frbs and the milky way - wordpress.com

8
MNRAS 000, 000–000 (2020) Preprint 14 December 2020 Compiled using MNRAS L A T E X style file v3.0 FRBs and the Milky Way N. Gohar 1 , C. Flynn 2,3 , et al. 1 Government College University, Lahore, Pakistan, 2 Centre for Astrophysics and Supercomputing, Swinburne University of Technology, Mail H30, PO Box 218, VIC 3122, Australia 3 ARC Centre of Excellence for Gravitational Wave Discovery (OzGrav), Australia Accepted XXX. Received YYY; in original form ZZZ ABSTRACT Fast Radio Bursts (FRBs) are highly energetic transient events, with durations of order of microseconds to milliseconds, recently shown to lie at cosmological distances. Recently, an FRB-like event was seen from the Milky Way magnetar SGR 1935+2154 by CHIME and STARE2, radio telescopes that have ongoing FRB search programs. Finding further events like that from SGR1935+2154 from other magnetars in the Milky Way will be strongly affected by the turbulent interstellar medium, their intrinsic energy distribution and to some extent their spatial locations within the plane of the Milky Way disk. Here, we examine these on searches for more such FRB-like events, using two models for the distribution of electrons in the ISM in order to estimate the dispersion measure and pulse scattering of mock events, and a range of models for the spatial distribution and luminosity functions, evaluating what fraction of FRB-like events in the Milky Way could be detected by all-sky experiments such as STARE-2. In all the models examined, only a fraction of burst events are detectible, mainly due to the scattering effects of the ISM. A similar experiment to CHIME-2, operating in the Southern Hemisphere, could increase the detection rate significantly. 1 INTRODUCTION Fast Radio Bursts (FRBs) are highly energetic transients, with durations of a few 10s of microseconds to 10s of mil- liseconds, and originating at cosmological distances. Since their discovery in archival data in 2007 (Lorimer et al. 2007), FRBs have been found at telescopes around the world, ini- tially with single dishes and in the last few years with radio telescope arrays, so that they have been localised to host galaxies (Macquart et al. 2019; Ravi et al. 2019). The primary distinguishing feature of FRBs is that the pulse arrival time is a function of frequency, having been dispersed in propagation through an intervening ionised medium. The dispersion measure (DM) characterises this propagation delay. DMs for FRBs far exceed the DM ex- pected for sources lying in the Milky Way, such that it was recognised early (Lorimer et al. 2007) they could be unique probes of the ionised Intergalactic Medium (IGM). A rela- tionship between DM and source luminosity was found by (Shannon et al. 2018), supported this case, and recently, the cosmic density of baryons Ω b , was measured for the first time using FRBs localised to host galaxies with follow-up redshifts (Macquart et al. 2019). Just over 100 FRBs have been published to date (Petroff et al. 2017). They can be broadly divided into repeating and non-repeating classes (which may partially or wholly over- lap), and there is evidence that the pulse profiles and pulse morphology, as well as the pulse durations of the classes differ somewhat (Cui et al. 2021). One-off FRBs require automatic detection algorithms operating live in order to capture the raw radio data and to localise them accurately to a host galaxy, while repeaters can be localised to hosts more readily once they have been es- tablished to repeat. The first such voltage capture via a live trigger was achieved at the Molonglo Radio Telescope (Farah et al. 2017), and is in operation there as well as at ASKAP (Bhandari et al. 2019), CHIME (Scholz & Chime/Frb Col- laboration 2020), DSA-10 (Ravi et al. 2019). All FRBs discovered until early 2020 are extra-galactic (based on their known association with a host galaxy, or their DMs) with the nearest FRB having been localised to a dwarf galaxy at a distance of approximately 120 Mpc (FRB171020) The search was successful in 2020, when an FRB-like event was found from a source, a magnetar, located in the Milky Way (Bochenek et al. 2020b; Scholz & Chime/Frb Collaboration 2020). This transient, detected by CHIME and STARE2, had a DM of 332.7 pc cm -3 and an intrin- sic width of 0.61 ms. The source was localised to a known Galactic magnetar (SGR1935), and with an estimated flu- ence of 300 kJyms, the would have been visible to a distance of a few Mpc, i.e. within the volume formed by galaxies in the Local Group. The event would have been considered an FRB were it detected in high sensitivity surveys (such as those conducted at Parkes which has a flux density limit of 0.5 Jyms – the most sensitive in the Southern Hemisphere). Located at a latitude of +30 degrees, STARE-2 has a an approximately 20×20 cm antenna with a 60 degree field of view, 300 MHz of spectrum centered at 1.4 GHz, and c 2020 The Authors

