ebook img

Searching for giga-Jansky fast radio bursts from the Milky Way with a global array of low-cost radio receivers PDF

0.14 MB·
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Searching for giga-Jansky fast radio bursts from the Milky Way with a global array of low-cost radio receivers

Mon.Not.R.Astron.Soc.000,000–000 (0000) Printed15February2017 (MNLATEXstylefilev2.2) Searching for giga-Jansky fast radio bursts from the Milky Way with a global array of low-cost radio receivers 7 Dan Maoz1, Abraham Loeb1,2 1 1 School of Physicsand Astronomy, Tel-AvivUniversity, Tel-Aviv69978, Israel 0 2 2Institute forTheory and Computation, Harvard University,Cambridge, MA 02138, USA b e F 15February2017 4 1 ABSTRACT If fast radio bursts (FRBs) originate from galaxies at cosmological distances, then ] M their all-sky rate implies that the Milky Way may host an FRB every 30–1500years, on average.If many FRBs persistently repeat for decades or more, a local giant FRB .I could be active now, with 1 GHz flux radio pulses of ∼ 3×1010 Jy, comparable to h the fluxes and frequencies detectable by cellular communication devices (cell phones, p Wi-Fi, GPS). We propose searching for Galactic FRBs using a global array of low- - o costradioreceivers.Onepossibilityisthe∼1GHz communicationchannelincellular r phones, through a Citizens-Science downloadable application. Participating phones t s would continuously listen for and record candidate FRBs and would periodically up- a loadinformationtoacentraldata-processingwebsitewhichwillidentifythesignature [ of a real, globe-encompassing, FRB from an astronomical distance. Triangulation of 3 theGPS-basedpulsearrivaltimesreportedfromdifferentEarthlocationswillprovide v theFRBskyposition,potentiallytoarcsecondaccuracy.Pulsearrivaltimesversusfre- 5 quency,fromreportsfromphones operatingat diversefrequencies,orfromfastsignal 7 de-dispersionbytheapplication,willyieldthedispersionmeasure(DM).Comparedto 4 aGalacticDMmodel,itwillindicatethesourcedistancewithintheGalaxy.Avariant 1 approach uses the built-in ∼ 100 MHz FM-radio receivers present in cell phones for 0 anFRB searchatlowerfrequencies.Alternatively,numerous“software-definedradio” 1. (SDR) devices,costing ∼$10US each,couldbe deployedandpluggedinto USB ports 0 of personal computers (particularly in radio-quiet locations) to establish the global 7 network of receivers. 1 : Key words: stars: radio continuum, variables v i X r a 1 INTRODUCTION cosmological distance hasbeen confirmed in thecase of the sole repeating FRB 121102, which has been localised to a The origin and nature of fast radio bursts (FRBs) have dwarf galaxy at redshift z = 0.19 (Chatterjee et al. 2017; remained enigmatic since the first FRB discovery by Tendulkaret al. 2017; Marcote et al. 2017). Lorimer et al.(2007).The17orsodistinctFRBsourcesthat have been reported so far are bright ( 0.1–1 Jy) and brief Therangeofexcess(aboveGalactic)DMsoftheknown ∼ ( 1 ms) pulses of 1 GHz radio emission (Lorimer et al. FRBs correspond to a range of co-moving distances of ∼ ∼ 2007; Keane et al. 2012; Thornton et al. 2013; Spitler et al. 0.9–4.4 Gpc, with a median of 2.4 Gpc (corresponding 2014;Burke-Spolaor & Bannister2014;Petroff et al.2015a; to a redshift of z = 0.64), under the assumption that Raviet al. 2015; Champion et al. 2016; Masui et al. 2015; most of the excess DM is contributed by the intergalactic Keane 2016). The pulse arrival times of FRBs show a ν−2 medium and using the standard cosmological parameters frequency dependence indicative of a passage through a (Planck Collaboration et al. 2016). The latest estimate of cold plasma, with the so-called dispersion measure (DM) the all-sky rate of FRBs with flux > 0.3 Jy is 1.1+3.8 measuring the line-of-sight column density of free elec- 104 day−1 at 95% confidence (Scholzet al. 2016). T−hi1s.0e×s- trons. FRBs are selected to have large measured DMs of timate is based on thesingle detection of FRB121102, and 300 1600pccm−3,inexcessofthevaluesexpectedfrom ignoring the fact that it has been detected repeatedly over ∼ − models of theinterstellar electron distribution in theMilky a period of 4 years. We note that some other recent rate Way