A fast radio burst with a low dispersion measure

Introduction

Fast radio bursts (FRBs) are observed as bright, millisecond radio transients of unknown origin (e.g. Lorimer et al., 2007; Thornton et al., 2013). FRBs are characterized by a high dispersion measure (DM) relative to the expected contribution due to the Galaxy, corresponding to a large electron column density along the line of sight. The entire population of more than 50 FRBs observed to-dateAll published FRBs are available on the FRB Catalogue; http://www.frbcat.org(Petroff et al., 2016) are believed to be extragalactic in origin. However, only one FRB source, FRB 121102, has been definitively localized to a host galaxy – a dwarf galaxy at $z=0.19273(8)$ (Spitler et al., 2014; Chatterjee et al., 2017; Tendulkar et al., 2017).

Due to their short durations ($\lesssim$ 50 ms), high flux densities ($\gtrsim$ 1 Jy), and high inferred brightness temperatures ($\gtrsim 10^{36}$ K), progenitor models involving beamed emission from compact objects are often invoked to explain FRBs. Favoured models include young, millisecond magnetars in dense progenitor environments (Metzger et al., 2017), young pulsars in nearby galaxies (Connor et al., 2016; Cordes & Wasserman, 2016), collapses of neutron stars to black holes (Falcke & Rezzolla, 2014), binary neutron star mergers (Totani, 2013), and energetic magnetars orbiting black holes (Michilli et al., 2018). However, current observations are insufficient to trace FRBs back to any of these progenitor scenarios with confidence. Ultimately, more well-localized sources with precise measurements of intrinsic flux (i.e. not convolved with an uncertain location in a telescope beam) and distance are needed to constrain theoretical models.

Shannon et al. (2018) and Macquart et al. (2018) have recently increased the FRB population by 22 sources from a large sample detected at the Australian Square Kilometre Array Pathfinder (ASKAP). These bursts, detected in a fly’s eye survey occupy a lower DM and higher fluence range than the sample from more sensitive telescopes with DMs from 114 pc cm-3 to 991 pc cm-3 and measured fluences from 34 – 420 Jy ms. If the FRB source count distribution is steep, these ultrabright events are expected to occur at a lower rate than the lower fluence events typically detected by telescopes such as Parkes.

Here we present a new FRB detected with the Parkes 64-m telescope in the High Time Resolution Universe (HTRU) South survey in 2011, FRB 110214. This FRB has one of the lowest measured FRB DMs to-date and was detected in the sidelobes of two outer beams of the Parkes multibeam receiver, implying a high intrinsic peak flux density. In § 2.1 we describe the observing system, in § 2.2 we present the burst properties, in § 3 we detail our efforts to localize FRB 110214 within the Parkes beam pattern, in § 4 we present results of follow-up at the possible locations of the FRB including searches for repeating pulses (§ 4.1) and attempts to identify a host galaxy (§ 4.2), and in § 5 we summarize these results and how they relate to the broader population of FRBs.

Observations

The discovery observations of FRB 110214 were part of the High Time Resolution Universe (HTRU) South survey conducted at the Parkes radio telescope in New South Wales, Australia (Keith et al., 2010) using the Parkes multibeam receiver (hereafter MB, Staveley-Smith et al., 1996). The MB has 13 circular feed horns, each of which forms an elliptical beam on the sky with a full-width half-maximum of approximately 14$\aas@@fstack{\prime}$4. For the HTRU survey, each beam has a separate data stream through the Berkeley Parkes Swinburne Recorder (BPSR) which records 2-bit data to disk in the form of 1024 frequency channels across 400 MHz of bandwidth from 1.182 $-$ 1.582 GHz with 64$\upmu$s time sampling (Keith et al., 2010). Only 340 MHz of the total bandwidth is used as the top 60 MHz of the band is highly contaminated with RFI from satellites.