Upload: others

Post on 08-Jun-2022

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: FRBs and the Milky Way - WordPress.com

MNRAS 000, 000–000 (2020) Preprint 14 December 2020 Compiled using MNRAS LATEX style file v3.0

FRBs and the Milky Way

N. Gohar1, C. Flynn2,3, et al.

1Government College University, Lahore, Pakistan,2Centre for Astrophysics and Supercomputing, Swinburne University of Technology, Mail H30, PO Box 218, VIC 3122, Australia3ARC Centre of Excellence for Gravitational Wave Discovery (OzGrav), Australia

Accepted XXX. Received YYY; in original form ZZZ

ABSTRACTFast Radio Bursts (FRBs) are highly energetic transient events, with durations oforder of microseconds to milliseconds, recently shown to lie at cosmological distances.Recently, an FRB-like event was seen from the Milky Way magnetar SGR 1935+2154by CHIME and STARE2, radio telescopes that have ongoing FRB search programs.Finding further events like that from SGR1935+2154 from other magnetars in theMilky Way will be strongly affected by the turbulent interstellar medium, their intrinsicenergy distribution and to some extent their spatial locations within the plane of theMilky Way disk. Here, we examine these on searches for more such FRB-like events,using two models for the distribution of electrons in the ISM in order to estimate thedispersion measure and pulse scattering of mock events, and a range of models for thespatial distribution and luminosity functions, evaluating what fraction of FRB-likeevents in the Milky Way could be detected by all-sky experiments such as STARE-2.In all the models examined, only a fraction of burst events are detectible, mainly dueto the scattering effects of the ISM. A similar experiment to CHIME-2, operating inthe Southern Hemisphere, could increase the detection rate significantly.

1 INTRODUCTION

Fast Radio Bursts (FRBs) are highly energetic transients,with durations of a few 10s of microseconds to 10s of mil-liseconds, and originating at cosmological distances. Sincetheir discovery in archival data in 2007 (Lorimer et al. 2007),FRBs have been found at telescopes around the world, ini-tially with single dishes and in the last few years with radiotelescope arrays, so that they have been localised to hostgalaxies (Macquart et al. 2019; Ravi et al. 2019).

The primary distinguishing feature of FRBs is that thepulse arrival time is a function of frequency, having beendispersed in propagation through an intervening ionisedmedium. The dispersion measure (DM) characterises thispropagation delay. DMs for FRBs far exceed the DM ex-pected for sources lying in the Milky Way, such that it wasrecognised early (Lorimer et al. 2007) they could be uniqueprobes of the ionised Intergalactic Medium (IGM). A rela-tionship between DM and source luminosity was found by(Shannon et al. 2018), supported this case, and recently, thecosmic density of baryons Ωb, was measured for the firsttime using FRBs localised to host galaxies with follow-upredshifts (Macquart et al. 2019).

Just over 100 FRBs have been published to date (Petroffet al. 2017). They can be broadly divided into repeating andnon-repeating classes (which may partially or wholly over-lap), and there is evidence that the pulse profiles and pulsemorphology, as well as the pulse durations of the classesdiffer somewhat (Cui et al. 2021).

One-off FRBs require automatic detection algorithms

operating live in order to capture the raw radio data and tolocalise them accurately to a host galaxy, while repeaters canbe localised to hosts more readily once they have been es-tablished to repeat. The first such voltage capture via a livetrigger was achieved at the Molonglo Radio Telescope (Farahet al. 2017), and is in operation there as well as at ASKAP(Bhandari et al. 2019), CHIME (Scholz & Chime/Frb Col-laboration 2020), DSA-10 (Ravi et al. 2019).

All FRBs discovered until early 2020 are extra-galactic(based on their known association with a host galaxy, ortheir DMs) with the nearest FRB having been localisedto a dwarf galaxy at a distance of approximately 120 Mpc(FRB171020)

The search was successful in 2020, when an FRB-likeevent was found from a source, a magnetar, located in theMilky Way (Bochenek et al. 2020b; Scholz & Chime/FrbCollaboration 2020). This transient, detected by CHIMEand STARE2, had a DM of 332.7 pc cm−3 and an intrin-sic width of 0.61 ms. The source was localised to a knownGalactic magnetar (SGR1935), and with an estimated flu-ence of 300 kJyms, the would have been visible to a distanceof a few Mpc, i.e. within the volume formed by galaxies inthe Local Group. The event would have been considered anFRB were it detected in high sensitivity surveys (such asthose conducted at Parkes which has a flux density limit of0.5 Jyms – the most sensitive in the Southern Hemisphere).

