The physics of fast radio bursts

Fast radio bursts (FRBs), millisecond-duration bursts prevailing in the radio sky, are the latest large puzzle in the Universe and have been a subject of intense observational and theoretical investigations in recent years. The rapid accumulation of observational data has painted the following sketch about the physical origin of FRBs: They predominantly originate from cosmological distances, so their sources produce the most extreme coherent radio emission in the Universe; at least some, probably most, FRBs are repeating sources that do not invoke cataclysmic events; and at least some FRBs are produced by magnetars, neutron stars with the strongest magnetic fields in the Universe. Many open questions regarding the physical origin(s) and mechanism(s) of FRBs remain. This review addresses the phenomenology and possible underlying physics of FRBs. Topics include a summary of the observational data, basic plasma physics, general constraints on FRB models from the data, radiation mechanisms, source and environment models, and propagation effects, as well as FRBs as cosmological probes. Current pressing problems and future prospects are also discussed.

Figure 1

Diverse light curves of FRBs. Upper left panel: first reported FRB, FRB 20010724, or the Lorimer burst. Main panel: frequency vs arrival time of the radio burst shown as a result of dispersion. Inset: light curve of the burst after the effect of dispersion is corrected for. From [229]. Upper right panel: FRB 20110220 depiction showing frequency-dependent widths. The convention is the same as in the upper left panel, but the light curves are constructed for three central frequencies. The decaying tail is wider in lower frequencies as a result of plasma scattering along the line of sight. From [392]. Lower left panel: five additional FRBs detected with the Parkes 64-m telescope. From [57]. Lower right panel: Galactic FRB 20200428 from SGR J1935+2154 as detected by the CHIME telescope. The upper panel is the light curve, while the lower panel shows the two-dimensional frequency-time distribution (also called the dynamic spectrum) of the emission after dispersion is corrected for. From [12].

Figure 2

An example of the dynamic spectra of individual bursts from rFRB 20121102A (R1) that show a down-drifting of pulses with frequency, which is also called the sad trombone effect. Horizontal solid red bars denote RFI excision. From [135].

Figure 3

The waiting time distributions of repeating FRBs. Left panel: the case of rFRB 20121102A during the 2019 active episode. The two peaks occur at a few milliseconds and $\sim 100\mathrm{s}$. From [210]. Right panel: the case of rFRB 20201124A during an active 4-d episode in September 2021. The second peak is at $\sim 10\mathrm{s}$, suggesting an active episode. In an earlier episode of the same source in April 2021, the second peak was at $\sim 100\mathrm{s}$, suggesting that the same source can have significantly different activity levels, and hence different waiting time distributions. From [488].

Figure 4

DM-$z$ relation of 21 FRBs with known redshifts, which is an updated version of the results of [257]. The NE2001 electron density model and ${\mathrm{DM}}_{\mathrm{halo}}=30\mathrm{pc}{\mathrm{cm}}^{-3}$ are adopted. The best-fit linear regression line is plotted. rFRB 20190520B has an abnormally large ${\mathrm{DM}}_{\mathrm{host}}$ ([298]), which is marked separately.

Figure 5

Examples of polarization angle (PA) variations across individual bursts from FRBs. In each subfigure, the upper panel is the polarization angle, the middle panel is the light curve, and the lower panel is the dynamic spectrum. Upper panels: constant PA in rFRB 20121102A bursts. From [283]. Lower panels: diverse PA swing patterns in rFRB 20180301A. From [245].

Figure 6

Examples of short-term rotation measure (RM) variations in active repeating FRBs. Upper panel: irregular RM variations of rFRB 20201124A observed during a 54-d compaign with FAST. From [436]. Lower panel: surprising RM reversal from rFRB 20190520B. From [16].

Figure 7

Energy distribution of FRBs. Upper panel: FRB isotropic energy distribution among different sources showing a rough $-1.8$ power-law distribution covering at least 7 orders of magnitude. From [235]. Lower panel: energy distribution of 1652 bursts detected from rFRB 20121102A showing a bimodal distribution. From [210].

Figure 8

Sketch of the beaming factor of individual bursts (with a solid angle $\delta \mathrm{\Omega }$) and global beaming (with a solid angle of $\mathrm{\Delta }\mathrm{\Omega }$). A fan beam from a magnetospheric rotator is illustrated as an example for global beaming, but a more general geometry is possible.

Figure 9

Illustration of the shape of bunch. The radial size is approximately limited by the wavelength, i.e., $\sim \lambda$. The maximum transverse size is at least $\sim {\gamma }_e\lambda$ but can be as large as the Fresnel length $\sim \sqrt{r_0\lambda }$. Note that in order to show the geometry clearly the bunch size is greatly exaggerated. In reality, the distances from the two edges of the bunch to the NS as well as $r_0$ are similar to each other.