The high implied brightness temperatures of FRBs (Tb > 1032 K) and their short intrinsic durations (milliseconds or less) require a coherent emission process from a compact region. The shortest-duration burst structures detected to date are ∼ 30 µs (Michilli et al., 2018a, Farah et al., 2018), implying an emission site of < 10 km (ignoring possible geometric and relativistic effects). What creates this coherent emission, and what is the underlying energy source? Does the same process that creates the radio burst also produce observable emission at other wavelengths? Different radiation mechanisms will produce different observed properties, and the better we can characterize the radio bursts and multi-wavelength emission, the better the chance of identifying the underlying emission mechanism. As described in Platts et al. (2018) 21 (Platts et al., 2018), one can consider the various radiation mechanisms relevant to astrophysics, as well as the necessary conditions for coherence, such as: bunched particles accelerating along electromagnetic field lines, simultaneous electron phase transitions (masers), and entangled particles collectively undergoing an atomic transition (Dicke’s superradiance).
Here we briefly consider the nature of FRB emission before giving a more general survey of progenitor models in the following section. The basic physical constraints on FRBs were investigated by Luan and Goldreich (2014). We also point the reader to Melrose (2017), who reviews the established coherent emission mechanisms in astrophysical plasmas in a general sense. The FRB emission mechanism, specifically, has been addressed, e.g., by Katz (2014), Romero et al. (2016) and Lu and Kumar (2018a). Given the possibility that there are multiple types of FRBs (Caleb et al., 2018a, Palaniswamy et al., 2018), we caution the reader that there could also be multiple types of emission mechanisms. Likewise, we caution the reader that separating intrinsic and extrinsic effects in the observed properties of FRBs adds significant uncertainty in investigating the emission mechanism, as we discuss below.
Pulsars (and magnetars) are well-established coherent radio emitters, and though the fundamental emission mechanism(s) are still a major open puzzle, they provide an important observational analogy. In other words, it would already be helpful to know if FRB emission is from a similar physical mechanism as pulsars, even if that physical mechanism is still not well understood. Firstly, it is important to note that radio pulsars show a wide range of observational properties that are similar to those seen in FRBs. Like FRBs, pulsars have a variety of circular and linear polarization fractions, pulse widths, pulse structure, and spectra. In canonical rotation-powered pulsars, emission is believed to originate a few tens to hundreds of kilometers above the neutron star polar caps (e.g. Hassall et al., 2012). Some neutron stars, of which the Crab is the best-studied example, also show so-called ‘giant’ pulses, which are brighter, shorter in duration, and may originate from a different region of the magnetosphere (Hankins et al., 2016). Magnetars also emit radio pulses (Camilo et al., 2006); their emission can be highly erratic, showing radio emission at a wide range of rotational phases, and with an average pulse profile that changes with time. Hence, there are at least three types of radio emission seen from magnetised neutron stars. Given the wide range of radio emission phenomena detected from Galactic neutron stars, it seems plausible that FRBs could be an even more extreme manifestation of one of these processes, or perhaps a fourth type of neutron star radio emission. Unfortunately, however, the mechanisms responsible for creating neutron star pulsed radio emission are still not well understood. Nonetheless, if we assume that FRBs originate in neutron star magnetospheres, or their near vicinity, the detection of a multi-wavelength counterpart (or lack thereof) could inform whether the bursts are rotationally or magnetically powered (Lyutikov and Lorimer, 2016).
FRBs and pulsar pulses have peak flux densities ∼ 1 Jy but the ∼ 106 times greater distance of the FRB population implies a ∼ 1012 times greater luminosity (assuming the same degree of beaming), which corresponds to burst energies
 |
(25) |
where δΩ is the solid angle of the emission (steradians),
DGpc the luminosity distance (Gpc),
ν the
fluence (Jy ms), and ΔνGHz the emission bandwidth
(GHz). All parameters are considered in the source frame. The
magnetospheres of canonical pulsars may have difficulty in providing
this much energy (e.g.
Cordes
and Wasserman, 2016).
FRBs might be powered instead by the strong ∼ 1014
− 1015 G magnetic fields in magnetars
(Popov
and Postnov, 2013,
Beloborodov, 2017).
A variety of works have considered whether FRBs could originate from rotationally powered super-giant pulses from rapidly spinning, highly magnetized young pulsars (e.g., Cordes and Wasserman, 2016, Lyutikov et al., 2016). Because the available spin-down luminosity scales with the magnetic field strength B and rotational period P as B2 / P4 (Lorimer and Kramer, 2012), it is conceivable that such a source could power giant-pulses that are orders of magnitude brighter than those seen from the Crab pulsar. Importantly, we have not yet seen a cut-off in the brightness distribution of Crab giant pulses. Beaming is also a critical consideration, and it is possible that the Crab giant pulses would appear substantially brighter if viewed from another angle. In summary, the maximum possible luminosity of radio emission from a neutron star is not well established.