galaxy, and have therefore been inferred to originate estimates have smaller uncertainties, but find, to varying from extragalactic sources at cosmological distances. The degreesofsignificance,adependenceofrateonGalacticlat- (cid:13)c 0000RAS 2 D. Maoz and A. Loeb itude (see, e.g., VanderWiel et al. (2016), and references ferentlocations,andtheabilitytofilteroutthatforeground therein). By adopting the Scholz et al. (2016) rate and its noise. uncertainty, we encompass this uncertainty in the latitude dependenceas well. 2.1 A cell phone communications channel IfFRBsoccurinMilky-Way-likegalaxies,wecandivide approach the observed cosmological rate by the number of galaxies within the FRB survey volume, to find the expected FRB There are currently an estimated 7 billion active cellular rate within a single such galaxy. We note that the recently phone accounts on our planet (similar to the number of localised FRB 121102 comes from an extreme-emission-line people), operating in several frequency bands, from 0.8 to dwarfgalaxy(Tendulkaret al. 2017),quiteunliketheMilky 2.4 GHz. Each of these phones is, as argued above, a ra- Way. However, this host galaxy is not necessarily repre- dio receiver that is in principle sensitive to a Galactic FRB sentative of all FRB hosts. The product between the co- signal. Furthermore, every smartphone is a programmable moving number density of L∗ galaxies, 10−2 Mpc−3 computer capable of analyzing the signal, of timing it up ∼ (Montero-Dorta & Prada 2009), and the cosmological vol- to ∆t 10−7 s precision with its global-positioning system umeouttothemediandistanceofknownFRBs,57comov- (GPS)∼module, of storing this information, and of diffusing ing Gpc3, implies that an FRB should occur in our galaxy it through theinternet. once per 140+−1141000 years (Lingam & Loeb 2017). A 0.3 Jy WeproposethereforetobuildaCitizens-Scienceproject FRBfromacomovingdistanceof2.4Gpc(aluminositydis- in which participants voluntarily download onto their tance of 3.9 Gpc), placed at a typical Galactic distance of phones an application that runs in the background some 10 kpc, would have an observed 1 GHz flux density of orallofthetime,monitoring thephone’santennainputfor ∼fν 3 1010 Jy or, equivalently, 3 10−16 W m−2Hz−1. candidate broad-band millisecond-timescale pulses that ap- ≈ × × FRB 121102 has been bursting repeatedly for at least 4 pear similar to an FRB. The application would record can- years. If most FRBs persist for decades or even centuries, didate FRB pulses (most of which originate from artificial aGalactic FRBcouldbeactivenow. Apowerfullocal FRB and natural noise sources) and would periodically upload may have already been detected in the far side-lobes of ra- the candidate pulse information (pulse profile, GPS-based dio telescope beams, but mistakenly ascribed to artificial arrivaltime),alongwithinformationaboutthephone(GPS- interference. basedlocation,operatingfrequency)toacentralprocessing Indeed,theradiofluxdensitylevelofthereceivedGHz- website. The central website will continuously correlate the band signals from commercial radio stations, cellular com- incoming information from all participants, to identify the munications and wireless networksis within afew orders of signature of a real, globe-encompassing, FRB. magnitude of the expected flux level from a Galactic FRB. Becauseofthereceivedsignal’sintegrationintomstime For example, a typical desktop Wi-Fi transmitter operat- bins (required to improve the sensitivity to FRB levels, see ing at 2.4 GHz under the 802.11b standard has a radiated above), every phone’s actual arrival time accuracy will be power of 100 mW over an 82 MHz bandpass with an out- no better than 10−3 s. However, improved time precision door range of 100 m, corresponding to a detected flux can be recovered by averaging the reported arrival times density of fν =∼1 10−14 W m−2Hz−1. Each of the trans- recorded by many participating phones at a similar loca- × mitter’sindividualchannelshasabandpassof22MHz,and tion. For example, averaging the ms-precision reports from therefore a time resolution of ∆t 2 10−8 s. By bin- 10,000phoneswithinacityofradius3km(lighttraveltime ning an incoming signal into