Located at a latitude of ≈ +30 degrees, STARE-2 has aan approximately 20×20 cm antenna with a 60 degree fieldof view, 300 MHz of spectrum centered at 1.4 GHz, and

c© 2020 The Authors

Page 2: FRBs and the Milky Way - WordPress.com

2 Gohar et al.

5 operating sites around Southern California, in order touse coincidencing to remove local sources of radio frequencyinterference (RFI) and search for FRBs. The limiting sensi-tivity is approximately 300 kJy (to a 1 ms burst). Its surveyfootprint is the Northern celestial sphere, as the 3 dB south-ern limit is at about the celestial equator.

The discovery by CHIME-2 of FRB-like pulses frommagnetars supports the case that they are closely related.Searching for more such bursts will be strongly affected bythe properties of the Interstellar Medium (ISM), since, as forpulsars, pulses travelling through the (turbulent) ISM arestrongly scattered as a function of DM (Bhat et al. 2004).

In this paper we investigate the effects of the ISM ofthe detectability of SGR1935-like events in the MilkWay.We use Monte-Carlo simulations, setting up various distri-butions of FRB sources in the Milky Way disk, following ei-ther a smooth exponential distribution or sources added inproportion to the ionised matter. We model the ionised ISMwith two publicly available codes (NE2001 and YMW16).This allows us to estimate the DM to the events. We modelthe scattering of the events using the Bhat et al. (2004)study of pulse scattering for pulsars as a function of theirDMs. We estimate what fraction of these Milky Way FRBsare detectable with a STARE2 like instrument in both hemi-spheres, taking into account its sky coverage, single pulsesensitivity, instrumental time resolution and the DM smear-ing due to the frequency channelisation. As expected, MilkyWay FRB-like pulses are strongly affected by the ISM, andto some extent by their spatial distribution. More than halfof the events are typically too scattered to be detectable at1.4 GHz, for all the models examined. Most events (≈ 75%)nevertheless are in the Southern Hemisphere, and a STARE-2 like experiment operating in the Southern Hemispherecould significantly boost the discovery rate of such events.

The paper is laid out as follows. In section 2, we de-scribe parameters of the STARE-2 experiment relevant toour simulations. In Section 3, we describe how we modelFRB progenitor sources in the Milky Way and the effects onthe dispersion measure and scattering of the pulses due tothe ISM. In Section 3.3, we give our results for a 2-D distri-bution of FRBs in the disk mid-plane, for which ISM effectsare maximised. In Section 3.4 we generalise the model to3-D distributions, where a vertical scaleheight for the FRBsources is taken into account. In section 4 we assume anintrinsic energy distribution for the FRBs. In section 5 wecompare radio pulses from magnetars in order to prove thevalidity of our simulations, and in Section 6 we describe howsearches might be improved in future, based on the modelingresults. We conclude in Section 7.

2 STARE2

The Survey for Transient Astronomical Radio Emission 2(STARE2) is an instrument aimed for detecting FRBs inthe Milky Way out to 7 kpc. Bochenek et al. (2020a) statethat the field of view is 3.6 steradians and is sensitive totransients of 1 ms and above 300 kJy. STARE2 has a timeresolution of 65.536 microseconds and frequency resolutionof 122.07 kHz. In April, STARE2 was successful in detectingan FRB associated with the magnetar SGR 1935+2154. Theaverage S/N measured for the event (ST 200428A) was ≈ 19.

We apply the sesnitivity, frequency and time resolutionof STARE2 to our models, to estimate the fraction of theMilky Way from which FRBs can be detected.

3 MODELLING FRB SOURCES IN THEMILKY WAY

3.1 Spatial distributions for the FRBs

In order to evaluate the effects of the ISM on FRBs, we setup a 2-D model of the FRB distributions in the Milky Wayfor the following two cases:

• The FRBs are exponentially distributed, similar to thegeneral stellar population in the Galactic disk (see Fig 1)• The FRBs are spirally distributed in the Milky Way,

following the electron densities in the ISM (see Fig 2).

