers9. GW detectors can ideally observe those binary systems up to a distance of ∼ 30 Mpc and ∼ 440 Mpc for initial and advanced detectors respectively (Abadie et al., 2010b). These distances coincide with the range where the HEN flux is thought to be large enough for detection with current HEN detectors. Note that these short GRBs are beamed and so is the expected HEN emission. Hence one can expect cases in which GWs will be observed from such sources without an observed GRB or HEN signal. However, an orphan afterglow (Levinson et al., 2002; Nakar and Piran, 2003), macronovae (Li and Paczynski, 1998) or radio flares (Nakar and Piran, 2011) might be observed in these cases.
Weaker GW signals are expected in any source that accelerates relativistic jets (Piran, 2013). A few mechanisms have been suggested for GW generation in long-soft GRBs, which arise during the collapse of a massive star; see e.g. Woosley and Bloom (2006) and references therein. According to the collapsar model proposed by Woosley and MacFadyen (1999) a long GRBs arises when a relativistic jet, produced by an central inner engine penetrates the stellar envelope of the collapsing star. The inner engine driving the jet can be either an accreting newborn black hole (Woosley and MacFadyen, 1999) or a newborn rapidly rotating magnetar (Usov, 1992). For accretion models, the high rotation rate required to form the accretion disk that powers the GRB may also lead to the production of GWs via bar or fragmentation instabilities in the accretion disks and also via the precession of the disks due to general relativistic effects10. Asymmetrically infalling matter produces the burst GW signals not only at the moment of the core bounce when the central density exceeds nuclear density (Kotake et al., 2006; Ott, 2009), but also at the moment of the black hole formation, followed by the subsequent ring-down phases (Ott et al., 2011). Optimistic estimates based on semi-analytical calculations suggest that the GW signals from some of these mechanisms are high enough to be visible in Advanced LIGO class detectors up to a 100 Mpc distance scale; see collective references in Kotake et al. (2012b). However, to obtain more quantitative predictions, full 3D simulations using general-relativistic magnetohydrodynamics and sophisticated neutrino transport schemes are needed. Much effort has been recently dedicated to this issue with encouraging results; see. e.g., Kuroda et al. (2012); Muller et al. (2012); Ott et al. (2012), and Kotake et al. (2012a) for a review. Unfortunately, such studies suggest a very weak GW emission and, given the fact that the long GRB population is distributed over cosmological distances, seriously challenge the prospects for long GRB detection even in the next generation of GW detectors.
“Low-luminosity GRBs” (llGRBs) are a subclass of long-soft GRBs. llGRBs are characterized by luminosities lower by a few orders of magnitude than typical gamma-ray luminosities, a smooth, single-peaked light curve, and a soft spectrum. These bursts are associated with particularly energetic type Ibc core-collapse supernovae as observed in GRB 980425/SN 1998bw (Galama et al., 1998; Kulkarni et al., 1998), GRB 031203/SN 2003lw (Malesani et al., 2004; Soderberg et al., 2004b), and GRB 060218/SN 2006aj (Campana et al., 2006; Cobb et al., 2006; Pian et al., 2006; Soderberg et al., 2006). It is interesting to note that most of the GRB-SNe associations are with llGRBs and not with regular long GRBs. Less luminous than typical long GRBs, these events are (not surprisingly) discovered at much smaller distances (SN 1998bw at redshift z = 0.0085, about 40 Mpc away from Earth; SN 2003lw at z = 0.105, and SN 2006aj at z = 0.033). Remarkably, the event rate of llGRBs per unit local volume is more than one order of magnitude larger than that of conventional long GRBs; see e.g. Coward (2005), Guetta and Della Valle (2007), Soderberg et al. (2006), Daigne and Mochkovitch (2007) and Liang et al. (2007). This makes this source population an interesting target of study from the GW+HEN point of view as well, as discussed by Razzaque et al. (2004), Murase et al. (2006), Gupta and Zhang (2007), and Wang et al. (2007).