Furthermore, if a significant fraction of the observed DM can be contributed from a surrounding supernova remnant, then FRBs may be closer than we would otherwise infer (Connor et al., 2016b), thereby reducing the energy requirement. However, the precise localization of FRB 121102 led to a redshift measurement that places it firmly at z = 0.193 (dL ∼ 1 Gpc) (Tendulkar et al., 2017). Lyutikov (2017) argue that this large distance rules out rotation-powered super-giant pulses like those from the Crab. The Crab pulsar is a singular source in our Galaxy, however, and we do not know whether its giant pulses are at the limit of what a neutron star can produce. Perhaps with more fortuitous beaming and a younger, more highly magnetized neutron star, the energy requirements imposed by FRB 121102 can be met with giant-pulse-like emission.
Magnetically powered bursts from neutron stars have also been considered in the literature. Flares from magnetars were first proposed by Popov and Postnov (2010) and Popov and Postnov (2013). A flaring magnetar model for FRB 121102 was proposed by Beloborodov (2017), and was partially motivated in order to explain the source’s compact, persistent radio counterpart (Chatterjee et al., 2017, Marcote et al., 2017). In this model, the FRBs are from a giga-Hertz maser and originate in shocks far from the neutron star itself.
Hessels et al. (2018) show that FRB 121102 bursts have complex time-frequency structures. This includes sub-bursts (∼ 0.5−1 ms wide) displaying finite bandwidths of 100−400 MHz at 1.4 GHz. Hessels et al. (2018) also find that the sub-bursts have characteristic frequencies that typically drift lower at later times in the total burst envelope, by ∼ 200 MHz/ms in the 1.1−1.7 GHz band. This differs from typical pulsars and radio-emitting magnetars, which have smooth, wide-band spectra (even in their single pulses, e.g., Kramer et al., 2003, Jankowski et al., 2018). In pulsars, the only narrow-band modulation seen is from diffractive interstellar scintillation, which is augmented in some cases by constructive and destructive interference from multiple imaging due to interstellar refraction.
The spectral behaviour of FRB 121102 may be intrinsic to the emission process. It could also be due to post-emission propagation processes, or some combination of intrinsic and extrinsic effects. Dynamic spectral structures are seen in other astrophysical sources that emit short-timescale radio bursts: e.g., the Sun (e.g., Kaneda et al., 2015), flare stars (e.g., Osten and Bastian, 2006, Osten and Bastian, 2008), and Solar System planets (e.g. Zarka, 1992, Ryabov et al., 2014). Time-frequency drifts, qualitatively similar to those seen from FRB 121102 and the CHIME/FRB repeater FRB 180814.J0422+73, have been detected from such sources. These drifts occur when the emission regions moves upwards to regions with lower plasma frequencies or cyclotron frequencies (these, in turn, are tied to the observed electromagnetic frequency). Fine time-frequency structure in the radio emission is related to variations in the particle density (e.g., Treumann, 2006). If we extrapolate similar processes to FRBs, it suggests that FRB 121102's (and FRB 180814.J0422+73's) emission could originate from cyclotron or synchrotron maser emission (Lyubarsky, 2014, Beloborodov, 2017, Waxman, 2017), in which case relatively narrow-band emission in the GHz range could be expected. Antenna mechanisms involving curvature radiation from charge bunches have also been considered (Cordes and Wasserman, 2016, Lu and Kumar, 2018a). However, it is not clear if the energetics can be satisfied or how time-frequency structure is produced in this case.
In the 100 MHz to 100 GHz radio frequency range, the Crab pulsar shows a remarkable range of emission features. The Crab’s rich and diverse phenomenology is thus potentially relevant to understanding FRB emission. For example, as discussed in Hessels et al. (2018), the giant pulse emission in the Crab pulsar's high-frequency interpulse (HFIP; Hankins et al., 2016), which is seen above ∼ 4 GHz radio frequencies, provides an interesting observational comparison to the burst features seen in FRB 121102. Note that the polarimetric and time-frequency properties of the HFIPs are highly specific and differ significantly from those of the main giant pulses (MP; Jessner et al., 2010, Hankins et al., 2016).
The Crab's HFIP spectra display periodic bands of increased brightness (Hankins and Eilek, 2007) with separations Δν that scale with frequency (Δν / ν = constant). In comparison, the drift rates in FRB 121102 potentially show a similar scaling (see Figure 3 of Hessels et al., 2018) but a larger sample is needed to be conclusive. While the Crab HFIPs are microseconds in duration, the burst envelopes of FRB 121102 are typically milliseconds – though with underlying ∼ 30 µs structure clearly visible in some cases (Michilli et al., 2018a). Searches for even finer-timescale structure in FRB 121102 should thus continue, using high observing frequencies to avoid smearing from scattering.
Lastly, the polarization angle of the ∼ 100% linearly polarized radiation from FRB 121102 at 4−8 GHz appears constant across individual bursts and is stable between bursts (Michilli et al., 2018a, Gajjar et al., 2018). This phenomenology is also similar to that of the Crab HFIPs, which are ∼ 80−100% linearly polarized and have a constant polarization position angle across the duration of each pulse – as well as between HFIPs that span ∼ 3% of the pulsar’s rotational phase (see Fig. 14 of Hankins et al., 2016). Lastly, the Crab HFIPs typically show no circular polarization, and thus far no circularly polarised emission has been detected from FRB 121102.
For now, the emission mechanism responsible for the coherent radio emission of FRBs remains a mystery. As with pulsars, however, regardless of whether we eventually understand the detailed physical emission it should still be possible to identify the progenitors of FRBs and to use them as astrophysical probes.