The first of these models supposes that FRB progenitorsources are distributed like the general stellar population asa whole, which the second supposes they are closely relatedto very young stars.

In order to set up mock FRBs in the Milky Way, wegenerate large numbers of (X,Y ) positions1 with a surfacedensity in proportion to the stellar density distribution (ex-ponential disk model)- or to the electrons in the ISM (spiralarm model). In these 2-D models, Z = 0 ie the FRBs areplaced at the exact mid-plane on in order to maximize theeffects of the ISM. In both 2-D and 3-D models, we take intoaccount the ISM of the Milky Way, using the NE2001 andYMW16 models of the ionised ISM.

The dispersion measure DM is computed for each eventusing these models, where:

DM =

∫ d

0

ne(l)dl, (1)

where d is the distance from the FRB and the observerand ne is the electron density along the line-of-sight.

3.1.1 FRBs in an exponential disk

The surface density of the FRBs generated in this model ex-ponentially decrease with increasing distance from the galac-tic center i.e. they are randomly distributed under an expo-nential curve.

ρ ∝ e(−|R|/Rh), (2)

where R = (X2 +Y 2)0.5 is the distance from the Galac-tic center, and Rh is the scale length of the disk (here, weadopt Rh = 4 kpc as representative of the old stellar disk).

In Figure 1, we show that the FRBs are more concen-trated near the Sun and Galactic Center and less as we movefurther away. Taking into account the electron density anddistance, the FRBs near the Sun are both dispersed andscattered less as compared to the ones further away in thegalaxy.

1 X, Y and Z are Galoctocentric coordinates in the disk with theSun on the X axis at a distance of 8 kpc, and Z perpendicular to

the disk

MNRAS 000, 000–000 (2020)

Page 3: FRBs and the Milky Way - WordPress.com

FRBs in the Milky Way 3

Figure 1. The positions of the FRBs in the (X,Y ) plane of a sim-

ulated exponential disk (blue dots). The Sun’s position is shownby the yellow circle. The exponential length scale of the disk is 4

kpc, ie the surface density of FRBs is scaled as e−R/4, where R

is the distance from the center of the disk.

Figure 2. The (X,Y ) positions of the FRBs in a simulated spiral

disk, where the red dots represent the bursts while the yellow dot

represents the position of the Sun. The black dot represents theGalactic Center.

3.1.2 FRBs spatially co-located in the ionised medium

In our second model for the FRB distribution, positions forthe events are generated with probability proportional tothe integrated surface density of ne in the YMW16 model.This gives a distribution which is more akin to the sources ofionisation in the disk (i.e. young, energetic stars) and, as canbe seen in Figure 2, strongly traces the spiral arms in thatmodel. The electron density map is expressed in a logarithmscale, where the highest density is found to be in the spiralarms.

Figure 3. Histograms showing the number of FRBs as a functionof pulse width, for two models of the FRB spatial distributions.

The coloured regions of the histograms show the numbers that are

detectable for pulse widths of up to 20, 50 and 100 ms. The upperpanel shows the results for a FRBs distributed like stars in the

Milky Way’s exponential disk (see Fig 1), while the lower panel

shows FRBs distributed proportionally to the surface density ofthe ionised ISM – ie concentrated into spiral arms as in Fig 2.

3.2 Scattering and smearing of the FRB pulses

In order to assign pulse scattering for the FRBs, we use thepulse broadening fit to pulsars from Bhat et al. (2004) :

log(td) = −6.46 + 0.154x+ 1.07x2 − 3.9(log(f0)), (3)

where x = log(DM), td is the pulse broadening time (mea-sured in ms) and f0 is the frequency of observation in Hz(1.4 GHz for the STARE-2). We assume that the bursts havean intrinsic width of 0.61 ms (tintrinsic), such that the thetotal pulse tpulse width is:

τpulse = (t2d + t2intrinsic + t2dm)1/2 (4)

where tdm is the dispersion measure smearing (DM smear-ing), caused by the finite width of the frequency channels ofthe instrument.

DM smearing is given by:

tdm = 8.3×B ×DM × ν−3µs (5)

where B is the bandwidth, and ν is the width of the fre-quency channels. The bandwidth is taken as 0.122 MHz andfrequency is taken as 1.4 GHz for STARE-2. In our calcula-tions, intrinsic width has negligible effect on the final pulsewidth. We performed several trials and analyse the numbersof FRBs for widths of up to 20, 50 and even 100 ms. Theresults are discussed in Section 3.3 and 3.4.