milli∼secon×d (∆t = 10−3 s) <10−5s),wouldimprovetheprecisionbyafactorof100,to time bins a Wi-Fi receiver would improve its sensitivity in 10−5s.TriangulationoftheGPS-timedpulsearrivaltimes proportion to √∆t, i.e. by a factor of 200, to a level of ∼fromdifferentEarthlocations wouldthengivetheFRBsky ifsνa∼fa5ct×or160−fa17inWtermth−a2nHtzh−e1t(y5p×ica1l0G9aJ∼lya,cti.iec.F5RGBJfly)u.xTdhisis- tpiomsietiboinnntoinagnofactchueraFcRyBofsiogrndaelr, ∼ancd∆stu/b2sReq⊕u∼ent1laorscsmoifnt.hIef cussed above, and means that such a Galactic FRB, and native10−7 sGPStimingprecisioncouldbeavoided(apos- evenfainterandperhaps-more-frequentFRBs,wouldbede- sibilityconsideredinsomeoftheothertechnicalframeworks tectable by existing communication devices. In the subse- thatweproposebelow),thennaturallythelocalisation pre- quent sections, we outline how an array of numerous low- cision can be improved down tothe sub-arcsecond level. cost radio receivers can beused todetect and localise giant Because of the ν−2 arrival-time dependence of a radio Galactic FRBs. pulsepropagating through theGalactic plasma, phones op- eratingatdiversefrequencies(multiplenetworksandphone models) will receive thesignal at a time delay, 2 A GLOBAL ARRAY OF CELLULAR DM ν −2 δt=0.144 s. (1) RECEIVERS FOR GALACTIC FRB ×(cid:18)200 pc cm−3(cid:19)(cid:16)2.4GHz(cid:17) DETECTION Over, e.g., a 22 MHz cellphone channel bandwidth Weconsiderbelowthreerelatedtechnicalapproachestothe at 2.4 GHz, a typical Galactic DM of 200 pc cm−3 assemblyofanarrayoflow-cost radioreceivers,suitablefor (Rane& Loeb 2016) will spreadtheFRBarrival timeover thedetectionofGalacticFRBs.Thechoiceofthemostprac- just 2.6 ms, comparable to typical FRB pulse widths. The ticalapproach will dependonseveral issuesthatneedtobe channel bandwidth therefore will not result in any signif- resolved, such as the ability to access and manipulate raw icant smearing of the pulse over time, which could have radiosignalspickedupbytheantennas,thefluxfromFRBs reduced the detection sensitivity and timing precision. By at sub-GHz frequencies, the level of terrestrial noise at dif- comparing the arrival times of different frequencies at the (cid:13)c 0000RAS,MNRAS000,000–000 Searching for Galactic FRBs with radio receivers 3 same locations, the central website will be able to solve 2.2 A cell phone FM radio channel approach for the FRB’s DM that, when compared to a Galactic DM Mostorallcellphoneshavebuilt-inFM-bandradioreceivers model (Cordes & Lazio 2002; Yaoet al. 2016),will indicate operating at around ν 100 MHz and enabling direct (i.e. the FRB source distance within the Galaxy. Alternatively, ∼ notthroughtheinternetortheserviceprovider)receptionof the application software itself could attempt to de-disperse radiobroadcasts.Interestingly,thishardwareisde-activated allcandidateincomingsignalsacrossthefullfrequencyrange by phone manufacturers in about 2 of all phones, in the available to each receiver. With efficient new algorithms, ∼ 3 service-providers’interestofhavingthecustomersdownload real-time de-dispersion of FRB signals is now feasible on andpayfortheradiobroadcasts,ratherthanreceivingthem small computers (Zackay & Ofek 2014; Zackay 2017), and for free. Nevertheless, about 1 of all phones (still a size- so is likely possible on smartphones as well. In such a sce- 3 nario, the identification of a ν−2 frequency sweep would be able number when considering the global number) do have the direct FM reception option activated. The raw, non- a real-time test of incoming signals, performed at the level demodulated radio signal from this channel is more likely of each individual receiver. to beaccessible to the application software in its search for Oneclearadvantageoftheaboveoperationplanisthat an FRB signal than in the preceeding approach using the itisessentially costfree—allofthenecessaryhardware(the 1GHzcellularcommunication channel.Ashortcomingof world’s cell phones) is already in place, and one needs only ∼thisoption,however,istheyet-unknownpropertiesofFRBs tocarryouttheplan’sorganizationalstepsinordertomake at 100 MHz frequencies. Current upper limits from FRB it work for the scientific program. Potential problems