The HTRU South survey data were collected between 2008 and 2014 and consisted of three survey regions at low, intermediate, and high Galactic latitudes with integration times of 4300s, 540s, and 270s, respectively. The HTRU high latitude data were partially processed by Thornton et al. (2013) leading to the discovery of FRBs 110220, 110626, 110723, and 120127. A full re-processing of the high latitude survey was done to search for FRBs using the heimdall single pulse search softwarehttps://sourceforge.net/projects/heimdall-astro/ for events that match the criteria of an FRB outlined in previous publications (Petroff et al., 2015; Champion et al., 2016; Bhandari et al., 2018).

In this processing all four FRBs detected by Thornton et al. (2013) were recovered as well as six others. Five new detections were reported in Champion et al. (2016): FRBs 090625, 121002, 130626, 130628, and 130729. A small fraction of the high latitude data, approximately 0.5%, were not processed at the time of the Champion et al. paper due to processing failures on the gSTAR supercomputer. In reprocessing these failed jobs, a new FRB was discovered. We describe the sixth detection in the following section.

2 FRB 110214

The spectral index of the burst in both detection beams is negative, with significantly more signal in the lower half of the band. Integrating over the bottom 50% of the Parkes bandwidth, from 1.182 $-$ 1.352 GHz, results in a higher significance detection in both beams with S/Nbeam2_lower = 17 and S/Nbeam8_lower = 7 (see Table 2) and non-detections in all other beams. The data for this observation are not bandpass corrected; BPSR sets the levels for an observation using the first ten seconds of data. However, the BPSR bandpass shapes are consistent across many observations and close to flat.

While the pulse may have an intrinsically negative spectral index either due to its emission process or Galactic scintillation, the detection of the burst in multiple beams of the receiver and stronger detections at lower frequencies leads to the conclusion that the source location might be far off-axis relative to both beams ($>$7′). Given the significantly reduced sensitivity of the telescope at off-axis positions, the intrinsic flux density of FRB 110214 must be high.

Localization of FRB 110214

To estimate the location of FRB 110214 in the beam pattern of the MB, the beam model developed for FRB 150807 by Ravi et al. (2016) was modified to reflect the case of this particular burst. Briefly, the model consists of radiation patterns for each individual beam of the MB accounting for the geometry of the dish and receiver, blockage from the focus cabin, and edge tapering for each beam calculated at each point on a 1000x1000 pixel rectangular grid covering an area of 3 deg2 centered on the central beam. Ravi et al. found this model closely approximated the observed response of the central, inner, and outer beams of the MB for bright Galactic pulsars. While any analytic model may be insufficient for the purposes of precisely pinpointing a location of the FRB on the sky, in the case of the MB it is the best approximation possible since the actual beam pattern is not fully mapped with real measurements. Nonetheless, it can provide useful information about the high probability region(s) where the burst may have originated and how bright it may have been intrinsically.

Within these constraints, three allowed regions in the beam pattern emergeContours of the localization regions are provided in a supplementary file of this manuscript., as shown in Figure 2. The three regions vary in distance from the primary detection beam; Region A is the closest, placing FRB 110214 in the first sidelobe of Beam 2. Region B is elongated away from the detection beams and lies along a line of constant S/N2:8 in the outer sidelobes. Region C is approximately 30$\arcmin$ away from the primary detection beam and would place FRB 110214 in an outer sidelobe of the MB beams.

For each region, we estimate the intrinsic flux density and fluence of FRB 110214 required to produce the observed signal in the data. We calculate the intrinsic peak flux density of FRB 110214 if it were located in each region. In each case we calculate an average and median peak flux density for each region, as estimated flux density increases rapidly with distance from the center of the primary beam and can significantly affect the average for an elongated region. We calculate a fluence $\mathcal{F}=S\times\Delta t$ where we use the observed pulse width $\Delta t$ = 1.95 ms since no substantial broadening of the pulse is expected in the case of an off-axis detection. The estimated flux densities and fluences for these regions are summarized in Table 3.