3.3 2-D distribution of FRBs

The FRBs are placed in a two dimensional distribution forboth cases in order to maximize the ISM’s effect.

In Figure 4, the NE2001 (Cordes et al., 2003) was em-ployed for the exponential distribution of the FRBs. TheDM of bursts near the Sun here is 0 pc/cm3 and around

MNRAS 000, 000–000 (2020)

Page 4: FRBs and the Milky Way - WordPress.com

4 Gohar et al.

Figure 4. Upper panel : the distribution of the FRBs shown overthe electron density map (NE2001). Detectible FRBs are shown in

red, with the remainder shown in blue. The intrinsic FRB width in

this run is 0.61 ms. Lower panel: Disperson Measure map for theNE2001 model of the ISM in the Milky Way, as seen from the Sun

(yellow dot), computed for lines-of-sight in the mid-plane (Z = 0).FRBs in this model are distributed in a smooth exponential disk

(cf Eqn 2). FRBs which are detectable with a STARE-2 type

experiment are shown in green, with the rest of the FRBs shownin white. FRBs are typically detectable in a few kpc region around

the Sun in this model.

1500 pc/cm3 near the Galactic Center. The simulations re-vealed that 38 ± 2 percent of bursts, less than 50 ms, canbe found. If we take 20 ms, around 32 ± 1 percent of thesecan be detected. If we increase our range to 100 ms, 42 ±1 percent of FRBs can be detected. These error bars reflectPoisson uncertainties in the simulations.

In Figure 5, we employed the YMW16 model (Yao et al.2017) and the obtained different probabilities for detection.Here, the DM of bursts near the Galactic Center is around2500 pc/cm3 and zero for bursts near the Sun. By taking 20ms, 50 ms and 100 ms, as the limits, we found that 23 ± 0.4,

Figure 5. Upper Panel:the distribution of the FRBs shown over

the electron density map (YMW16). Detectible FRBs are shown

in blue, with the rest shown in red. The intrinsic FRB widthis taken as 0.61 ms. Lower panel: Disperson Measure map for

the YMW16 model of the ISM in the Milky Way, as seen from

the Sun (yellow dot), computed for lines-of-sight in the mid-plane(Z = 0). FRBs in this model are distributed in the spiral arm disk.

FRBs which are detectable with a STARE-2 type experiment are

shown in white, with the rest of the FRBs shown in red. FRBsare typically detectable in a few kpc region around the Sun inthis model too.

25 ± 0.3 and 29 ± 0.5 percent of the total FRBs fell withindetectable range.

For both cases, we added a factor of 10 scatter, in addi-tion to the pulse broadening equation 3. This intrinsic scat-ter in the Bhat relation at any given DM for different linesof sight will also apply to FRB like pulses.

These results imply that the ISM imposes a major con-straint on finding these transients. We inferred from our find-ings that due to the turbulent nature of this ISM, FRBs inthe Milky Way are likely to be affected, They tend to scat-ter and are dispersed to the extent where they cannot bedetected anymore. In addition, the results mean that de-tectable FRBs might be nearby. As distance plays a huge

MNRAS 000, 000–000 (2020)

Page 5: FRBs and the Milky Way - WordPress.com

FRBs in the Milky Way 5

Figure 6. Histograms of the distances of from the Sun of de-

tectable FRBs. The spatial model is the “exponential disk” distri-bution as seen in Fig 1. The three histogram colours show FRBs

which are less scattered than 20, 50 and 100 ms respectively.

Figure 7. Histograms of the distances from the Sun of detectable

FRBs for all three pulse widths. The spatial distribution of the

FRBs in the disk plane is as in Fig 2. The three histogram coloursshow FRBs which are less scattered than 20, 50 and 100 ms re-

spectively.

role in the DM. majority of them can be found out to 5-10kpc as shown in Figures 6 and 7. Therefore, the GalacticCenter is not a good option for searching these bursts.

3.4 3-D Distributions for the FRBs

We went further with our code and decided to model 3-Ddistribution of FRBs by adding a scale-height by introducingthe Z coordinate for FRBs. This was done to examine theeffects of ISM as the distance of bursts from the mid-planeis increased. The values of scale-height, hZ , were modeled inthe range 0 to 1 kpc, and a Gaussian distribution for Z wasconsidered. For both models, FRBs distributed in variouslevels of scale height were analysed.