with sea∼rches at 145 MHz (Karastergiou et al. 2015) and 139- thisproposedmodeare,first,thatcellphonesmaybehard- 170MHz(Tingay et al.2015),limittheFRBspectralslope wired at the basic electronics level to demodulate and digi- to > +0.1. As with the 1 GHz cellular-communications tizeincomingcommunicationssignals,andthereforetheraw option,discussedabove,h∼eretoointegrationovertimecould broad-band radio signal containing the FRB may be inac- make a typical Galactic FRB detectable at 100 MHz, even cessible to software. Furthermore, mobile phone communi- for more positive slopes as high as +1, such that the FRB cations are encoded so as to allow many users to share the wouldhave&5GJyat100MHz.Theforegroundnoiseques- frequency band, and this encoding permits thedetection of tioninthisoptionissimilar(thoughinadifferentfrequency communicationsignalsatsub-noiselevels(asopposedtothe band)to that in theprevious, cellular-communications, op- un-encodedFRB signal). tion. The sought-after millisecond-timescale FRB signal will need to be disentangled from the foreground noise of cel- lular and other communications emissions, as well as from 2.3 A software-defined radio approach natural radio noise from atmospheric processes and from Asoftware-definedradio(SDR)isaradiosystemwherecom- the sun. Although the feasibility of this requires further ponents such as filters, amplifiers, demodulators, etc., that study, the prospects look promising based on a number of aretypicallyimplementedinhardware,areimplementedin- past attempts. Katz et al. (2003) review a handful of ex- stead in software on a personal computer. SDR devices are periments, and decribetheir own experiment,which is sim- widely available for $10 US a piece, and they are popular ilar to our proposal. The basic concept consisted of a num- ∼ with radio amateurs. They are often the size of a memory berofwide-angle,geographically distributedradioreceivers stick and likewise can be USB plugged. An SDR device in- that searched for short radio bursts, separating astronom- cludesanantennathancandetectthefullrawambientradio ical signals from noise by requiring coincident detections. emissions over some frequency range and can input them Katz et al. (2003)usedthree611MHzreceiversintheeast- with minimal processing into a computer, where the sig- ernUS,sensitiveto&3 104 Jybursts(2 105 fainterthan × × nals can be software-processed at will. Our third approach considered here) on timescales & 125 ms (50 times longer is therefore to deploy a large number (depending on the than here). The recorded, GPS-time-stamped, bursts were available budget) of such SDR devices, to be plugged into periodicallyuploadedtoacentralprocessingstation,exactly participating personal computers around the globe, or base as in our proposed plan. thenetworkondevicesalreadyinusebyparticipatingradio Over 18 months of continuous operation, Katz et al. amateurs. As with the phone option, the participants will (2003) detected a burst roughly every 10 s, but 99.9% of download and install software that will continuously moni- thesesignalscouldberejectedaslocalinterferencebasedon tor the input from the SDR. As before, the computers will their non-coincidence between the three receivers. The re- upload theinformation on candidateMilky WayFRBstoa maining 4000coincident signalscould allbetracedtoso- central data-processing website. ∼ larradiobursts,bycomparisontoreportsfromasolarradio A disadvantage of this approach is the need to actu- observatory. No other astronomical sources were detected ally buy and send the SDR hardware to the selected par- by Katzet al. (2003) nor by previous experiments. Inter- ticipating individuals of the network (unless one takes the estingly, Katzet al. (2003) succeeded in using their GPS existing-amateur-SDR approach). The advantages involve signal, with its 10−7 s accuracy, to time-stamp their de- having an accessible FRB signal, uniformly processed and tected bursts to the accuracy of their 20 µs-long individual fullyanalysableatwill(includingspectralinformationfrom time samples, and they note that, in principle, they could every station). Every SDR could be supplemented with a have used a multiple-time-sample