A Population of Fast Radio Bursts at Cosmological Distances

See allHide authors and affiliations

Science  05 Jul 2013:
Vol. 341, Issue 6141, pp. 53-56
DOI: 10.1126/science.1236789

This article has a correction. Please see:

Mysterious Radio Bursts

It has been uncertain whether single, short, and bright bursts of radio emission that have been observed are celestial or terrestrial in origin. Thornton et al. (p. 53; see the Perspective by Cordes) report the detection of four nonrepeating radio transient events with millisecond duration in data from the 64-meter Parkes radio telescope in Australia. The properties of these radio bursts indicate that they had their origin outside our galaxy, but it is not possible to tell what caused them. Because the intergalactic medium affects the characteristics of the bursts, it will be possible to use them to study its properties.


Searches for transient astrophysical sources often reveal unexpected classes of objects that are useful physical laboratories. In a recent survey for pulsars and fast transients, we have uncovered four millisecond-duration radio transients all more than 40° from the Galactic plane. The bursts’ properties indicate that they are of celestial rather than terrestrial origin. Host galaxy and intergalactic medium models suggest that they have cosmological redshifts of 0.5 to 1 and distances of up to 3 gigaparsecs. No temporally coincident x- or gamma-ray signature was identified in association with the bursts. Characterization of the source population and identification of host galaxies offers an opportunity to determine the baryonic content of the universe.

The four fast radio bursts (FRBs) (Fig. 1) reported here were detected in the high Galactic latitude region of the High Time Resolution Universe (HTRU) survey (1), which was designed to detect short-time-scale radio transients and pulsars (Galactic pulsed radio sources). The survey uses the 64-m Parkes radio telescope and its 13-beam receiver to acquire data across a bandwidth of 400 MHz centered at 1.382 GHz (table S1). We measured minimum fluences for the FRBs of F = 0.6 to 8.0 Jy ms (1 Jy = 10–26 W m–2 Hz–1) (2). At cosmological distances, this indicates that they are more luminous than bursts from any known transient radio source (3). Follow-up observations at the original beam positions have not detected any repeat events, indicating that the FRBs are likely cataclysmic in nature.

Fig. 1 The frequency-integrated flux densities for the four FRBs.

The time resolutions match the level of dispersive smearing in the central frequency channel (0.8, 0.6, 0.9, and 0.5 ms, respectively).

Candidate extragalactic bursts have previously been reported with varying degrees of plausibility (47), along with a suggestion that FRB 010724 (the “Lorimer burst”) is similar to other signals that may be of local origin (8, 9). To be consistent with a celestial origin, FRBs should exhibit certain pulse properties. In particular, observations of radio pulsars in the Milky Way (MW) have confirmed that radio emission is delayed by propagation through the ionized interstellar medium (ISM), which can be considered a cold plasma. This delay has a power law dependence of δtDMν2 and a typical frequency-dependent width of Wν4. The dispersion measure (DM) is related to the integrated column density of free electrons along the line of sight to the source and is a proxy for distance. The frequency-dependent pulse broadening occurs as an astrophysical pulse is scattered by an inhomogeneous turbulent medium, causing a characteristic exponential tail. Parameterizing the frequency dependence of δt and W as α and β, respectively, we measured α=2.003±0.006 and β=4.0±0.4 for FRB 110220 (Table 1 and Fig. 2), as expected for propagation through a cold plasma. Although FRB 110703 shows no evidence of scattering, we determined α=2.000±0.006. The other FRBs do not have sufficient signal-to-noise ratios (SNRs) to yield astrophysically interesting constraints for either parameter and show no evidence of scattering.

Table 1 Parameters for the four FRBs.

The position given is the center of the gain pattern of the beam in which the FRB was detected (half-power beam width ~ 14 arc min). The UTC corresponds to the arrival time at 1581.804688 MHz. The DM uncertainties depend not only on SNR but also on whether α and β are assumed (Embedded Image; no scattering) or fit for; where fitted, α and β are given. The comoving distance was calculated by using DMHost = 100 cm−3 pc (in the rest frame of the host) and a standard, flat-universe ΛCDM cosmology, which describes the expansion of the universe with baryonic and dark matter and dark energy [H0 = 71 km s−1 Mpc−1, ΩM = 0.27, ΩΛ = 0.73; H0 is the Hubble constant and ΩM and ΩΛ are fractions of the critical density of matter and dark energy, respectively (29)]. α and β are from a series of fits using intrinsic pulse widths of 0.87 to 3.5 ms; the uncertainties reflect the spread of values obtained (2). The observed widths are shown; FRBs 110627, 110703, and 120127 are limited by the temporal resolution due to dispersion smearing. The energy released is calculated for the observing band in the rest frame of the source (2).