FRBs lying on the mid-plane that are gradually ele-vated to a height 0.5 kpc increase the chances of observa-tion. Increasing this 1 kpc increases this likelihood further.We tested this out for pulse widths 20, 50 and 100 ms. Theresults are illustrated on Figures 8a and 8b.

In the NE2001 model, we increased the height from 0to 0.3 kpc, 42 ± 5 percent of the bursts were detectable for20 ms, 56 ± 3 percent for 50 ms, and 62 ± 5 percent for100 ms. Raising the height to 0.5 kpc, we obtained 59 ± 3percent for 20 ms, 66 ± 4 percent for 50 ms, and 72 ± 4percent for 100 ms. Escalating the height further to 1 kpc

(a)

(b)

Figure 8. Fractional FRB detection rates, versus the scale-heightof the progenitor populations. Panel (a): results for the NE2001

model and panel (b) shows reslts for the YMW16 model. Both

figures show an increasing trend in the detection as the scale-height increases. Results are shown for pulse widths of up to 20,

50 and 100 ms, as expected.

gave us 74 ± 3 percent for 20 ms, 76 ± 1 for 50 ms, 86 ± 2percent for 100 ms.

In the YMW16 model, we performed the same method.Increasing the scale-height to 0 to 0.3 kpc gave us 55 ± 1percent for 20 ms, 62 ± 3 percent for 50 ms, and 69 ± 5percent for 100 ms. Again, a further increase gave to 0.5 kpcshowed 71 ± 1 percent for 20 ms, 79 ± 2 percent for 50 ms,and 83 ± 1 percent for 100 ms. Elevating this to 1 kpc gaveus 91 ± 1 percent for 20 ms, 92 ± 1 percent for 50 ms, and96 ± 0.3 percent for 100 ms.

As the vertical height of bursts is increased from thecenter, the ISM’s effect gradually decreases. The reason isthat the Milky Way has an approximate thickness of 0.3kpc. The ISM’s effect tends to diminish as we plot the mockbursts farther from the mid-plane. FRBs are less likely toscatter and elongate in a relatively less ionized medium.

4 INTRINSIC ENERGY DISTRIBUTION

Other than pulse width, intrinsic energy is also an impor-tant factor when calculating our observational probabilities.According to Marcote et al. (2020), ST 200428A, the FRBlike event from the Milky Way, had an isotropic-equivalentenergy release of 2.2×1035 erg. By using a power law for en-ergy, we computed multiple results for various Z scaleheightsand power law slopes.

MNRAS 000, 000–000 (2020)

Page 6: FRBs and the Milky Way - WordPress.com

6 Gohar et al.

Figure 9. The pulse width distribution for the YMW16 model.

Power law index, α is taken as −1.5 and the energy cutoff is 1027J.Here, the S/N is also taken into account for FRBs. Detectable

FRBs have an S/N > 10 and pulse widths less than 20, 50 and

100 ms.

We adopt a power law energy distribution for the pulsesof:

dN(E) = (E/E0)αdE, (6)

where E is the intrinsic energy of the bursts, E0 is alower energy cutoff to the distribution, and α is the powerindex. α was allotted values of −0.05,−0.5,−1.5 and −2.5.We ran tests for both 2-D and 3-D distribution, and filteredout bursts with S/N < 10.

For 2-D exponential spatial distribution, if the lowerlimit was taken as E0 = 1029 J, for 20 and 50 ms as thisgave us similar values to 32 ± 1 % of the generated FRBs.Similarly, E0 =1030J gave us similar results for 100 ms. Forsteeper power laws, fewer FRBs are detected. Increasing α toa steeper luminosity function decreases the FRB detectionfractions into the range of 1 to 30 percent. For energy valuesbelow 1025J, power law indices of −2.5 and −1.5 suggestedthat approximately 1 % of FRBs are detectable. Energies be-low 1024J corresponded to −0.05 only and gave us detectableFRB rates of below 25% for all three pulse widths.

The same tests were carried out for the 2-D spiral dis-tribution, with the the same energy limits for 20, 50 and 100ms. Increasing α by 1 increased the detection in the rangeof 1 to 20 %. Energy below 1026J corresponded to −0.5 and−0.05 and increasing the value from −0.5 to −0.05 for theseenergies changes the detectablity rate of FRBs from 2 to24 %. It can be inferred from these results that as the en-ergy decreases, the chances of observation also decreases, nomatter what the energy distribution.