averaging period shorter simple exterior antenna or antennabooster (wireless recep- than0.125s(attheexpenseofsensitivity).Thiswouldhave tionboostersarealsowidelyandinexpensivelyavailablefor allowed them to triangulate their source localisations, just cell phones and laptops) that would considerably enhance as we propose to do. its sensitivity, lowering or fully avoiding the need for time (cid:13)c 0000RAS,MNRAS000,000–000 4 D. Maoz and A. Loeb integration,andhenceforthesacrificeoftimingprecision,or e.g. by adding simple antennas in the SRD option, would simplyprobingforfainterandmorefrequentbursts(seebe- potentially allow for the detection of Galactic FRBs on a low).Furthermore,theabilitytochoosethestationssitesat yearly to weekly basis, and for the direct determination of will in a well-spaced global network, specifically in “radio- theirluminosity function. quiet” locations with minimal artificial and natural radio interference, may proveto bethe most important benefit. 4 CONCLUSIONS 3 LOWER-FLUX, MORE-COMMON The first, and so-far only, FRB that has been localised, GALACTIC FRBS FRB 121102, is at a cosmological distance, it has been re- peating for at least 4 years, and its host galaxy is a low- A major practical problem of the schemes described above metallicity dwarf. We have argued that if most FRBs are are the long and uncertain timescales—decades to many cosmological,buttheirhostsarenotnecessarilydwarfgalax- centuries—expected for the detection of a single, Galactic ies like the host of FRB 121102, then their all-sky rate im- 3 1010 Jy FRB, unless typical FRBs persistently repeat × plies that the Milky Way hosts an FRB every 30 to 1500 for decades or centuries (which is a real possibility, given years.If,furthermore,manyFRBsrepeatlikeFRB121102, the case of FRB 121102). If FRBs typically do not repeat, andforlongenough,thentheoccurrencefrequencycouldbe then even for the more optimistic end of the rate estimate, higher, and alocal FRB may evenbe activenow. A typical broadcasting standards, phone models and other technical Galactic FRB will be a millisecond broad-band radio pulse factors, may change over a decade, not to mention the lim- with 1 GHz flux density of 3 1010 Jy, not much differ- ited patience of the participants and the experiment man- ∼ × ent from the radio flux levels and frequencies detectable by agers.Aresolution ofthisconcern,however,couldbebased cellularcommunicationdevices(cellphones,WiFi,GPS).If on the fact that FRBs must have a distribution of lumi- the Milky Way has a currently active and repeating FRB nosities. Indeed, if the known FRBs are at the cosmologi- source, then some Local Group galaxies would have them cal distances indicated by their excess DMs, then they are too, at MJy flux-density levels, which could be detected by clearly not “standard candles”. Areasonable expectation is monitoring nearbygalaxies with dedicated small radio tele- thenthatFRBnumbersincreaseatdecreasingluminosities. scopes. If so, lower-luminosity FRBs should be detected more fre- An argument against frequent Galactic FRBs could be quentlyby theglobal cellular network. that FRBs require some kind of exotic and energetic phys- LetusassumethatwecanparameterizetheFRBnum- ical event, such as a super-flare from a magnetar, and that ber perunit luminosity with a Schechterform, irradiationoftheEarthbysuchaneventoncepercenturyor dN L−α+1e−Lν/LF⋆, (2) milleniumwouldbeaccompaniedbyclearsignatures,orper- d(logLν) ∝ ν hapsevenbymassextinctions.However,thisargumentrelies onastill-speculativeconnectionbetweentheradioemission where L corresponds to the characteristic specific lumi- F⋆ ofFRBsandtheiremissionsinotherbands.Observationally, nosity of an FRB source (namely, the one that yields an an upper limit of 108 Jy has been set on any FRB-like ra- observedfluxdensityof 0.3Jyataluminositydistanceof ∼ diofluxaccompanyingthegiant2004Decemberγ-rayburst 4Gpc).Onewaytocalibrateαisbyspeculatingthatthe ∼ fromthemagnetarSGR1806-20,andnoγ-raycounterparts Galactic population of rotating radio transients (RRATs), havebeen detected for any FRB (Tendulkaret al. 2016). which have some properties in common with FRBs, con- Our proposed search