View this table:
Fig. 2 A dynamic spectrum showing the frequency-dependent delay of FRB 110220.

Time is measured relative to the time of arrival in the highest frequency channel. For clarity we have integrated 30 time samples, corresponding to the dispersion smearing in the lowest frequency channel. (Inset) The top, middle, and bottom 25-MHz-wide dedispersed subband used in the pulse-fitting analysis (2); the peaks of the pulses are aligned to time = 0. The data are shown as solid gray lines and the best-fit profiles by dashed black lines.

Our FRBs were detected with DMs in the range from 553 to 1103 cm−3 pc. Their high Galactic latitudes (|b|>41°, Table 1) correspond to lines of sight through the low column density Galactic ISM corresponding to just 3 to 6% of the DM measured (10). These small Galactic DM contributions are highly supportive of an extragalactic origin and are substantially smaller fractions than those of previously reported bursts, which were 15% of DM = 375 cm−3 pc for FRB 010724 (4) and 70% of DM = 746 cm−3 pc for FRB 010621 (5).

The non-Galactic DM contribution, DME, is the sum of two components: the intergalactic medium (IGM; DMIGM) and a possible host galaxy (DMHost). The intervening medium could be purely intergalactic and could also include a contribution from an intervening galaxy. Two options are considered according to the proximity of the source to the center of a host galaxy.

If located at the center of a galaxy, this may be a highly dispersive region; for example, lines of sight passing through the central regions of Milky Way–like galaxies could lead to DMs in excess of 700 cm−3 pc in the central ~100 pc (11), independent of the line-of-sight inclination. In this case, DME is dominated by DMHost and requires FRBs to be emitted by an unknown mechanism in the central region, possibly associated with the supermassive black hole located there.

If outside a central region, then elliptical host galaxies (which are expected to have a low electron density) will not contribute to DME substantially, and DMHost for a spiral galaxy will only contribute substantially to DME if viewed close to edge-on [inclination, i>87° for DM>700 cm3 pc; probability(i>87°)0.05]. The chance of all four FRBs coming from edge-on spiral galaxies is therefore negligible (10−6). Consequently, if the sources are not located in a galactic center, DMHost would likely be small, and DMIGM dominates. Assuming an IGM free-electron distribution, which takes into account cosmological redshift and assumes a universal ionization fraction of 1 (12, 13), the sources are inferred to be at redshifts z = 0.45 to 0.96, corresponding to comoving distances of 1.7 to 3.2 Gpc (Table 1).

In principle, pulse scatter-broadening measurements can constrain the location and strength of an intervening scattering screen (14). FRBs 110627, 110703, and 120127 are too weak to enable the determination of any scattering; however, FRB 110220 exhibits an exponential scattering tail (Fig. 1). There are at least two possible sources and locations for the responsible scattering screens: a host galaxy or the IGM. It is possible that both contribute to varying degrees.

For screen-source, Dsrc, and screen-observer, Dobs, distances, the magnitude of the pulse broadening resulting from scattering is multiplied by the factor DsrcDobs(Dsrc+Dobs)2. For a screen and source located in a distant galaxy, this effect probably requires the source to be in a high-scattering region, for example, a galactic center.

The second possibility is scattering because of turbulence in the ionized IGM, unassociated with any galaxy. There is a weakly constrained empirical relationship between DM and measured scattering for pulsars in the MW. If applicable to the IGM, then the observed scattering implies DMIGM>100 cm3pc (2, 15). With use of the aforementioned model of the ionized IGM, this DM equates to z>0.11 (2, 12, 13). The probability of an intervening galaxy located along the line of sight within z ≈ 1 is ≤0.05 (16). Such a galaxy could be a source of scattering and dispersion, but the magnitude would be subject to the same inclination dependence as described for a source located in the disk of a spiral galaxy.

It is important to be sure that FRBs are not a terrestrial source of interference. Observations at Parkes have previously shown swept frequency pulses of terrestrial origin, dubbed “perytons.” These are symmetric W > 20 ms pulses, which imperfectly mimic a dispersive sweep (2, 8). Although perytons peak in apparent DM near 375 cm3pc (range from ~200 to 420 cm–3 pc), close to that of FRB 010724, the FRBs presented here have much higher and randomly distributed DMs. Three of these FRBs are factors of >3 narrower than any documented peryton. Last, the characteristic scattering shape and strong dispersion delay adherence of FRB 110220 make a case for cold plasma propagation.

The Sun is known to emit frequency-swept radio bursts at 1 to 3 GHz [typeIIIdm (17)]. These bursts have typical widths of 0.2 to 10 s and positive frequency sweeps, entirely inconsistent with measurements of W and α for the FRBs. Whereas FRB 110220 was separated from the Sun by 5.6°, FRB 110703 was detected at night and the others so far from the Sun that any solar radiation should have appeared in multiple beams. These FRBs were only detected in a single beam; it is therefore unlikely they are of solar origin.