In 3-D distribution, increasing the scale-height im-pacted the detection. Since FRBs are already less scatteredand elongated as the height is increased, imposing an S/Nlimit of 10 further filters out the bursts, showing that thelikelihood of finding bursts, with these conditions, is reducedeven further.

5 COMPARISON OF MODELS WITHMAGNETARS IN THE MILKY WAY

Magnetar pulses from our galaxy were analyzed and includedin the simulations. Radio magnetars PSR J1745−2900, PSR

Figure 10. Plot of DM versus pulse widths for mock FRBs,

which includes magnetars’ values as well. Radio Magnetars wereincluded because the distribution of FRBs in YMW16 resembles

young stars. The red dots represents observable FRBs (with S/N

> 10 and a detectable pulse width), while the black dots representnon-observable FRBs. 2/5 of the magnetars’ pulse widths are too

scattered to be observed at a frequency of 1.4 GHz. These were

originally found at higher frequencies. This result validates themodel.

J119−6127, PSR J1622−4950 and XTE J1810−197 pulsewidths were calculated at a frequency of 1.4 GHz by theformula:

τ/τ0 = (ν0/ν4) (7)

where τ is the pulse width of the magnetar at our frequencyof observation, τ0 is the initial pulse width, ν is our frequencyof observation (1.4 GHz), and ν0 is the original frequency ofobservation. For PSR J1745−2900, (Pearlman et al. 2018)detected a single pulse broadening timescale of 6.9 ± 0.2 msat 8.4 GHz. The DM for this magnetar is 1778 ± 3 cm−3pc(Eatough et al. 2013). For PSR J119−6127, the DM is 707.4± 1.3 pc cm−3 (He et al. 2013), and a spin period of 410ms (Majid et al. 2017). Measuring the pulse profile for thismagnetar from (Majid et al. 2017), the pulse width is 8.2ms in 2.3 GHz. Magnetar PSR J1622−4950 has a rotationperiod of 4.3 secs and DM of 820 pc cm−3 (Levin et al. 2010).By measuring the pulse profile in (Pearlman et al. 2019), wegot 86.6 ms in 2.3 GHz. Magnetar XTE J1810−197 has a DMof 178 ± 9 pc cm−3 (Pearlman et al. 2019). Magnetar XTEJ1810−197’s pulse width, 3 ms, was measured by takingdata from UTMOST. Applying Equation 7, we calculatedthe widths of these pulses at a frequency of 1.4 GHz. Thecalculated widths were 8942 ms, 60 ms, and 630 ms.

The results in Figure 10 indicate that two out of fivemagnetars are too scattered to be observed at STARE2’sfrequency of observation (1.4 GHz). This means that thesepulses are consistent with our simulations. The magnetardistances were also included in Figure 11. All of them fallare less than 10 kpc away from us, meaning they are ina detectable distance. PSR J1745−2900 is the only candi-date here with a higher DM than the other candidates. Thismeans that FRB like events from this star are probably toobroad to be detected.

MNRAS 000, 000–000 (2020)

Page 7: FRBs and the Milky Way - WordPress.com

FRBs in the Milky Way 7

Figure 11. The distance versus DM plot for mock FRBs, and for

the 5 radio magnetars for which pulses have been detected in theradio spectrum. As shown in the figure, the distance of all these

magnetars are within the detectable range of 5-10 kpc.

Figure 12. Positions of 5 magnetars for which radio detectionshave been made, shown in Galactic coordinates. The DM of the

YMW16 model is shown in the background, and the green line

marks the divide between the northern and southern hemipsheres.

6 IMPROVED DETECTION STRATEGIES

We have used the results of our models to suggest how wecould improve the detection rate of FRB-like pulses fromsources in the Milky Way.

6.1 Southern Hemisphere search and Field ofView

The large field of at STARE-2 was crucial to finding thebright FRB-like event from SGR1935.

In Figs 12 and 13 we show the distribution of Milky Waymagnetars and mock FRBs in our 3-D model (scale height0.1 kpc) in Galactic coordinates, overlaid on the DM fromthe YMW2016 model to a distance of 30 kpc. The green linein the two figures shows the Northern and Southern hemi-spheres. For all of the models run, 60 to 70% of detectableFRBs can be found in the Southern Hemisphere. Figure 12shows that aside from SGR 1935+2154, the other (known)radio magnetars are located in the South. Of approximately30 magnetars, only 6 are known radio sources.