2 Region B

3 Region C

4 Spectral properties

The presence of three regions here is not particularly surprising. The detection significance of FRB 110214 is much lower than in the case of of other multi-beam FRBs such as FRB 010724 and FRB 150807 (Lorimer et al., 2007; Ravi et al., 2016), making the triangulation of the burst location on the sky more challenging. In the cases of FRB 010724 and FRB 150807 an additional localization constraint could be made using multiple sub-bands. No regions were identified in the beam pattern matching the constraints outlined in Section 3 that also satisfied the condition of non-detection in the top half of the band assuming a flat spectrum. Thus the FRB itself must have a negative spectral index either due to intrinsic emission or propagation effects along the line of sight. Only weak constraints on the spectral index or localization can be derived due to the complete lack of signal at these frequencies. All calculations for the intrinsic peak flux density and fluence of FRB 110214, therefore, only consider emission over 50% of the Parkes observing band, as this is where signal was present for analysis.

Follow-up observations

Given the low DM and the large implied fluence of 50 – 2,000 Jy ms for FRB 110214, significant time was spent on follow-up to search for repeating pulses. Repeating pulses from FRB 121102 are several orders of magnitude fainter in peak flux density (Spitler et al., 2016; Chatterjee et al., 2017), but the distance to the host galaxy of the repeating FRB is almost twice that of the estimated distance to FRB 110214 (Tendulkar et al., 2017). The burst was identified in the HTRU data in February 2016. Due to the commissioning of the Effelsberg phased array feed (PAF; Chippendale et al., 2016) at the time, the only available receiver at the Parkes focus was the single pixel H-OH receiverhttps://www.parkes.atnf.csiro.au/observing/documentation/ user_guide/pks_ug_3.html#Receiver-Fleet. The H-OH receiver was used over a 256 MHz bandwidth centered at 1.386 GHz with a beam full-width half-maximum (FWHM) of 14$\aas@@fstack{\prime}$8. A total of 62.5 hours of follow-up were conducted with the H-OH receiver with 96$\upmu$s time resolution centered at the position RA 01:20:13 Dec –49:49:47, in Region A. No single pulses were found at any DM $\leq$ 5000 pc cm-3 above S/N$>$5 over the entire bandwidth, corresponding to a flux density threshold of 0.15 Jy for a 2-ms pulse. These data were heavily affected by RFI since no multi-beam coincidence could be used for candidate rejection and we estimate that approximately 5% of the data were ruined by interference. This estimate is derived from the total number of time samples which had to be masked due to the presence of impulsive broadband RFI across all observations.

In December 2016 the MB was re-installed in the Parkes focus cabin and additional follow-up efforts were undertaken. An additional 32.7 hours of follow-up were conducted with the MB with the central beam centered on Region A such that outer beams covered the majority of Regions B and C. These data were searched with heimdall for pulses matching the same criteria as above and no pulses were found at any DM above S/N$>$5 over the entire bandwidth, corresponding to a flux density threshold of 0.13 Jy for a 2-ms pulse.

2 Identifying a host galaxy

Each region was matched to several catalogues with the aim of identifying a possible host galaxy. No sources were found in the Chandra Source Catalog or the XMM-Newton Serendipitous Source Catalogue (Evans et al., 2010; Rosen et al., 2016). The near-infrared Vista Hemisphere Survey (VHS; McMahon et al., 2013) only covers region C and two thirds of region B, however it already identifies about 250 objects as possible galaxies in those regions. As redshifts are not available for these sources, we are unable to reduce this number by considering the maximum expected redshift of the FRB.

All regions are fully covered by the 2-micron All-Sky Survey (2MASS; Skrutskie et al., 2006). It reports one known galaxy in region A at RA 01:20:13.411 DEC –49:49:47.64. This galaxy, FRL 692, is an elliptical galaxy at a redshift of $z\approx 0.025$ (Fairall, 1984). At this redshift, the IGM is expected to contribute $\sim$ 17 pc cm-3 – 25 pc cm-3 to the total DM from the Yao et al. (2017) and Ioka (2003) models, respectively. The total DM in the host galaxy (i.e. the excess from the Galaxy and the IGM) would then be 131 pc cm-3 using YMW16 and 113 pc cm-3 using NE2001 and the IGM model from Ioka (2003). We do not expect a significant contribution to the DM due to an interstellar component in elliptical galaxies (Xu & Han, 2015); however, if the source of the FRB were embedded in an ionized progenitor region like FRB 121102, such a host contribution could be feasible.