Uncertainty in the true position of the FRBs within the frequency-dependent gain pattern of the telescope makes inferring a spectral index, and hence flux densities outside the observing band, difficult. A likely off-axis position changes the intrinsic spectral index substantially. The spectral energy distribution across the band in FRB 110220 is characterized by bright bands ~100 MHz wide (Fig. 2); the SNRs are too low in the other three FRBs to quantify this behavior (2). Similar spectral characteristics are commonly observed in the emission of high-|b| pulsars.

With four FRBs, it is possible to calculate an approximate event rate. The high-latitude HTRU survey region is 24% complete, resulting in 4500 square degrees observed for 270 s. This corresponds to an FRB rate of RFRB(F3Jy ms)=1.00.5+0.6×104sky1day1, where the 1-σ uncertainty assumes Poissonian statistics. The MW foreground would reduce this rate, with increased sky temperature, scattering, and dispersion for surveys close to the Galactic plane. In the absence of these conditions, our rate implies that 177+9, 73+4, and 125+6 FRBs should be found in the completed high- and medium-latitude parts of the HTRU (1) and Parkes multibeam pulsar (PMPS) surveys (18).

One candidate FRB with DM>DMMW has been detected in the PMPS [|b|<5° (5, 19)]. This burst could be explained by neutron star emission, given a small scale-height error; however, observations have not detected any repetition. No excess-DM FRBs were detected in a burst search of the first 23% of the medium-latitude HTRU survey [|b|<15° (20)].

The event rate originally suggested for FRB 010724, R010724=225sky1day1 (4), is consistent with our event rate given a Euclidean universe and a population with distance-independent intrinsic luminosities (source count, NF32) yielding RFRB(F3Jyms)102RFRB(F010724150 Jy ms).

There are no known transients detected at gamma-ray, x-ray, or optical wavelengths or gravitational wave triggers that can be temporally associated with any FRBs. In particular there is no known gamma-ray burst (GRB) with a coincident position on a time scale commensurate with previous tentative detections of short-duration radio emission (6). GRBs have highly beamed gamma-ray emission (21), and, if FRBs are associated with them, the radio emission must be beamed differently. By using the distances in Table 1, we found that the comoving volume contains ~109 late-type galaxies (22), and the FRB rate is therefore RFRB103year1 per galaxy. RFRB is thus inconsistent with RGRB106year1 per galaxy, even when beaming of emission is accounted for (21). Soft gamma-ray repeaters (SGRs) undergo giant bursts at a rate consistent with FRBs (23), and the energy within our band is well within the budget of the few known SGR giant burst cases (24).

Another postulated source class is the interaction of the magnetic fields of two coalescing neutron stars (25). However, the large implied FRB luminosities indicate that coalescing neutron stars may not be responsible for FRBs. Furthermore, RFRB is substantially higher than the predicted rate for neutron star mergers. Black hole evaporation has also been postulated as a source of FRBs; however, the predicted luminosity within our observing band far exceeds the energy budget of an evaporation event (26).

The core-collapse supernova (ccSN) rate of RccSN102year1 per galaxy (27) is consistent with RFRB. There is no known mechanism to generate an FRB from a lone ccSN. It may, however, be possible that a ccSN with an orbiting neutron star can produce millisecond-duration radio bursts during the interaction of the ccSN explosion and the magnetic field of the neutron star (28), although the need for an orbiting neutron star will make these rarer.

As extragalactic sources, FRBs represent a probe of the ionized IGM. Real-time detections and immediate follow-up at other wavelengths may identify a host galaxy with an independent redshift measurement, thus enabling the IGM baryon content to be determined (12). Even without host identifications, further bright FRB detections will be a unique probe of the magneto-ionic properties of the IGM.

Supplementary Materials

Materials and Methods

Figs. S1 to S4

Table S1

References (30, 31)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. The Lorimer burst is designated as FRB 010724; this date is a correction to that in the original paper.
  3. Acknowledgments: This research has made use of the NASA/IPAC (Infrared Processing and Analysis Center) Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with NASA. This research has made use of data obtained from the High Energy Astrophysics Science Archive Research Center, provided by NASA’s Goddard Space Flight Center. Part of this research was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA. The Parkes radio telescope is part of the Australia Telescope National Facility, which is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO. Part of this research was conducted because of the support of CAASTRO through project number CE110001020. D.T. gratefully acknowledges the support of the Science and Technology Facilities Council and CSIRO Astronomy and Space Science in his Ph.D. studentship. N.D.R.B. is supported by a Curtin Research Fellowship (CRF12228).
View Abstract

Stay Connected to Science

Navigate This Article