Bromberg et al. (2011) have recently argued that, given their apparently low power, these llGRBs cannot arise from the regular collapsar model because the time needed for the jet to bore an escape channel through the host envelope would, in most reported cases, exceed the GRB duration. Rather, they may be gamma rays from shock break-out imparted to the host envelope by jets that fail to emerge (“choked jets”)11. The smooth light curve and soft spectra of these events are indeed expected from shock breakout; see Katz et al. (2010); Nakar and Sari (2012); Waxman et al. (2007). Other suggested models, which produce smooth, soft emission, include scattering of the gamma-rays off an accelerating envelope or wind material (Eichler and Levinson, 1999), or gamma rays that are released from baryon-rich jet material (dirty fireballs) only after some adiabatic loss (Mandal and Eichler, 2010). It has also been suggested that llGRBs are associated with the formation of magnetars rather than black holes, as argued
9 This mechanism has been described e.g. by Eichler et al. (1989); Kochanek and Piran (1993); Nakar (2007); Bloom et al. (2007); Lee and Ramirez-Ruiz (2007); Etienne et al. (2009).
10 See e.g. Fryer and Woosley (1998); Davies et al. (2002); Fryer et al. (2002); Kobayashi and Me´sza´ros (2003); Piro and Pfahl (2007), Romero et al. (2010), Sun et al. (2012).
11 Relevant references include MacFadyen et al. (2001); Tan et al. (2001); Campana et al. (2006); Wang et al. (2007); Waxman et al. (2007); Katz et al. (2010); Nakar and Sari (2012).
8
central engine active
central engine active
(a) 100s
(b)
(c) 250s gamma HEN GW
(a)
(e) 500s
100s
150s
(b) (d) gamma > 100 MeV
FIG. 2 Summary of the upper bounds on the duration of GRB emission processes taken into account in the total GW+HEN coincidence time window. (a) active central engine before the relativistic jet has broken out of the stellar envelope (note that recent estimates of jet propagation within a stellar envelope suggest that this phase lasts typically 10-20 sec; see Bromberg et al. (2011, 2012)); (b) active central engine with the relativistic jet broken out of the envelope; (c) delay between the onset of the precursor and the main burst; (d) duration corresponding to 90% of GeV photon emission; (e) time span of central engine activity. The top of the figure shows a schematic drawing of a plausible emission scenario. Figure taken from Baret et al. (2011).
for GRB060218 by Mazzali et al. (2006), a scenario that might give rise to somewhat longer GW signals (Corsi and Me´sza´ros, 2009; Piro and Ott, 2011).
Regardless of the question of whether or not they produce llGRBs, such “choked jet” events are interesting objects on their own, as pointed out e.g. by Eichler and Levinson (1999), Me´sza´ros and Waxman (2001) and Ando and Beacom (2005). In fact, late-time radio emission of some type Ic supernovae indeed suggests the presence of mildly relativistic outflow (Granot and Ramirez-Ruiz, 2004; Mazzali et al., 2005; Soderberg et al., 2004a, 2010) that may indicate the activity of a jet in these cases. The expected overall energy budget of choked jets is comparable to the energy budget observed in regular GRBs12. Therefore, these choked jets could produce GWs and HENs13 in a way comparable with regular GRBs; see Eichler and Levinson (1999), Me´sza´ros and Waxman (2001), Ando and Beacom (2005), Koers and Wijers (2007), Horiuchi and Ando (2008). As they are likely to be more numerous than regular GRBs (as is for example the case of llGRBs), they could be observed from nearer distances. In such a case, optimistic estimates predict potentially detectable levels of both GW and HEN signals, and an observable occurrence rate in the volume probed by current GW and HEN detectors. For example, according to Ando and Beacom (2005), an ejected mass with a kinetic energy of 3 × 1051 erg and a Lorentz factor of 3 at 10 Mpc would generate ∼30 neutrino events detected in a km3 detector. These events should be seen accompanying some specific local core collapse SNe. In this context, HEN and GW could play an interesting role in revealing the inner “choked jets” nature of these sources.
B. Bounds on the GW+HEN time delay
The possible time delay between the arrival of GWs and HENs from a given source defines the coincidence time window to apply in a multimessenger search algorithm. This window should not be too small, which could lead to the exclusion of potential emission mechanisms, nor too large, which would decrease the detection sensitivity by including non-physical coincidences. Upon detection, the difference between the times of arrival of GW and HEN signals can give us important clues about the emission mechanism. For instance detecting a HEN prior to a GW signal may indicate that the strongest GW emission from the source is not connected to the onset of the activity of the central engine that one might expect from core-collapse models.