6.2 Higher frequency search

Another strategy to increase the detection rate of FRBswould be to increase to frequency of the search experiment,

Figure 13. The position of the detectable FRBs, represented

by white dots, as viewed in Galactic coordinates.The YMW16model was used here, with a spiral distribution of FRBs in 3-D.

The scale-height was taken as 0.5 kpc. The majority (60 to 70

%) of the FRBs are located in the Southern Hemisphere as onewould expect, bolstering the case for a STARE-2 like experiment

in the south.

since the scattering of the FRB pulses is highly dependenton the frequency ((Bhat et al. 2004), see Eqn 3 and Eqn 7):

Our base simulations run at 1.4 GHz, which is the op-erating frequency of STARE-2. We ran simulations at a fre-quency of 3 GHz to estimate the improvement in the detec-tion rate, finding an increase of only 3 to 4% more FRBs (inthe Southern Hemisphere). Increasing the frequency, how-ever, reduces the field of view, so a small array of telescopestracking the Milky Way in that area would help.

7 CONCLUSIONS

We conclude that the ISM has a major effect on the pulsewidths of the FRBs. Our chances of observing these ra-dio transients in the Milky Way are severely affected. Thisfact was established by creating 3-D spatial distributions ofFRBs that exhibited a gradual decrease in pulse width elon-gation as the presence of the ISM became less prominent.

We adopted a power law luminosity function and exam-ined the effects of a lower energy cutoff and different powerlaw slopes on the numbers of detected FRBs. The luminosityfunction is not a major driver of FRB detection rates, whichis still mostly affected by the wide range of DMs throughwhich the bursts travel, at least for the power laws thatwere modeled.

We used radio magnetar pulse profiles to validated oursimulations. At 1.4 GHz, about half of the radio magnetarshave pulse widths which are considered detectible with cur-rent FRB searches, consistent with the simulations.

We discuss improved strategies for finding FRB likepulses in the Milky Way. The clearest is to build a STARE-2like experiment in the Southern Hemisphere, where mostof the modelled FRBs arise. Smaller improvements can bemade by improving the software that searches for FRBs,from 20 ms out to approximately 100 ms pulse widths,and/or moving to high frequencies of detection (to reducepulse scattering effects).

MNRAS 000, 000–000 (2020)

Page 8: FRBs and the Milky Way - WordPress.com

8 Gohar et al.

REFERENCES

Bhandari S., Bannister K. W., James C. W., Shannon R. M.,

Flynn C. M., Caleb M., Bunton J. D., 2019, MNRAS, 486, 70

Bhat N. D. R., Cordes J. M., Camilo F., Nice D. J., LorimerD. R., 2004, ApJ, 605, 759

Bochenek C. D., McKenna D. L., Belov K. V., Kocz J., Kulkarni

S. R., Lamb J., Ravi V., Woody D., 2020a, PASP, 132, 034202Bochenek C. D., Ravi V., Belov K. V., Hallinan G., Kocz J.,

Kulkarni S. R., McKenna D. L., 2020b, Nature, 587, 59

Cui X.-H., et al., 2021, MNRAS, 500, 3275Eatough R. P., et al., 2013, Nature, 501, 391

Farah W., et al., 2017, The Astronomer’s Telegram, 10697

He C., Ng C. Y., Kaspi V. M., 2013, ApJ, 768, 64Levin L., et al., 2010, ApJ, 721, L33

Lorimer D. R., Bailes M., McLaughlin M. A., Narkevic D. J.,

Crawford F., 2007, Science, 318, 777Macquart J.-P., Shannon R. M., Bannister K. W., James C. W.,

Ekers R. D., Bunton J. D., 2019, ApJ, 872, L19Majid W. A., Pearlman A. B., Dobreva T., Horiuchi S., Kocz J.,

Lippuner J., Prince T. A., 2017, ApJ, 834, L2

Pearlman A. B., Majid W. A., Prince T. A., Kocz J., Horiuchi S.,2018, ApJ, 866, 160

Pearlman A. B., Majid W. A., Prince T. A., 2019, Advances in

Astronomy, 2019, 6325183Petroff E., et al., 2017, preprint, (arXiv:1710.08155)

Ravi V., et al., 2019, Nature, 572, 352

Scholz P., Chime/Frb Collaboration 2020, The Astronomer’s Tele-gram, 13681, 1

Shannon R. M., et al., 2018, Nature, 562, 386

Yao J. M., Manchester R. N., Wang N., 2017, ApJ, 835, 29

MNRAS 000, 000–000 (2020)