Discussion and Conclusions

However, the true flux distribution of FRBs remains unknown. A shallower distribution of $\alpha=-0.6$ has been suggested by Vedantham et al. (2016) and early results from ASKAP suggest a steeper than Euclidean distribution of $\alpha=-2.1^{+0.6}_{-0.5}$ (Shannon et al., 2018). The latter case would further favor Region A. A larger statistical sample is needed, however, to determine the true value.

An ultra-bright FRB originating in one of the outer regions is still possible, but in either case an origin closer to the detection beam is more likely. Despite significant follow-up efforts with the Parkes telescope, FRB 110214 has not been seen to repeat; most follow-up presented here was focused on searches in Region A. The low DM (and thus small implied distance) and the high intrinsic fluence make this source an excellent candidate for further follow up. A sensitive telescope such as Parkes or MeerKAT centered on or near the true sky position should be able to detect fainter repeating pulses, even if the source pulse energy distribution were steep. In our monitoring observations, no pulses were detected above a flux density of 0.15 Jy in almost 100 hours of follow-up.

There was a bright galaxy in the innermost localization region of FRB 110214, the elliptical galaxy FRL 692 at $z\simeq 0.025$. This is a similar case to the lowest DM FRB of the ASKAP sample, FRB 171020 which was found to have one bright and potentially interesting field galaxy in the error region, ESO 601$–$G036 (Mahony et al., 2018). The error region for FRB 171020 (0.38 deg2) was much larger than that of FRB 110214 (0.014 deg2) and thus the large positional uncertainty made it similarly difficult to precisely identify a host. Ultimately, more precise localization can only be achieved through the detection of repeating pulses.

If FRB 110214 is found to repeat and can be localized through its single pulses it could provide us with one of the closest FRB host galaxies available for study. A host galaxy in the local Universe, particularly if FRB 110214 is found to reside in a dwarf galaxy like FRB 121102, would provide a rich opportunity to study the structure and composition of the host that have been limited for FRB 121102 due to low signal-to-noise and the long integration times necessary to obtain spectra. Thus, although the total available observing time for follow-up of archival FRBs is limited, we argue that FRB 110214 should be one of the top priorities for monitoring campaigns in the future.

The localization regions presented here are not exact, as the MB beam model used is only an approximation. The sky around the regions identified here should also be monitored, such as with a phased array feed on a single dish, or with a number of overlapping or adjacent beams formed on the field with an interferometer. FRB surveys are sensitive to bright bursts over more of the sky, and high fluence events like FRB 110214 will be detectable in the telescope sidelobes. Thus, future surveys with multi-beam instruments should take great care to model and understand the beam and sidelobe patterns of their instruments in order to more accurately localize bright bursts on the sky.

Acknowledgements

The authors thank the anonymous referee for their comments and feedback which improved the quality of the manuscript. 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. Parts of this research were conducted by the Australian Research Council Centre of Excellence for All-sky Astrophysics (CAASTRO), through project number CE110001020 and the ARC Laureate Fellowship project FL150100148. This work was performed on the gSTAR national facility at Swinburne University of Technology. gSTAR is funded by Swinburne and the Australian Government’s Education Investment Fund. EP and LCO acknowledge funding from the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement No. 617199. SBS is supported by NSF award #1458952. BWS acknowledges funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 694745]). This research made use of data obtained from the Chandra Source Catalog, provided by the Chandra X-ray Center (CXC) as part of the Chandra Data Archive. This research made use of Astropy,http://www.astropy.org a community-developed core Python package for Astronomy (Astropy Collaboration et al., 2013; Price-Whelan et al., 2018).

References