for Galactic FRBs using a global stitute the low-luminosity counterparts of FRBs. The rate array of low-cost (possibly already existing) radio receivers of RRATs over the entire sky at a flux of 0.3 Jy is 106 day−1, based on the estimated numbe∼r of sources would enable triangulation of the GPS-timed pulse arrival i∼n the Galaxy, 105, and their individual repetition rates, timesfromdifferentEarthlocations,localisingtheFRBsky 10day−1(Mc∼Laughlin et al.2006).TheRRATrateisthus positiontoarcminuteorevenarcsecondprecision.Pulsear- ∼1011 timestheGalacticFRBrate(ofonceper300yr,i.e. rivaltimesfrom devicesoperatingat diversefrequencies, or ∼10−5 day−1). The RRAT flux of 0.3 Jy, in turn, corre- from de-dispersion calculations on the devices themselves, sponds to 10−11 the typical flu∼x of a Galactic FRB. If will yield the DM that, when compared to a Galactic DM ∼ model, will indicate the FRB source distance within the thesetwopopulationsoftransientradiosourcesarerelated, Galaxy. Fainter FRBs could potentially be detected on a thenα 2.Interestingly,thisvaluecorrespondstoanequal ≈ yearly or even weekly basis, enabling a direct measurement luminosity contribution from transients per logarithmic in- of theFRB luminosity function. terval of luminosity. The flux distribution from a Galactic FRB population having a particular luminosity will be (dN/d(logf )) f−3/2 for a spherically distributed population (e.gν. i|nLνth∝e ν ACKNOWLEDGMENTS Galactic halo), or f−1 for a planar distribution (e.g. the ∝ ν Galacticdisk)—coincidentallymatchingthepower-lawscal- WethankC.Carlsson,A.Fialkov,J.Guillochon,Z.Manch- ing at low fluxes in the luminosity function for α = 2. At ester, E.Ofek, M. Reid,B. Zackay,and theanonymousref- a 5 GJy flux level, still detectable by our proposed arrays, eree, for useful advice and comments. This work was sup- one might then expect to find Galactic FRBs 6 times more ported in part by Grant 1829/12 of the I-CORE program frequently than at 30 GJy, i.e. once per 5 to 250 years. In- of the PBC and the Israel Science Foundation (D.M.) and creasing thesensitivitybyoneortwoordersofmagnitudes, byagrantfrom theBreakthroughPrizeFoundation(A.L.). (cid:13)c 0000RAS,MNRAS000,000–000 Searching for Galactic FRBs with radio receivers 5 A.L. acknowledges support from the Sackler Professorship by Special Appointmentat Tel Aviv University. REFERENCES Burke-Spolaor, S.,& Bannister, K.W. 2014, ApJ,792, 19 Chatterjee, S., et al. 2017, Nature, 541, 58 Champion,D.J., Petroff,E., Kramer,M.,et al.2016, MN- RAS,460, L30 Cordes, J. M., & Lazio, T. J. W. 2002, arXiv:astro-ph/0207156 Karastergiou,A.,Chennamangalam,J.,Armour,W.,etal. 2015, MNRAS,452, 1254 Katz, C.A., Hewitt, J.N., Corey, B.E., Moore, C.B 2003, PASP,115, 675 Keane,E.F.,Stappers,B.W.,Kramer,M.,&Lyne,A.G. 2012, MNRAS,425, L71 Keane, E. F. 2016, MNRAS,459, 1360 Lingam, M. & Loeb, A.2017, arXiv:1701.01109 Lorimer, D. R., Bailes, M., McLaughlin, M. A., Narkevic, D.J., & Crawford, F. 2007, Science, 318, 777 Marcote, B., et al. 2017, ApJ, 834, L8 Masui, K.,Lin,H.-H.,Sievers,J.,et al.2015, Nature,528, 523 McLaughlin, M. A., Lyne, A. G., Lorimer, D. R., et al. 2006, Nature, 439, 817 Montero-Dorta, A. D., & Prada, F. 2009, MNRAS, 399, 1106 Petroff, E., Bailes, M., Barr, E. D., et al. 2015, MNRAS, 447, 246 Planck Collaboration, Ade, P. A. R., Aghanim, N., et al. 2016, A&A,594, A13 Rane, A., Loeb, A. 2016, arXiv:1608.06952 Ravi, V., Shannon,R.M., & Jameson, A. 2015, ApJ, 799, L5 Scholz,P.,Spitler,L.G.,Hessels,J.W.T.etal.2016,ApJ, 833, 177 Spitler,L.G.,Cordes,J.M.,Hessels,J.W.T.,etal.2014, ApJ,790, 101 Tendulkar, S.P., Kaspi, V.M., & Patel, C. 2016, ApJ, 827, 59 Tendulkar, S.P., et al. 2017, ApJ, 834, L7 Tingay, S. J., Trott, C. M., Wayth, R. B., et al. 2015, AJ, 150, 199 Thornton,D.,Stappers,B.,Bailes,M.,etal.2013,Science, 341, 53 Vander Wiel, S., Burke-Spolaor, S., Lawrence, E., Law, C.J., Bower, G.C. 2016, arXiv:1612.00896 Yao, J. M., Manchester, R. N., & Wang, N. 2016, arXiv:1610.09448 Zackay,B., Ofek, E. 2014, arXiv:1411.5373 Zackay, B. 2017, American Astronomical Society Meeting Abstracts, 229, 330.06 (cid:13)c 0000RAS,MNRAS000,000–000

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.