12 Note that if indeed a choked jet produces a llGRB via shock breakout, the prompt γ-rays involve only a small fraction of the total energy (Bromberg et al., 2011; Nakar and Sari, 2012).
13 The production of HEN is closely related to the efficiency of proton acceleration inside the jet, an issue which is still debated considering that the relevant shocks in these choked GRBs are expected to be radiation-dominated (Levinson and Bromberg, 2008).
9
FIG. 3 Optical scheme of a gravitational-wave interferometric detector, consisting of two twin laser beams propagating in km-long arms oriented at 90◦ to each other. The Fabry-Pe´rot cavities enable the storage of the beams, thereby increasing by a significant factor their effective path length; the suspended, highly reflective mirrors play the role of test masses. A gravitational wave would cause a difference in the optical pathlengths in each arm, which can be inferred by measuring the interference pattern at the photodiode. Figure adapted from Y. Aso (GECo, Columbia University).
Baret et al. (2011) used model-motivated comparisons with GRB observations to derive a conservative coincidence time window for joint GW+HEN searches. Various GRB emission processes were considered, assuming that GW and HEN emission are connected to the activity of the central engine. Considered processes include prompt gamma-ray emission of GRBs, with a duration upper limit (∼ 150 s) based on BATSE observations (Paciesas et al., 1999), as well as GRB precursor activity, with an upper limit on the time difference (as compared to the onset of the main burst) of ∼ 250 s, following the analysis of Burlon et al. (2009). Further processes considered include precursor neutrino emission, as well as >∼ 100 MeV photon emission from some GRBs, as detected by Fermi LAT (Atwood et al., 2009). The authors conclude that GW and HEN signals are likely to arrive within a time window of ± 500 s, as illustrated in Fig. 2.
The time-delay between HENs and GWs could be much smaller for binary mergers which are often mentioned as the possible progenitor of short-hard GRBs. The amount of accreted/ejected matter involved in such case is very small, and the outflowing matter can expand unhindered, adding almost nothing to the time delay. A semi-analytical description of the final stage of such mergers indicates that most of the matter is accreted within 1 second (Davies et al., 2005), and numerical simulations of the mass transfer suggest time scales of milliseconds (Shibata and Taniguchi, 2008) to a few seconds maximum (Faber et al., 2006). Therefore, the GW signal is expected to arrive very close to HENs. A window of [−5, +1] seconds around the trigger time, as used for (short) GRB-GW searches in Abbott et al. (2008) and Abadie et al. (2010c), seems reasonable.
III. GW AND HEN DETECTION: STATUS AND PROSPECTS
A. Interferometric Gravitational Wave detectors
1. Detection principle and state of the art
The first generation of interferometric GW detectors included a total of six large-scale instruments. The US-based Laser Interferometer Gravitational-Wave Observatory (LIGO, see Abbott et al. (2009)) was comprised of three kilometer-scale instruments located in Livingston, Louisiana and Hanford, Washington (the latter hosted two interferometers in the same vacuum enclosure). The French-Italian project Virgo (Accadia et al., 2012) had one instrument of the same class located in Cascina near Pisa, Italy. This set of kilometer-scale instruments was complemented by a couple of detectors with more modest dimensions (several hundreds of meters): GEO (Grote, 2010), a German-British detector in operation near Hanover, Germany and the Japanese prototype CLIO (Agatsuma et al., 2010) located in the Kamioka mine.
Despite major differences in the technologies, all past and upcoming ground-based km-scale detectors measure gravitational waves through the same principle (see Fig. 3 for an illustration of the general scheme). They all sense the strain that a passing GW exerts on space-time by monitoring the differential length δ of the optical path followed by two laser beams propagating along orthogonal directions. Measurement noises (mainly the thermal noise due to the Brownian agitation of the atoms constitutive of the main optics and the shot noise due to the quantum nature of light) can be reduced to reach the level of h ≡ δ /L ∼ 10−21,